Try fixing segfault by not being as aggressive with freeing Persisted Resolvers

allow `HTMLEvents` in `createEvent`

.github/actions/install/action.yml

@@ -0,0 +1,94 @@
+name: "Browsercore install"
+description: "Install deps for the project browsercore"
+
+inputs:
+  zig:
+    description: 'Zig version to install'
+    required: false
+    default: '0.15.1'
+  arch:
+    description: 'CPU arch used to select the v8 lib'
+    required: false
+    default: 'x86_64'
+  os:
+    description: 'OS used to select the v8 lib'
+    required: false
+    default: 'linux'
+  zig-v8:
+    description: 'zig v8 version to install'
+    required: false
+    default: 'v0.1.30'
+  v8:
+    description: 'v8 version to install'
+    required: false
+    default: '14.0.365.4'
+  cache-dir:
+    description: 'cache dir to use'
+    required: false
+    default: '~/.cache'
+
+runs:
+  using: "composite"
+
+  steps:
+    - name: Install apt deps
+      if: ${{ inputs.os == 'linux' }}
+      shell: bash
+      run: |
+        sudo apt-get update
+        sudo apt-get install -y wget xz-utils python3 ca-certificates git pkg-config libglib2.0-dev gperf libexpat1-dev cmake clang
+
+    - uses: mlugg/setup-zig@v2
+      with:
+        version: ${{ inputs.zig }}
+
+    - name: Cache v8
+      id: cache-v8
+      uses: actions/cache@v4
+      env:
+        cache-name: cache-v8
+      with:
+        path: ${{ inputs.cache-dir }}/v8
+        key: libc_v8_${{ inputs.v8 }}_${{ inputs.os }}_${{ inputs.arch }}_${{ inputs.zig-v8 }}.a
+
+    - if: ${{ steps.cache-v8.outputs.cache-hit != 'true' }}
+      shell: bash
+      run: |
+        mkdir -p ${{ inputs.cache-dir }}/v8
+
+        wget -O ${{ inputs.cache-dir }}/v8/libc_v8.a https://github.com/lightpanda-io/zig-v8-fork/releases/download/${{ inputs.zig-v8 }}/libc_v8_${{ inputs.v8 }}_${{ inputs.os }}_${{ inputs.arch }}.a
+
+    - name: install v8
+      shell: bash
+      run: |
+        mkdir -p v8/out/${{ inputs.os }}/debug/obj/zig/
+        ln -s ${{ inputs.cache-dir }}/v8/libc_v8.a v8/out/${{ inputs.os }}/debug/obj/zig/libc_v8.a
+
+        mkdir -p v8/out/${{ inputs.os }}/release/obj/zig/
+        ln -s ${{ inputs.cache-dir }}/v8/libc_v8.a v8/out/${{ inputs.os }}/release/obj/zig/libc_v8.a
+
+    - name: Cache libiconv
+      id: cache-libiconv
+      uses: actions/cache@v4
+      env:
+        cache-name: cache-libiconv
+      with:
+        path: ${{ inputs.cache-dir }}/libiconv
+        key: vendor/libiconv/libiconv-1.17
+
+    - name: download libiconv
+      if: ${{ steps.cache-libiconv.outputs.cache-hit != 'true' }}
+      shell: bash
+      run: make download-libiconv
+
+    - name: build libiconv
+      shell: bash
+      run: make build-libiconv
+
+    - name: build mimalloc
+      shell: bash
+      run: make install-mimalloc
+
+    - name: build netsurf
+      shell: bash
+      run: make install-netsurf

.github/workflows/build-deps.yml

@@ -1,60 +0,0 @@
-name: build-deps
-
-on:
-  push:
-    branches:
-      - "main"
-    paths:
-      - "vendor/lexbor-src"
-      - "vendor/netsurf/**"
-  # Allows you to run this workflow manually from the Actions tab
-  workflow_dispatch:
-
-env:
-  REGISTRY: ghcr.io
-  IMAGE_NAME: lightpanda-io/browsercore-deps
-  ZIG_DOCKER_VERSION: 0.12.0-dev.1773-8a8fd47d2
-
-jobs:
-  build-deps:
-    strategy:
-      matrix:
-        include:
-          - os: linux
-            build_arch: amd64
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      packages: write
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          fetch-depth: 0
-          token: ${{ secrets.GH_CI_PAT }}
-          submodules: true
-
-      - name: Docker connect
-        uses: docker/login-action@65b78e6e13532edd9afa3aa52ac7964289d1a9c1
-        with:
-          registry: ${{ env.REGISTRY }}
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@9ec57ed1fcdbf14dcef7dfbe97b2010124a938b7
-        with:
-          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
-
-      - name: Docker build
-        uses: docker/build-push-action@f2a1d5e99d037542a71f64918e516c093c6f3fc4
-        with:
-          context: .
-          file: Dockerfile.deps
-          push: true
-          tags: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:${{env.ZIG_DOCKER_VERSION}}
-          labels: ${{ steps.meta.outputs.labels }}
-          build-args: |
-            ZIG_DOCKER_VERSION=${{ env.ZIG_DOCKER_VERSION }}
-          platforms: ${{matrix.os}}/${{matrix.build_arch}}

.github/workflows/build.yml

@@ -0,0 +1,177 @@
+name: nightly build
+
+env:
+  AWS_ACCESS_KEY_ID: ${{ vars.NIGHTLY_BUILD_AWS_ACCESS_ID }}
+  AWS_SECRET_ACCESS_KEY: ${{ secrets.NIGHTLY_BUILD_AWS_SECRET_ACCESS_KEY }}
+  AWS_BUCKET: ${{ vars.NIGHTLY_BUILD_AWS_BUCKET }}
+  AWS_REGION: ${{ vars.NIGHTLY_BUILD_AWS_REGION }}
+
+on:
+  schedule:
+    - cron: "2 2 * * *"
+
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+permissions:
+  contents: write
+
+jobs:
+  build-linux-x86_64:
+    env:
+      ARCH: x86_64
+      OS: linux
+
+    runs-on: ubuntu-22.04
+    timeout-minutes: 15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+        with:
+          os: ${{env.OS}}
+          arch: ${{env.ARCH}}
+
+      - name: zig build
+        run: zig build --release=safe -Doptimize=ReleaseSafe -Dcpu=x86_64 -Dgit_commit=$(git rev-parse --short ${{ github.sha }})
+
+      - name: Rename binary
+        run: mv zig-out/bin/lightpanda lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: upload on s3
+        run: |
+          export DIR=`git show --no-patch --no-notes --pretty='%cs_%h'`
+          aws s3 cp --storage-class=GLACIER_IR lightpanda-${{ env.ARCH }}-${{ env.OS }} s3://lpd-nightly-build/${DIR}/lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: Upload the build
+        uses: ncipollo/release-action@v1
+        with:
+          allowUpdates: true
+          artifacts: lightpanda-${{ env.ARCH }}-${{ env.OS }}
+          tag: nightly
+
+  build-linux-aarch64:
+    env:
+      ARCH: aarch64
+      OS: linux
+
+    runs-on: ubuntu-22.04-arm
+    timeout-minutes: 15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+        with:
+          os: ${{env.OS}}
+          arch: ${{env.ARCH}}
+
+      - name: zig build
+        run: zig build --release=safe -Doptimize=ReleaseSafe -Dcpu=generic -Dgit_commit=$(git rev-parse --short ${{ github.sha }})
+
+      - name: Rename binary
+        run: mv zig-out/bin/lightpanda lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: upload on s3
+        run: |
+          export DIR=`git show --no-patch --no-notes --pretty='%cs_%h'`
+          aws s3 cp --storage-class=GLACIER_IR lightpanda-${{ env.ARCH }}-${{ env.OS }} s3://lpd-nightly-build/${DIR}/lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: Upload the build
+        uses: ncipollo/release-action@v1
+        with:
+          allowUpdates: true
+          artifacts: lightpanda-${{ env.ARCH }}-${{ env.OS }}
+          tag: nightly
+
+  build-macos-aarch64:
+    env:
+      ARCH: aarch64
+      OS: macos
+
+    # macos-14 runs on arm CPU. see
+    # https://github.com/actions/runner-images?tab=readme-ov-file
+    runs-on: macos-14
+    timeout-minutes: 15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+        with:
+          os: ${{env.OS}}
+          arch: ${{env.ARCH}}
+
+      - name: zig build
+        run: zig build --release=safe -Doptimize=ReleaseSafe -Dgit_commit=$(git rev-parse --short ${{ github.sha }})
+
+      - name: Rename binary
+        run: mv zig-out/bin/lightpanda lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: upload on s3
+        run: |
+          export DIR=`git show --no-patch --no-notes --pretty='%cs_%h'`
+          aws s3 cp --storage-class=GLACIER_IR lightpanda-${{ env.ARCH }}-${{ env.OS }} s3://lpd-nightly-build/${DIR}/lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: Upload the build
+        uses: ncipollo/release-action@v1
+        with:
+          allowUpdates: true
+          artifacts: lightpanda-${{ env.ARCH }}-${{ env.OS }}
+          tag: nightly
+
+  build-macos-x86_64:
+    env:
+      ARCH: x86_64
+      OS: macos
+
+    # macos-13 runs on x86 CPU. see
+    # https://github.com/actions/runner-images?tab=readme-ov-file
+    # If we want to build for macos-14 or superior, we need to switch to
+    # macos-14-large.
+    # No need for now, but maybe we will need it in the short term.
+    runs-on: macos-13
+    timeout-minutes: 15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+        with:
+          os: ${{env.OS}}
+          arch: ${{env.ARCH}}
+
+      - name: zig build
+        run: zig build --release=safe -Doptimize=ReleaseSafe -Dgit_commit=$(git rev-parse --short ${{ github.sha }})
+
+      - name: Rename binary
+        run: mv zig-out/bin/lightpanda lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: upload on s3
+        run: |
+          export DIR=`git show --no-patch --no-notes --pretty='%cs_%h'`
+          aws s3 cp --storage-class=GLACIER_IR lightpanda-${{ env.ARCH }}-${{ env.OS }} s3://lpd-nightly-build/${DIR}/lightpanda-${{ env.ARCH }}-${{ env.OS }}
+
+      - name: Upload the build
+        uses: ncipollo/release-action@v1
+        with:
+          allowUpdates: true
+          artifacts: lightpanda-${{ env.ARCH }}-${{ env.OS }}
+          tag: nightly

.github/workflows/cla.yml

@@ -0,0 +1,34 @@
+name: "CLA Assistant"
+on:
+  issue_comment:
+    types: [created]
+  pull_request_target:
+    types: [opened,closed,synchronize]
+
+permissions:
+  actions: write
+  contents: read
+  pull-requests: write
+  statuses: write
+
+jobs:
+  CLAAssistant:
+    runs-on: ubuntu-latest
+    timeout-minutes: 5
+
+    steps:
+      - name: "CLA Assistant"
+        if: (github.event.comment.body == 'recheck' || github.event.comment.body == 'I have read the CLA Document and I hereby sign the CLA') || github.event_name == 'pull_request_target'
+        uses: contributor-assistant/github-action@v2.6.1
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          PERSONAL_ACCESS_TOKEN: ${{ secrets.CLA_GH_PAT }}
+        with:
+          path-to-signatures: 'signatures/browser/version1/cla.json'
+          path-to-document: 'https://github.com/lightpanda-io/browser/blob/main/CLA.md'
+          # branch should not be protected
+          branch: 'main'
+          allowlist: krichprollsch,francisbouvier,katie-lpd,sjorsdonkers,bornlex
+
+          remote-organization-name: lightpanda-io
+          remote-repository-name: cla

.github/workflows/e2e-test.yml

@@ -0,0 +1,232 @@
+name: e2e-test
+
+env:
+  AWS_ACCESS_KEY_ID: ${{ vars.LPD_PERF_AWS_ACCESS_KEY_ID }}
+  AWS_SECRET_ACCESS_KEY: ${{ secrets.LPD_PERF_AWS_SECRET_ACCESS_KEY }}
+  AWS_BUCKET: ${{ vars.LPD_PERF_AWS_BUCKET }}
+  AWS_REGION: ${{ vars.LPD_PERF_AWS_REGION }}
+  LIGHTPANDA_DISABLE_TELEMETRY: true
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - "build.zig"
+      - "src/**/*.zig"
+      - "src/*.zig"
+      - "vendor/zig-js-runtime"
+      - ".github/**"
+      - "vendor/**"
+  pull_request:
+
+    # By default GH trigger on types opened, synchronize and reopened.
+    # see https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows#pull_request
+    # Since we skip the job when the PR is in draft state, we want to force CI
+    # running when the PR is marked ready_for_review w/o other change.
+    # see https://github.com/orgs/community/discussions/25722#discussioncomment-3248917
+    types: [opened, synchronize, reopened, ready_for_review]
+
+    paths:
+      - ".github/**"
+      - "build.zig"
+      - "src/**/*.zig"
+      - "src/*.zig"
+      - "vendor/**"
+      - ".github/**"
+      - "vendor/**"
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+jobs:
+  zig-build-release:
+    name: zig build release
+
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+
+    # Don't run the CI with draft PR.
+    if: github.event.pull_request.draft == false
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+
+      - name: zig build release
+        run: zig build -Doptimize=ReleaseFast -Dcpu=x86_64 -Dgit_commit=$(git rev-parse --short ${{ github.sha }})
+
+      - name: upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: lightpanda-build-release
+          path: |
+            zig-out/bin/lightpanda
+          retention-days: 1
+
+  demo-scripts:
+    name: demo-scripts
+    needs: zig-build-release
+
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          repository: 'lightpanda-io/demo'
+          fetch-depth: 0
+
+      - run: npm install
+
+      - name: download artifact
+        uses: actions/download-artifact@v4
+        with:
+          name: lightpanda-build-release
+
+      - run: chmod a+x ./lightpanda
+
+      - name: run end to end tests
+        run: |
+          ./lightpanda serve & echo $! > LPD.pid
+          go run runner/main.go
+          kill `cat LPD.pid`
+
+      - name: build proxy
+        run: |
+          cd proxy
+          go build
+
+      - name: run end to end tests through proxy
+        run: |
+          ./proxy/proxy & echo $! > PROXY.id
+          ./lightpanda serve --http_proxy 'http://127.0.0.1:3000' & echo $! > LPD.pid
+          go run runner/main.go
+          kill `cat LPD.pid` `cat PROXY.id`
+
+      - name: run request interception through proxy
+        run: |
+          export PROXY_USERNAME=username PROXY_PASSWORD=password
+          ./proxy/proxy & echo $! > PROXY.id
+          ./lightpanda serve & echo $! > LPD.pid
+          URL=https://demo-browser.lightpanda.io/campfire-commerce/ node puppeteer/proxy_auth.js
+          BASE_URL=https://demo-browser.lightpanda.io/ node playwright/proxy_auth.js
+          kill `cat LPD.pid` `cat PROXY.id`
+
+  cdp-and-hyperfine-bench:
+    name: cdp-and-hyperfine-bench
+    needs: zig-build-release
+
+    env:
+      MAX_MEMORY: 27000
+      MAX_AVG_DURATION: 23
+      LIGHTPANDA_DISABLE_TELEMETRY: true
+
+    # use a self host runner.
+    runs-on: lpd-bench-hetzner
+    timeout-minutes: 15
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          repository: 'lightpanda-io/demo'
+          fetch-depth: 0
+
+      - run: npm install
+
+      - name: download artifact
+        uses: actions/download-artifact@v4
+        with:
+          name: lightpanda-build-release
+
+      - run: chmod a+x ./lightpanda
+
+      - name: start http
+        run: |
+          go run ws/main.go & echo $! > WS.pid
+          sleep 2
+
+      - name: run puppeteer
+        run: |
+          ./lightpanda serve & echo $! > LPD.pid
+          sleep 2
+          RUNS=100 npm run bench-puppeteer-cdp > puppeteer.out || exit 1
+          cat /proc/`cat LPD.pid`/status |grep VmHWM|grep -oP '\d+' > LPD.VmHWM
+          kill `cat LPD.pid`
+
+      - name: puppeteer result
+        run: cat puppeteer.out
+
+      - name: memory regression
+        run: |
+          export LPD_VmHWM=`cat LPD.VmHWM`
+          echo "Peak resident set size: $LPD_VmHWM"
+          test "$LPD_VmHWM" -le "$MAX_MEMORY"
+
+      - name: duration regression
+        run: |
+          export PUPPETEER_AVG_DURATION=`cat puppeteer.out|grep 'avg run'|sed 's/avg run duration (ms) //'`
+          echo "puppeteer avg duration: $PUPPETEER_AVG_DURATION"
+          test "$PUPPETEER_AVG_DURATION" -le "$MAX_AVG_DURATION"
+
+      - name: json output
+        run: |
+          export AVG_DURATION=`cat puppeteer.out|grep 'avg run'|sed 's/avg run duration (ms) //'`
+          export TOTAL_DURATION=`cat puppeteer.out|grep 'total duration'|sed 's/total duration (ms) //'`
+          export LPD_VmHWM=`cat LPD.VmHWM`
+          echo "{\"duration_total\":${TOTAL_DURATION},\"duration_avg\":${AVG_DURATION},\"mem_peak\":${LPD_VmHWM}}" > bench.json
+          cat bench.json
+
+      - name: run hyperfine
+        run: |
+          hyperfine --export-json=hyperfine.json --warmup 3 --runs 20 --shell=none "./lightpanda --dump http://127.0.0.1:1234/campfire-commerce/"
+
+      - name: stop http
+        run: kill `cat WS.pid`
+
+      - name: write commit
+        run: |
+          echo "${{github.sha}}" > commit.txt
+
+      - name: upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: bench-results
+          path: |
+            bench.json
+            hyperfine.json
+            commit.txt
+          retention-days: 10
+
+
+  perf-fmt:
+    name: perf-fmt
+    needs: cdp-and-hyperfine-bench
+
+    # Don't execute on PR
+    if: github.event_name != 'pull_request'
+
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+
+    container:
+      image: ghcr.io/lightpanda-io/perf-fmt:latest
+      credentials:
+       username: ${{ github.actor }}
+       password: ${{ secrets.GITHUB_TOKEN }}
+
+    steps:
+      - name: download artifact
+        uses: actions/download-artifact@v4
+        with:
+          name: bench-results
+
+      - name: format and send json result
+        run: /perf-fmt cdp ${{ github.sha }} bench.json
+
+      - name: format and send json result
+        run: /perf-fmt hyperfine ${{ github.sha }} hyperfine.json

.github/workflows/wpt.yml

@@ -1,94 +1,44 @@
 name: wpt
 
 env:
-  ARCH: x86_64-linux
   AWS_ACCESS_KEY_ID: ${{ vars.LPD_PERF_AWS_ACCESS_KEY_ID }}
   AWS_SECRET_ACCESS_KEY: ${{ secrets.LPD_PERF_AWS_SECRET_ACCESS_KEY }}
   AWS_BUCKET: ${{ vars.LPD_PERF_AWS_BUCKET }}
   AWS_REGION: ${{ vars.LPD_PERF_AWS_REGION }}
+  LIGHTPANDA_DISABLE_TELEMETRY: true
 
 on:
-  push:
-    branches:
-      - main
-    paths:
-      - "src/**/*.zig"
-      - "src/*.zig"
-      - "tests/wpt/**"
-      - "vendor/**"
-  pull_request:
+  schedule:
+    - cron: "23 2 * * *"
 
-    # By default GH trigger on types opened, synchronize and reopened.
-    # see https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows#pull_request
-    # Since we skip the job when the PR is in draft state, we want to force CI
-    # running when the PR is marked ready_for_review w/o other change.
-    # see https://github.com/orgs/community/discussions/25722#discussioncomment-3248917
-    types: [opened, synchronize, reopened, ready_for_review]
-
-    paths:
-      - "src/**/*.zig"
-      - "src/*.zig"
-      - "tests/wpt/**"
-      - "vendor/**"
   # Allows you to run this workflow manually from the Actions tab
   workflow_dispatch:
 
 jobs:
   wpt:
-    name: web platform tests
-
-    # Don't run the CI with draft PR.
-    if: github.event.pull_request.draft == false
+    name: web platform tests json output
 
     runs-on: ubuntu-latest
-    container:
-      image: ghcr.io/lightpanda-io/zig-browsercore:0.12.0-dev.1773-8a8fd47d2
-      credentials:
-       username: ${{ github.actor }}
-       password: ${{ secrets.GITHUB_TOKEN }}
+    timeout-minutes: 90
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
-          token: ${{ secrets.GH_CI_PAT }}
-          # fetch submodules recusively, to get jsruntime-lib submodules also.
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
           submodules: recursive
 
-      - name: install v8
-        run: |
-          mkdir -p vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/debug
-          ln -s /usr/local/lib/libc_v8.a vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/debug/libc_v8.a
-
-          mkdir -p vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/release
-          ln -s /usr/local/lib/libc_v8.a vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/release/libc_v8.a
-
-      - name: install deps
-        run: |
-          ln -s /usr/local/lib/lexbor vendor/lexbor
-
-          ln -s /usr/local/lib/libiconv vendor/libiconv
-
-          ln -s /usr/local/lib/netsurf/build vendor/netsurf/build
-          ln -s /usr/local/lib/netsurf/lib vendor/netsurf/lib
-          ln -s /usr/local/lib/netsurf/include vendor/netsurf/include
-
-      - run: zig build wpt -Dengine=v8 -- --safe --summary
-
-        # For now WPT tests doesn't pass at all.
-        # We accept then to continue the job on failure.
-        # TODO remove the continue-on-error when tests will pass.
-        continue-on-error: true
+      - uses: ./.github/actions/install
 
       - name: json output
-        run: zig build wpt -Dengine=v8 -- --safe --json > wpt.json
+        run: zig build -Doptimize=ReleaseFast wpt -- --json > wpt.json
 
       - name: write commit
         run: |
           echo "${{github.sha}}" > commit.txt
 
       - name: upload artifact
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: wpt-results
           path: |
@@ -100,10 +50,9 @@ jobs:
     name: perf-fmt
     needs: wpt
 
-    # Don't execute on PR
-    if: github.event_name != 'pull_request'
-
     runs-on: ubuntu-latest
+    timeout-minutes: 15
+
     container:
       image: ghcr.io/lightpanda-io/perf-fmt:latest
       credentials:
@@ -112,7 +61,7 @@ jobs:
 
     steps:
       - name: download artifact
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
         with:
           name: wpt-results
 

.github/workflows/zig-fmt.yml

@@ -1,5 +1,8 @@
 name: zig-fmt
 
+env:
+  ZIG_VERSION: 0.15.1
+
 on:
   pull_request:
 
@@ -11,6 +14,8 @@ on:
     types: [opened, synchronize, reopened, ready_for_review]
 
     paths:
+      - ".github/**"
+      - "build.zig"
       - "src/**/*.zig"
       - "src/*.zig"
   # Allows you to run this workflow manually from the Actions tab
@@ -24,16 +29,14 @@ jobs:
     if: github.event.pull_request.draft == false
 
     runs-on: ubuntu-latest
-    container:
-      image: ghcr.io/lightpanda-io/zig:0.12.0-dev.1773-8a8fd47d2
-      credentials:
-       username: ${{ github.actor }}
-       password: ${{ secrets.GITHUB_TOKEN }}
-    outputs:
-      zig_fmt_errs: ${{ steps.fmt.outputs.zig_fmt_errs }}
+    timeout-minutes: 15
 
     steps:
-      - uses: actions/checkout@v3
+      - uses: mlugg/setup-zig@v2
+        with:
+          version: ${{ env.ZIG_VERSION }}
+
+      - uses: actions/checkout@v4
         with:
           fetch-depth: 0
 

.github/workflows/zig-test.yml

@@ -1,16 +1,22 @@
 name: zig-test
 
 env:
-  ARCH: x86_64-linux
+  AWS_ACCESS_KEY_ID: ${{ vars.LPD_PERF_AWS_ACCESS_KEY_ID }}
+  AWS_SECRET_ACCESS_KEY: ${{ secrets.LPD_PERF_AWS_SECRET_ACCESS_KEY }}
+  AWS_BUCKET: ${{ vars.LPD_PERF_AWS_BUCKET }}
+  AWS_REGION: ${{ vars.LPD_PERF_AWS_REGION }}
 
 on:
   push:
     branches:
       - main
     paths:
+      - "build.zig"
       - "src/**/*.zig"
       - "src/*.zig"
-      - "vendor/jsruntime-lib"
+      - "vendor/zig-js-runtime"
+      - ".github/**"
+      - "vendor/**"
   pull_request:
 
     # By default GH trigger on types opened, synchronize and reopened.
@@ -21,57 +27,119 @@ on:
     types: [opened, synchronize, reopened, ready_for_review]
 
     paths:
+      - ".github/**"
+      - "build.zig"
       - "src/**/*.zig"
       - "src/*.zig"
       - "vendor/**"
+      - ".github/**"
+      - "vendor/**"
   # Allows you to run this workflow manually from the Actions tab
   workflow_dispatch:
 
 jobs:
-  zig-test:
-    name: zig test
+  zig-build-dev:
+    name: zig build dev
 
     # Don't run the CI with draft PR.
     if: github.event.pull_request.draft == false
 
     runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+
+      - name: zig build debug
+        run: zig build
+
+      - name: upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: lightpanda-build-dev
+          path: |
+            zig-out/bin/lightpanda
+          retention-days: 1
+
+  browser-fetch:
+    name: browser fetch
+    needs: zig-build-dev
+
+    # Don't run the CI with draft PR.
+    if: github.event.pull_request.draft == false
+
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: download artifact
+        uses: actions/download-artifact@v4
+        with:
+          name: lightpanda-build-dev
+
+      - run: chmod a+x ./lightpanda
+
+      - run: ./lightpanda fetch https://httpbin.io/xhr/get
+
+  zig-test:
+    name: zig test
+    timeout-minutes: 15
+
+    # Don't run the CI with draft PR.
+    if: github.event.pull_request.draft == false
+
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+          # fetch submodules recusively, to get zig-js-runtime submodules also.
+          submodules: recursive
+
+      - uses: ./.github/actions/install
+
+      - name: zig build test
+        run: zig build test -- --json > bench.json
+
+      - name: write commit
+        run: |
+          echo "${{github.sha}}" > commit.txt
+
+      - name: upload artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: bench-results
+          path: |
+            bench.json
+            commit.txt
+          retention-days: 10
+
+  bench-fmt:
+    name: perf-fmt
+    needs: zig-test
+
+    # Don't execute on PR
+    if: github.event_name != 'pull_request'
+
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+
     container:
-      image: ghcr.io/lightpanda-io/zig-browsercore:0.12.0-dev.1773-8a8fd47d2
+      image: ghcr.io/lightpanda-io/perf-fmt:latest
       credentials:
        username: ${{ github.actor }}
        password: ${{ secrets.GITHUB_TOKEN }}
 
     steps:
-      - uses: actions/checkout@v3
+      - name: download artifact
+        uses: actions/download-artifact@v4
         with:
-          fetch-depth: 0
-          token: ${{ secrets.GH_CI_PAT }}
-          # fetch submodules recusively, to get jsruntime-lib submodules also.
-          submodules: recursive
+          name: bench-results
 
-      - name: install v8
-        run: |
-          mkdir -p vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/debug
-          ln -s /usr/local/lib/libc_v8.a vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/debug/libc_v8.a
-
-          mkdir -p vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/release
-          ln -s /usr/local/lib/libc_v8.a vendor/jsruntime-lib/vendor/v8/${{env.ARCH}}/release/libc_v8.a
-
-      - name: install deps
-        run: |
-          ln -s /usr/local/lib/lexbor vendor/lexbor
-
-          ln -s /usr/local/lib/libiconv vendor/libiconv
-
-          ln -s /usr/local/lib/netsurf/build vendor/netsurf/build
-          ln -s /usr/local/lib/netsurf/lib vendor/netsurf/lib
-          ln -s /usr/local/lib/netsurf/include vendor/netsurf/include
-
-      - name: zig build debug
-        run: zig build -Dengine=v8
-
-      - name: zig build test
-        run: zig build test -Dengine=v8
-
-      - name: zig build release
-        run: zig build -Doptimize=ReleaseSafe -Dengine=v8
+      - name: format and send json result
+        run: /perf-fmt bench-browser ${{ github.sha }} bench.json

.gitignore

@@ -1,7 +1,7 @@
 zig-cache
+/.zig-cache/
 zig-out
-/vendor/lexbor/
-/vendor/netsurf/build/
-/vendor/netsurf/lib/
-/vendor/netsurf/include/
+/vendor/netsurf/out
 /vendor/libiconv/
+lightpanda.id
+/v8/

.gitmodules

@@ -1,24 +1,33 @@
-[submodule "vendor/jsruntime-lib"]
-	path = vendor/jsruntime-lib
-	url = git@github.com:lightpanda-io/jsruntime-lib.git
-[submodule "vendor/lexbor-src"]
-	path = vendor/lexbor-src
-	url = https://github.com/lexbor/lexbor
 [submodule "vendor/netsurf/libwapcaplet"]
 	path = vendor/netsurf/libwapcaplet
-	url = https://source.netsurf-browser.org/libwapcaplet.git
+	url = https://github.com/lightpanda-io/libwapcaplet.git/
 [submodule "vendor/netsurf/libparserutils"]
 	path = vendor/netsurf/libparserutils
-	url = https://source.netsurf-browser.org/libparserutils.git
+	url = https://github.com/lightpanda-io/libparserutils.git/
 [submodule "vendor/netsurf/libdom"]
 	path = vendor/netsurf/libdom
-	url = git@github.com:lightpanda-io/libdom.git
+	url = https://github.com/lightpanda-io/libdom.git/
 [submodule "vendor/netsurf/share/netsurf-buildsystem"]
 	path = vendor/netsurf/share/netsurf-buildsystem
-	url = https://source.netsurf-browser.org/buildsystem.git
+	url = https://github.com/lightpanda-io/netsurf-buildsystem.git
 [submodule "vendor/netsurf/libhubbub"]
 	path = vendor/netsurf/libhubbub
-	url = https://source.netsurf-browser.org/libhubbub.git
+	url = https://github.com/lightpanda-io/libhubbub.git/
 [submodule "tests/wpt"]
 	path = tests/wpt
 	url = https://github.com/lightpanda-io/wpt
+[submodule "vendor/mimalloc"]
+	path = vendor/mimalloc
+	url = https://github.com/microsoft/mimalloc.git/
+[submodule "vendor/nghttp2"]
+	path = vendor/nghttp2
+	url = https://github.com/nghttp2/nghttp2.git
+[submodule "vendor/mbedtls"]
+	path = vendor/mbedtls
+	url = https://github.com/Mbed-TLS/mbedtls.git
+[submodule "vendor/zlib"]
+	path = vendor/zlib
+	url = https://github.com/madler/zlib.git
+[submodule "vendor/curl"]
+	path = vendor/curl
+	url = https://github.com/curl/curl.git

CLA.md

@@ -0,0 +1,93 @@
+# Lightpanda (Selecy SAS) Grant and Contributor License Agreement (“Agreement”)
+
+This agreement is based on the Apache Software Foundation Contributor License
+Agreement. (v r190612)
+
+Thank you for your interest in software projects stewarded by Lightpanda
+(Selecy SAS) (“Lightpanda”). In order to clarify the intellectual property
+license granted with Contributions from any person or entity, Lightpanda must
+have a Contributor License Agreement (CLA) on file that has been agreed to by
+each Contributor, indicating agreement to the license terms below. This license
+is for your protection as a Contributor as well as the protection of Lightpanda
+and its users; it does not change your rights to use your own Contributions for
+any other purpose. This Agreement allows an individual to contribute to
+Lightpanda on that individual’s own behalf, or an entity (the “Corporation”) to
+submit Contributions to Lightpanda, to authorize Contributions submitted by its
+designated employees to Lightpanda, and to grant copyright and patent licenses
+thereto.
+
+You accept and agree to the following terms and conditions for Your present and
+future Contributions submitted to Lightpanda. Except for the license granted
+herein to Lightpanda and recipients of software distributed by Lightpanda, You
+reserve all right, title, and interest in and to Your Contributions.
+
+1. Definitions. “You” (or “Your”) shall mean the copyright owner or legal
+   entity authorized by the copyright owner that is making this Agreement with
+   Lightpanda. For legal entities, the entity making a Contribution and all
+   other entities that control, are controlled by, or are under common control
+   with that entity are considered to be a single Contributor. For the purposes
+   of this definition, “control” means (i) the power, direct or indirect, to
+   cause the direction or management of such entity, whether by contract or
+   otherwise, or (ii) ownership of fifty percent (50%) or more of the
+   outstanding shares, or (iii) beneficial ownership of such entity.
+   “Contribution” shall mean any work, as well as any modifications or
+   additions to an existing work, that is intentionally submitted by You to
+   Lightpanda for inclusion in, or documentation of, any of the products owned
+   or managed by Lightpanda (the “Work”). For the purposes of this definition,
+   “submitted” means any form of electronic, verbal, or written communication
+   sent to Lightpanda or its representatives, including but not limited to
+   communication on electronic mailing lists, source code control systems (such
+   as GitHub), and issue tracking systems that are managed by, or on behalf of,
+   Lightpanda for the purpose of discussing and improving the Work, but
+   excluding communication that is conspicuously marked or otherwise designated
+   in writing by You as “Not a Contribution.”
+
+2. Grant of Copyright License. Subject to the terms and conditions of this
+   Agreement, You hereby grant to Lightpanda and to recipients of software
+   distributed by Lightpanda a perpetual, worldwide, non-exclusive, no-charge,
+   royalty-free, irrevocable copyright license to reproduce, prepare derivative
+   works of, publicly display, publicly perform, sublicense, and distribute
+   Your Contributions and such derivative works.
+
+3. Grant of Patent License. Subject to the terms and conditions of this
+   Agreement, You hereby grant to Lightpanda and to recipients of software
+   distributed by Lightpanda a perpetual, worldwide, non-exclusive, no-charge,
+   royalty-free, irrevocable (except as stated in this section) patent license
+   to make, have made, use, offer to sell, sell, import, and otherwise transfer
+   the Work, where such license applies only to those patent claims licensable
+   by You that are necessarily infringed by Your Contribution(s) alone or by
+   combination of Your Contribution(s) with the Work to which such
+   Contribution(s) were submitted. If any entity institutes patent litigation
+   against You or any other entity (including a cross-claim or counterclaim in
+   a lawsuit) alleging that your Contribution, or the Work to which you have
+   contributed, constitutes direct or contributory patent infringement, then
+   any patent licenses granted to that entity under this Agreement for that
+   Contribution or Work shall terminate as of the date such litigation is
+   filed.
+
+4. You represent that You are legally entitled to grant the above license. If
+   You are an individual, and if Your employer(s) has rights to intellectual
+   property that you create that includes Your Contributions, you represent
+   that You have received permission to make Contributions on behalf of that
+   employer, or that Your employer has waived such rights for your
+   Contributions to Lightpanda. If You are a Corporation, any individual who
+   makes a contribution from an account associated with You will be considered
+   authorized to Contribute on Your behalf.
+
+5. You represent that each of Your Contributions is Your original creation (see
+   section 7 for submissions on behalf of others).
+
+6. You are not expected to provide support for Your Contributions,except to the
+   extent You desire to provide support. You may provide support for free, for
+   a fee, or not at all. Unless required by applicable law or agreed to in
+   writing, You provide Your Contributions on an “AS IS” BASIS, WITHOUT
+   WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including,
+   without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT,
+   MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE.
+
+7. Should You wish to submit work that is not Your original creation, You may
+   submit it to Lightpanda separately from any Contribution, identifying the
+   complete details of its source and of any license or other restriction
+   (including, but not limited to, related patents, trademarks, and license
+   agreements) of which you are personally aware, and conspicuously marking the
+   work as “Submitted on behalf of a third-party: [named here]”.

CONTRIBUTING.md

@@ -0,0 +1,10 @@
+# Contributing
+
+Lightpanda accepts pull requests through GitHub.
+
+You have to sign our [CLA](CLA.md) during your first pull request process
+otherwise we're not able to accept your contributions.
+
+The process signature uses the [CLA assistant
+lite](https://github.com/marketplace/actions/cla-assistant-lite). You can see
+an example of the process in [#303](https://github.com/lightpanda-io/browser/pull/303).

Dockerfile

@@ -0,0 +1,68 @@
+FROM debian:stable
+
+ARG MINISIG=0.12
+ARG ZIG=0.15.1
+ARG ZIG_MINISIG=RWSGOq2NVecA2UPNdBUZykf1CCb147pkmdtYxgb3Ti+JO/wCYvhbAb/U
+ARG V8=14.0.365.4
+ARG ZIG_V8=v0.1.30
+ARG TARGETPLATFORM
+
+RUN apt-get update -yq && \
+    apt-get install -yq xz-utils \
+        python3 ca-certificates git \
+        pkg-config libglib2.0-dev \
+        gperf libexpat1-dev \
+        cmake clang \
+        curl git
+
+# install minisig
+RUN curl --fail -L -O https://github.com/jedisct1/minisign/releases/download/${MINISIG}/minisign-${MINISIG}-linux.tar.gz && \
+    tar xvzf minisign-${MINISIG}-linux.tar.gz
+
+# install zig
+RUN case $TARGETPLATFORM in \
+    "linux/arm64") ARCH="aarch64" ;; \
+    *) ARCH="x86_64" ;; \
+    esac && \
+    curl --fail -L -O https://ziglang.org/download/${ZIG}/zig-${ARCH}-linux-${ZIG}.tar.xz && \
+    curl --fail -L -O https://ziglang.org/download/${ZIG}/zig-${ARCH}-linux-${ZIG}.tar.xz.minisig && \
+    minisign-linux/${ARCH}/minisign -Vm zig-${ARCH}-linux-${ZIG}.tar.xz -P ${ZIG_MINISIG} && \
+    tar xvf zig-${ARCH}-linux-${ZIG}.tar.xz && \
+    mv zig-${ARCH}-linux-${ZIG} /usr/local/lib && \
+    ln -s /usr/local/lib/zig-${ARCH}-linux-${ZIG}/zig /usr/local/bin/zig
+
+# clone lightpanda
+RUN git clone https://github.com/lightpanda-io/browser.git
+
+WORKDIR /browser
+
+# install deps
+RUN git submodule init && \
+    git submodule update --recursive
+
+RUN make install-libiconv && \
+    make install-netsurf && \
+    make install-mimalloc
+
+# download and install v8
+RUN case $TARGETPLATFORM in \
+    "linux/arm64") ARCH="aarch64" ;; \
+    *) ARCH="x86_64" ;; \
+    esac && \
+    curl --fail -L -o libc_v8.a https://github.com/lightpanda-io/zig-v8-fork/releases/download/${ZIG_V8}/libc_v8_${V8}_linux_${ARCH}.a && \
+    mkdir -p v8/out/linux/release/obj/zig/ && \
+    mv libc_v8.a v8/out/linux/release/obj/zig/libc_v8.a
+
+# build release
+RUN make build
+
+FROM debian:stable-slim
+
+# copy ca certificates
+COPY --from=0 /etc/ssl/certs/ca-certificates.crt /etc/ssl/certs/ca-certificates.crt
+
+COPY --from=0 /browser/zig-out/bin/lightpanda /bin/lightpanda
+
+EXPOSE 9222/tcp
+
+CMD ["/bin/lightpanda", "serve", "--host", "0.0.0.0", "--port", "9222"]

Dockerfile.deps

@@ -1,34 +0,0 @@
-# This dockerfile is used to build browsercore vendor dependencies except
-# jsruntime-lib v8.
-# jsruntime-lib v8 is built via zig-v8-fork/Dockerfile.
-ARG ZIG_DOCKER_VERSION=0.11.0
-FROM ghcr.io/lightpanda-io/zig:${ZIG_DOCKER_VERSION} as build
-
-# Install required dependencies
-RUN apt update && \
-    apt install -y git curl bash xz-utils python3 ca-certificates pkg-config \
-    libglib2.0-dev gperf libexpat1-dev cmake build-essential
-
-COPY ./Makefile /src/
-WORKDIR /src
-
-# build lexbor
-ADD ./vendor/lexbor-src /src/vendor/lexbor-src
-RUN make install-lexbor
-
-# build libiconv
-RUN make install-libiconv
-
-# build netsurf
-ADD ./vendor/netsurf /src/vendor/netsurf
-RUN make install-netsurf
-
-FROM scratch as artifact
-
-COPY --from=build /src/vendor/libiconv /usr/local/lib/libiconv
-
-COPY --from=build /src/vendor/lexbor /usr/local/lib/lexbor
-
-COPY --from=build /src/vendor/netsurf/build /usr/local/lib/netsurf/build
-COPY --from=build /src/vendor/netsurf/lib /usr/local/lib/netsurf/lib
-COPY --from=build /src/vendor/netsurf/include /usr/local/lib/netsurf/include

LICENSE

@@ -0,0 +1,661 @@
+                    GNU AFFERO GENERAL PUBLIC LICENSE
+                       Version 3, 19 November 2007
+
+ Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
+ Everyone is permitted to copy and distribute verbatim copies
+ of this license document, but changing it is not allowed.
+
+                            Preamble
+
+  The GNU Affero General Public License is a free, copyleft license for
+software and other kinds of works, specifically designed to ensure
+cooperation with the community in the case of network server software.
+
+  The licenses for most software and other practical works are designed
+to take away your freedom to share and change the works.  By contrast,
+our General Public Licenses are intended to guarantee your freedom to
+share and change all versions of a program--to make sure it remains free
+software for all its users.
+
+  When we speak of free software, we are referring to freedom, not
+price.  Our General Public Licenses are designed to make sure that you
+have the freedom to distribute copies of free software (and charge for
+them if you wish), that you receive source code or can get it if you
+want it, that you can change the software or use pieces of it in new
+free programs, and that you know you can do these things.
+
+  Developers that use our General Public Licenses protect your rights
+with two steps: (1) assert copyright on the software, and (2) offer
+you this License which gives you legal permission to copy, distribute
+and/or modify the software.
+
+  A secondary benefit of defending all users' freedom is that
+improvements made in alternate versions of the program, if they
+receive widespread use, become available for other developers to
+incorporate.  Many developers of free software are heartened and
+encouraged by the resulting cooperation.  However, in the case of
+software used on network servers, this result may fail to come about.
+The GNU General Public License permits making a modified version and
+letting the public access it on a server without ever releasing its
+source code to the public.
+
+  The GNU Affero General Public License is designed specifically to
+ensure that, in such cases, the modified source code becomes available
+to the community.  It requires the operator of a network server to
+provide the source code of the modified version running there to the
+users of that server.  Therefore, public use of a modified version, on
+a publicly accessible server, gives the public access to the source
+code of the modified version.
+
+  An older license, called the Affero General Public License and
+published by Affero, was designed to accomplish similar goals.  This is
+a different license, not a version of the Affero GPL, but Affero has
+released a new version of the Affero GPL which permits relicensing under
+this license.
+
+  The precise terms and conditions for copying, distribution and
+modification follow.
+
+                       TERMS AND CONDITIONS
+
+  0. Definitions.
+
+  "This License" refers to version 3 of the GNU Affero General Public License.
+
+  "Copyright" also means copyright-like laws that apply to other kinds of
+works, such as semiconductor masks.
+
+  "The Program" refers to any copyrightable work licensed under this
+License.  Each licensee is addressed as "you".  "Licensees" and
+"recipients" may be individuals or organizations.
+
+  To "modify" a work means to copy from or adapt all or part of the work
+in a fashion requiring copyright permission, other than the making of an
+exact copy.  The resulting work is called a "modified version" of the
+earlier work or a work "based on" the earlier work.
+
+  A "covered work" means either the unmodified Program or a work based
+on the Program.
+
+  To "propagate" a work means to do anything with it that, without
+permission, would make you directly or secondarily liable for
+infringement under applicable copyright law, except executing it on a
+computer or modifying a private copy.  Propagation includes copying,
+distribution (with or without modification), making available to the
+public, and in some countries other activities as well.
+
+  To "convey" a work means any kind of propagation that enables other
+parties to make or receive copies.  Mere interaction with a user through
+a computer network, with no transfer of a copy, is not conveying.
+
+  An interactive user interface displays "Appropriate Legal Notices"
+to the extent that it includes a convenient and prominently visible
+feature that (1) displays an appropriate copyright notice, and (2)
+tells the user that there is no warranty for the work (except to the
+extent that warranties are provided), that licensees may convey the
+work under this License, and how to view a copy of this License.  If
+the interface presents a list of user commands or options, such as a
+menu, a prominent item in the list meets this criterion.
+
+  1. Source Code.
+
+  The "source code" for a work means the preferred form of the work
+for making modifications to it.  "Object code" means any non-source
+form of a work.
+
+  A "Standard Interface" means an interface that either is an official
+standard defined by a recognized standards body, or, in the case of
+interfaces specified for a particular programming language, one that
+is widely used among developers working in that language.
+
+  The "System Libraries" of an executable work include anything, other
+than the work as a whole, that (a) is included in the normal form of
+packaging a Major Component, but which is not part of that Major
+Component, and (b) serves only to enable use of the work with that
+Major Component, or to implement a Standard Interface for which an
+implementation is available to the public in source code form.  A
+"Major Component", in this context, means a major essential component
+(kernel, window system, and so on) of the specific operating system
+(if any) on which the executable work runs, or a compiler used to
+produce the work, or an object code interpreter used to run it.
+
+  The "Corresponding Source" for a work in object code form means all
+the source code needed to generate, install, and (for an executable
+work) run the object code and to modify the work, including scripts to
+control those activities.  However, it does not include the work's
+System Libraries, or general-purpose tools or generally available free
+programs which are used unmodified in performing those activities but
+which are not part of the work.  For example, Corresponding Source
+includes interface definition files associated with source files for
+the work, and the source code for shared libraries and dynamically
+linked subprograms that the work is specifically designed to require,
+such as by intimate data communication or control flow between those
+subprograms and other parts of the work.
+
+  The Corresponding Source need not include anything that users
+can regenerate automatically from other parts of the Corresponding
+Source.
+
+  The Corresponding Source for a work in source code form is that
+same work.
+
+  2. Basic Permissions.
+
+  All rights granted under this License are granted for the term of
+copyright on the Program, and are irrevocable provided the stated
+conditions are met.  This License explicitly affirms your unlimited
+permission to run the unmodified Program.  The output from running a
+covered work is covered by this License only if the output, given its
+content, constitutes a covered work.  This License acknowledges your
+rights of fair use or other equivalent, as provided by copyright law.
+
+  You may make, run and propagate covered works that you do not
+convey, without conditions so long as your license otherwise remains
+in force.  You may convey covered works to others for the sole purpose
+of having them make modifications exclusively for you, or provide you
+with facilities for running those works, provided that you comply with
+the terms of this License in conveying all material for which you do
+not control copyright.  Those thus making or running the covered works
+for you must do so exclusively on your behalf, under your direction
+and control, on terms that prohibit them from making any copies of
+your copyrighted material outside their relationship with you.
+
+  Conveying under any other circumstances is permitted solely under
+the conditions stated below.  Sublicensing is not allowed; section 10
+makes it unnecessary.
+
+  3. Protecting Users' Legal Rights From Anti-Circumvention Law.
+
+  No covered work shall be deemed part of an effective technological
+measure under any applicable law fulfilling obligations under article
+11 of the WIPO copyright treaty adopted on 20 December 1996, or
+similar laws prohibiting or restricting circumvention of such
+measures.
+
+  When you convey a covered work, you waive any legal power to forbid
+circumvention of technological measures to the extent such circumvention
+is effected by exercising rights under this License with respect to
+the covered work, and you disclaim any intention to limit operation or
+modification of the work as a means of enforcing, against the work's
+users, your or third parties' legal rights to forbid circumvention of
+technological measures.
+
+  4. Conveying Verbatim Copies.
+
+  You may convey verbatim copies of the Program's source code as you
+receive it, in any medium, provided that you conspicuously and
+appropriately publish on each copy an appropriate copyright notice;
+keep intact all notices stating that this License and any
+non-permissive terms added in accord with section 7 apply to the code;
+keep intact all notices of the absence of any warranty; and give all
+recipients a copy of this License along with the Program.
+
+  You may charge any price or no price for each copy that you convey,
+and you may offer support or warranty protection for a fee.
+
+  5. Conveying Modified Source Versions.
+
+  You may convey a work based on the Program, or the modifications to
+produce it from the Program, in the form of source code under the
+terms of section 4, provided that you also meet all of these conditions:
+
+    a) The work must carry prominent notices stating that you modified
+    it, and giving a relevant date.
+
+    b) The work must carry prominent notices stating that it is
+    released under this License and any conditions added under section
+    7.  This requirement modifies the requirement in section 4 to
+    "keep intact all notices".
+
+    c) You must license the entire work, as a whole, under this
+    License to anyone who comes into possession of a copy.  This
+    License will therefore apply, along with any applicable section 7
+    additional terms, to the whole of the work, and all its parts,
+    regardless of how they are packaged.  This License gives no
+    permission to license the work in any other way, but it does not
+    invalidate such permission if you have separately received it.
+
+    d) If the work has interactive user interfaces, each must display
+    Appropriate Legal Notices; however, if the Program has interactive
+    interfaces that do not display Appropriate Legal Notices, your
+    work need not make them do so.
+
+  A compilation of a covered work with other separate and independent
+works, which are not by their nature extensions of the covered work,
+and which are not combined with it such as to form a larger program,
+in or on a volume of a storage or distribution medium, is called an
+"aggregate" if the compilation and its resulting copyright are not
+used to limit the access or legal rights of the compilation's users
+beyond what the individual works permit.  Inclusion of a covered work
+in an aggregate does not cause this License to apply to the other
+parts of the aggregate.
+
+  6. Conveying Non-Source Forms.
+
+  You may convey a covered work in object code form under the terms
+of sections 4 and 5, provided that you also convey the
+machine-readable Corresponding Source under the terms of this License,
+in one of these ways:
+
+    a) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by the
+    Corresponding Source fixed on a durable physical medium
+    customarily used for software interchange.
+
+    b) Convey the object code in, or embodied in, a physical product
+    (including a physical distribution medium), accompanied by a
+    written offer, valid for at least three years and valid for as
+    long as you offer spare parts or customer support for that product
+    model, to give anyone who possesses the object code either (1) a
+    copy of the Corresponding Source for all the software in the
+    product that is covered by this License, on a durable physical
+    medium customarily used for software interchange, for a price no
+    more than your reasonable cost of physically performing this
+    conveying of source, or (2) access to copy the
+    Corresponding Source from a network server at no charge.
+
+    c) Convey individual copies of the object code with a copy of the
+    written offer to provide the Corresponding Source.  This
+    alternative is allowed only occasionally and noncommercially, and
+    only if you received the object code with such an offer, in accord
+    with subsection 6b.
+
+    d) Convey the object code by offering access from a designated
+    place (gratis or for a charge), and offer equivalent access to the
+    Corresponding Source in the same way through the same place at no
+    further charge.  You need not require recipients to copy the
+    Corresponding Source along with the object code.  If the place to
+    copy the object code is a network server, the Corresponding Source
+    may be on a different server (operated by you or a third party)
+    that supports equivalent copying facilities, provided you maintain
+    clear directions next to the object code saying where to find the
+    Corresponding Source.  Regardless of what server hosts the
+    Corresponding Source, you remain obligated to ensure that it is
+    available for as long as needed to satisfy these requirements.
+
+    e) Convey the object code using peer-to-peer transmission, provided
+    you inform other peers where the object code and Corresponding
+    Source of the work are being offered to the general public at no
+    charge under subsection 6d.
+
+  A separable portion of the object code, whose source code is excluded
+from the Corresponding Source as a System Library, need not be
+included in conveying the object code work.
+
+  A "User Product" is either (1) a "consumer product", which means any
+tangible personal property which is normally used for personal, family,
+or household purposes, or (2) anything designed or sold for incorporation
+into a dwelling.  In determining whether a product is a consumer product,
+doubtful cases shall be resolved in favor of coverage.  For a particular
+product received by a particular user, "normally used" refers to a
+typical or common use of that class of product, regardless of the status
+of the particular user or of the way in which the particular user
+actually uses, or expects or is expected to use, the product.  A product
+is a consumer product regardless of whether the product has substantial
+commercial, industrial or non-consumer uses, unless such uses represent
+the only significant mode of use of the product.
+
+  "Installation Information" for a User Product means any methods,
+procedures, authorization keys, or other information required to install
+and execute modified versions of a covered work in that User Product from
+a modified version of its Corresponding Source.  The information must
+suffice to ensure that the continued functioning of the modified object
+code is in no case prevented or interfered with solely because
+modification has been made.
+
+  If you convey an object code work under this section in, or with, or
+specifically for use in, a User Product, and the conveying occurs as
+part of a transaction in which the right of possession and use of the
+User Product is transferred to the recipient in perpetuity or for a
+fixed term (regardless of how the transaction is characterized), the
+Corresponding Source conveyed under this section must be accompanied
+by the Installation Information.  But this requirement does not apply
+if neither you nor any third party retains the ability to install
+modified object code on the User Product (for example, the work has
+been installed in ROM).
+
+  The requirement to provide Installation Information does not include a
+requirement to continue to provide support service, warranty, or updates
+for a work that has been modified or installed by the recipient, or for
+the User Product in which it has been modified or installed.  Access to a
+network may be denied when the modification itself materially and
+adversely affects the operation of the network or violates the rules and
+protocols for communication across the network.
+
+  Corresponding Source conveyed, and Installation Information provided,
+in accord with this section must be in a format that is publicly
+documented (and with an implementation available to the public in
+source code form), and must require no special password or key for
+unpacking, reading or copying.
+
+  7. Additional Terms.
+
+  "Additional permissions" are terms that supplement the terms of this
+License by making exceptions from one or more of its conditions.
+Additional permissions that are applicable to the entire Program shall
+be treated as though they were included in this License, to the extent
+that they are valid under applicable law.  If additional permissions
+apply only to part of the Program, that part may be used separately
+under those permissions, but the entire Program remains governed by
+this License without regard to the additional permissions.
+
+  When you convey a copy of a covered work, you may at your option
+remove any additional permissions from that copy, or from any part of
+it.  (Additional permissions may be written to require their own
+removal in certain cases when you modify the work.)  You may place
+additional permissions on material, added by you to a covered work,
+for which you have or can give appropriate copyright permission.
+
+  Notwithstanding any other provision of this License, for material you
+add to a covered work, you may (if authorized by the copyright holders of
+that material) supplement the terms of this License with terms:
+
+    a) Disclaiming warranty or limiting liability differently from the
+    terms of sections 15 and 16 of this License; or
+
+    b) Requiring preservation of specified reasonable legal notices or
+    author attributions in that material or in the Appropriate Legal
+    Notices displayed by works containing it; or
+
+    c) Prohibiting misrepresentation of the origin of that material, or
+    requiring that modified versions of such material be marked in
+    reasonable ways as different from the original version; or
+
+    d) Limiting the use for publicity purposes of names of licensors or
+    authors of the material; or
+
+    e) Declining to grant rights under trademark law for use of some
+    trade names, trademarks, or service marks; or
+
+    f) Requiring indemnification of licensors and authors of that
+    material by anyone who conveys the material (or modified versions of
+    it) with contractual assumptions of liability to the recipient, for
+    any liability that these contractual assumptions directly impose on
+    those licensors and authors.
+
+  All other non-permissive additional terms are considered "further
+restrictions" within the meaning of section 10.  If the Program as you
+received it, or any part of it, contains a notice stating that it is
+governed by this License along with a term that is a further
+restriction, you may remove that term.  If a license document contains
+a further restriction but permits relicensing or conveying under this
+License, you may add to a covered work material governed by the terms
+of that license document, provided that the further restriction does
+not survive such relicensing or conveying.
+
+  If you add terms to a covered work in accord with this section, you
+must place, in the relevant source files, a statement of the
+additional terms that apply to those files, or a notice indicating
+where to find the applicable terms.
+
+  Additional terms, permissive or non-permissive, may be stated in the
+form of a separately written license, or stated as exceptions;
+the above requirements apply either way.
+
+  8. Termination.
+
+  You may not propagate or modify a covered work except as expressly
+provided under this License.  Any attempt otherwise to propagate or
+modify it is void, and will automatically terminate your rights under
+this License (including any patent licenses granted under the third
+paragraph of section 11).
+
+  However, if you cease all violation of this License, then your
+license from a particular copyright holder is reinstated (a)
+provisionally, unless and until the copyright holder explicitly and
+finally terminates your license, and (b) permanently, if the copyright
+holder fails to notify you of the violation by some reasonable means
+prior to 60 days after the cessation.
+
+  Moreover, your license from a particular copyright holder is
+reinstated permanently if the copyright holder notifies you of the
+violation by some reasonable means, this is the first time you have
+received notice of violation of this License (for any work) from that
+copyright holder, and you cure the violation prior to 30 days after
+your receipt of the notice.
+
+  Termination of your rights under this section does not terminate the
+licenses of parties who have received copies or rights from you under
+this License.  If your rights have been terminated and not permanently
+reinstated, you do not qualify to receive new licenses for the same
+material under section 10.
+
+  9. Acceptance Not Required for Having Copies.
+
+  You are not required to accept this License in order to receive or
+run a copy of the Program.  Ancillary propagation of a covered work
+occurring solely as a consequence of using peer-to-peer transmission
+to receive a copy likewise does not require acceptance.  However,
+nothing other than this License grants you permission to propagate or
+modify any covered work.  These actions infringe copyright if you do
+not accept this License.  Therefore, by modifying or propagating a
+covered work, you indicate your acceptance of this License to do so.
+
+  10. Automatic Licensing of Downstream Recipients.
+
+  Each time you convey a covered work, the recipient automatically
+receives a license from the original licensors, to run, modify and
+propagate that work, subject to this License.  You are not responsible
+for enforcing compliance by third parties with this License.
+
+  An "entity transaction" is a transaction transferring control of an
+organization, or substantially all assets of one, or subdividing an
+organization, or merging organizations.  If propagation of a covered
+work results from an entity transaction, each party to that
+transaction who receives a copy of the work also receives whatever
+licenses to the work the party's predecessor in interest had or could
+give under the previous paragraph, plus a right to possession of the
+Corresponding Source of the work from the predecessor in interest, if
+the predecessor has it or can get it with reasonable efforts.
+
+  You may not impose any further restrictions on the exercise of the
+rights granted or affirmed under this License.  For example, you may
+not impose a license fee, royalty, or other charge for exercise of
+rights granted under this License, and you may not initiate litigation
+(including a cross-claim or counterclaim in a lawsuit) alleging that
+any patent claim is infringed by making, using, selling, offering for
+sale, or importing the Program or any portion of it.
+
+  11. Patents.
+
+  A "contributor" is a copyright holder who authorizes use under this
+License of the Program or a work on which the Program is based.  The
+work thus licensed is called the contributor's "contributor version".
+
+  A contributor's "essential patent claims" are all patent claims
+owned or controlled by the contributor, whether already acquired or
+hereafter acquired, that would be infringed by some manner, permitted
+by this License, of making, using, or selling its contributor version,
+but do not include claims that would be infringed only as a
+consequence of further modification of the contributor version.  For
+purposes of this definition, "control" includes the right to grant
+patent sublicenses in a manner consistent with the requirements of
+this License.
+
+  Each contributor grants you a non-exclusive, worldwide, royalty-free
+patent license under the contributor's essential patent claims, to
+make, use, sell, offer for sale, import and otherwise run, modify and
+propagate the contents of its contributor version.
+
+  In the following three paragraphs, a "patent license" is any express
+agreement or commitment, however denominated, not to enforce a patent
+(such as an express permission to practice a patent or covenant not to
+sue for patent infringement).  To "grant" such a patent license to a
+party means to make such an agreement or commitment not to enforce a
+patent against the party.
+
+  If you convey a covered work, knowingly relying on a patent license,
+and the Corresponding Source of the work is not available for anyone
+to copy, free of charge and under the terms of this License, through a
+publicly available network server or other readily accessible means,
+then you must either (1) cause the Corresponding Source to be so
+available, or (2) arrange to deprive yourself of the benefit of the
+patent license for this particular work, or (3) arrange, in a manner
+consistent with the requirements of this License, to extend the patent
+license to downstream recipients.  "Knowingly relying" means you have
+actual knowledge that, but for the patent license, your conveying the
+covered work in a country, or your recipient's use of the covered work
+in a country, would infringe one or more identifiable patents in that
+country that you have reason to believe are valid.
+
+  If, pursuant to or in connection with a single transaction or
+arrangement, you convey, or propagate by procuring conveyance of, a
+covered work, and grant a patent license to some of the parties
+receiving the covered work authorizing them to use, propagate, modify
+or convey a specific copy of the covered work, then the patent license
+you grant is automatically extended to all recipients of the covered
+work and works based on it.
+
+  A patent license is "discriminatory" if it does not include within
+the scope of its coverage, prohibits the exercise of, or is
+conditioned on the non-exercise of one or more of the rights that are
+specifically granted under this License.  You may not convey a covered
+work if you are a party to an arrangement with a third party that is
+in the business of distributing software, under which you make payment
+to the third party based on the extent of your activity of conveying
+the work, and under which the third party grants, to any of the
+parties who would receive the covered work from you, a discriminatory
+patent license (a) in connection with copies of the covered work
+conveyed by you (or copies made from those copies), or (b) primarily
+for and in connection with specific products or compilations that
+contain the covered work, unless you entered into that arrangement,
+or that patent license was granted, prior to 28 March 2007.
+
+  Nothing in this License shall be construed as excluding or limiting
+any implied license or other defenses to infringement that may
+otherwise be available to you under applicable patent law.
+
+  12. No Surrender of Others' Freedom.
+
+  If conditions are imposed on you (whether by court order, agreement or
+otherwise) that contradict the conditions of this License, they do not
+excuse you from the conditions of this License.  If you cannot convey a
+covered work so as to satisfy simultaneously your obligations under this
+License and any other pertinent obligations, then as a consequence you may
+not convey it at all.  For example, if you agree to terms that obligate you
+to collect a royalty for further conveying from those to whom you convey
+the Program, the only way you could satisfy both those terms and this
+License would be to refrain entirely from conveying the Program.
+
+  13. Remote Network Interaction; Use with the GNU General Public License.
+
+  Notwithstanding any other provision of this License, if you modify the
+Program, your modified version must prominently offer all users
+interacting with it remotely through a computer network (if your version
+supports such interaction) an opportunity to receive the Corresponding
+Source of your version by providing access to the Corresponding Source
+from a network server at no charge, through some standard or customary
+means of facilitating copying of software.  This Corresponding Source
+shall include the Corresponding Source for any work covered by version 3
+of the GNU General Public License that is incorporated pursuant to the
+following paragraph.
+
+  Notwithstanding any other provision of this License, you have
+permission to link or combine any covered work with a work licensed
+under version 3 of the GNU General Public License into a single
+combined work, and to convey the resulting work.  The terms of this
+License will continue to apply to the part which is the covered work,
+but the work with which it is combined will remain governed by version
+3 of the GNU General Public License.
+
+  14. Revised Versions of this License.
+
+  The Free Software Foundation may publish revised and/or new versions of
+the GNU Affero General Public License from time to time.  Such new versions
+will be similar in spirit to the present version, but may differ in detail to
+address new problems or concerns.
+
+  Each version is given a distinguishing version number.  If the
+Program specifies that a certain numbered version of the GNU Affero General
+Public License "or any later version" applies to it, you have the
+option of following the terms and conditions either of that numbered
+version or of any later version published by the Free Software
+Foundation.  If the Program does not specify a version number of the
+GNU Affero General Public License, you may choose any version ever published
+by the Free Software Foundation.
+
+  If the Program specifies that a proxy can decide which future
+versions of the GNU Affero General Public License can be used, that proxy's
+public statement of acceptance of a version permanently authorizes you
+to choose that version for the Program.
+
+  Later license versions may give you additional or different
+permissions.  However, no additional obligations are imposed on any
+author or copyright holder as a result of your choosing to follow a
+later version.
+
+  15. Disclaimer of Warranty.
+
+  THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
+APPLICABLE LAW.  EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
+HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
+OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
+THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
+PURPOSE.  THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
+IS WITH YOU.  SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
+ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
+
+  16. Limitation of Liability.
+
+  IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
+WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
+THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
+GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
+USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
+DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
+PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
+EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
+SUCH DAMAGES.
+
+  17. Interpretation of Sections 15 and 16.
+
+  If the disclaimer of warranty and limitation of liability provided
+above cannot be given local legal effect according to their terms,
+reviewing courts shall apply local law that most closely approximates
+an absolute waiver of all civil liability in connection with the
+Program, unless a warranty or assumption of liability accompanies a
+copy of the Program in return for a fee.
+
+                     END OF TERMS AND CONDITIONS
+
+            How to Apply These Terms to Your New Programs
+
+  If you develop a new program, and you want it to be of the greatest
+possible use to the public, the best way to achieve this is to make it
+free software which everyone can redistribute and change under these terms.
+
+  To do so, attach the following notices to the program.  It is safest
+to attach them to the start of each source file to most effectively
+state the exclusion of warranty; and each file should have at least
+the "copyright" line and a pointer to where the full notice is found.
+
+    <one line to give the program's name and a brief idea of what it does.>
+    Copyright (C) <year>  <name of author>
+
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU Affero General Public License as published
+    by the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU Affero General Public License for more details.
+
+    You should have received a copy of the GNU Affero General Public License
+    along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+Also add information on how to contact you by electronic and paper mail.
+
+  If your software can interact with users remotely through a computer
+network, you should also make sure that it provides a way for users to
+get its source.  For example, if your program is a web application, its
+interface could display a "Source" link that leads users to an archive
+of the code.  There are many ways you could offer source, and different
+solutions will be better for different programs; see section 13 for the
+specific requirements.
+
+  You should also get your employer (if you work as a programmer) or school,
+if any, to sign a "copyright disclaimer" for the program, if necessary.
+For more information on this, and how to apply and follow the GNU AGPL, see
+<https://www.gnu.org/licenses/>.

LICENSING.md

@@ -0,0 +1,22 @@
+# Licensing
+
+License names used in this document are as per [SPDX License
+List](https://spdx.org/licenses/).
+
+The default license for this project is [AGPL-3.0-only](LICENSE).
+
+## MIT
+
+The following files are licensed under MIT:
+
+```
+src/polyfill/fetch.js
+```
+
+The following directories and their subdirectories are licensed under their
+original upstream licenses:
+
+```
+vendor/
+tests/wpt/
+```

Makefile

@@ -3,6 +3,30 @@
 
 ZIG := zig
 BC := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
+# option test filter make test F="server"
+F=
+
+# OS and ARCH
+kernel = $(shell uname -ms)
+ifeq ($(kernel), Darwin arm64)
+	OS := macos
+	ARCH := aarch64
+else ifeq ($(kernel), Darwin x86_64)
+	OS := macos
+	ARCH := x86_64
+else ifeq ($(kernel), Linux aarch64)
+	OS := linux
+	ARCH := aarch64
+else ifeq ($(kernel), Linux arm64)
+	OS := linux
+	ARCH := aarch64
+else ifeq ($(kernel), Linux x86_64)
+	OS := linux
+	ARCH := x86_64
+else
+	$(error "Unhandled kernel: $(kernel)")
+endif
+
 
 # Infos
 # -----
@@ -23,85 +47,90 @@ help:
 
 # $(ZIG) commands
 # ------------
-.PHONY: build build-release run run-release shell test bench download-zig wpt
+.PHONY: build build-dev run run-release shell test bench download-zig wpt data get-v8 build-v8 build-v8-dev
+.PHONY: end2end
 
-zig_version = $(shell grep 'recommended_zig_version = "' "vendor/jsruntime-lib/build.zig" | cut -d'"' -f2)
-kernel = $(shell uname -ms)
+zig_version = $(shell grep 'recommended_zig_version = "' "vendor/zig-js-runtime/build.zig" | cut -d'"' -f2)
 
 ## Download the zig recommended version
 download-zig:
-ifeq ($(kernel), Darwin x86_64)
-	$(eval target="macos")
-	$(eval arch="x86_64")
-else ifeq ($(kernel), Darwin arm64)
-	$(eval target="macos")
-	$(eval arch="aarch64")
-else ifeq ($(kernel), Linux aarch64)
-	$(eval target="linux")
-	$(eval arch="aarch64")
-else ifeq ($(kernel), Linux arm64)
-	$(eval target="linux")
-	$(eval arch="aarch64")
-else ifeq ($(kernel), Linux x86_64)
-	$(eval target="linux")
-	$(eval arch="x86_64")
-else
-	$(error "Unhandled kernel: $(kernel)")
-endif
-	$(eval url = "https://ziglang.org/builds/zig-$(target)-$(arch)-$(zig_version).tar.xz")
-	$(eval dest = "/tmp/zig-$(target)-$(arch)-$(zig_version).tar.xz")
+	$(eval url = "https://ziglang.org/download/$(zig_version)/zig-$(OS)-$(ARCH)-$(zig_version).tar.xz")
+	$(eval dest = "/tmp/zig-$(OS)-$(ARCH)-$(zig_version).tar.xz")
 	@printf "\e[36mDownload zig version $(zig_version)...\e[0m\n"
 	@curl -o "$(dest)" -L "$(url)" || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
 	@printf "\e[33mDownloaded $(dest)\e[0m\n"
 
-## Build in debug mode
+## Build in release-safe mode
 build:
-	@printf "\e[36mBuilding (debug)...\e[0m\n"
-	@$(ZIG) build -Dengine=v8 || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
-	@printf "\e[33mBuild OK\e[0m\n"
-
-build-release:
 	@printf "\e[36mBuilding (release safe)...\e[0m\n"
-	@$(ZIG) build -Doptimize=ReleaseSafe -Dengine=v8 || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
+	$(ZIG) build -Doptimize=ReleaseSafe -Dgit_commit=$$(git rev-parse --short HEAD) || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
 	@printf "\e[33mBuild OK\e[0m\n"
 
-## Run the server
+## Build in debug mode
+build-dev:
+	@printf "\e[36mBuilding (debug)...\e[0m\n"
+	@$(ZIG) build -Dgit_commit=$$(git rev-parse --short HEAD) || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
+	@printf "\e[33mBuild OK\e[0m\n"
+
+## Run the server in release mode
 run: build
 	@printf "\e[36mRunning...\e[0m\n"
-	@./zig-out/bin/browsercore || (printf "\e[33mRun ERROR\e[0m\n"; exit 1;)
+	@./zig-out/bin/lightpanda || (printf "\e[33mRun ERROR\e[0m\n"; exit 1;)
 
-## Run a JS shell in release-safe mode
+## Run the server in debug mode
+run-debug: build-dev
+	@printf "\e[36mRunning...\e[0m\n"
+	@./zig-out/bin/lightpanda || (printf "\e[33mRun ERROR\e[0m\n"; exit 1;)
+
+## Run a JS shell in debug mode
 shell:
 	@printf "\e[36mBuilding shell...\e[0m\n"
-	@$(ZIG) build shell -Dengine=v8 || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
+	@$(ZIG) build shell || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
 
 ## Run WPT tests
 wpt:
 	@printf "\e[36mBuilding wpt...\e[0m\n"
-	@$(ZIG) build wpt -Dengine=v8 -- --safe $(filter-out $@,$(MAKECMDGOALS)) || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
+	@$(ZIG) build wpt -- $(filter-out $@,$(MAKECMDGOALS)) || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
 
 wpt-summary:
 	@printf "\e[36mBuilding wpt...\e[0m\n"
-	@$(ZIG) build wpt -Dengine=v8 -- --safe --summary $(filter-out $@,$(MAKECMDGOALS)) || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
+	@$(ZIG) build wpt -- --summary $(filter-out $@,$(MAKECMDGOALS)) || (printf "\e[33mBuild ERROR\e[0m\n"; exit 1;)
 
 ## Test
 test:
-	@printf "\e[36mTesting...\e[0m\n"
-	@$(ZIG) build test -Dengine=v8 || (printf "\e[33mTest ERROR\e[0m\n"; exit 1;)
-	@printf "\e[33mTest OK\e[0m\n"
+	@TEST_FILTER='${F}' $(ZIG) build test -freference-trace --summary all
+
+## Run demo/runner end to end tests
+end2end:
+	@test -d ../demo
+	cd ../demo && go run runner/main.go
+
+## v8
+get-v8:
+	@printf "\e[36mGetting v8 source...\e[0m\n"
+	@$(ZIG) build get-v8
+
+build-v8-dev:
+	@printf "\e[36mBuilding v8 (dev)...\e[0m\n"
+	@$(ZIG) build build-v8
+
+build-v8:
+	@printf "\e[36mBuilding v8...\e[0m\n"
+	@$(ZIG) build -Doptimize=ReleaseSafe build-v8
 
 # Install and build required dependencies commands
 # ------------
 .PHONY: install-submodule
-.PHONY: install-lexbor install-jsruntime install-jsruntime-dev install-libiconv
+.PHONY: install-libiconv
 .PHONY: _install-netsurf install-netsurf clean-netsurf test-netsurf install-netsurf-dev
+.PHONY: install-mimalloc install-mimalloc-dev clean-mimalloc
 .PHONY: install-dev install
 
 ## Install and build dependencies for release
-install: install-submodule install-lexbor install-jsruntime install-netsurf
+install: install-submodule install-libiconv install-netsurf install-mimalloc
 
 ## Install and build dependencies for dev
-install-dev: install-submodule install-lexbor install-jsruntime-dev install-netsurf-dev
+install-dev: install-submodule install-libiconv install-netsurf-dev install-mimalloc-dev
 
 install-netsurf-dev: _install-netsurf
 install-netsurf-dev: OPTCFLAGS := -O0 -g -DNDEBUG
@@ -109,14 +138,16 @@ install-netsurf-dev: OPTCFLAGS := -O0 -g -DNDEBUG
 install-netsurf: _install-netsurf
 install-netsurf: OPTCFLAGS := -DNDEBUG
 
-BC_NS := $(BC)vendor/netsurf
-ICONV := $(BC)vendor/libiconv
+BC_NS := $(BC)vendor/netsurf/out/$(OS)-$(ARCH)
+ICONV := $(BC)vendor/libiconv/out/$(OS)-$(ARCH)
 # TODO: add Linux iconv path (I guess it depends on the distro)
 # TODO: this way of linking libiconv is not ideal. We should have a more generic way
 # and stick to a specif version. Maybe build from source. Anyway not now.
-_install-netsurf: install-libiconv
+_install-netsurf: clean-netsurf
 	@printf "\e[36mInstalling NetSurf...\e[0m\n" && \
-	ls $(ICONV) 1> /dev/null || (printf "\e[33mERROR: you need to install libiconv in your system (on MacOS on with Homebrew)\e[0m\n"; exit 1;) && \
+	ls $(ICONV)/lib/libiconv.a 1> /dev/null || (printf "\e[33mERROR: you need to execute 'make install-libiconv'\e[0m\n"; exit 1;) && \
+	mkdir -p $(BC_NS) && \
+	cp -R vendor/netsurf/share $(BC_NS) && \
 	export PREFIX=$(BC_NS) && \
 	export OPTLDFLAGS="-L$(ICONV)/lib" && \
 	export OPTCFLAGS="$(OPTCFLAGS) -I$(ICONV)/include" && \
@@ -135,7 +166,7 @@ _install-netsurf: install-libiconv
 	BUILDDIR=$(BC_NS)/build/libdom make install && \
 	printf "\e[33mRunning libdom example...\e[0m\n" && \
 	cd examples && \
-	zig cc \
+	$(ZIG) cc \
 	-I$(ICONV)/include \
 	-I$(BC_NS)/include \
 	-L$(ICONV)/lib \
@@ -154,10 +185,7 @@ _install-netsurf: install-libiconv
 
 clean-netsurf:
 	@printf "\e[36mCleaning NetSurf build...\e[0m\n" && \
-	cd vendor/netsurf && \
-	rm -R build && \
-	rm -R lib && \
-	rm -R include
+	rm -Rf $(BC_NS)
 
 test-netsurf:
 	@printf "\e[36mTesting NetSurf...\e[0m\n" && \
@@ -167,29 +195,52 @@ test-netsurf:
 	cd vendor/netsurf/libdom && \
 	BUILDDIR=$(BC_NS)/build/libdom make test
 
-install-libiconv:
-ifeq ("$(wildcard vendor/libiconv/lib/libiconv.a)","")
+download-libiconv:
+ifeq ("$(wildcard vendor/libiconv/libiconv-1.17)","")
 	@mkdir -p vendor/libiconv
 	@cd vendor/libiconv && \
-	curl https://ftp.gnu.org/pub/gnu/libiconv/libiconv-1.17.tar.gz | tar -xvzf -
-	@cd vendor/libiconv/libiconv-1.17 && \
-	./configure --prefix=$(BC)vendor/libiconv --enable-static && \
-	make && make install
+	curl -L https://github.com/lightpanda-io/libiconv/releases/download/1.17/libiconv-1.17.tar.gz | tar -xvzf -
 endif
 
-install-lexbor:
-	@mkdir -p vendor/lexbor
-	@cd vendor/lexbor && \
-	cmake ../lexbor-src -DLEXBOR_BUILD_SHARED=OFF && \
-	make
+build-libiconv: clean-libiconv
+	@cd vendor/libiconv/libiconv-1.17 && \
+	./configure --prefix=$(ICONV) --enable-static && \
+	make && make install
 
-install-jsruntime-dev:
-	@cd vendor/jsruntime-lib && \
-	make install-dev
+install-libiconv: download-libiconv build-libiconv
 
-install-jsruntime:
-	@cd vendor/jsruntime-lib && \
-	make install
+clean-libiconv:
+ifneq ("$(wildcard vendor/libiconv/libiconv-1.17/Makefile)","")
+	@cd vendor/libiconv/libiconv-1.17 && \
+	make clean
+endif
+
+data:
+	cd src/data && go run public_suffix_list_gen.go > public_suffix_list.zig
+
+.PHONY: _build_mimalloc
+
+MIMALLOC := $(BC)vendor/mimalloc/out/$(OS)-$(ARCH)
+_build_mimalloc: clean-mimalloc
+	@mkdir -p $(MIMALLOC)/build && \
+	cd $(MIMALLOC)/build && \
+	cmake -DMI_BUILD_SHARED=OFF -DMI_BUILD_OBJECT=OFF -DMI_BUILD_TESTS=OFF -DMI_OVERRIDE=OFF $(OPTS) ../../.. && \
+	make && \
+	mkdir -p $(MIMALLOC)/lib
+
+install-mimalloc-dev: _build_mimalloc
+install-mimalloc-dev: OPTS=-DCMAKE_BUILD_TYPE=Debug
+install-mimalloc-dev:
+	@cd $(MIMALLOC) && \
+	mv build/libmimalloc-debug.a lib/libmimalloc.a
+
+install-mimalloc: _build_mimalloc
+install-mimalloc:
+	@cd $(MIMALLOC) && \
+	mv build/libmimalloc.a lib/libmimalloc.a
+
+clean-mimalloc:
+	@rm -Rf $(MIMALLOC)/build
 
 ## Init and update git submodule
 install-submodule:

README.md

@@ -1,115 +1,350 @@
-# Browsercore
+<p align="center">
+  <a href="https://lightpanda.io"><img src="https://cdn.lightpanda.io/assets/images/logo/lpd-logo.png" alt="Logo" height=170></a>
+</p>
 
-## Build
+<h1 align="center">Lightpanda Browser</h1>
+
+<p align="center"><a href="https://lightpanda.io/">lightpanda.io</a></p>
+
+<div align="center">
+
+[![License](https://img.shields.io/github/license/lightpanda-io/browser)](https://github.com/lightpanda-io/browser/blob/main/LICENSE)
+[![Twitter Follow](https://img.shields.io/twitter/follow/lightpanda_io)](https://twitter.com/lightpanda_io)
+[![GitHub stars](https://img.shields.io/github/stars/lightpanda-io/browser)](https://github.com/lightpanda-io/browser)
+
+</div>
+
+Lightpanda is the open-source browser made for headless usage:
+
+- Javascript execution
+- Support of Web APIs (partial, WIP)
+- Compatible with Playwright[^1], Puppeteer, chromedp through CDP
+
+Fast web automation for AI agents, LLM training, scraping and testing:
+
+- Ultra-low memory footprint (9x less than Chrome)
+- Exceptionally fast execution (11x faster than Chrome)
+- Instant startup
+
+[<img width="350px" src="https://cdn.lightpanda.io/assets/images/github/execution-time.svg">
+](https://github.com/lightpanda-io/demo)
+&emsp;
+[<img width="350px" src="https://cdn.lightpanda.io/assets/images/github/memory-frame.svg">
+](https://github.com/lightpanda-io/demo)
+</div>
+
+_Puppeteer requesting 100 pages from a local website on a AWS EC2 m5.large instance.
+See [benchmark details](https://github.com/lightpanda-io/demo)._
+
+[^1]: **Playwright support disclaimer:**
+Due to the nature of Playwright, a script that works with the current version of the browser may not function correctly with a future version. Playwright uses an intermediate JavaScript layer that selects an execution strategy based on the browser's available features. If Lightpanda adds a new [Web API](https://developer.mozilla.org/en-US/docs/Web/API), Playwright may choose to execute different code for the same script. This new code path could attempt to use features that are not yet implemented. Lightpanda makes an effort to add compatibility tests, but we can't cover all scenarios. If you encounter an issue, please create a [GitHub issue](https://github.com/lightpanda-io/browser/issues) and include the last known working version of the script.
+
+## Quick start
+
+### Install
+**Install from the nightly builds**
+
+You can download the last binary from the [nightly
+builds](https://github.com/lightpanda-io/browser/releases/tag/nightly) for
+Linux x86_64 and MacOS aarch64.
+
+*For Linux*
+```console
+curl -L -o lightpanda https://github.com/lightpanda-io/browser/releases/download/nightly/lightpanda-x86_64-linux && \
+chmod a+x ./lightpanda
+```
+
+*For MacOS*
+```console
+curl -L -o lightpanda https://github.com/lightpanda-io/browser/releases/download/nightly/lightpanda-aarch64-macos && \
+chmod a+x ./lightpanda
+```
+
+*For Windows + WSL2*
+
+The Lightpanda browser is compatible to run on windows inside WSL. Follow the Linux instruction for installation from a WSL terminal.
+It is recommended to install clients like Puppeteer on the Windows host.
+
+**Install from Docker**
+
+Lightpanda provides [official Docker
+images](https://hub.docker.com/r/lightpanda/browser) for both Linux amd64 and
+arm64 architectures.
+The following command fetches the Docker image and starts a new container exposing Lightpanda's CDP server on port `9222`.
+```console
+docker run -d --name lightpanda -p 9222:9222 lightpanda/browser:nightly
+```
+
+### Dump a URL
+
+```console
+./lightpanda fetch --dump https://lightpanda.io
+```
+```console
+info(browser): GET https://lightpanda.io/ http.Status.ok
+info(browser): fetch script https://api.website.lightpanda.io/js/script.js: http.Status.ok
+info(browser): eval remote https://api.website.lightpanda.io/js/script.js: TypeError: Cannot read properties of undefined (reading 'pushState')
+<!DOCTYPE html>
+```
+
+### Start a CDP server
+
+```console
+./lightpanda serve --host 127.0.0.1 --port 9222
+```
+```console
+info(websocket): starting blocking worker to listen on 127.0.0.1:9222
+info(server): accepting new conn...
+```
+
+Once the CDP server started, you can run a Puppeteer script by configuring the
+`browserWSEndpoint`.
+
+```js
+'use strict'
+
+import puppeteer from 'puppeteer-core';
+
+// use browserWSEndpoint to pass the Lightpanda's CDP server address.
+const browser = await puppeteer.connect({
+  browserWSEndpoint: "ws://127.0.0.1:9222",
+});
+
+// The rest of your script remains the same.
+const context = await browser.createBrowserContext();
+const page = await context.newPage();
+
+// Dump all the links from the page.
+await page.goto('https://wikipedia.com/');
+
+const links = await page.evaluate(() => {
+  return Array.from(document.querySelectorAll('a')).map(row => {
+    return row.getAttribute('href');
+  });
+});
+
+console.log(links);
+
+await page.close();
+await context.close();
+await browser.disconnect();
+```
+
+### Telemetry
+By default, Lightpanda collects and sends usage telemetry. This can be disabled by setting an environment variable `LIGHTPANDA_DISABLE_TELEMETRY=true`. You can read Lightpanda's privacy policy at: [https://lightpanda.io/privacy-policy](https://lightpanda.io/privacy-policy).
+
+## Status
+
+Lightpanda is in Beta and currently a work in progress. Stability and coverage are improving and many websites now work.
+You may still encounter errors or crashes. Please open an issue with specifics if so.
+
+Here are the key features we have implemented:
+
+- [x] HTTP loader (based on Libcurl)
+- [x] HTML parser and DOM tree (based on Netsurf libs)
+- [x] Javascript support (v8)
+- [x] DOM APIs
+- [x] Ajax
+  - [x] XHR API
+  - [x] Fetch API (polyfill)
+- [x] DOM dump
+- [x] CDP/websockets server
+- [x] Click
+- [x] Input form
+- [x] Cookies
+- [x] Custom HTTP headers
+- [x] Proxy support
+- [x] Network interception
+
+NOTE: There are hundreds of Web APIs. Developing a browser (even just for headless mode) is a huge task. Coverage will increase over time.
+
+You can also follow the progress of our Javascript support in our dedicated [zig-js-runtime](https://github.com/lightpanda-io/zig-js-runtime#development) project.
+
+## Build from sources
 
 ### Prerequisites
 
-Browsercore is written with [Zig](https://ziglang.org/) `0.11.0`. You have to
+Lightpanda is written with [Zig](https://ziglang.org/) `0.15.1`. You have to
 install it with the right version in order to build the project.
 
-Browsercore also depends on
-[js-runtimelib](https://github.com/francisbouvier/jsruntime-lib/) and
-[lexbor](https://github.com/lexbor/lexbor) libs.
+Lightpanda also depends on
+[zig-js-runtime](https://github.com/lightpanda-io/zig-js-runtime/) (with v8),
+[Libcurl](https://curl.se/libcurl/),
+[Netsurf libs](https://www.netsurf-browser.org/) and
+[Mimalloc](https://microsoft.github.io/mimalloc).
 
-To be able to build the v8 engine for js-runtimelib, you have to install some libs:
+To be able to build the v8 engine for zig-js-runtime, you have to install some libs:
 
 For Debian/Ubuntu based Linux:
+
 ```
 sudo apt install xz-utils \
     python3 ca-certificates git \
     pkg-config libglib2.0-dev \
-    gperf libexpat1-dev \
+    gperf libexpat1-dev unzip rsync \
     cmake clang
 ```
 
-For MacOS, you only need Python 3 and cmake.
+For systems with [Nix](https://nixos.org/download/), you can use the devShell:
+```
+nix develop
+```
 
-To be able to build lexbor, you need to install also `cmake`.
+For MacOS, you only need cmake:
+
+```
+brew install cmake
+```
 
 ### Install and build dependencies
 
-The project uses git submodule for dependencies.
-The `make install-submodule` will init and update the submodules in the `vendor/`
-directory.
+#### All in one build
+
+You can run `make install` to install deps all in one (or `make install-dev` if you need the development versions).
+
+Be aware that the build task is very long and cpu consuming, as you will build from sources all dependencies, including the v8 Javascript engine.
+
+#### Step by step build dependency
+
+The project uses git submodules for dependencies.
+
+To init or update the submodules in the `vendor/` directory:
 
 ```
 make install-submodule
 ```
 
-### Build netsurf
+**iconv**
+
+libiconv is an internationalization library used by Netsurf.
+
+```
+make install-libiconv
+```
+
+**Netsurf libs**
+
+Netsurf libs are used for HTML parsing and DOM tree generation.
 
-The command `make install-netsurf` will build netsurf libs used by browsercore.
 ```
 make install-netsurf
 ```
 
-### Build lexbor
+For dev env, use `make install-netsurf-dev`.
 
-The command `make install-lexbor` will build lexbor lib used by browsercore.
-```
-make install-lexbor
-```
+**Mimalloc**
 
-### Build jsruntime-lib
-
-The command `make install-jsruntime-dev` uses jsruntime-lib's `zig-v8` dependency to build v8 engine lib.
-Be aware the build task is very long and cpu consuming.
-
-Build v8 engine for debug/dev version, it creates
-`vendor/jsruntime-lib/vendor/v8/$ARCH/debug/libc_v8.a` file.
+Mimalloc is used as a C memory allocator.
 
 ```
-make install-jsruntime-dev
+make install-mimalloc
 ```
 
-You should also build a release vesion of v8 with:
+For dev env, use `make install-mimalloc-dev`.
+
+Note: when Mimalloc is built in dev mode, you can dump memory stats with the
+env var `MIMALLOC_SHOW_STATS=1`. See
+[https://microsoft.github.io/mimalloc/environment.html](https://microsoft.github.io/mimalloc/environment.html).
+
+**v8**
+
+First, get the tools necessary for building V8, as well as the V8 source code:
 
 ```
-make install-jsruntime
+make get-v8
 ```
 
-### All in one build
+Next, build v8. This build task is very long and cpu consuming, as you will build v8 from sources.
 
-You can run `make intall` and `make install-dev` to install deps all in one.
+```
+make build-v8
+```
+
+For dev env, use `make build-v8-dev`.
 
 ## Test
 
 ### Unit Tests
 
-You can test browsercore by running `make test`.
+You can test Lightpanda by running `make test`.
+
+### End to end tests
+
+To run end to end tests, you need to clone the [demo
+repository](https://github.com/lightpanda-io/demo) into `../demo` dir.
+
+You have to install the [demo's node
+requirements](https://github.com/lightpanda-io/demo?tab=readme-ov-file#dependencies-1)
+
+You also need to install [Go](https://go.dev) > v1.24.
+
+```
+make end2end
+```
 
 ### Web Platform Tests
 
-Browsercore is tested against the standardized [Web Platform
+Lightpanda is tested against the standardized [Web Platform
 Tests](https://web-platform-tests.org/).
 
-The relevant tests cases for Browsercore are commit with the project.
-All the tests cases executed are located in `tests/wpt` dir and come from an
-external repository: https://github.com/lightpanda-io/wpt
+The relevant tests cases are committed in a [dedicated repository](https://github.com/lightpanda-io/wpt) which is fetched by the `make install-submodule` command.
+
+All the tests cases executed are located in the `tests/wpt` sub-directory.
 
 For reference, you can easily execute a WPT test case with your browser via
 [wpt.live](https://wpt.live).
 
-*Run WPT test suite*
+#### Run WPT test suite
+
+To run all the tests:
 
-You can run all the test.
-The runner execute all the tests ending with `.html`.
 ```
 make wpt
 ```
 
-Or one specific test by using a suffix.
+Or one specific test:
+
 ```
 make wpt Node-childNodes.html
 ```
 
-*Add a new WPT test case*
+#### Add a new WPT test case
 
-We add new tests cases files with implemented changes in Browsercore.
+We add new relevant tests cases files when we implemented changes in Lightpanda.
 
-Copy the test case you want to add from the [WPT
-repo](https://github.com/web-platform-tests/wpt) into `tests/wpt` dir, commit
-the files in the https://github.com/lightpanda-io/wpt repository and update the
-git submodule in browsercore.
+To add a new test, copy the file you want from the [WPT
+repo](https://github.com/web-platform-tests/wpt) into the `tests/wpt` directory.
 
-:warning: Please keep the original directory tree structure into `tests/wpt`.
+:warning: Please keep the original directory tree structure of `tests/wpt`.
+
+## Contributing
+
+Lightpanda accepts pull requests through GitHub.
+
+You have to sign our [CLA](CLA.md) during the pull request process otherwise
+we're not able to accept your contributions.
+
+## Why?
+
+### Javascript execution is mandatory for the modern web
+
+In the good old days, scraping a webpage was as easy as making an HTTP request, cURL-like. It’s not possible anymore, because Javascript is everywhere, like it or not:
+
+- Ajax, Single Page App, infinite loading, “click to display”, instant search, etc.
+- JS web frameworks: React, Vue, Angular & others
+
+### Chrome is not the right tool
+
+If we need Javascript, why not use a real web browser? Take a huge desktop application, hack it, and run it on the server. Hundreds or thousands of instances of Chrome if you use it at scale. Are you sure it’s such a good idea?
+
+- Heavy on RAM and CPU, expensive to run
+- Hard to package, deploy and maintain at scale
+- Bloated, lots of features are not useful in headless usage
+
+### Lightpanda is built for performance
+
+If we want both Javascript and performance in a true headless browser, we need to start from scratch. Not another iteration of Chromium, really from a blank page. Crazy right? But that’s what we did:
+
+- Not based on Chromium, Blink or WebKit
+- Low-level system programming language (Zig) with optimisations in mind
+- Opinionated: without graphical rendering

build.zig

@@ -1,16 +1,31 @@
-const std = @import("std");
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
+const std = @import("std");
 const builtin = @import("builtin");
 
-const jsruntime_path = "vendor/jsruntime-lib/";
-const jsruntime = @import("vendor/jsruntime-lib/build.zig");
-const jsruntime_pkgs = jsruntime.packages(jsruntime_path);
+const Build = std.Build;
 
 /// Do not rename this constant. It is scanned by some scripts to determine
 /// which zig version to install.
-const recommended_zig_version = jsruntime.recommended_zig_version;
+const recommended_zig_version = "0.15.1";
 
-pub fn build(b: *std.build.Builder) !void {
+pub fn build(b: *Build) !void {
     switch (comptime builtin.zig_version.order(std.SemanticVersion.parse(recommended_zig_version) catch unreachable)) {
         .eq => {},
         .lt => {
@@ -24,146 +39,402 @@ pub fn build(b: *std.build.Builder) !void {
         },
     }
 
+    var opts = b.addOptions();
+    opts.addOption(
+        []const u8,
+        "git_commit",
+        b.option([]const u8, "git_commit", "Current git commit") orelse "dev",
+    );
+
     const target = b.standardTargetOptions(.{});
-    const mode = b.standardOptimizeOption(.{});
+    const optimize = b.standardOptimizeOption(.{});
 
-    const options = try jsruntime.buildOptions(b);
+    // We're still using llvm because the new x86 backend seems to crash
+    // with v8. This can be reproduced in zig-v8-fork.
 
-    // browser
-    // -------
-
-    // compile and install
-    const exe = b.addExecutable(.{
-        .name = "browsercore",
-        .root_source_file = .{ .path = "src/main.zig" },
+    const lightpanda_module = b.addModule("lightpanda", .{
+        .root_source_file = b.path("src/main.zig"),
         .target = target,
-        .optimize = mode,
+        .optimize = optimize,
+        .link_libc = true,
+        .link_libcpp = true,
     });
-    try common(exe, options);
-    b.installArtifact(exe);
+    try addDependencies(b, lightpanda_module, opts);
 
-    // run
-    const run_cmd = b.addRunArtifact(exe);
-    run_cmd.step.dependOn(b.getInstallStep());
-    if (b.args) |args| {
-        run_cmd.addArgs(args);
+    {
+        // browser
+        // -------
+
+        // compile and install
+        const exe = b.addExecutable(.{
+            .name = "lightpanda",
+            .use_llvm = true,
+            .root_module = lightpanda_module,
+        });
+        b.installArtifact(exe);
+
+        // run
+        const run_cmd = b.addRunArtifact(exe);
+        if (b.args) |args| {
+            run_cmd.addArgs(args);
+        }
+
+        // step
+        const run_step = b.step("run", "Run the app");
+        run_step.dependOn(&run_cmd.step);
     }
 
-    // step
-    const run_step = b.step("run", "Run the app");
-    run_step.dependOn(&run_cmd.step);
+    {
+        // tests
+        // ----
 
-    // shell
-    // -----
+        // compile
+        const tests = b.addTest(.{
+            .root_module = lightpanda_module,
+            .use_llvm = true,
+            .test_runner = .{ .path = b.path("src/test_runner.zig"), .mode = .simple },
+        });
 
-    // compile and install
-    const shell = b.addExecutable(.{
-        .name = "browsercore-shell",
-        .root_source_file = .{ .path = "src/main_shell.zig" },
-        .target = target,
-        .optimize = mode,
-    });
-    try common(shell, options);
-    try jsruntime_pkgs.add_shell(shell);
-    // do not install shell binary
-    b.installArtifact(shell);
+        const run_tests = b.addRunArtifact(tests);
+        if (b.args) |args| {
+            run_tests.addArgs(args);
+        }
 
-    // run
-    const shell_cmd = b.addRunArtifact(shell);
-    shell_cmd.step.dependOn(b.getInstallStep());
-    if (b.args) |args| {
-        shell_cmd.addArgs(args);
+        // step
+        const tests_step = b.step("test", "Run unit tests");
+        tests_step.dependOn(&run_tests.step);
     }
 
-    // step
-    const shell_step = b.step("shell", "Run JS shell");
-    shell_step.dependOn(&shell_cmd.step);
+    {
+        // wpt
+        // -----
+        const wpt_module = b.createModule(.{
+            .root_source_file = b.path("src/main_wpt.zig"),
+            .target = target,
+            .optimize = optimize,
+        });
+        try addDependencies(b, wpt_module, opts);
 
-    // test
-    // ----
+        // compile and install
+        const wpt = b.addExecutable(.{
+            .name = "lightpanda-wpt",
+            .use_llvm = true,
+            .root_module = wpt_module,
+        });
 
-    // compile
-    const tests = b.addTest(.{ .root_source_file = .{ .path = "src/run_tests.zig" } });
-    try common(tests, options);
-    tests.single_threaded = true;
-    tests.test_runner = "src/test_runner.zig";
-    const run_tests = b.addRunArtifact(tests);
-
-    // step
-    const test_step = b.step("test", "Run unit tests");
-    test_step.dependOn(&run_tests.step);
-
-    // wpt
-    // -----
-
-    // compile and install
-    const wpt = b.addExecutable(.{
-        .name = "browsercore-wpt",
-        .root_source_file = .{ .path = "src/main_wpt.zig" },
-        .target = target,
-        .optimize = mode,
-    });
-    try common(wpt, options);
-    b.installArtifact(wpt);
-
-    // run
-    const wpt_cmd = b.addRunArtifact(wpt);
-    wpt_cmd.step.dependOn(b.getInstallStep());
-    if (b.args) |args| {
-        wpt_cmd.addArgs(args);
+        // run
+        const wpt_cmd = b.addRunArtifact(wpt);
+        if (b.args) |args| {
+            wpt_cmd.addArgs(args);
+        }
+        // step
+        const wpt_step = b.step("wpt", "WPT tests");
+        wpt_step.dependOn(&wpt_cmd.step);
     }
-    // step
-    const wpt_step = b.step("wpt", "WPT tests");
-    wpt_step.dependOn(&wpt_cmd.step);
 
-    // get
-    // -----
-
-    // compile and install
-    const get = b.addExecutable(.{
-        .name = "browsercore-get",
-        .root_source_file = .{ .path = "src/main_get.zig" },
-        .target = target,
-        .optimize = mode,
-    });
-    try common(get, options);
-    b.installArtifact(get);
-
-    // run
-    const get_cmd = b.addRunArtifact(get);
-    get_cmd.step.dependOn(b.getInstallStep());
-    if (b.args) |args| {
-        get_cmd.addArgs(args);
+    {
+        // get v8
+        // -------
+        const v8 = b.dependency("v8", .{ .target = target, .optimize = optimize });
+        const get_v8 = b.addRunArtifact(v8.artifact("get-v8"));
+        const get_step = b.step("get-v8", "Get v8");
+        get_step.dependOn(&get_v8.step);
+    }
+
+    {
+        // build v8
+        // -------
+        const v8 = b.dependency("v8", .{ .target = target, .optimize = optimize });
+        const build_v8 = b.addRunArtifact(v8.artifact("build-v8"));
+        const build_step = b.step("build-v8", "Build v8");
+        build_step.dependOn(&build_v8.step);
     }
-    // step
-    const get_step = b.step("get", "request URL");
-    get_step.dependOn(&get_cmd.step);
 }
 
-fn common(
-    step: *std.Build.CompileStep,
-    options: jsruntime.Options,
-) !void {
-    try jsruntime_pkgs.add(step, options);
-    linkLexbor(step);
-    linkNetSurf(step);
+fn addDependencies(b: *Build, mod: *Build.Module, opts: *Build.Step.Options) !void {
+    try moduleNetSurf(b, mod);
+    mod.addImport("build_config", opts.createModule());
+
+    const target = mod.resolved_target.?;
+    const dep_opts = .{
+        .target = target,
+        .optimize = mod.optimize.?,
+    };
+
+    mod.addIncludePath(b.path("vendor/lightpanda"));
+
+    {
+        // v8
+        const v8_opts = b.addOptions();
+        v8_opts.addOption(bool, "inspector_subtype", false);
+
+        const v8_mod = b.dependency("v8", dep_opts).module("v8");
+        v8_mod.addOptions("default_exports", v8_opts);
+        mod.addImport("v8", v8_mod);
+
+        const release_dir = if (mod.optimize.? == .Debug) "debug" else "release";
+        const os = switch (target.result.os.tag) {
+            .linux => "linux",
+            .macos => "macos",
+            else => return error.UnsupportedPlatform,
+        };
+        var lib_path = try std.fmt.allocPrint(
+            mod.owner.allocator,
+            "v8/out/{s}/{s}/obj/zig/libc_v8.a",
+            .{ os, release_dir },
+        );
+        std.fs.cwd().access(lib_path, .{}) catch {
+            // legacy path
+            lib_path = try std.fmt.allocPrint(
+                mod.owner.allocator,
+                "v8/out/{s}/obj/zig/libc_v8.a",
+                .{release_dir},
+            );
+        };
+        mod.addObjectFile(mod.owner.path(lib_path));
+
+        switch (target.result.os.tag) {
+            .macos => {
+                // v8 has a dependency, abseil-cpp, which, on Mac, uses CoreFoundation
+                mod.addSystemFrameworkPath(.{ .cwd_relative = "/System/Library/Frameworks" });
+                mod.linkFramework("CoreFoundation", .{});
+            },
+            else => {},
+        }
+    }
+
+    {
+        //curl
+        {
+            const is_linux = target.result.os.tag == .linux;
+            if (is_linux) {
+                mod.addCMacro("HAVE_LINUX_TCP_H", "1");
+                mod.addCMacro("HAVE_MSG_NOSIGNAL", "1");
+                mod.addCMacro("HAVE_GETHOSTBYNAME_R", "1");
+            }
+            mod.addCMacro("_FILE_OFFSET_BITS", "64");
+            mod.addCMacro("BUILDING_LIBCURL", "1");
+            mod.addCMacro("CURL_DISABLE_AWS", "1");
+            mod.addCMacro("CURL_DISABLE_DICT", "1");
+            mod.addCMacro("CURL_DISABLE_DOH", "1");
+            mod.addCMacro("CURL_DISABLE_FILE", "1");
+            mod.addCMacro("CURL_DISABLE_FTP", "1");
+            mod.addCMacro("CURL_DISABLE_GOPHER", "1");
+            mod.addCMacro("CURL_DISABLE_KERBEROS", "1");
+            mod.addCMacro("CURL_DISABLE_IMAP", "1");
+            mod.addCMacro("CURL_DISABLE_IPFS", "1");
+            mod.addCMacro("CURL_DISABLE_LDAP", "1");
+            mod.addCMacro("CURL_DISABLE_LDAPS", "1");
+            mod.addCMacro("CURL_DISABLE_MQTT", "1");
+            mod.addCMacro("CURL_DISABLE_NTLM", "1");
+            mod.addCMacro("CURL_DISABLE_PROGRESS_METER", "1");
+            mod.addCMacro("CURL_DISABLE_POP3", "1");
+            mod.addCMacro("CURL_DISABLE_RTSP", "1");
+            mod.addCMacro("CURL_DISABLE_SMB", "1");
+            mod.addCMacro("CURL_DISABLE_SMTP", "1");
+            mod.addCMacro("CURL_DISABLE_TELNET", "1");
+            mod.addCMacro("CURL_DISABLE_TFTP", "1");
+            mod.addCMacro("CURL_EXTERN_SYMBOL", "__attribute__ ((__visibility__ (\"default\"))");
+            mod.addCMacro("CURL_OS", if (is_linux) "\"Linux\"" else "\"mac\"");
+            mod.addCMacro("CURL_STATICLIB", "1");
+            mod.addCMacro("ENABLE_IPV6", "1");
+            mod.addCMacro("HAVE_ALARM", "1");
+            mod.addCMacro("HAVE_ALLOCA_H", "1");
+            mod.addCMacro("HAVE_ARPA_INET_H", "1");
+            mod.addCMacro("HAVE_ARPA_TFTP_H", "1");
+            mod.addCMacro("HAVE_ASSERT_H", "1");
+            mod.addCMacro("HAVE_BASENAME", "1");
+            mod.addCMacro("HAVE_BOOL_T", "1");
+            mod.addCMacro("HAVE_BUILTIN_AVAILABLE", "1");
+            mod.addCMacro("HAVE_CLOCK_GETTIME_MONOTONIC", "1");
+            mod.addCMacro("HAVE_DLFCN_H", "1");
+            mod.addCMacro("HAVE_ERRNO_H", "1");
+            mod.addCMacro("HAVE_FCNTL", "1");
+            mod.addCMacro("HAVE_FCNTL_H", "1");
+            mod.addCMacro("HAVE_FCNTL_O_NONBLOCK", "1");
+            mod.addCMacro("HAVE_FREEADDRINFO", "1");
+            mod.addCMacro("HAVE_FSETXATTR", "1");
+            mod.addCMacro("HAVE_FSETXATTR_5", "1");
+            mod.addCMacro("HAVE_FTRUNCATE", "1");
+            mod.addCMacro("HAVE_GETADDRINFO", "1");
+            mod.addCMacro("HAVE_GETEUID", "1");
+            mod.addCMacro("HAVE_GETHOSTBYNAME", "1");
+            mod.addCMacro("HAVE_GETHOSTBYNAME_R_6", "1");
+            mod.addCMacro("HAVE_GETHOSTNAME", "1");
+            mod.addCMacro("HAVE_GETPEERNAME", "1");
+            mod.addCMacro("HAVE_GETPPID", "1");
+            mod.addCMacro("HAVE_GETPPID", "1");
+            mod.addCMacro("HAVE_GETPROTOBYNAME", "1");
+            mod.addCMacro("HAVE_GETPWUID", "1");
+            mod.addCMacro("HAVE_GETPWUID_R", "1");
+            mod.addCMacro("HAVE_GETRLIMIT", "1");
+            mod.addCMacro("HAVE_GETSOCKNAME", "1");
+            mod.addCMacro("HAVE_GETTIMEOFDAY", "1");
+            mod.addCMacro("HAVE_GMTIME_R", "1");
+            mod.addCMacro("HAVE_IDN2_H", "1");
+            mod.addCMacro("HAVE_IF_NAMETOINDEX", "1");
+            mod.addCMacro("HAVE_IFADDRS_H", "1");
+            mod.addCMacro("HAVE_INET_ADDR", "1");
+            mod.addCMacro("HAVE_INET_PTON", "1");
+            mod.addCMacro("HAVE_INTTYPES_H", "1");
+            mod.addCMacro("HAVE_IOCTL", "1");
+            mod.addCMacro("HAVE_IOCTL_FIONBIO", "1");
+            mod.addCMacro("HAVE_IOCTL_SIOCGIFADDR", "1");
+            mod.addCMacro("HAVE_LDAP_URL_PARSE", "1");
+            mod.addCMacro("HAVE_LIBGEN_H", "1");
+            mod.addCMacro("HAVE_LIBZ", "1");
+            mod.addCMacro("HAVE_LL", "1");
+            mod.addCMacro("HAVE_LOCALE_H", "1");
+            mod.addCMacro("HAVE_LOCALTIME_R", "1");
+            mod.addCMacro("HAVE_LONGLONG", "1");
+            mod.addCMacro("HAVE_MALLOC_H", "1");
+            mod.addCMacro("HAVE_MEMORY_H", "1");
+            mod.addCMacro("HAVE_NET_IF_H", "1");
+            mod.addCMacro("HAVE_NETDB_H", "1");
+            mod.addCMacro("HAVE_NETINET_IN_H", "1");
+            mod.addCMacro("HAVE_NETINET_TCP_H", "1");
+            mod.addCMacro("HAVE_PIPE", "1");
+            mod.addCMacro("HAVE_POLL", "1");
+            mod.addCMacro("HAVE_POLL_FINE", "1");
+            mod.addCMacro("HAVE_POLL_H", "1");
+            mod.addCMacro("HAVE_POSIX_STRERROR_R", "1");
+            mod.addCMacro("HAVE_PTHREAD_H", "1");
+            mod.addCMacro("HAVE_PWD_H", "1");
+            mod.addCMacro("HAVE_RECV", "1");
+            mod.addCMacro("HAVE_SA_FAMILY_T", "1");
+            mod.addCMacro("HAVE_SELECT", "1");
+            mod.addCMacro("HAVE_SEND", "1");
+            mod.addCMacro("HAVE_SETJMP_H", "1");
+            mod.addCMacro("HAVE_SETLOCALE", "1");
+            mod.addCMacro("HAVE_SETRLIMIT", "1");
+            mod.addCMacro("HAVE_SETSOCKOPT", "1");
+            mod.addCMacro("HAVE_SIGACTION", "1");
+            mod.addCMacro("HAVE_SIGINTERRUPT", "1");
+            mod.addCMacro("HAVE_SIGNAL", "1");
+            mod.addCMacro("HAVE_SIGNAL_H", "1");
+            mod.addCMacro("HAVE_SIGSETJMP", "1");
+            mod.addCMacro("HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID", "1");
+            mod.addCMacro("HAVE_SOCKET", "1");
+            mod.addCMacro("HAVE_STDBOOL_H", "1");
+            mod.addCMacro("HAVE_STDINT_H", "1");
+            mod.addCMacro("HAVE_STDIO_H", "1");
+            mod.addCMacro("HAVE_STDLIB_H", "1");
+            mod.addCMacro("HAVE_STRCASECMP", "1");
+            mod.addCMacro("HAVE_STRDUP", "1");
+            mod.addCMacro("HAVE_STRERROR_R", "1");
+            mod.addCMacro("HAVE_STRING_H", "1");
+            mod.addCMacro("HAVE_STRINGS_H", "1");
+            mod.addCMacro("HAVE_STRSTR", "1");
+            mod.addCMacro("HAVE_STRTOK_R", "1");
+            mod.addCMacro("HAVE_STRTOLL", "1");
+            mod.addCMacro("HAVE_STRUCT_SOCKADDR_STORAGE", "1");
+            mod.addCMacro("HAVE_STRUCT_TIMEVAL", "1");
+            mod.addCMacro("HAVE_SYS_IOCTL_H", "1");
+            mod.addCMacro("HAVE_SYS_PARAM_H", "1");
+            mod.addCMacro("HAVE_SYS_POLL_H", "1");
+            mod.addCMacro("HAVE_SYS_RESOURCE_H", "1");
+            mod.addCMacro("HAVE_SYS_SELECT_H", "1");
+            mod.addCMacro("HAVE_SYS_SOCKET_H", "1");
+            mod.addCMacro("HAVE_SYS_STAT_H", "1");
+            mod.addCMacro("HAVE_SYS_TIME_H", "1");
+            mod.addCMacro("HAVE_SYS_TYPES_H", "1");
+            mod.addCMacro("HAVE_SYS_UIO_H", "1");
+            mod.addCMacro("HAVE_SYS_UN_H", "1");
+            mod.addCMacro("HAVE_TERMIO_H", "1");
+            mod.addCMacro("HAVE_TERMIOS_H", "1");
+            mod.addCMacro("HAVE_TIME_H", "1");
+            mod.addCMacro("HAVE_UNAME", "1");
+            mod.addCMacro("HAVE_UNISTD_H", "1");
+            mod.addCMacro("HAVE_UTIME", "1");
+            mod.addCMacro("HAVE_UTIME_H", "1");
+            mod.addCMacro("HAVE_UTIMES", "1");
+            mod.addCMacro("HAVE_VARIADIC_MACROS_C99", "1");
+            mod.addCMacro("HAVE_VARIADIC_MACROS_GCC", "1");
+            mod.addCMacro("HAVE_ZLIB_H", "1");
+            mod.addCMacro("RANDOM_FILE", "\"/dev/urandom\"");
+            mod.addCMacro("RECV_TYPE_ARG1", "int");
+            mod.addCMacro("RECV_TYPE_ARG2", "void *");
+            mod.addCMacro("RECV_TYPE_ARG3", "size_t");
+            mod.addCMacro("RECV_TYPE_ARG4", "int");
+            mod.addCMacro("RECV_TYPE_RETV", "ssize_t");
+            mod.addCMacro("SEND_QUAL_ARG2", "const");
+            mod.addCMacro("SEND_TYPE_ARG1", "int");
+            mod.addCMacro("SEND_TYPE_ARG2", "void *");
+            mod.addCMacro("SEND_TYPE_ARG3", "size_t");
+            mod.addCMacro("SEND_TYPE_ARG4", "int");
+            mod.addCMacro("SEND_TYPE_RETV", "ssize_t");
+            mod.addCMacro("SIZEOF_CURL_OFF_T", "8");
+            mod.addCMacro("SIZEOF_INT", "4");
+            mod.addCMacro("SIZEOF_LONG", "8");
+            mod.addCMacro("SIZEOF_OFF_T", "8");
+            mod.addCMacro("SIZEOF_SHORT", "2");
+            mod.addCMacro("SIZEOF_SIZE_T", "8");
+            mod.addCMacro("SIZEOF_TIME_T", "8");
+            mod.addCMacro("STDC_HEADERS", "1");
+            mod.addCMacro("TIME_WITH_SYS_TIME", "1");
+            mod.addCMacro("USE_NGHTTP2", "1");
+            mod.addCMacro("USE_MBEDTLS", "1");
+            mod.addCMacro("USE_THREADS_POSIX", "1");
+            mod.addCMacro("USE_UNIX_SOCKETS", "1");
+        }
+
+        try buildZlib(b, mod);
+        try buildMbedtls(b, mod);
+        try buildNghttp2(b, mod);
+        try buildCurl(b, mod);
+
+        switch (target.result.os.tag) {
+            .macos => {
+                // needed for proxying on mac
+                mod.addSystemFrameworkPath(.{ .cwd_relative = "/System/Library/Frameworks" });
+                mod.linkFramework("CoreFoundation", .{});
+                mod.linkFramework("SystemConfiguration", .{});
+            },
+            else => {},
+        }
+    }
 }
 
-fn linkLexbor(step: *std.build.LibExeObjStep) void {
-    // cmake . -DLEXBOR_BUILD_SHARED=OFF
-    const lib_path = "vendor/lexbor/liblexbor_static.a";
-    step.addObjectFile(.{ .path = lib_path });
-    step.addIncludePath(.{ .path = "vendor/lexbor-src/source" });
-}
-
-fn linkNetSurf(step: *std.build.LibExeObjStep) void {
+fn moduleNetSurf(b: *Build, mod: *Build.Module) !void {
+    const target = mod.resolved_target.?;
+    const os = target.result.os.tag;
+    const arch = target.result.cpu.arch;
 
     // iconv
-    step.addObjectFile(.{ .path = "vendor/libiconv/lib/libiconv.a" });
-    step.addIncludePath(.{ .path = "vendor/libiconv/include" });
+    const libiconv_lib_path = try std.fmt.allocPrint(
+        b.allocator,
+        "vendor/libiconv/out/{s}-{s}/lib/libiconv.a",
+        .{ @tagName(os), @tagName(arch) },
+    );
+    const libiconv_include_path = try std.fmt.allocPrint(
+        b.allocator,
+        "vendor/libiconv/out/{s}-{s}/lib/libiconv.a",
+        .{ @tagName(os), @tagName(arch) },
+    );
+    mod.addObjectFile(b.path(libiconv_lib_path));
+    mod.addIncludePath(b.path(libiconv_include_path));
+
+    {
+        // mimalloc
+        const mimalloc = "vendor/mimalloc";
+        const lib_path = try std.fmt.allocPrint(
+            b.allocator,
+            mimalloc ++ "/out/{s}-{s}/lib/libmimalloc.a",
+            .{ @tagName(os), @tagName(arch) },
+        );
+        mod.addObjectFile(b.path(lib_path));
+        mod.addIncludePath(b.path(mimalloc ++ "/include"));
+    }
 
     // netsurf libs
-    const ns = "vendor/netsurf/";
+    const ns = "vendor/netsurf";
+    const ns_include_path = try std.fmt.allocPrint(
+        b.allocator,
+        ns ++ "/out/{s}-{s}/include",
+        .{ @tagName(os), @tagName(arch) },
+    );
+    mod.addIncludePath(b.path(ns_include_path));
+
     const libs: [4][]const u8 = .{
         "libdom",
         "libhubbub",
@@ -171,8 +442,384 @@ fn linkNetSurf(step: *std.build.LibExeObjStep) void {
         "libwapcaplet",
     };
     inline for (libs) |lib| {
-        step.addObjectFile(.{ .path = ns ++ "/lib/" ++ lib ++ ".a" });
-        step.addIncludePath(.{ .path = ns ++ lib ++ "/src" });
+        const ns_lib_path = try std.fmt.allocPrint(
+            b.allocator,
+            ns ++ "/out/{s}-{s}/lib/" ++ lib ++ ".a",
+            .{ @tagName(os), @tagName(arch) },
+        );
+        mod.addObjectFile(b.path(ns_lib_path));
+        mod.addIncludePath(b.path(ns ++ "/" ++ lib ++ "/src"));
     }
-    step.addIncludePath(.{ .path = ns ++ "/include" });
+}
+
+fn buildZlib(b: *Build, m: *Build.Module) !void {
+    const zlib = b.addLibrary(.{
+        .name = "zlib",
+        .root_module = m,
+    });
+
+    const root = "vendor/zlib/";
+    zlib.installHeader(b.path(root ++ "zlib.h"), "zlib.h");
+    zlib.installHeader(b.path(root ++ "zconf.h"), "zconf.h");
+    zlib.addCSourceFiles(.{ .flags = &.{
+        "-DHAVE_SYS_TYPES_H",
+        "-DHAVE_STDINT_H",
+        "-DHAVE_STDDEF_H",
+    }, .files = &.{
+        root ++ "adler32.c",
+        root ++ "compress.c",
+        root ++ "crc32.c",
+        root ++ "deflate.c",
+        root ++ "gzclose.c",
+        root ++ "gzlib.c",
+        root ++ "gzread.c",
+        root ++ "gzwrite.c",
+        root ++ "inflate.c",
+        root ++ "infback.c",
+        root ++ "inftrees.c",
+        root ++ "inffast.c",
+        root ++ "trees.c",
+        root ++ "uncompr.c",
+        root ++ "zutil.c",
+    } });
+}
+
+fn buildMbedtls(b: *Build, m: *Build.Module) !void {
+    const mbedtls = b.addLibrary(.{
+        .name = "mbedtls",
+        .root_module = m,
+    });
+
+    const root = "vendor/mbedtls/";
+    mbedtls.addIncludePath(b.path(root ++ "include"));
+    mbedtls.addIncludePath(b.path(root ++ "library"));
+
+    mbedtls.addCSourceFiles(.{ .flags = &.{}, .files = &.{
+        root ++ "library/aes.c",
+        root ++ "library/aesni.c",
+        root ++ "library/aesce.c",
+        root ++ "library/aria.c",
+        root ++ "library/asn1parse.c",
+        root ++ "library/asn1write.c",
+        root ++ "library/base64.c",
+        root ++ "library/bignum.c",
+        root ++ "library/bignum_core.c",
+        root ++ "library/bignum_mod.c",
+        root ++ "library/bignum_mod_raw.c",
+        root ++ "library/camellia.c",
+        root ++ "library/ccm.c",
+        root ++ "library/chacha20.c",
+        root ++ "library/chachapoly.c",
+        root ++ "library/cipher.c",
+        root ++ "library/cipher_wrap.c",
+        root ++ "library/constant_time.c",
+        root ++ "library/cmac.c",
+        root ++ "library/ctr_drbg.c",
+        root ++ "library/des.c",
+        root ++ "library/dhm.c",
+        root ++ "library/ecdh.c",
+        root ++ "library/ecdsa.c",
+        root ++ "library/ecjpake.c",
+        root ++ "library/ecp.c",
+        root ++ "library/ecp_curves.c",
+        root ++ "library/entropy.c",
+        root ++ "library/entropy_poll.c",
+        root ++ "library/error.c",
+        root ++ "library/gcm.c",
+        root ++ "library/hkdf.c",
+        root ++ "library/hmac_drbg.c",
+        root ++ "library/lmots.c",
+        root ++ "library/lms.c",
+        root ++ "library/md.c",
+        root ++ "library/md5.c",
+        root ++ "library/memory_buffer_alloc.c",
+        root ++ "library/nist_kw.c",
+        root ++ "library/oid.c",
+        root ++ "library/padlock.c",
+        root ++ "library/pem.c",
+        root ++ "library/pk.c",
+        root ++ "library/pk_ecc.c",
+        root ++ "library/pk_wrap.c",
+        root ++ "library/pkcs12.c",
+        root ++ "library/pkcs5.c",
+        root ++ "library/pkparse.c",
+        root ++ "library/pkwrite.c",
+        root ++ "library/platform.c",
+        root ++ "library/platform_util.c",
+        root ++ "library/poly1305.c",
+        root ++ "library/psa_crypto.c",
+        root ++ "library/psa_crypto_aead.c",
+        root ++ "library/psa_crypto_cipher.c",
+        root ++ "library/psa_crypto_client.c",
+        root ++ "library/psa_crypto_ffdh.c",
+        root ++ "library/psa_crypto_driver_wrappers_no_static.c",
+        root ++ "library/psa_crypto_ecp.c",
+        root ++ "library/psa_crypto_hash.c",
+        root ++ "library/psa_crypto_mac.c",
+        root ++ "library/psa_crypto_pake.c",
+        root ++ "library/psa_crypto_rsa.c",
+        root ++ "library/psa_crypto_se.c",
+        root ++ "library/psa_crypto_slot_management.c",
+        root ++ "library/psa_crypto_storage.c",
+        root ++ "library/psa_its_file.c",
+        root ++ "library/psa_util.c",
+        root ++ "library/ripemd160.c",
+        root ++ "library/rsa.c",
+        root ++ "library/rsa_alt_helpers.c",
+        root ++ "library/sha1.c",
+        root ++ "library/sha3.c",
+        root ++ "library/sha256.c",
+        root ++ "library/sha512.c",
+        root ++ "library/threading.c",
+        root ++ "library/timing.c",
+        root ++ "library/version.c",
+        root ++ "library/version_features.c",
+        root ++ "library/pkcs7.c",
+        root ++ "library/x509.c",
+        root ++ "library/x509_create.c",
+        root ++ "library/x509_crl.c",
+        root ++ "library/x509_crt.c",
+        root ++ "library/x509_csr.c",
+        root ++ "library/x509write.c",
+        root ++ "library/x509write_crt.c",
+        root ++ "library/x509write_csr.c",
+        root ++ "library/debug.c",
+        root ++ "library/mps_reader.c",
+        root ++ "library/mps_trace.c",
+        root ++ "library/net_sockets.c",
+        root ++ "library/ssl_cache.c",
+        root ++ "library/ssl_ciphersuites.c",
+        root ++ "library/ssl_client.c",
+        root ++ "library/ssl_cookie.c",
+        root ++ "library/ssl_debug_helpers_generated.c",
+        root ++ "library/ssl_msg.c",
+        root ++ "library/ssl_ticket.c",
+        root ++ "library/ssl_tls.c",
+        root ++ "library/ssl_tls12_client.c",
+        root ++ "library/ssl_tls12_server.c",
+        root ++ "library/ssl_tls13_keys.c",
+        root ++ "library/ssl_tls13_server.c",
+        root ++ "library/ssl_tls13_client.c",
+        root ++ "library/ssl_tls13_generic.c",
+    } });
+}
+
+fn buildNghttp2(b: *Build, m: *Build.Module) !void {
+    const nghttp2 = b.addLibrary(.{
+        .name = "nghttp2",
+        .root_module = m,
+    });
+
+    const root = "vendor/nghttp2/";
+    nghttp2.addIncludePath(b.path(root ++ "lib"));
+    nghttp2.addIncludePath(b.path(root ++ "lib/includes"));
+    nghttp2.addCSourceFiles(.{ .flags = &.{
+        "-DNGHTTP2_STATICLIB",
+        "-DHAVE_NETINET_IN",
+        "-DHAVE_TIME_H",
+    }, .files = &.{
+        root ++ "lib/sfparse.c",
+        root ++ "lib/nghttp2_alpn.c",
+        root ++ "lib/nghttp2_buf.c",
+        root ++ "lib/nghttp2_callbacks.c",
+        root ++ "lib/nghttp2_debug.c",
+        root ++ "lib/nghttp2_extpri.c",
+        root ++ "lib/nghttp2_frame.c",
+        root ++ "lib/nghttp2_hd.c",
+        root ++ "lib/nghttp2_hd_huffman.c",
+        root ++ "lib/nghttp2_hd_huffman_data.c",
+        root ++ "lib/nghttp2_helper.c",
+        root ++ "lib/nghttp2_http.c",
+        root ++ "lib/nghttp2_map.c",
+        root ++ "lib/nghttp2_mem.c",
+        root ++ "lib/nghttp2_option.c",
+        root ++ "lib/nghttp2_outbound_item.c",
+        root ++ "lib/nghttp2_pq.c",
+        root ++ "lib/nghttp2_priority_spec.c",
+        root ++ "lib/nghttp2_queue.c",
+        root ++ "lib/nghttp2_rcbuf.c",
+        root ++ "lib/nghttp2_session.c",
+        root ++ "lib/nghttp2_stream.c",
+        root ++ "lib/nghttp2_submit.c",
+        root ++ "lib/nghttp2_version.c",
+        root ++ "lib/nghttp2_ratelim.c",
+        root ++ "lib/nghttp2_time.c",
+    } });
+}
+
+fn buildCurl(b: *Build, m: *Build.Module) !void {
+    const curl = b.addLibrary(.{
+        .name = "curl",
+        .root_module = m,
+    });
+
+    const root = "vendor/curl/";
+
+    curl.addIncludePath(b.path(root ++ "lib"));
+    curl.addIncludePath(b.path(root ++ "include"));
+    curl.addCSourceFiles(.{
+        .flags = &.{},
+        .files = &.{
+            root ++ "lib/altsvc.c",
+            root ++ "lib/amigaos.c",
+            root ++ "lib/asyn-ares.c",
+            root ++ "lib/asyn-base.c",
+            root ++ "lib/asyn-thrdd.c",
+            root ++ "lib/bufq.c",
+            root ++ "lib/bufref.c",
+            root ++ "lib/cf-h1-proxy.c",
+            root ++ "lib/cf-h2-proxy.c",
+            root ++ "lib/cf-haproxy.c",
+            root ++ "lib/cf-https-connect.c",
+            root ++ "lib/cf-socket.c",
+            root ++ "lib/cfilters.c",
+            root ++ "lib/conncache.c",
+            root ++ "lib/connect.c",
+            root ++ "lib/content_encoding.c",
+            root ++ "lib/cookie.c",
+            root ++ "lib/cshutdn.c",
+            root ++ "lib/curl_addrinfo.c",
+            root ++ "lib/curl_des.c",
+            root ++ "lib/curl_endian.c",
+            root ++ "lib/curl_fnmatch.c",
+            root ++ "lib/curl_get_line.c",
+            root ++ "lib/curl_gethostname.c",
+            root ++ "lib/curl_gssapi.c",
+            root ++ "lib/curl_memrchr.c",
+            root ++ "lib/curl_ntlm_core.c",
+            root ++ "lib/curl_range.c",
+            root ++ "lib/curl_rtmp.c",
+            root ++ "lib/curl_sasl.c",
+            root ++ "lib/curl_sha512_256.c",
+            root ++ "lib/curl_sspi.c",
+            root ++ "lib/curl_threads.c",
+            root ++ "lib/curl_trc.c",
+            root ++ "lib/cw-out.c",
+            root ++ "lib/cw-pause.c",
+            root ++ "lib/dict.c",
+            root ++ "lib/doh.c",
+            root ++ "lib/dynhds.c",
+            root ++ "lib/easy.c",
+            root ++ "lib/easygetopt.c",
+            root ++ "lib/easyoptions.c",
+            root ++ "lib/escape.c",
+            root ++ "lib/fake_addrinfo.c",
+            root ++ "lib/file.c",
+            root ++ "lib/fileinfo.c",
+            root ++ "lib/fopen.c",
+            root ++ "lib/formdata.c",
+            root ++ "lib/ftp.c",
+            root ++ "lib/ftplistparser.c",
+            root ++ "lib/getenv.c",
+            root ++ "lib/getinfo.c",
+            root ++ "lib/gopher.c",
+            root ++ "lib/hash.c",
+            root ++ "lib/headers.c",
+            root ++ "lib/hmac.c",
+            root ++ "lib/hostip.c",
+            root ++ "lib/hostip4.c",
+            root ++ "lib/hostip6.c",
+            root ++ "lib/hsts.c",
+            root ++ "lib/http.c",
+            root ++ "lib/http1.c",
+            root ++ "lib/http2.c",
+            root ++ "lib/http_aws_sigv4.c",
+            root ++ "lib/http_chunks.c",
+            root ++ "lib/http_digest.c",
+            root ++ "lib/http_negotiate.c",
+            root ++ "lib/http_ntlm.c",
+            root ++ "lib/http_proxy.c",
+            root ++ "lib/httpsrr.c",
+            root ++ "lib/idn.c",
+            root ++ "lib/if2ip.c",
+            root ++ "lib/imap.c",
+            root ++ "lib/krb5.c",
+            root ++ "lib/ldap.c",
+            root ++ "lib/llist.c",
+            root ++ "lib/macos.c",
+            root ++ "lib/md4.c",
+            root ++ "lib/md5.c",
+            root ++ "lib/memdebug.c",
+            root ++ "lib/mime.c",
+            root ++ "lib/mprintf.c",
+            root ++ "lib/mqtt.c",
+            root ++ "lib/multi.c",
+            root ++ "lib/multi_ev.c",
+            root ++ "lib/netrc.c",
+            root ++ "lib/noproxy.c",
+            root ++ "lib/openldap.c",
+            root ++ "lib/parsedate.c",
+            root ++ "lib/pingpong.c",
+            root ++ "lib/pop3.c",
+            root ++ "lib/progress.c",
+            root ++ "lib/psl.c",
+            root ++ "lib/rand.c",
+            root ++ "lib/rename.c",
+            root ++ "lib/request.c",
+            root ++ "lib/rtsp.c",
+            root ++ "lib/select.c",
+            root ++ "lib/sendf.c",
+            root ++ "lib/setopt.c",
+            root ++ "lib/sha256.c",
+            root ++ "lib/share.c",
+            root ++ "lib/slist.c",
+            root ++ "lib/smb.c",
+            root ++ "lib/smtp.c",
+            root ++ "lib/socketpair.c",
+            root ++ "lib/socks.c",
+            root ++ "lib/socks_gssapi.c",
+            root ++ "lib/socks_sspi.c",
+            root ++ "lib/speedcheck.c",
+            root ++ "lib/splay.c",
+            root ++ "lib/strcase.c",
+            root ++ "lib/strdup.c",
+            root ++ "lib/strequal.c",
+            root ++ "lib/strerror.c",
+            root ++ "lib/system_win32.c",
+            root ++ "lib/telnet.c",
+            root ++ "lib/tftp.c",
+            root ++ "lib/transfer.c",
+            root ++ "lib/uint-bset.c",
+            root ++ "lib/uint-hash.c",
+            root ++ "lib/uint-spbset.c",
+            root ++ "lib/uint-table.c",
+            root ++ "lib/url.c",
+            root ++ "lib/urlapi.c",
+            root ++ "lib/version.c",
+            root ++ "lib/ws.c",
+            root ++ "lib/curlx/base64.c",
+            root ++ "lib/curlx/dynbuf.c",
+            root ++ "lib/curlx/inet_ntop.c",
+            root ++ "lib/curlx/nonblock.c",
+            root ++ "lib/curlx/strparse.c",
+            root ++ "lib/curlx/timediff.c",
+            root ++ "lib/curlx/timeval.c",
+            root ++ "lib/curlx/wait.c",
+            root ++ "lib/curlx/warnless.c",
+            root ++ "lib/vquic/curl_ngtcp2.c",
+            root ++ "lib/vquic/curl_osslq.c",
+            root ++ "lib/vquic/curl_quiche.c",
+            root ++ "lib/vquic/vquic.c",
+            root ++ "lib/vquic/vquic-tls.c",
+            root ++ "lib/vauth/cleartext.c",
+            root ++ "lib/vauth/cram.c",
+            root ++ "lib/vauth/digest.c",
+            root ++ "lib/vauth/digest_sspi.c",
+            root ++ "lib/vauth/gsasl.c",
+            root ++ "lib/vauth/krb5_gssapi.c",
+            root ++ "lib/vauth/krb5_sspi.c",
+            root ++ "lib/vauth/ntlm.c",
+            root ++ "lib/vauth/ntlm_sspi.c",
+            root ++ "lib/vauth/oauth2.c",
+            root ++ "lib/vauth/spnego_gssapi.c",
+            root ++ "lib/vauth/spnego_sspi.c",
+            root ++ "lib/vauth/vauth.c",
+            root ++ "lib/vtls/cipher_suite.c",
+            root ++ "lib/vtls/mbedtls.c",
+            root ++ "lib/vtls/mbedtls_threadlock.c",
+            root ++ "lib/vtls/vtls.c",
+            root ++ "lib/vtls/vtls_scache.c",
+            root ++ "lib/vtls/x509asn1.c",
+        },
+    });
 }

build.zig.zon

@@ -0,0 +1,13 @@
+.{
+    .name = .browser,
+    .paths = .{""},
+    .version = "0.0.0",
+    .fingerprint = 0xda130f3af836cea0,
+    .dependencies = .{
+        .v8 = .{
+            .url = "https://github.com/lightpanda-io/zig-v8-fork/archive/7177ee1ae267a44751a0e7e012e257177699a375.tar.gz",
+            .hash = "v8-0.0.0-xddH63TCAwC1D1hEiOtbEnLBbtz9ZPHrdiGWLcBcYQB7",
+        },
+        // .v8 = .{ .path = "../zig-v8-fork" }
+    },
+}

flake.lock

@@ -0,0 +1,180 @@
+{
+  "nodes": {
+    "flake-compat": {
+      "flake": false,
+      "locked": {
+        "lastModified": 1696426674,
+        "narHash": "sha256-kvjfFW7WAETZlt09AgDn1MrtKzP7t90Vf7vypd3OL1U=",
+        "owner": "edolstra",
+        "repo": "flake-compat",
+        "rev": "0f9255e01c2351cc7d116c072cb317785dd33b33",
+        "type": "github"
+      },
+      "original": {
+        "owner": "edolstra",
+        "repo": "flake-compat",
+        "type": "github"
+      }
+    },
+    "flake-utils": {
+      "inputs": {
+        "systems": "systems"
+      },
+      "locked": {
+        "lastModified": 1731533236,
+        "narHash": "sha256-l0KFg5HjrsfsO/JpG+r7fRrqm12kzFHyUHqHCVpMMbI=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "11707dc2f618dd54ca8739b309ec4fc024de578b",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "flake-utils_2": {
+      "inputs": {
+        "systems": "systems_2"
+      },
+      "locked": {
+        "lastModified": 1705309234,
+        "narHash": "sha256-uNRRNRKmJyCRC/8y1RqBkqWBLM034y4qN7EprSdmgyA=",
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "rev": "1ef2e671c3b0c19053962c07dbda38332dcebf26",
+        "type": "github"
+      },
+      "original": {
+        "owner": "numtide",
+        "repo": "flake-utils",
+        "type": "github"
+      }
+    },
+    "gitignore": {
+      "inputs": {
+        "nixpkgs": [
+          "zlsPkg",
+          "nixpkgs"
+        ]
+      },
+      "locked": {
+        "lastModified": 1709087332,
+        "narHash": "sha256-HG2cCnktfHsKV0s4XW83gU3F57gaTljL9KNSuG6bnQs=",
+        "owner": "hercules-ci",
+        "repo": "gitignore.nix",
+        "rev": "637db329424fd7e46cf4185293b9cc8c88c95394",
+        "type": "github"
+      },
+      "original": {
+        "owner": "hercules-ci",
+        "repo": "gitignore.nix",
+        "type": "github"
+      }
+    },
+    "nixpkgs": {
+      "locked": {
+        "lastModified": 1756822655,
+        "narHash": "sha256-xQAk8xLy7srAkR5NMZFsQFioL02iTHuuEIs3ohGpgdk=",
+        "owner": "nixos",
+        "repo": "nixpkgs",
+        "rev": "4bdac60bfe32c41103ae500ddf894c258291dd61",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nixos",
+        "ref": "release-25.05",
+        "repo": "nixpkgs",
+        "type": "github"
+      }
+    },
+    "root": {
+      "inputs": {
+        "flake-utils": "flake-utils",
+        "nixpkgs": "nixpkgs",
+        "zigPkgs": "zigPkgs",
+        "zlsPkg": "zlsPkg"
+      }
+    },
+    "systems": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    },
+    "systems_2": {
+      "locked": {
+        "lastModified": 1681028828,
+        "narHash": "sha256-Vy1rq5AaRuLzOxct8nz4T6wlgyUR7zLU309k9mBC768=",
+        "owner": "nix-systems",
+        "repo": "default",
+        "rev": "da67096a3b9bf56a91d16901293e51ba5b49a27e",
+        "type": "github"
+      },
+      "original": {
+        "owner": "nix-systems",
+        "repo": "default",
+        "type": "github"
+      }
+    },
+    "zigPkgs": {
+      "inputs": {
+        "flake-compat": "flake-compat",
+        "flake-utils": "flake-utils_2",
+        "nixpkgs": [
+          "nixpkgs"
+        ]
+      },
+      "locked": {
+        "lastModified": 1756555914,
+        "narHash": "sha256-7yoSPIVEuL+3Wzf6e7NHuW3zmruHizRrYhGerjRHTLI=",
+        "owner": "mitchellh",
+        "repo": "zig-overlay",
+        "rev": "d0df3a2fd0f11134409d6d5ea0e510e5e477f7d6",
+        "type": "github"
+      },
+      "original": {
+        "owner": "mitchellh",
+        "repo": "zig-overlay",
+        "type": "github"
+      }
+    },
+    "zlsPkg": {
+      "inputs": {
+        "gitignore": "gitignore",
+        "nixpkgs": [
+          "nixpkgs"
+        ],
+        "zig-overlay": [
+          "zigPkgs"
+        ]
+      },
+      "locked": {
+        "lastModified": 1756048867,
+        "narHash": "sha256-GFzSHUljcxy7sM1PaabbkQUdUnLwpherekPWJFxXtnk=",
+        "owner": "zigtools",
+        "repo": "zls",
+        "rev": "ce6c8f02c78e622421cfc2405c67c5222819ec03",
+        "type": "github"
+      },
+      "original": {
+        "owner": "zigtools",
+        "ref": "0.15.0",
+        "repo": "zls",
+        "type": "github"
+      }
+    }
+  },
+  "root": "root",
+  "version": 7
+}

flake.nix

@@ -0,0 +1,77 @@
+{
+  description = "headless browser designed for AI and automation";
+
+  inputs = {
+    nixpkgs.url = "github:nixos/nixpkgs/release-25.05";
+
+    zigPkgs.url = "github:mitchellh/zig-overlay";
+    zigPkgs.inputs.nixpkgs.follows = "nixpkgs";
+
+    zlsPkg.url = "github:zigtools/zls/0.15.0";
+    zlsPkg.inputs.zig-overlay.follows = "zigPkgs";
+    zlsPkg.inputs.nixpkgs.follows = "nixpkgs";
+
+    flake-utils.url = "github:numtide/flake-utils";
+  };
+
+  outputs =
+    {
+      nixpkgs,
+      zigPkgs,
+      zlsPkg,
+      flake-utils,
+      ...
+    }:
+    flake-utils.lib.eachDefaultSystem (
+      system:
+      let
+        overlays = [
+          (final: prev: {
+            zigpkgs = zigPkgs.packages.${prev.system};
+            zls = zlsPkg.packages.${prev.system}.default;
+          })
+        ];
+
+        pkgs = import nixpkgs {
+          inherit system overlays;
+        };
+
+        # We need crtbeginS.o for building.
+        crtFiles = pkgs.runCommand "crt-files" { } ''
+          mkdir -p $out/lib
+          cp -r ${pkgs.gcc.cc}/lib/gcc $out/lib/gcc
+        '';
+
+        # This build pipeline is very unhappy without an FHS-compliant env.
+        fhs = pkgs.buildFHSEnv {
+          name = "fhs-shell";
+          multiArch = true;
+          targetPkgs =
+            pkgs: with pkgs; [
+              # Build Tools
+              zigpkgs."0.15.1"
+              zls
+              python3
+              pkg-config
+              cmake
+              gperf
+
+              # GCC
+              gcc
+              gcc.cc.lib
+              crtFiles
+
+              # Libaries
+              expat.dev
+              glib.dev
+              glibc.dev
+              zlib
+              zlib.dev
+            ];
+        };
+      in
+      {
+        devShells.default = fhs.env;
+      }
+    );
+}

src/TestHTTPServer.zig

@@ -0,0 +1,118 @@
+const std = @import("std");
+
+const TestHTTPServer = @This();
+
+shutdown: bool,
+listener: ?std.net.Server,
+handler: Handler,
+
+const Handler = *const fn (req: *std.http.Server.Request) anyerror!void;
+
+pub fn init(handler: Handler) TestHTTPServer {
+    return .{
+        .shutdown = true,
+        .listener = null,
+        .handler = handler,
+    };
+}
+
+pub fn deinit(self: *TestHTTPServer) void {
+    self.shutdown = true;
+    if (self.listener) |*listener| {
+        listener.deinit();
+    }
+}
+
+pub fn run(self: *TestHTTPServer, wg: *std.Thread.WaitGroup) !void {
+    const address = try std.net.Address.parseIp("127.0.0.1", 9582);
+
+    self.listener = try address.listen(.{ .reuse_address = true });
+    var listener = &self.listener.?;
+
+    wg.finish();
+
+    while (true) {
+        const conn = listener.accept() catch |err| {
+            if (self.shutdown) {
+                return;
+            }
+            return err;
+        };
+        const thrd = try std.Thread.spawn(.{}, handleConnection, .{ self, conn });
+        thrd.detach();
+    }
+}
+
+fn handleConnection(self: *TestHTTPServer, conn: std.net.Server.Connection) !void {
+    defer conn.stream.close();
+
+    var req_buf: [2048]u8 = undefined;
+    var conn_reader = conn.stream.reader(&req_buf);
+    var conn_writer = conn.stream.writer(&req_buf);
+
+    var http_server = std.http.Server.init(conn_reader.interface(), &conn_writer.interface);
+
+    while (true) {
+        var req = http_server.receiveHead() catch |err| switch (err) {
+            error.ReadFailed => continue,
+            error.HttpConnectionClosing => continue,
+            else => {
+                std.debug.print("Test HTTP Server error: {}\n", .{err});
+                return err;
+            },
+        };
+        self.handler(&req) catch |err| {
+            std.debug.print("test http error '{s}': {}\n", .{ req.head.target, err });
+            try req.respond("server error", .{ .status = .internal_server_error });
+            return;
+        };
+    }
+}
+
+pub fn sendFile(req: *std.http.Server.Request, file_path: []const u8) !void {
+    var file = std.fs.cwd().openFile(file_path, .{}) catch |err| switch (err) {
+        error.FileNotFound => return req.respond("server error", .{ .status = .not_found }),
+        else => return err,
+    };
+    defer file.close();
+
+    const stat = try file.stat();
+    var send_buffer: [4096]u8 = undefined;
+
+    var res = try req.respondStreaming(&send_buffer, .{
+        .content_length = stat.size,
+        .respond_options = .{
+            .extra_headers = &.{
+                .{ .name = "content-type", .value = getContentType(file_path) },
+            },
+        },
+    });
+
+    var read_buffer: [4096]u8 = undefined;
+    var reader = file.reader(&read_buffer);
+    _ = try res.writer.sendFileAll(&reader, .unlimited);
+    try res.writer.flush();
+    try res.end();
+}
+
+fn getContentType(file_path: []const u8) []const u8 {
+    if (std.mem.endsWith(u8, file_path, ".js")) {
+        return "application/json";
+    }
+
+    if (std.mem.endsWith(u8, file_path, ".html")) {
+        return "text/html";
+    }
+
+    if (std.mem.endsWith(u8, file_path, ".htm")) {
+        return "text/html";
+    }
+
+    if (std.mem.endsWith(u8, file_path, ".xml")) {
+        // some wpt tests do this
+        return "text/xml";
+    }
+
+    std.debug.print("TestHTTPServer asked to serve an unknown file type: {s}\n", .{file_path});
+    return "text/html";
+}

src/apiweb.zig

@@ -1,19 +0,0 @@
-const generate = @import("generate.zig");
-
-const Console = @import("jsruntime").Console;
-
-const DOM = @import("dom/dom.zig");
-const HTML = @import("html/html.zig");
-const Events = @import("events/event.zig");
-const XHR = @import("xhr/xhr.zig");
-
-pub const HTMLDocument = @import("html/document.zig").HTMLDocument;
-
-// Interfaces
-pub const Interfaces = generate.Tuple(.{
-    Console,
-    DOM.Interfaces,
-    Events.Interfaces,
-    HTML.Interfaces,
-    XHR.Interfaces,
-});

src/app.zig

@@ -0,0 +1,113 @@
+const std = @import("std");
+
+const Allocator = std.mem.Allocator;
+
+const log = @import("log.zig");
+const Http = @import("http/Http.zig");
+const Platform = @import("runtime/js.zig").Platform;
+
+const Telemetry = @import("telemetry/telemetry.zig").Telemetry;
+const Notification = @import("notification.zig").Notification;
+
+// Container for global state / objects that various parts of the system
+// might need.
+pub const App = struct {
+    http: Http,
+    config: Config,
+    platform: Platform,
+    allocator: Allocator,
+    telemetry: Telemetry,
+    app_dir_path: ?[]const u8,
+    notification: *Notification,
+
+    pub const RunMode = enum {
+        help,
+        fetch,
+        serve,
+        version,
+    };
+
+    pub const Config = struct {
+        run_mode: RunMode,
+        tls_verify_host: bool = true,
+        http_proxy: ?[:0]const u8 = null,
+        proxy_bearer_token: ?[:0]const u8 = null,
+        http_timeout_ms: ?u31 = null,
+        http_connect_timeout_ms: ?u31 = null,
+        http_max_host_open: ?u8 = null,
+        http_max_concurrent: ?u8 = null,
+    };
+
+    pub fn init(allocator: Allocator, config: Config) !*App {
+        const app = try allocator.create(App);
+        errdefer allocator.destroy(app);
+
+        const notification = try Notification.init(allocator, null);
+        errdefer notification.deinit();
+
+        var http = try Http.init(allocator, .{
+            .max_host_open = config.http_max_host_open orelse 4,
+            .max_concurrent = config.http_max_concurrent orelse 10,
+            .timeout_ms = config.http_timeout_ms orelse 5000,
+            .connect_timeout_ms = config.http_connect_timeout_ms orelse 0,
+            .http_proxy = config.http_proxy,
+            .tls_verify_host = config.tls_verify_host,
+            .proxy_bearer_token = config.proxy_bearer_token,
+        });
+        errdefer http.deinit();
+
+        const platform = try Platform.init();
+        errdefer platform.deinit();
+
+        const app_dir_path = getAndMakeAppDir(allocator);
+
+        app.* = .{
+            .http = http,
+            .allocator = allocator,
+            .telemetry = undefined,
+            .platform = platform,
+            .app_dir_path = app_dir_path,
+            .notification = notification,
+            .config = config,
+        };
+
+        app.telemetry = try Telemetry.init(app, config.run_mode);
+        errdefer app.telemetry.deinit();
+
+        try app.telemetry.register(app.notification);
+
+        return app;
+    }
+
+    pub fn deinit(self: *App) void {
+        const allocator = self.allocator;
+        if (self.app_dir_path) |app_dir_path| {
+            allocator.free(app_dir_path);
+        }
+        self.telemetry.deinit();
+        self.notification.deinit();
+        self.http.deinit();
+        self.platform.deinit();
+        allocator.destroy(self);
+    }
+};
+
+fn getAndMakeAppDir(allocator: Allocator) ?[]const u8 {
+    if (@import("builtin").is_test) {
+        return allocator.dupe(u8, "/tmp") catch unreachable;
+    }
+    const app_dir_path = std.fs.getAppDataDir(allocator, "lightpanda") catch |err| {
+        log.warn(.app, "get data dir", .{ .err = err });
+        return null;
+    };
+
+    std.fs.cwd().makePath(app_dir_path) catch |err| switch (err) {
+        error.PathAlreadyExists => return app_dir_path,
+        else => {
+            allocator.free(app_dir_path);
+            log.warn(.app, "create data dir", .{ .err = err, .path = app_dir_path });
+            return null;
+        },
+    };
+    return app_dir_path;
+}

src/async/stream.zig

@@ -1,115 +0,0 @@
-const std = @import("std");
-const builtin = @import("builtin");
-const os = std.os;
-const io = std.io;
-const assert = std.debug.assert;
-
-const tcp = @import("tcp.zig");
-
-pub const Stream = struct {
-    alloc: std.mem.Allocator,
-    conn: *tcp.Conn,
-
-    handle: std.os.socket_t,
-
-    pub fn close(self: Stream) void {
-        os.closeSocket(self.handle);
-        self.alloc.destroy(self.conn);
-    }
-
-    pub const ReadError = os.ReadError;
-    pub const WriteError = os.WriteError;
-
-    pub const Reader = io.Reader(Stream, ReadError, read);
-    pub const Writer = io.Writer(Stream, WriteError, write);
-
-    pub fn reader(self: Stream) Reader {
-        return .{ .context = self };
-    }
-
-    pub fn writer(self: Stream) Writer {
-        return .{ .context = self };
-    }
-
-    pub fn read(self: Stream, buffer: []u8) ReadError!usize {
-        return self.conn.receive(self.handle, buffer) catch |err| switch (err) {
-            else => return error.Unexpected,
-        };
-    }
-
-    pub fn readv(s: Stream, iovecs: []const os.iovec) ReadError!usize {
-        return os.readv(s.handle, iovecs);
-    }
-
-    /// Returns the number of bytes read. If the number read is smaller than
-    /// `buffer.len`, it means the stream reached the end. Reaching the end of
-    /// a stream is not an error condition.
-    pub fn readAll(s: Stream, buffer: []u8) ReadError!usize {
-        return readAtLeast(s, buffer, buffer.len);
-    }
-
-    /// Returns the number of bytes read, calling the underlying read function
-    /// the minimal number of times until the buffer has at least `len` bytes
-    /// filled. If the number read is less than `len` it means the stream
-    /// reached the end. Reaching the end of the stream is not an error
-    /// condition.
-    pub fn readAtLeast(s: Stream, buffer: []u8, len: usize) ReadError!usize {
-        assert(len <= buffer.len);
-        var index: usize = 0;
-        while (index < len) {
-            const amt = try s.read(buffer[index..]);
-            if (amt == 0) break;
-            index += amt;
-        }
-        return index;
-    }
-
-    /// TODO in evented I/O mode, this implementation incorrectly uses the event loop's
-    /// file system thread instead of non-blocking. It needs to be reworked to properly
-    /// use non-blocking I/O.
-    pub fn write(self: Stream, buffer: []const u8) WriteError!usize {
-        return self.conn.send(self.handle, buffer) catch |err| switch (err) {
-            error.AccessDenied => error.AccessDenied,
-            error.WouldBlock => error.WouldBlock,
-            error.ConnectionResetByPeer => error.ConnectionResetByPeer,
-            error.MessageTooBig => error.FileTooBig,
-            error.BrokenPipe => error.BrokenPipe,
-            else => return error.Unexpected,
-        };
-    }
-
-    pub fn writeAll(self: Stream, bytes: []const u8) WriteError!void {
-        var index: usize = 0;
-        while (index < bytes.len) {
-            index += try self.write(bytes[index..]);
-        }
-    }
-
-    /// See https://github.com/ziglang/zig/issues/7699
-    /// See equivalent function: `std.fs.File.writev`.
-    pub fn writev(self: Stream, iovecs: []const os.iovec_const) WriteError!usize {
-        if (iovecs.len == 0) return 0;
-        const first_buffer = iovecs[0].iov_base[0..iovecs[0].iov_len];
-        return try self.write(first_buffer);
-    }
-
-    /// The `iovecs` parameter is mutable because this function needs to mutate the fields in
-    /// order to handle partial writes from the underlying OS layer.
-    /// See https://github.com/ziglang/zig/issues/7699
-    /// See equivalent function: `std.fs.File.writevAll`.
-    pub fn writevAll(self: Stream, iovecs: []os.iovec_const) WriteError!void {
-        if (iovecs.len == 0) return;
-
-        var i: usize = 0;
-        while (true) {
-            var amt = try self.writev(iovecs[i..]);
-            while (amt >= iovecs[i].iov_len) {
-                amt -= iovecs[i].iov_len;
-                i += 1;
-                if (i >= iovecs.len) return;
-            }
-            iovecs[i].iov_base += amt;
-            iovecs[i].iov_len -= amt;
-        }
-    }
-};

src/async/tcp.zig

@@ -1,94 +0,0 @@
-const std = @import("std");
-const net = std.net;
-const Stream = @import("stream.zig").Stream;
-const Loop = @import("jsruntime").Loop;
-const NetworkImpl = Loop.Network(Conn.Command);
-
-// Conn is a TCP connection using jsruntime Loop async I/O.
-// connect, send and receive are blocking, but use async I/O in the background.
-// Client doesn't own the socket used for the connection, the caller is
-// responsible for closing it.
-pub const Conn = struct {
-    const Command = struct {
-        impl: NetworkImpl,
-
-        done: bool = false,
-        err: ?anyerror = null,
-        ln: usize = 0,
-
-        fn ok(self: *Command, err: ?anyerror, ln: usize) void {
-            self.err = err;
-            self.ln = ln;
-            self.done = true;
-        }
-
-        fn wait(self: *Command) !usize {
-            while (!self.done) try self.impl.tick();
-
-            if (self.err) |err| return err;
-            return self.ln;
-        }
-        pub fn onConnect(self: *Command, err: ?anyerror) void {
-            self.ok(err, 0);
-        }
-        pub fn onSend(self: *Command, ln: usize, err: ?anyerror) void {
-            self.ok(err, ln);
-        }
-        pub fn onReceive(self: *Command, ln: usize, err: ?anyerror) void {
-            self.ok(err, ln);
-        }
-    };
-
-    loop: *Loop,
-
-    pub fn connect(self: *Conn, socket: std.os.socket_t, address: std.net.Address) !void {
-        var cmd = Command{ .impl = NetworkImpl.init(self.loop) };
-        cmd.impl.connect(&cmd, socket, address);
-        _ = try cmd.wait();
-    }
-
-    pub fn send(self: *Conn, socket: std.os.socket_t, buffer: []const u8) !usize {
-        var cmd = Command{ .impl = NetworkImpl.init(self.loop) };
-        cmd.impl.send(&cmd, socket, buffer);
-        return try cmd.wait();
-    }
-
-    pub fn receive(self: *Conn, socket: std.os.socket_t, buffer: []u8) !usize {
-        var cmd = Command{ .impl = NetworkImpl.init(self.loop) };
-        cmd.impl.receive(&cmd, socket, buffer);
-        return try cmd.wait();
-    }
-};
-
-pub fn tcpConnectToHost(alloc: std.mem.Allocator, loop: *Loop, name: []const u8, port: u16) !Stream {
-    // TODO async resolve
-    const list = try net.getAddressList(alloc, name, port);
-    defer list.deinit();
-
-    if (list.addrs.len == 0) return error.UnknownHostName;
-
-    for (list.addrs) |addr| {
-        return tcpConnectToAddress(alloc, loop, addr) catch |err| switch (err) {
-            error.ConnectionRefused => {
-                continue;
-            },
-            else => return err,
-        };
-    }
-    return std.os.ConnectError.ConnectionRefused;
-}
-
-pub fn tcpConnectToAddress(alloc: std.mem.Allocator, loop: *Loop, addr: net.Address) !Stream {
-    const sockfd = try std.os.socket(addr.any.family, std.os.SOCK.STREAM, std.os.IPPROTO.TCP);
-    errdefer std.os.closeSocket(sockfd);
-
-    var conn = try alloc.create(Conn);
-    conn.* = Conn{ .loop = loop };
-    try conn.connect(sockfd, addr);
-
-    return Stream{
-        .alloc = alloc,
-        .conn = conn,
-        .handle = sockfd,
-    };
-}

src/async/test.zig

@@ -1,172 +0,0 @@
-const std = @import("std");
-const http = std.http;
-const Client = @import("Client.zig");
-const Request = @import("Client.zig").Request;
-
-pub const Loop = @import("jsruntime").Loop;
-
-const url = "https://w3.org";
-
-test "blocking mode fetch API" {
-    const alloc = std.testing.allocator;
-
-    var loop = try Loop.init(alloc);
-    defer loop.deinit();
-
-    var client: Client = .{
-        .allocator = alloc,
-        .loop = &loop,
-    };
-    defer client.deinit();
-
-    // force client's CA cert scan from system.
-    try client.ca_bundle.rescan(client.allocator);
-
-    var res = try client.fetch(alloc, .{
-        .location = .{ .uri = try std.Uri.parse(url) },
-        .payload = .none,
-    });
-    defer res.deinit();
-
-    try std.testing.expect(res.status == .ok);
-}
-
-test "blocking mode open/send/wait API" {
-    const alloc = std.testing.allocator;
-
-    var loop = try Loop.init(alloc);
-    defer loop.deinit();
-
-    var client: Client = .{
-        .allocator = alloc,
-        .loop = &loop,
-    };
-    defer client.deinit();
-
-    // force client's CA cert scan from system.
-    try client.ca_bundle.rescan(client.allocator);
-
-    var headers = try std.http.Headers.initList(alloc, &[_]std.http.Field{});
-    defer headers.deinit();
-
-    var req = try client.open(.GET, try std.Uri.parse(url), headers, .{});
-    defer req.deinit();
-
-    try req.send(.{});
-    try req.finish();
-    try req.wait();
-
-    try std.testing.expect(req.response.status == .ok);
-}
-
-// Example how to write an async http client using the modified standard client.
-const AsyncClient = struct {
-    cli: Client,
-
-    const YieldImpl = Loop.Yield(AsyncRequest);
-    const AsyncRequest = struct {
-        const State = enum { new, open, send, finish, wait, done };
-
-        cli: *Client,
-        uri: std.Uri,
-        headers: std.http.Headers,
-
-        req: ?Request = undefined,
-        state: State = .new,
-
-        impl: YieldImpl,
-        err: ?anyerror = null,
-
-        pub fn deinit(self: *AsyncRequest) void {
-            if (self.req) |*r| r.deinit();
-            self.headers.deinit();
-        }
-
-        pub fn fetch(self: *AsyncRequest) void {
-            self.state = .new;
-            return self.impl.yield(self);
-        }
-
-        fn onerr(self: *AsyncRequest, err: anyerror) void {
-            self.state = .done;
-            self.err = err;
-        }
-
-        pub fn onYield(self: *AsyncRequest, err: ?anyerror) void {
-            if (err) |e| return self.onerr(e);
-
-            switch (self.state) {
-                .new => {
-                    self.state = .open;
-                    self.req = self.cli.open(.GET, self.uri, self.headers, .{}) catch |e| return self.onerr(e);
-                },
-                .open => {
-                    self.state = .send;
-                    self.req.?.send(.{}) catch |e| return self.onerr(e);
-                },
-                .send => {
-                    self.state = .finish;
-                    self.req.?.finish() catch |e| return self.onerr(e);
-                },
-                .finish => {
-                    self.state = .wait;
-                    self.req.?.wait() catch |e| return self.onerr(e);
-                },
-                .wait => {
-                    self.state = .done;
-                    return;
-                },
-                .done => return,
-            }
-
-            return self.impl.yield(self);
-        }
-
-        pub fn wait(self: *AsyncRequest) !void {
-            while (self.state != .done) try self.impl.tick();
-            if (self.err) |err| return err;
-        }
-    };
-
-    pub fn init(alloc: std.mem.Allocator, loop: *Loop) AsyncClient {
-        return .{
-            .cli = .{
-                .allocator = alloc,
-                .loop = loop,
-            },
-        };
-    }
-
-    pub fn deinit(self: *AsyncClient) void {
-        self.cli.deinit();
-    }
-
-    pub fn createRequest(self: *AsyncClient, uri: std.Uri) !AsyncRequest {
-        return .{
-            .impl = YieldImpl.init(self.cli.loop),
-            .cli = &self.cli,
-            .uri = uri,
-            .headers = .{ .allocator = self.cli.allocator, .owned = false },
-        };
-    }
-};
-
-test "non blocking client" {
-    const alloc = std.testing.allocator;
-
-    var loop = try Loop.init(alloc);
-    defer loop.deinit();
-
-    var client = AsyncClient.init(alloc, &loop);
-    defer client.deinit();
-
-    var reqs: [3]AsyncClient.AsyncRequest = undefined;
-    for (0..reqs.len) |i| {
-        reqs[i] = try client.createRequest(try std.Uri.parse(url));
-        reqs[i].fetch();
-    }
-    for (0..reqs.len) |i| {
-        try reqs[i].wait();
-        reqs[i].deinit();
-    }
-}

src/browser/DataURI.zig

@@ -0,0 +1,52 @@
+const std = @import("std");
+const Allocator = std.mem.Allocator;
+
+// Parses data:[<media-type>][;base64],<data>
+pub fn parse(allocator: Allocator, src: []const u8) !?[]const u8 {
+    if (!std.mem.startsWith(u8, src, "data:")) {
+        return null;
+    }
+
+    const uri = src[5..];
+    const data_starts = std.mem.indexOfScalar(u8, uri, ',') orelse return null;
+
+    var data = uri[data_starts + 1 ..];
+
+    // Extract the encoding.
+    const metadata = uri[0..data_starts];
+    if (std.mem.endsWith(u8, metadata, ";base64")) {
+        const decoder = std.base64.standard.Decoder;
+        const decoded_size = try decoder.calcSizeForSlice(data);
+
+        const buffer = try allocator.alloc(u8, decoded_size);
+        errdefer allocator.free(buffer);
+
+        try decoder.decode(buffer, data);
+        data = buffer;
+    }
+
+    return data;
+}
+
+const testing = @import("../testing.zig");
+test "DataURI: parse valid" {
+    try test_valid("data:text/javascript; charset=utf-8;base64,Zm9v", "foo");
+    try test_valid("data:text/javascript; charset=utf-8;,foo", "foo");
+    try test_valid("data:,foo", "foo");
+}
+
+test "DataURI: parse invalid" {
+    try test_cannot_parse("atad:,foo");
+    try test_cannot_parse("data:foo");
+    try test_cannot_parse("data:");
+}
+
+fn test_valid(uri: []const u8, expected: []const u8) !void {
+    defer testing.reset();
+    const data_uri = try parse(testing.arena_allocator, uri) orelse return error.TestFailed;
+    try testing.expectEqual(expected, data_uri);
+}
+
+fn test_cannot_parse(uri: []const u8) !void {
+    try testing.expectEqual(null, parse(undefined, uri));
+}

src/browser/Scheduler.zig

@@ -0,0 +1,162 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../log.zig");
+const Allocator = std.mem.Allocator;
+
+const Scheduler = @This();
+
+high_priority: Queue,
+
+// For repeating tasks. We only want to run these if there are other things to
+// do. We don't, for example, want a window.setInterval or the page.runMicrotasks
+// to block the page.wait.
+low_priority: Queue,
+
+// we expect allocator to be the page arena, hence we never call high_priority.deinit
+pub fn init(allocator: Allocator) Scheduler {
+    return .{
+        .high_priority = Queue.init(allocator, {}),
+        .low_priority = Queue.init(allocator, {}),
+    };
+}
+
+pub fn reset(self: *Scheduler) void {
+    self.high_priority.clearRetainingCapacity();
+    self.low_priority.clearRetainingCapacity();
+}
+
+const AddOpts = struct {
+    name: []const u8 = "",
+    low_priority: bool = false,
+};
+pub fn add(self: *Scheduler, ctx: *anyopaque, func: Task.Func, ms: u32, opts: AddOpts) !void {
+    var low_priority = opts.low_priority;
+    if (ms > 5_000) {
+        // we don't want tasks in the far future to block page.wait from
+        // completing. However, if page.wait is called multiple times (maybe
+        // a CDP driver is wait for something to happen), then we do want
+        // to [eventually] run these when their time is up.
+        low_priority = true;
+    }
+
+    var q = if (low_priority) &self.low_priority else &self.high_priority;
+    return q.add(.{
+        .ms = std.time.milliTimestamp() + ms,
+        .ctx = ctx,
+        .func = func,
+        .name = opts.name,
+    });
+}
+
+pub fn run(self: *Scheduler) !?i32 {
+    _ = try self.runQueue(&self.low_priority);
+    return self.runQueue(&self.high_priority);
+}
+
+fn runQueue(self: *Scheduler, queue: *Queue) !?i32 {
+    // this is O(1)
+    if (queue.count() == 0) {
+        return null;
+    }
+
+    const now = std.time.milliTimestamp();
+
+    var next = queue.peek();
+    while (next) |task| {
+        const time_to_next = task.ms - now;
+        if (time_to_next > 0) {
+            // @intCast is petty safe since we limit tasks to just 5 seconds
+            // in the future
+            return @intCast(time_to_next);
+        }
+
+        if (task.func(task.ctx)) |repeat_delay| {
+            // if we do (now + 0) then our WHILE loop will run endlessly.
+            // no task should ever return 0
+            std.debug.assert(repeat_delay != 0);
+
+            var copy = task;
+            copy.ms = now + repeat_delay;
+            try self.low_priority.add(copy);
+        }
+        _ = queue.remove();
+        next = queue.peek();
+    }
+    return null;
+}
+
+const Task = struct {
+    ms: i64,
+    func: Func,
+    ctx: *anyopaque,
+    name: []const u8,
+
+    const Func = *const fn (ctx: *anyopaque) ?u32;
+};
+
+const Queue = std.PriorityQueue(Task, void, struct {
+    fn compare(_: void, a: Task, b: Task) std.math.Order {
+        return std.math.order(a.ms, b.ms);
+    }
+}.compare);
+
+const testing = @import("../testing.zig");
+test "Scheduler" {
+    defer testing.reset();
+
+    var task = TestTask{ .allocator = testing.arena_allocator };
+
+    var s = Scheduler.init(testing.arena_allocator);
+    try testing.expectEqual(null, s.run());
+    try testing.expectEqual(0, task.calls.items.len);
+
+    try s.add(&task, TestTask.run1, 3, .{});
+
+    try testing.expectDelta(3, try s.run(), 1);
+    try testing.expectEqual(0, task.calls.items.len);
+
+    std.Thread.sleep(std.time.ns_per_ms * 5);
+    try testing.expectEqual(null, s.run());
+    try testing.expectEqualSlices(u32, &.{1}, task.calls.items);
+
+    try s.add(&task, TestTask.run2, 3, .{});
+    try s.add(&task, TestTask.run1, 2, .{});
+
+    std.Thread.sleep(std.time.ns_per_ms * 5);
+    try testing.expectDelta(null, try s.run(), 1);
+    try testing.expectEqualSlices(u32, &.{ 1, 1, 2 }, task.calls.items);
+}
+
+const TestTask = struct {
+    allocator: Allocator,
+    calls: std.ArrayListUnmanaged(u32) = .{},
+
+    fn run1(ctx: *anyopaque) ?u32 {
+        var self: *TestTask = @ptrCast(@alignCast(ctx));
+        self.calls.append(self.allocator, 1) catch unreachable;
+        return null;
+    }
+
+    fn run2(ctx: *anyopaque) ?u32 {
+        var self: *TestTask = @ptrCast(@alignCast(ctx));
+        self.calls.append(self.allocator, 2) catch unreachable;
+        return 2;
+    }
+};

src/browser/ScriptManager.zig

@@ -0,0 +1,832 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../log.zig");
+const parser = @import("netsurf.zig");
+
+const Env = @import("env.zig").Env;
+const Page = @import("page.zig").Page;
+const DataURI = @import("DataURI.zig");
+const Http = @import("../http/Http.zig");
+const Browser = @import("browser.zig").Browser;
+const URL = @import("../url.zig").URL;
+
+const Allocator = std.mem.Allocator;
+const ArrayListUnmanaged = std.ArrayListUnmanaged;
+
+const ScriptManager = @This();
+
+page: *Page,
+
+// used to prevent recursive evalutaion
+is_evaluating: bool,
+
+// used to prevent executing scripts while we're doing a blocking load
+is_blocking: bool = false,
+
+// Only once this is true can deferred scripts be run
+static_scripts_done: bool,
+
+// List of async scripts. We don't care about the execution order of these, but
+// on shutdown/abort, we need to cleanup any pending ones.
+asyncs: OrderList,
+
+// When an async script is ready to be evaluated, it's moved from asyncs to
+// this list. You might think we can evaluate an async script as soon as it's
+// done, but we can only evaluate scripts when `is_blocking == false`. So this
+// becomes a list of scripts to execute on the next evaluate().
+asyncs_ready: OrderList,
+
+// Normal scripts (non-deferred & non-async). These must be executed in order
+scripts: OrderList,
+
+// List of deferred scripts. These must be executed in order, but only once
+// dom_loaded == true,
+deferreds: OrderList,
+
+shutdown: bool = false,
+
+client: *Http.Client,
+allocator: Allocator,
+buffer_pool: BufferPool,
+script_pool: std.heap.MemoryPool(PendingScript),
+
+const OrderList = std.DoublyLinkedList;
+
+pub fn init(browser: *Browser, page: *Page) ScriptManager {
+    // page isn't fully initialized, we can setup our reference, but that's it.
+    const allocator = browser.allocator;
+    return .{
+        .page = page,
+        .asyncs = .{},
+        .scripts = .{},
+        .deferreds = .{},
+        .asyncs_ready = .{},
+        .is_evaluating = false,
+        .allocator = allocator,
+        .client = browser.http_client,
+        .static_scripts_done = false,
+        .buffer_pool = BufferPool.init(allocator, 5),
+        .script_pool = std.heap.MemoryPool(PendingScript).init(allocator),
+    };
+}
+
+pub fn deinit(self: *ScriptManager) void {
+    self.reset();
+    self.buffer_pool.deinit();
+    self.script_pool.deinit();
+}
+
+pub fn reset(self: *ScriptManager) void {
+    self.clearList(&self.asyncs);
+    self.clearList(&self.scripts);
+    self.clearList(&self.deferreds);
+    self.clearList(&self.asyncs_ready);
+    self.static_scripts_done = false;
+}
+
+fn clearList(_: *const ScriptManager, list: *OrderList) void {
+    while (list.first) |node| {
+        const pending_script: *PendingScript = @fieldParentPtr("node", node);
+        // this removes it from the list
+        pending_script.deinit();
+    }
+    std.debug.assert(list.first == null);
+}
+
+pub fn addFromElement(self: *ScriptManager, element: *parser.Element) !void {
+    if (try parser.elementGetAttribute(element, "nomodule") != null) {
+        // these scripts should only be loaded if we don't support modules
+        // but since we do support modules, we can just skip them.
+        return;
+    }
+
+    // If a script tag gets dynamically created and added to the dom:
+    //    document.getElementsByTagName('head')[0].appendChild(script)
+    // that script tag will immediately get executed by our scriptAddedCallback.
+    // However, if the location where the script tag is inserted happens to be
+    // below where processHTMLDoc currently is, then we'll re-run that same script
+    // again in processHTMLDoc. This flag is used to let us know if a specific
+    // <script> has already been processed.
+    if (try parser.scriptGetProcessed(@ptrCast(element))) {
+        return;
+    }
+    try parser.scriptSetProcessed(@ptrCast(element), true);
+
+    const kind: Script.Kind = blk: {
+        const script_type = try parser.elementGetAttribute(element, "type") orelse break :blk .javascript;
+        if (script_type.len == 0) {
+            break :blk .javascript;
+        }
+        if (std.ascii.eqlIgnoreCase(script_type, "application/javascript")) {
+            break :blk .javascript;
+        }
+        if (std.ascii.eqlIgnoreCase(script_type, "text/javascript")) {
+            break :blk .javascript;
+        }
+        if (std.ascii.eqlIgnoreCase(script_type, "module")) {
+            break :blk .module;
+        }
+
+        // "type" could be anything, but only the above are ones we need to process.
+        // Common other ones are application/json, application/ld+json, text/template
+
+        return;
+    };
+
+    const page = self.page;
+    var source: Script.Source = undefined;
+    var remote_url: ?[:0]const u8 = null;
+    if (try parser.elementGetAttribute(element, "src")) |src| {
+        if (try DataURI.parse(page.arena, src)) |data_uri| {
+            source = .{ .@"inline" = data_uri };
+        }
+        remote_url = try URL.stitch(page.arena, src, page.url.raw, .{ .null_terminated = true });
+        source = .{ .remote = .{} };
+    } else {
+        const inline_source = try parser.nodeTextContent(@ptrCast(element)) orelse return;
+        source = .{ .@"inline" = inline_source };
+    }
+
+    var script = Script{
+        .kind = kind,
+        .element = element,
+        .source = source,
+        .url = remote_url orelse page.url.raw,
+        .is_defer = if (remote_url == null) false else try parser.elementGetAttribute(element, "defer") != null,
+        .is_async = if (remote_url == null) false else try parser.elementGetAttribute(element, "async") != null,
+    };
+
+    if (source == .@"inline" and self.scripts.first == null) {
+        // inline script with no pending scripts, execute it immediately.
+        // (if there is a pending script, then we cannot execute this immediately
+        // as it needs to best executed in order)
+        return script.eval(page);
+    }
+
+    const pending_script = try self.script_pool.create();
+    errdefer self.script_pool.destroy(pending_script);
+    pending_script.* = .{
+        .script = script,
+        .complete = false,
+        .manager = self,
+        .node = .{},
+    };
+
+    if (source == .@"inline") {
+        // if we're here, it means that we have pending scripts (i.e. self.scripts
+        // is not empty). Because the script is inline, it's complete/ready, but
+        // we need to process them in order
+        pending_script.complete = true;
+        self.scripts.append(&pending_script.node);
+        return;
+    } else {
+        log.debug(.http, "script queue", .{ .url = remote_url.? });
+    }
+
+    pending_script.getList().append(&pending_script.node);
+
+    errdefer pending_script.deinit();
+
+    var headers = try Http.Headers.init();
+    try page.requestCookie(.{}).headersForRequest(page.arena, remote_url.?, &headers);
+
+    try self.client.request(.{
+        .url = remote_url.?,
+        .ctx = pending_script,
+        .method = .GET,
+        .headers = headers,
+        .cookie_jar = page.cookie_jar,
+        .resource_type = .script,
+        .start_callback = if (log.enabled(.http, .debug)) startCallback else null,
+        .header_callback = headerCallback,
+        .data_callback = dataCallback,
+        .done_callback = doneCallback,
+        .error_callback = errorCallback,
+    });
+}
+
+// @TODO: Improving this would have the simplest biggest performance improvement
+// for most sites.
+//
+// For JS imports (both static and dynamic), we currently block to get the
+// result (the content of the file).
+//
+// For static imports, this is necessary, since v8 is expecting the compiled module
+// as part of the function return. (we should try to pre-load the JavaScript
+// source via module.GetModuleRequests(), but that's for a later time).
+//
+// For dynamic dynamic imports, this is not strictly necessary since the v8
+// call returns a Promise; we could make this a normal get call, associated with
+// the promise, and when done, resolve the promise.
+//
+// In both cases, for now at least, we just issue a "blocking" request. We block
+// by ticking the http client until the script is complete.
+//
+// This uses the client.blockingRequest call which has a dedicated handle for
+// these blocking requests. Because they are blocking, we're guaranteed to have
+// only 1 at a time, thus the 1 reserved handle.
+//
+// You almost don't need the http client's blocking handle. In most cases, you
+// should always have 1 free handle whenever you get here, because we always
+// release the handle before executing the doneCallback. So, if a module does:
+//    import * as x from 'blah'
+// And we need to load 'blah', there should always be 1 free handle - the handle
+// of the http GET we just completed before executing the module.
+// The exception to this, and the reason we need a special blocking handle, is
+// for inline modules within the HTML page itself:
+//    <script type=module>import ....</script>
+// Unlike external modules which can only ever be executed after releasing an
+// http handle, these are executed without there necessarily being a free handle.
+// Thus, Http/Client.zig maintains a dedicated handle for these calls.
+pub fn blockingGet(self: *ScriptManager, url: [:0]const u8) !BlockingResult {
+    std.debug.assert(self.is_blocking == false);
+
+    self.is_blocking = true;
+    defer {
+        self.is_blocking = false;
+
+        // we blocked evaluation while loading this script, there could be
+        // scripts ready to process.
+        self.evaluate();
+    }
+
+    var blocking = Blocking{
+        .allocator = self.allocator,
+        .buffer_pool = &self.buffer_pool,
+    };
+
+    var headers = try Http.Headers.init();
+    try self.page.requestCookie(.{}).headersForRequest(self.page.arena, url, &headers);
+
+    var client = self.client;
+    try client.blockingRequest(.{
+        .url = url,
+        .method = .GET,
+        .headers = headers,
+        .cookie_jar = self.page.cookie_jar,
+        .ctx = &blocking,
+        .resource_type = .script,
+        .start_callback = if (log.enabled(.http, .debug)) Blocking.startCallback else null,
+        .header_callback = Blocking.headerCallback,
+        .data_callback = Blocking.dataCallback,
+        .done_callback = Blocking.doneCallback,
+        .error_callback = Blocking.errorCallback,
+    });
+
+    // rely on http's timeout settings to avoid an endless/long loop.
+    while (true) {
+        _ = try client.tick(200);
+        switch (blocking.state) {
+            .running => {},
+            .done => |result| return result,
+            .err => |err| return err,
+        }
+    }
+}
+
+pub fn staticScriptsDone(self: *ScriptManager) void {
+    std.debug.assert(self.static_scripts_done == false);
+    self.static_scripts_done = true;
+}
+
+// try to evaluate completed scripts (in order). This is called whenever a script
+// is completed.
+fn evaluate(self: *ScriptManager) void {
+    if (self.is_evaluating) {
+        // It's possible for a script.eval to cause evaluate to be called again.
+        // This is particularly true with blockingGet, but even without this,
+        // it's theoretically possible (but unlikely). We could make this work
+        // but there's little reason to support the complexity.
+        return;
+    }
+
+    if (self.is_blocking) {
+        // Cannot evaluate scripts while a blocking-load is in progress. Not
+        // only could that result in incorrect evaluation order, it could
+        // trigger another blocking request, while we're doing a blocking request.
+        return;
+    }
+
+    const page = self.page;
+    self.is_evaluating = true;
+    defer self.is_evaluating = false;
+
+    // every script in asyncs_ready is ready to be evaluated.
+    while (self.asyncs_ready.first) |n| {
+        var pending_script: *PendingScript = @fieldParentPtr("node", n);
+        defer pending_script.deinit();
+        pending_script.script.eval(page);
+    }
+
+    while (self.scripts.first) |n| {
+        var pending_script: *PendingScript = @fieldParentPtr("node", n);
+        if (pending_script.complete == false) {
+            return;
+        }
+        defer pending_script.deinit();
+        pending_script.script.eval(page);
+    }
+
+    if (self.static_scripts_done == false) {
+        // We can only execute deferred scripts if
+        // 1 - all the normal scripts are done
+        // 2 - we've finished parsing the HTML and at least queued all the scripts
+        // The last one isn't obvious, but it's possible for self.scripts to
+        // be empty not because we're done executing all the normal scripts
+        // but because we're done executing some (or maybe none), but we're still
+        // parsing the HTML.
+        return;
+    }
+
+    while (self.deferreds.first) |n| {
+        var pending_script: *PendingScript = @fieldParentPtr("node", n);
+        if (pending_script.complete == false) {
+            return;
+        }
+        defer pending_script.deinit();
+        pending_script.script.eval(page);
+    }
+
+    // When all scripts (normal and deferred) are done loading, the document
+    // state changes (this ultimately triggers the DOMContentLoaded event)
+    page.documentIsLoaded();
+
+    if (self.asyncs.first == null) {
+        // 1 - there are no async scripts pending
+        // 2 - we checkecked static_scripts_done == true above
+        // 3 - we drained self.scripts above
+        // 4 - we drained self.deferred above
+        page.documentIsComplete();
+    }
+}
+
+pub fn isDone(self: *const ScriptManager) bool {
+    return self.asyncs.first == null and // there are no more async scripts
+        self.static_scripts_done and // and we've finished parsing the HTML to queue all <scripts>
+        self.scripts.first == null and // and there are no more <script src=> to wait for
+        self.deferreds.first == null; // and there are no more <script defer src=> to wait for
+}
+
+fn startCallback(transfer: *Http.Transfer) !void {
+    const script: *PendingScript = @ptrCast(@alignCast(transfer.ctx));
+    script.startCallback(transfer) catch |err| {
+        log.err(.http, "SM.startCallback", .{ .err = err, .transfer = transfer });
+        return err;
+    };
+}
+
+fn headerCallback(transfer: *Http.Transfer) !void {
+    const script: *PendingScript = @ptrCast(@alignCast(transfer.ctx));
+    try script.headerCallback(transfer);
+}
+
+fn dataCallback(transfer: *Http.Transfer, data: []const u8) !void {
+    const script: *PendingScript = @ptrCast(@alignCast(transfer.ctx));
+    script.dataCallback(transfer, data) catch |err| {
+        log.err(.http, "SM.dataCallback", .{ .err = err, .transfer = transfer, .len = data.len });
+        return err;
+    };
+}
+
+fn doneCallback(ctx: *anyopaque) !void {
+    const script: *PendingScript = @ptrCast(@alignCast(ctx));
+    script.doneCallback();
+}
+
+fn errorCallback(ctx: *anyopaque, err: anyerror) void {
+    const script: *PendingScript = @ptrCast(@alignCast(ctx));
+    script.errorCallback(err);
+}
+
+// A script which is pending execution.
+// It could be pending because:
+//   (a) we're still downloading its content or
+//   (b) this is a non-async script that has to be executed in order
+pub const PendingScript = struct {
+    script: Script,
+    complete: bool,
+    node: OrderList.Node,
+    manager: *ScriptManager,
+
+    fn deinit(self: *PendingScript) void {
+        const script = &self.script;
+        const manager = self.manager;
+
+        if (script.source == .remote) {
+            manager.buffer_pool.release(script.source.remote);
+        }
+        self.getList().remove(&self.node);
+    }
+
+    fn remove(self: *PendingScript) void {
+        if (self.node) |*node| {
+            self.getList().remove(node);
+            self.node = null;
+        }
+    }
+
+    fn startCallback(self: *PendingScript, transfer: *Http.Transfer) !void {
+        _ = self;
+        log.debug(.http, "script fetch start", .{ .req = transfer });
+    }
+
+    fn headerCallback(self: *PendingScript, transfer: *Http.Transfer) !void {
+        const header = &transfer.response_header.?;
+        if (header.status != 200) {
+            log.info(.http, "script header", .{
+                .req = transfer,
+                .status = header.status,
+                .content_type = header.contentType(),
+            });
+            return;
+        }
+
+        log.debug(.http, "script header", .{
+            .req = transfer,
+            .status = header.status,
+            .content_type = header.contentType(),
+        });
+
+        // If this isn't true, then we'll likely leak memory. If you don't
+        // set `CURLOPT_SUPPRESS_CONNECT_HEADERS` and CONNECT to a proxy, this
+        // will fail. This assertion exists to catch incorrect assumptions about
+        // how libcurl works, or about how we've configured it.
+        std.debug.assert(self.script.source.remote.capacity == 0);
+        var buffer = self.manager.buffer_pool.get();
+        if (transfer.getContentLength()) |cl| {
+            try buffer.ensureTotalCapacity(self.manager.allocator, cl);
+        }
+        self.script.source = .{ .remote = buffer };
+    }
+
+    fn dataCallback(self: *PendingScript, transfer: *Http.Transfer, data: []const u8) !void {
+        _ = transfer;
+        // too verbose
+        // log.debug(.http, "script data chunk", .{
+        //     .req = transfer,
+        //     .len = data.len,
+        // });
+
+        try self.script.source.remote.appendSlice(self.manager.allocator, data);
+    }
+
+    fn doneCallback(self: *PendingScript) void {
+        log.debug(.http, "script fetch complete", .{ .req = self.script.url });
+
+        const manager = self.manager;
+        self.complete = true;
+        if (self.script.is_async) {
+            manager.asyncs.remove(&self.node);
+            manager.asyncs_ready.append(&self.node);
+        }
+        manager.evaluate();
+    }
+
+    fn errorCallback(self: *PendingScript, err: anyerror) void {
+        log.warn(.http, "script fetch error", .{ .req = self.script.url, .err = err });
+
+        const manager = self.manager;
+
+        self.deinit();
+
+        if (manager.shutdown) {
+            return;
+        }
+
+        manager.evaluate();
+    }
+
+    fn getList(self: *const PendingScript) *OrderList {
+        // When a script has both the async and defer flag set, it should be
+        // treated as async. Async is newer, so some websites use both so that
+        // if async isn't known, it'll fallback to defer.
+
+        const script = &self.script;
+        if (script.is_async) {
+            return if (self.complete) &self.manager.asyncs_ready else &self.manager.asyncs;
+        }
+
+        if (script.is_defer) {
+            return &self.manager.deferreds;
+        }
+
+        return &self.manager.scripts;
+    }
+};
+
+const Script = struct {
+    kind: Kind,
+    url: []const u8,
+    is_async: bool,
+    is_defer: bool,
+    source: Source,
+    element: *parser.Element,
+
+    const Kind = enum {
+        module,
+        javascript,
+    };
+
+    const Callback = union(enum) {
+        string: []const u8,
+        function: Env.Function,
+    };
+
+    const Source = union(enum) {
+        @"inline": []const u8,
+        remote: std.ArrayListUnmanaged(u8),
+
+        fn content(self: Source) []const u8 {
+            return switch (self) {
+                .remote => |buf| buf.items,
+                .@"inline" => |c| c,
+            };
+        }
+    };
+
+    fn eval(self: *Script, page: *Page) void {
+        page.setCurrentScript(@ptrCast(self.element)) catch |err| {
+            log.err(.browser, "set document script", .{ .err = err });
+            return;
+        };
+
+        defer page.setCurrentScript(null) catch |err| {
+            log.err(.browser, "clear document script", .{ .err = err });
+        };
+
+        // inline scripts aren't cached. remote ones are.
+        const cacheable = self.source == .remote;
+
+        const url = self.url;
+
+        log.info(.browser, "executing script", .{
+            .src = url,
+            .kind = self.kind,
+            .cacheable = cacheable,
+        });
+
+        const js_context = page.main_context;
+        var try_catch: Env.TryCatch = undefined;
+        try_catch.init(js_context);
+        defer try_catch.deinit();
+
+        const success = blk: {
+            const content = self.source.content();
+            switch (self.kind) {
+                .javascript => _ = js_context.eval(content, url) catch break :blk false,
+                .module => {
+                    // We don't care about waiting for the evaluation here.
+                    _ = js_context.module(content, url, cacheable) catch break :blk false;
+                },
+            }
+            break :blk true;
+        };
+
+        if (success) {
+            self.executeCallback("onload", page);
+            return;
+        }
+
+        if (page.delayed_navigation) {
+            // If we're navigating to another page, an error is expected
+            // since we probably terminated the script forcefully.
+            return;
+        }
+
+        const msg = try_catch.err(page.arena) catch |err| @errorName(err) orelse "unknown";
+        log.warn(.user_script, "eval script", .{
+            .url = url,
+            .err = msg,
+            .cacheable = cacheable,
+        });
+
+        self.executeCallback("onerror", page);
+    }
+
+    fn executeCallback(self: *const Script, comptime typ: []const u8, page: *Page) void {
+        const callback = self.getCallback(typ, page) orelse return;
+
+        switch (callback) {
+            .string => |str| {
+                var try_catch: Env.TryCatch = undefined;
+                try_catch.init(page.main_context);
+                defer try_catch.deinit();
+
+                _ = page.main_context.exec(str, typ) catch |err| {
+                    const msg = try_catch.err(page.arena) catch @errorName(err) orelse "unknown";
+                    log.warn(.user_script, "script callback", .{
+                        .url = self.url,
+                        .err = msg,
+                        .type = typ,
+                        .@"inline" = true,
+                    });
+                };
+            },
+            .function => |f| {
+                const Event = @import("events/event.zig").Event;
+                const loadevt = parser.eventCreate() catch |err| {
+                    log.err(.browser, "SM event creation", .{ .err = err });
+                    return;
+                };
+                defer parser.eventDestroy(loadevt);
+
+                var result: Env.Function.Result = undefined;
+                const iface = Event.toInterface(loadevt);
+                f.tryCall(void, .{iface}, &result) catch {
+                    log.warn(.user_script, "script callback", .{
+                        .url = self.url,
+                        .type = typ,
+                        .err = result.exception,
+                        .stack = result.stack,
+                        .@"inline" = false,
+                    });
+                };
+            },
+        }
+    }
+
+    fn getCallback(self: *const Script, comptime typ: []const u8, page: *Page) ?Callback {
+        const element = self.element;
+        // first we check if there was an el.onload set directly on the
+        // element in JavaScript (if so, it'd be stored in the node state)
+        if (page.getNodeState(@ptrCast(element))) |se| {
+            if (@field(se, typ)) |function| {
+                return .{ .function = function };
+            }
+        }
+        // if we have no node state, or if the node state has no onload/onerror
+        // then check for the onload/onerror attribute
+        if (parser.elementGetAttribute(element, typ) catch null) |string| {
+            return .{ .string = string };
+        }
+        return null;
+    }
+};
+
+const BufferPool = struct {
+    count: usize,
+    available: List = .{},
+    allocator: Allocator,
+    max_concurrent_transfers: u8,
+    mem_pool: std.heap.MemoryPool(Container),
+
+    const List = std.DoublyLinkedList;
+
+    const Container = struct {
+        node: List.Node,
+        buf: std.ArrayListUnmanaged(u8),
+    };
+
+    fn init(allocator: Allocator, max_concurrent_transfers: u8) BufferPool {
+        return .{
+            .available = .{},
+            .count = 0,
+            .allocator = allocator,
+            .max_concurrent_transfers = max_concurrent_transfers,
+            .mem_pool = std.heap.MemoryPool(Container).init(allocator),
+        };
+    }
+
+    fn deinit(self: *BufferPool) void {
+        const allocator = self.allocator;
+
+        var node = self.available.first;
+        while (node) |n| {
+            const container: *Container = @fieldParentPtr("node", n);
+            container.buf.deinit(allocator);
+            node = n.next;
+        }
+        self.mem_pool.deinit();
+    }
+
+    fn get(self: *BufferPool) std.ArrayListUnmanaged(u8) {
+        const node = self.available.popFirst() orelse {
+            // return a new buffer
+            return .{};
+        };
+
+        self.count -= 1;
+        const container: *Container = @fieldParentPtr("node", node);
+        defer self.mem_pool.destroy(container);
+        return container.buf;
+    }
+
+    fn release(self: *BufferPool, buffer: ArrayListUnmanaged(u8)) void {
+        // create mutable copy
+        var b = buffer;
+
+        if (self.count == self.max_concurrent_transfers) {
+            b.deinit(self.allocator);
+            return;
+        }
+
+        const container = self.mem_pool.create() catch |err| {
+            b.deinit(self.allocator);
+            log.err(.http, "SM BufferPool release", .{ .err = err });
+            return;
+        };
+
+        b.clearRetainingCapacity();
+        container.* = .{ .buf = b, .node = .{} };
+        self.count += 1;
+        self.available.append(&container.node);
+    }
+};
+
+const Blocking = struct {
+    allocator: Allocator,
+    buffer_pool: *BufferPool,
+    state: State = .{ .running = {} },
+    buffer: std.ArrayListUnmanaged(u8) = .{},
+
+    const State = union(enum) {
+        running: void,
+        err: anyerror,
+        done: BlockingResult,
+    };
+
+    fn startCallback(transfer: *Http.Transfer) !void {
+        log.debug(.http, "script fetch start", .{ .req = transfer, .blocking = true });
+    }
+
+    fn headerCallback(transfer: *Http.Transfer) !void {
+        const header = &transfer.response_header.?;
+        log.debug(.http, "script header", .{
+            .req = transfer,
+            .blocking = true,
+            .status = header.status,
+            .content_type = header.contentType(),
+        });
+
+        if (header.status != 200) {
+            return error.InvalidStatusCode;
+        }
+
+        var self: *Blocking = @ptrCast(@alignCast(transfer.ctx));
+        self.buffer = self.buffer_pool.get();
+    }
+
+    fn dataCallback(transfer: *Http.Transfer, data: []const u8) !void {
+        // too verbose
+        // log.debug(.http, "script data chunk", .{
+        //     .req = transfer,
+        //     .blocking = true,
+        // });
+
+        var self: *Blocking = @ptrCast(@alignCast(transfer.ctx));
+        self.buffer.appendSlice(self.allocator, data) catch |err| {
+            log.err(.http, "SM.dataCallback", .{
+                .err = err,
+                .len = data.len,
+                .blocking = true,
+                .transfer = transfer,
+            });
+            return err;
+        };
+    }
+
+    fn doneCallback(ctx: *anyopaque) !void {
+        var self: *Blocking = @ptrCast(@alignCast(ctx));
+        self.state = .{ .done = .{
+            .buffer = self.buffer,
+            .buffer_pool = self.buffer_pool,
+        } };
+    }
+
+    fn errorCallback(ctx: *anyopaque, err: anyerror) void {
+        var self: *Blocking = @ptrCast(@alignCast(ctx));
+        self.state = .{ .err = err };
+        self.buffer_pool.release(self.buffer);
+    }
+};
+
+pub const BlockingResult = struct {
+    buffer: std.ArrayListUnmanaged(u8),
+    buffer_pool: *BufferPool,
+
+    pub fn deinit(self: *BlockingResult) void {
+        self.buffer_pool.release(self.buffer);
+    }
+
+    pub fn src(self: *const BlockingResult) []const u8 {
+        return self.buffer.items;
+    }
+};

src/browser/State.zig

@@ -0,0 +1,78 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+// Sometimes we need to extend libdom. For example, its HTMLDocument doesn't
+// have a readyState. We have a couple different options, such as making the
+// correction in libdom directly. Another option stems from the fact that every
+// libdom node has an opaque embedder_data field. This is the struct that we
+// lazily load into that field.
+//
+// It didn't originally start off as a collection of every single extension, but
+// this quickly proved necessary, since different fields are needed on the same
+// data at different levels of the prototype chain. This isn't memory efficient.
+
+const Env = @import("env.zig").Env;
+const parser = @import("netsurf.zig");
+const DataSet = @import("html/DataSet.zig");
+const ShadowRoot = @import("dom/shadow_root.zig").ShadowRoot;
+const StyleSheet = @import("cssom/StyleSheet.zig");
+const CSSStyleSheet = @import("cssom/CSSStyleSheet.zig");
+const CSSStyleDeclaration = @import("cssom/CSSStyleDeclaration.zig");
+
+// for HTMLScript (but probably needs to be added to more)
+onload: ?Env.Function = null,
+onerror: ?Env.Function = null,
+
+// for HTMLElement
+style: CSSStyleDeclaration = .empty,
+dataset: ?DataSet = null,
+template_content: ?*parser.DocumentFragment = null,
+
+// For dom/element
+shadow_root: ?*ShadowRoot = null,
+
+// for html/document
+ready_state: ReadyState = .loading,
+
+// for html/HTMLStyleElement
+style_sheet: ?*StyleSheet = null,
+
+// for dom/document
+active_element: ?*parser.Element = null,
+adopted_style_sheets: ?Env.JsObject = null,
+
+// for HTMLSelectElement
+// By default, if no option is explicitly selected, the first option should
+// be selected. However, libdom doesn't do this, and it sets the
+// selectedIndex to -1, which is a valid value for "nothing selected".
+// Therefore, when libdom says the selectedIndex == -1, we don't know if
+// it means that nothing is selected, or if the first option is selected by
+// default.
+// There are cases where this won't work, but when selectedIndex is
+// explicitly set, we set this boolean flag. Then, when we're getting then
+// selectedIndex, if this flag is == false, which is to say that if
+// selectedIndex hasn't been explicitly set AND if we have at least 1 option
+// AND if it isn't a multi select, we can make the 1st item selected by
+// default (by returning selectedIndex == 0).
+explicit_index_set: bool = false,
+
+const ReadyState = enum {
+    loading,
+    interactive,
+    complete,
+};

src/browser/browser.zig

@@ -1,450 +1,116 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
 const std = @import("std");
 
-const Types = @import("root").Types;
+const Allocator = std.mem.Allocator;
+const ArenaAllocator = std.heap.ArenaAllocator;
 
-const parser = @import("../netsurf.zig");
-const Loader = @import("loader.zig").Loader;
-const Dump = @import("dump.zig");
-const Mime = @import("mime.zig");
+const State = @import("State.zig");
+const Env = @import("env.zig").Env;
+const App = @import("../app.zig").App;
+const Session = @import("session.zig").Session;
+const Notification = @import("../notification.zig").Notification;
 
-const jsruntime = @import("jsruntime");
-const Loop = jsruntime.Loop;
-const Env = jsruntime.Env;
-
-const apiweb = @import("../apiweb.zig");
-
-const Window = @import("../html/window.zig").Window;
-const Walker = @import("../dom/walker.zig").WalkerDepthFirst;
-
-const FetchResult = std.http.Client.FetchResult;
-
-const log = std.log.scoped(.browser);
+const log = @import("../log.zig");
+const HttpClient = @import("../http/Client.zig");
 
 // Browser is an instance of the browser.
 // You can create multiple browser instances.
 // A browser contains only one session.
-// TODO allow multiple sessions per browser.
 pub const Browser = struct {
-    session: *Session,
+    env: *Env,
+    app: *App,
+    session: ?Session,
+    allocator: Allocator,
+    http_client: *HttpClient,
+    page_arena: ArenaAllocator,
+    session_arena: ArenaAllocator,
+    transfer_arena: ArenaAllocator,
+    notification: *Notification,
+    state_pool: std.heap.MemoryPool(State),
 
-    pub fn init(alloc: std.mem.Allocator, vm: jsruntime.VM) !Browser {
-        // We want to ensure the caller initialised a VM, but the browser
-        // doesn't use it directly...
-        _ = vm;
+    pub fn init(app: *App) !Browser {
+        const allocator = app.allocator;
 
-        return Browser{
-            .session = try Session.init(alloc, "about:blank"),
+        const env = try Env.init(allocator, &app.platform, .{});
+        errdefer env.deinit();
+
+        const notification = try Notification.init(allocator, app.notification);
+        app.http.client.notification = notification;
+        app.http.client.next_request_id = 0; // Should we track ids in CDP only?
+        errdefer notification.deinit();
+
+        return .{
+            .app = app,
+            .env = env,
+            .session = null,
+            .allocator = allocator,
+            .notification = notification,
+            .http_client = app.http.client,
+            .page_arena = ArenaAllocator.init(allocator),
+            .session_arena = ArenaAllocator.init(allocator),
+            .transfer_arena = ArenaAllocator.init(allocator),
+            .state_pool = std.heap.MemoryPool(State).init(allocator),
         };
     }
 
     pub fn deinit(self: *Browser) void {
-        self.session.deinit();
-    }
-
-    pub fn currentSession(self: *Browser) *Session {
-        return self.session;
-    }
-};
-
-// Session is like a browser's tab.
-// It owns the js env and the loader for all the pages of the session.
-// You can create successively multiple pages for a session, but you must
-// deinit a page before running another one.
-pub const Session = struct {
-    // allocator used to init the arena.
-    alloc: std.mem.Allocator,
-
-    // The arena is used only to bound the js env init b/c it leaks memory.
-    // see https://github.com/lightpanda-io/jsruntime-lib/issues/181
-    //
-    // The arena is initialised with self.alloc allocator.
-    // all others Session deps use directly self.alloc and not the arena.
-    arena: std.heap.ArenaAllocator,
-
-    uri: []const u8,
-
-    // TODO handle proxy
-    loader: Loader,
-    env: Env = undefined,
-    loop: Loop,
-    window: Window,
-
-    jstypes: [Types.len]usize = undefined,
-
-    fn init(alloc: std.mem.Allocator, uri: []const u8) !*Session {
-        var self = try alloc.create(Session);
-        self.* = Session{
-            .uri = uri,
-            .alloc = alloc,
-            .arena = std.heap.ArenaAllocator.init(alloc),
-            .window = Window.create(null),
-            .loader = Loader.init(alloc),
-            .loop = try Loop.init(alloc),
-        };
-
-        self.env = try Env.init(self.arena.allocator(), &self.loop);
-        try self.env.load(&self.jstypes);
-
-        return self;
-    }
-
-    fn deinit(self: *Session) void {
+        self.closeSession();
         self.env.deinit();
-        self.arena.deinit();
-
-        self.loader.deinit();
-        self.loop.deinit();
-        self.alloc.destroy(self);
+        self.page_arena.deinit();
+        self.session_arena.deinit();
+        self.transfer_arena.deinit();
+        self.http_client.notification = null;
+        self.notification.deinit();
+        self.state_pool.deinit();
     }
 
-    pub fn createPage(self: *Session) !Page {
-        return Page.init(self.alloc, self);
+    pub fn newSession(self: *Browser) !*Session {
+        self.closeSession();
+        self.session = @as(Session, undefined);
+        const session = &self.session.?;
+        try Session.init(session, self);
+        return session;
+    }
+
+    pub fn closeSession(self: *Browser) void {
+        if (self.session) |*session| {
+            session.deinit();
+            self.session = null;
+            _ = self.session_arena.reset(.{ .retain_with_limit = 1 * 1024 * 1024 });
+            self.env.lowMemoryNotification();
+        }
+    }
+
+    pub fn runMicrotasks(self: *const Browser) void {
+        self.env.runMicrotasks();
+    }
+
+    pub fn runMessageLoop(self: *const Browser) void {
+        while (self.env.pumpMessageLoop()) {
+            log.debug(.browser, "pumpMessageLoop", .{});
+        }
+        self.env.runIdleTasks();
     }
 };
 
-// Page navigates to an url.
-// You can navigates multiple urls with the same page, but you have to call
-// end() to stop the previous navigation before starting a new one.
-// The page handle all its memory in an arena allocator. The arena is reseted
-// when end() is called.
-pub const Page = struct {
-    arena: std.heap.ArenaAllocator,
-    session: *Session,
-    doc: ?*parser.Document = null,
-
-    // handle url
-    rawuri: ?[]const u8 = null,
-    uri: std.Uri = undefined,
-
-    raw_data: ?[]const u8 = null,
-
-    fn init(
-        alloc: std.mem.Allocator,
-        session: *Session,
-    ) Page {
-        return Page{
-            .arena = std.heap.ArenaAllocator.init(alloc),
-            .session = session,
-        };
-    }
-
-    // reset js env and mem arena.
-    pub fn end(self: *Page) void {
-        self.session.env.stop();
-        // TODO unload document: https://html.spec.whatwg.org/#unloading-documents
-
-        _ = self.arena.reset(.free_all);
-    }
-
-    pub fn deinit(self: *Page) void {
-        self.arena.deinit();
-    }
-
-    // dump writes the page content into the given file.
-    pub fn dump(self: *Page, out: std.fs.File) !void {
-
-        // if no HTML document pointer available, dump the data content only.
-        if (self.doc == null) {
-            // no data loaded, nothing to do.
-            if (self.raw_data == null) return;
-            return try out.writeAll(self.raw_data.?);
-        }
-
-        // if the page has a pointer to a document, dumps the HTML.
-        try Dump.htmlFile(self.doc.?, out);
-    }
-
-    // spec reference: https://html.spec.whatwg.org/#document-lifecycle
-    pub fn navigate(self: *Page, uri: []const u8) !void {
-        const alloc = self.arena.allocator();
-
-        log.debug("starting GET {s}", .{uri});
-
-        // own the url
-        if (self.rawuri) |prev| alloc.free(prev);
-        self.rawuri = try alloc.dupe(u8, uri);
-        self.uri = std.Uri.parse(self.rawuri.?) catch try std.Uri.parseWithoutScheme(self.rawuri.?);
-
-        // TODO handle fragment in url.
-
-        // load the data
-        var resp = try self.session.loader.get(alloc, self.uri);
-        defer resp.deinit();
-
-        const req = resp.req;
-
-        log.info("GET {any} {d}", .{ self.uri, req.response.status });
-
-        // TODO handle redirection
-        if (req.response.status != .ok) {
-            log.debug("{?} {d} {s}\n{any}", .{
-                req.response.version,
-                req.response.status,
-                req.response.reason,
-                req.response.headers,
-            });
-            return error.BadStatusCode;
-        }
-
-        // TODO handle charset
-        // https://html.spec.whatwg.org/#content-type
-        const ct = req.response.headers.getFirstValue("Content-Type") orelse {
-            // no content type in HTTP headers.
-            // TODO try to sniff mime type from the body.
-            log.info("no content-type HTTP header", .{});
-            return;
-        };
-        log.debug("header content-type: {s}", .{ct});
-        const mime = try Mime.parse(ct);
-        if (mime.eql(Mime.HTML)) {
-            try self.loadHTMLDoc(req.reader(), mime.charset orelse "utf-8");
-        } else {
-            log.info("non-HTML document: {s}", .{ct});
-
-            // save the body into the page.
-            self.raw_data = try req.reader().readAllAlloc(alloc, 16 * 1024 * 1024);
-        }
-    }
-
-    // https://html.spec.whatwg.org/#read-html
-    fn loadHTMLDoc(self: *Page, reader: anytype, charset: []const u8) !void {
-        const alloc = self.arena.allocator();
-
-        log.debug("parse html with charset {s}", .{charset});
-
-        const ccharset = try alloc.dupeZ(u8, charset);
-        defer alloc.free(ccharset);
-
-        const html_doc = try parser.documentHTMLParse(reader, ccharset);
-        const doc = parser.documentHTMLToDocument(html_doc);
-
-        // save a document's pointer in the page.
-        self.doc = doc;
-
-        // TODO set document.readyState to interactive
-        // https://html.spec.whatwg.org/#reporting-document-loading-status
-
-        // TODO inject the URL to the document including the fragment.
-        // TODO set the referrer to the document.
-
-        self.session.window.replaceDocument(doc);
-
-        // https://html.spec.whatwg.org/#read-html
-
-        // start JS env
-        // TODO load the js env concurrently with the HTML parsing.
-        log.debug("start js env", .{});
-        try self.session.env.start(alloc);
-
-        // add global objects
-        log.debug("setup global env", .{});
-        try self.session.env.addObject(self.session.window, "window");
-        try self.session.env.addObject(self.session.window, "self");
-        try self.session.env.addObject(html_doc, "document");
-
-        // browse the DOM tree to retrieve scripts
-        // TODO execute the synchronous scripts during the HTL parsing.
-        // TODO fetch the script resources concurrently but execute them in the
-        // declaration order for synchronous ones.
-
-        // sasync stores scripts which can be run asynchronously.
-        // for now they are just run after the non-async one in order to
-        // dispatch DOMContentLoaded the sooner as possible.
-        var sasync = std.ArrayList(*parser.Element).init(alloc);
-        defer sasync.deinit();
-
-        const root = parser.documentToNode(doc);
-        const walker = Walker{};
-        var next: ?*parser.Node = null;
-        while (true) {
-            next = try walker.get_next(root, next) orelse break;
-
-            // ignore non-elements nodes.
-            if (try parser.nodeType(next.?) != .element) {
-                continue;
-            }
-
-            const e = parser.nodeToElement(next.?);
-            const tag = try parser.elementHTMLGetTagType(@as(*parser.ElementHTML, @ptrCast(e)));
-
-            // ignore non-script tags
-            if (tag != .script) continue;
-
-            // ignore non-js script.
-            // > type
-            // > Attribute is not set (default), an empty string, or a JavaScript MIME
-            // > type indicates that the script is a "classic script", containing
-            // > JavaScript code.
-            // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#attribute_is_not_set_default_an_empty_string_or_a_javascript_mime_type
-            const stype = try parser.elementGetAttribute(e, "type");
-            if (!isJS(stype)) {
-                continue;
-            }
-
-            // Ignore the defer attribute b/c we analyze all script
-            // after the document has been parsed.
-            // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#defer
-
-            // TODO use fetchpriority
-            // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#fetchpriority
-
-            // > async
-            // > For classic scripts, if the async attribute is present,
-            // > then the classic script will be fetched in parallel to
-            // > parsing and evaluated as soon as it is available.
-            // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#async
-            if (try parser.elementGetAttribute(e, "async") != null) {
-                try sasync.append(e);
-                continue;
-            }
-
-            // TODO handle for attribute
-            // TODO handle event attribute
-
-            // TODO defer
-            // > This Boolean attribute is set to indicate to a browser
-            // > that the script is meant to be executed after the
-            // > document has been parsed, but before firing
-            // > DOMContentLoaded.
-            // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#defer
-            // defer allow us to load a script w/o blocking the rest of
-            // evaluations.
-
-            // > Scripts without async, defer or type="module"
-            // > attributes, as well as inline scripts without the
-            // > type="module" attribute, are fetched and executed
-            // > immediately before the browser continues to parse the
-            // > page.
-            // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#notes
-            self.evalScript(e) catch |err| log.warn("evaljs: {any}", .{err});
-        }
-
-        // TODO wait for deferred scripts
-
-        // dispatch DOMContentLoaded before the transition to "complete",
-        // at the point where all subresources apart from async script elements
-        // have loaded.
-        // https://html.spec.whatwg.org/#reporting-document-loading-status
-        const evt = try parser.eventCreate();
-        try parser.eventInit(evt, "DOMContentLoaded", .{ .bubbles = true, .cancelable = true });
-        _ = try parser.eventTargetDispatchEvent(parser.toEventTarget(parser.DocumentHTML, html_doc), evt);
-
-        // eval async scripts.
-        for (sasync.items) |e| {
-            self.evalScript(e) catch |err| log.warn("evaljs: {any}", .{err});
-        }
-
-        // TODO wait for async scripts
-
-        // TODO set document.readyState to complete
-
-        // dispatch window.load event
-        const loadevt = try parser.eventCreate();
-        try parser.eventInit(loadevt, "load", .{});
-        _ = try parser.eventTargetDispatchEvent(parser.toEventTarget(Window, &self.session.window), loadevt);
-    }
-
-    // evalScript evaluates the src in priority.
-    // if no src is present, we evaluate the text source.
-    // https://html.spec.whatwg.org/multipage/scripting.html#script-processing-model
-    fn evalScript(self: *Page, e: *parser.Element) !void {
-        const alloc = self.arena.allocator();
-
-        // https://html.spec.whatwg.org/multipage/webappapis.html#fetch-a-classic-script
-        const opt_src = try parser.elementGetAttribute(e, "src");
-        if (opt_src) |src| {
-            log.debug("starting GET {s}", .{src});
-
-            self.fetchScript(src) catch |err| {
-                switch (err) {
-                    FetchError.BadStatusCode => return err,
-
-                    // TODO If el's result is null, then fire an event named error at
-                    // el, and return.
-                    FetchError.NoBody => return,
-
-                    FetchError.JsErr => {}, // nothing to do here.
-                    else => return err,
-                }
-            };
-
-            // TODO If el's from an external file is true, then fire an event
-            // named load at el.
-
-            return;
-        }
-
-        const opt_text = try parser.nodeTextContent(parser.elementToNode(e));
-        if (opt_text) |text| {
-            // TODO handle charset attribute
-            var res = jsruntime.JSResult{};
-            try self.session.env.run(alloc, text, "", &res, null);
-            defer res.deinit(alloc);
-
-            if (res.success) {
-                log.debug("eval inline: {s}", .{res.result});
-            } else {
-                log.info("eval inline: {s}", .{res.result});
-            }
-
-            return;
-        }
-
-        // nothing has been loaded.
-        // TODO If el's result is null, then fire an event named error at
-        // el, and return.
-    }
-
-    const FetchError = error{
-        BadStatusCode,
-        NoBody,
-        JsErr,
-    };
-
-    // fetchScript senf a GET request to the src and execute the script
-    // received.
-    fn fetchScript(self: *Page, src: []const u8) !void {
-        const alloc = self.arena.allocator();
-
-        log.debug("starting fetch script {s}", .{src});
-
-        const u = std.Uri.parse(src) catch try std.Uri.parseWithoutScheme(src);
-        const ru = try std.Uri.resolve(self.uri, u, false, alloc);
-
-        var fetchres = try self.session.loader.fetch(alloc, ru);
-        defer fetchres.deinit();
-
-        log.info("fech script {any}: {d}", .{ ru, fetchres.status });
-
-        if (fetchres.status != .ok) return FetchError.BadStatusCode;
-
-        // TODO check content-type
-
-        // check no body
-        if (fetchres.body == null) return FetchError.NoBody;
-
-        var res = jsruntime.JSResult{};
-        try self.session.env.run(alloc, fetchres.body.?, src, &res, null);
-        defer res.deinit(alloc);
-
-        if (res.success) {
-            log.debug("eval remote {s}: {s}", .{ src, res.result });
-        } else {
-            log.info("eval remote {s}: {s}", .{ src, res.result });
-            return FetchError.JsErr;
-        }
-    }
-
-    // > type
-    // > Attribute is not set (default), an empty string, or a JavaScript MIME
-    // > type indicates that the script is a "classic script", containing
-    // > JavaScript code.
-    // https://developer.mozilla.org/en-US/docs/Web/HTML/Element/script#attribute_is_not_set_default_an_empty_string_or_a_javascript_mime_type
-    fn isJS(stype: ?[]const u8) bool {
-        if (stype == null or stype.?.len == 0) return true;
-        if (std.mem.eql(u8, stype.?, "application/javascript")) return true;
-        if (!std.mem.eql(u8, stype.?, "module")) return true;
-
-        return false;
-    }
-};
+const testing = @import("../testing.zig");
+test "Browser" {
+    try testing.htmlRunner("browser.html");
+}

src/browser/console/console.zig

@@ -0,0 +1,168 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const builtin = @import("builtin");
+const log = @import("../../log.zig");
+
+const Allocator = std.mem.Allocator;
+const Page = @import("../page.zig").Page;
+const JsObject = @import("../env.zig").Env.JsObject;
+
+pub const Console = struct {
+    // TODO: configurable writer
+    timers: std.StringHashMapUnmanaged(u32) = .{},
+    counts: std.StringHashMapUnmanaged(u32) = .{},
+
+    pub fn _lp(values: []JsObject, page: *Page) !void {
+        if (values.len == 0) {
+            return;
+        }
+        log.fatal(.console, "lightpanda", .{ .args = try serializeValues(values, page) });
+    }
+
+    pub fn _log(values: []JsObject, page: *Page) !void {
+        if (values.len == 0) {
+            return;
+        }
+        log.info(.console, "info", .{ .args = try serializeValues(values, page) });
+    }
+
+    pub fn _info(values: []JsObject, page: *Page) !void {
+        return _log(values, page);
+    }
+
+    pub fn _debug(values: []JsObject, page: *Page) !void {
+        if (values.len == 0) {
+            return;
+        }
+        log.debug(.console, "debug", .{ .args = try serializeValues(values, page) });
+    }
+
+    pub fn _warn(values: []JsObject, page: *Page) !void {
+        if (values.len == 0) {
+            return;
+        }
+        log.warn(.console, "warn", .{ .args = try serializeValues(values, page) });
+    }
+
+    pub fn _error(values: []JsObject, page: *Page) !void {
+        if (values.len == 0) {
+            return;
+        }
+
+        log.warn(.console, "error", .{
+            .args = try serializeValues(values, page),
+            .stack = page.stackTrace() catch "???",
+        });
+    }
+
+    pub fn _clear() void {}
+
+    pub fn _count(self: *Console, label_: ?[]const u8, page: *Page) !void {
+        const label = label_ orelse "default";
+        const gop = try self.counts.getOrPut(page.arena, label);
+
+        var current: u32 = 0;
+        if (gop.found_existing) {
+            current = gop.value_ptr.*;
+        } else {
+            gop.key_ptr.* = try page.arena.dupe(u8, label);
+        }
+
+        const count = current + 1;
+        gop.value_ptr.* = count;
+
+        log.info(.console, "count", .{ .label = label, .count = count });
+    }
+
+    pub fn _countReset(self: *Console, label_: ?[]const u8) !void {
+        const label = label_ orelse "default";
+        const kv = self.counts.fetchRemove(label) orelse {
+            log.info(.console, "invalid counter", .{ .label = label });
+            return;
+        };
+        log.info(.console, "count reset", .{ .label = label, .count = kv.value });
+    }
+
+    pub fn _time(self: *Console, label_: ?[]const u8, page: *Page) !void {
+        const label = label_ orelse "default";
+        const gop = try self.timers.getOrPut(page.arena, label);
+
+        if (gop.found_existing) {
+            log.info(.console, "duplicate timer", .{ .label = label });
+            return;
+        }
+        gop.key_ptr.* = try page.arena.dupe(u8, label);
+        gop.value_ptr.* = timestamp();
+    }
+
+    pub fn _timeLog(self: *Console, label_: ?[]const u8) void {
+        const elapsed = timestamp();
+        const label = label_ orelse "default";
+        const start = self.timers.get(label) orelse {
+            log.info(.console, "invalid timer", .{ .label = label });
+            return;
+        };
+        log.info(.console, "timer", .{ .label = label, .elapsed = elapsed - start });
+    }
+
+    pub fn _timeStop(self: *Console, label_: ?[]const u8) void {
+        const elapsed = timestamp();
+        const label = label_ orelse "default";
+        const kv = self.timers.fetchRemove(label) orelse {
+            log.info(.console, "invalid timer", .{ .label = label });
+            return;
+        };
+
+        log.warn(.console, "timer stop", .{ .label = label, .elapsed = elapsed - kv.value });
+    }
+
+    pub fn _assert(assertion: JsObject, values: []JsObject, page: *Page) !void {
+        if (assertion.isTruthy()) {
+            return;
+        }
+        var serialized_values: []const u8 = "";
+        if (values.len > 0) {
+            serialized_values = try serializeValues(values, page);
+        }
+        log.info(.console, "assertion failed", .{ .values = serialized_values });
+    }
+
+    fn serializeValues(values: []JsObject, page: *Page) ![]const u8 {
+        if (values.len == 0) {
+            return "";
+        }
+
+        const arena = page.call_arena;
+        const separator = log.separator();
+        var arr: std.ArrayListUnmanaged(u8) = .{};
+
+        for (values, 1..) |value, i| {
+            try arr.appendSlice(arena, separator);
+            try arr.writer(arena).print("{d}: ", .{i});
+            const serialized = if (builtin.mode == .Debug) value.toDetailString() else value.toString();
+            try arr.appendSlice(arena, try serialized);
+        }
+        return arr.items;
+    }
+};
+
+fn timestamp() u32 {
+    return @import("../../datetime.zig").timestamp();
+}

src/browser/crypto/crypto.zig

@@ -0,0 +1,71 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const Env = @import("../env.zig").Env;
+const uuidv4 = @import("../../id.zig").uuidv4;
+
+// https://w3c.github.io/webcrypto/#crypto-interface
+pub const Crypto = struct {
+    _not_empty: bool = true,
+
+    pub fn _getRandomValues(_: *const Crypto, js_obj: Env.JsObject) !Env.JsObject {
+        var into = try js_obj.toZig(Crypto, "getRandomValues", RandomValues);
+        const buf = into.asBuffer();
+        if (buf.len > 65_536) {
+            return error.QuotaExceededError;
+        }
+        std.crypto.random.bytes(buf);
+        return js_obj;
+    }
+
+    pub fn _randomUUID(_: *const Crypto) [36]u8 {
+        var hex: [36]u8 = undefined;
+        uuidv4(&hex);
+        return hex;
+    }
+};
+
+const RandomValues = union(enum) {
+    int8: []i8,
+    uint8: []u8,
+    int16: []i16,
+    uint16: []u16,
+    int32: []i32,
+    uint32: []u32,
+    int64: []i64,
+    uint64: []u64,
+
+    fn asBuffer(self: RandomValues) []u8 {
+        return switch (self) {
+            .int8 => |b| (@as([]u8, @ptrCast(b)))[0..b.len],
+            .uint8 => |b| (@as([]u8, @ptrCast(b)))[0..b.len],
+            .int16 => |b| (@as([]u8, @ptrCast(b)))[0 .. b.len * 2],
+            .uint16 => |b| (@as([]u8, @ptrCast(b)))[0 .. b.len * 2],
+            .int32 => |b| (@as([]u8, @ptrCast(b)))[0 .. b.len * 4],
+            .uint32 => |b| (@as([]u8, @ptrCast(b)))[0 .. b.len * 4],
+            .int64 => |b| (@as([]u8, @ptrCast(b)))[0 .. b.len * 8],
+            .uint64 => |b| (@as([]u8, @ptrCast(b)))[0 .. b.len * 8],
+        };
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: Crypto" {
+    try testing.htmlRunner("crypto.html");
+}

src/browser/css/README.md

@@ -0,0 +1,218 @@
+# css
+
+Lightpanda css implements CSS selectors parsing and matching in Zig.
+This package is a port of the Go lib [andybalholm/cascadia](https://github.com/andybalholm/cascadia).
+
+## Usage
+
+### Query parser
+
+```zig
+const css = @import("css.zig");
+
+const selector = try css.parse(alloc, "h1", .{});
+defer selector.deinit(alloc);
+```
+
+### DOM tree match
+
+The lib expects a `Node` interface implementation to match your DOM tree.
+
+```zig
+pub const Node = struct {
+    pub fn firstChild(_: Node) !?Node {
+        return error.TODO;
+    }
+
+    pub fn lastChild(_: Node) !?Node {
+        return error.TODO;
+    }
+
+    pub fn nextSibling(_: Node) !?Node {
+        return error.TODO;
+    }
+
+    pub fn prevSibling(_: Node) !?Node {
+        return error.TODO;
+    }
+
+    pub fn parent(_: Node) !?Node {
+        return error.TODO;
+    }
+
+    pub fn isElement(_: Node) bool {
+        return false;
+    }
+
+    pub fn isDocument(_: Node) bool {
+        return false;
+    }
+
+    pub fn isComment(_: Node) bool {
+        return false;
+    }
+
+    pub fn isText(_: Node) bool {
+        return false;
+    }
+
+    pub fn isEmptyText(_: Node) !bool {
+        return error.TODO;
+    }
+
+    pub fn tag(_: Node) ![]const u8 {
+        return error.TODO;
+    }
+
+    pub fn attr(_: Node, _: []const u8) !?[]const u8 {
+        return error.TODO;
+    }
+
+    pub fn eql(_: Node, _: Node) bool {
+        return false;
+    }
+};
+```
+
+You also need do define a `Matcher` implementing a `match` function to
+accumulate the results.
+
+```zig
+const Matcher = struct {
+    const Nodes = std.ArrayList(Node);
+
+    nodes: Nodes,
+
+    fn init(alloc: std.mem.Allocator) Matcher {
+        return .{ .nodes = Nodes.init(alloc) };
+    }
+
+    fn deinit(m: *Matcher) void {
+        m.nodes.deinit();
+    }
+
+    pub fn match(m: *Matcher, n: Node) !void {
+        try m.nodes.append(n);
+    }
+};
+```
+
+Then you can use the lib itself.
+
+```zig
+var matcher = Matcher.init(alloc);
+defer matcher.deinit();
+
+try css.matchAll(selector, node, &matcher);
+_ = try css.matchFirst(selector, node, &matcher); // returns true if a node matched.
+```
+
+## Features
+
+* [x] parse query selector
+* [x] `matchAll`
+* [x] `matchFirst`
+* [ ] specificity
+
+### Selectors implemented
+
+#### Selectors
+
+* [x] Class selectors
+* [x] Id selectors
+* [x] Type selectors
+* [x] Universal selectors
+* [ ] Nesting selectors
+
+#### Combinators
+
+* [x] Child combinator
+* [ ] Column combinator
+* [x] Descendant combinator
+* [ ] Namespace combinator
+* [x] Next-sibling combinator
+* [x] Selector list combinator
+* [x] Subsequent-sibling combinator
+
+#### Attribute
+
+* [x] `[attr]`
+* [x] `[attr=value]`
+* [x] `[attr|=value]`
+* [x] `[attr^=value]`
+* [x] `[attr$=value]`
+* [ ] `[attr*=value]`
+* [x] `[attr operator value i]`
+* [ ] `[attr operator value s]`
+
+#### Pseudo classes
+
+* [ ] `:active`
+* [ ] `:any-link`
+* [ ] `:autofill`
+* [ ] `:blank Experimental`
+* [x] `:checked`
+* [ ] `:current Experimental`
+* [ ] `:default`
+* [ ] `:defined`
+* [ ] `:dir() Experimental`
+* [x] `:disabled`
+* [x] `:empty`
+* [x] `:enabled`
+* [ ] `:first`
+* [x] `:first-child`
+* [x] `:first-of-type`
+* [ ] `:focus`
+* [ ] `:focus-visible`
+* [ ] `:focus-within`
+* [ ] `:fullscreen`
+* [ ] `:future Experimental`
+* [x] `:has() Experimental`
+* [ ] `:host`
+* [ ] `:host()`
+* [ ] `:host-context() Experimental`
+* [ ] `:hover`
+* [ ] `:indeterminate`
+* [ ] `:in-range`
+* [ ] `:invalid`
+* [ ] `:is()`
+* [x] `:lang()`
+* [x] `:last-child`
+* [x] `:last-of-type`
+* [ ] `:left`
+* [x] `:link`
+* [ ] `:local-link Experimental`
+* [ ] `:modal`
+* [x] `:not()`
+* [x] `:nth-child()`
+* [x] `:nth-last-child()`
+* [x] `:nth-last-of-type()`
+* [x] `:nth-of-type()`
+* [x] `:only-child`
+* [x] `:only-of-type`
+* [ ] `:optional`
+* [ ] `:out-of-range`
+* [ ] `:past Experimental`
+* [ ] `:paused`
+* [ ] `:picture-in-picture`
+* [ ] `:placeholder-shown`
+* [ ] `:playing`
+* [ ] `:read-only`
+* [ ] `:read-write`
+* [ ] `:required`
+* [ ] `:right`
+* [x] `:root`
+* [ ] `:scope`
+* [ ] `:state() Experimental`
+* [ ] `:target`
+* [ ] `:target-within Experimental`
+* [ ] `:user-invalid Experimental`
+* [ ] `:valid`
+* [ ] `:visited`
+* [ ] `:where()`
+* [ ] `:contains()`
+* [ ] `:containsown()`
+* [ ] `:matched()`
+* [ ] `:matchesown()`
+* [x] `:root`
+

src/browser/css/css.zig

@@ -0,0 +1,195 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+// CSS Selector parser and query
+// This package is a rewrite in Zig of Cascadia CSS Selector parser.
+// see https://github.com/andybalholm/cascadia
+const std = @import("std");
+const Selector = @import("selector.zig").Selector;
+const parser = @import("parser.zig");
+
+pub const Interfaces = .{
+    Css,
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/CSS
+pub const Css = struct {
+    _not_empty: bool = true,
+
+    pub fn _supports(_: *Css, _: []const u8, _: ?[]const u8) bool {
+        // TODO: Actually respond with which CSS features we support.
+        return true;
+    }
+};
+
+// parse parse a selector string and returns the parsed result or an error.
+pub fn parse(alloc: std.mem.Allocator, s: []const u8, opts: parser.ParseOptions) parser.ParseError!Selector {
+    var p = parser.Parser{ .s = s, .i = 0, .opts = opts };
+    return p.parse(alloc);
+}
+
+// matchFirst call m.match with the first node that matches the selector s, from the
+// descendants of n and returns true. If none matches, it returns false.
+pub fn matchFirst(s: *const Selector, node: anytype, m: anytype) !bool {
+    var c = try node.firstChild();
+    while (true) {
+        if (c == null) break;
+
+        if (try s.match(c.?)) {
+            try m.match(c.?);
+            return true;
+        }
+
+        if (try matchFirst(s, c.?, m)) return true;
+        c = try c.?.nextSibling();
+    }
+    return false;
+}
+
+// matchAll call m.match with the all the nodes that matches the selector s, from the
+// descendants of n.
+pub fn matchAll(s: *const Selector, node: anytype, m: anytype) !void {
+    var c = try node.firstChild();
+    while (true) {
+        if (c == null) break;
+
+        if (try s.match(c.?)) try m.match(c.?);
+        try matchAll(s, c.?, m);
+        c = try c.?.nextSibling();
+    }
+}
+
+test "parse" {
+    const alloc = std.testing.allocator;
+
+    const testcases = [_][]const u8{
+        "address",
+        "*",
+        "#foo",
+        "li#t1",
+        "*#t4",
+        ".t1",
+        "p.t1",
+        "div.teST",
+        ".t1.fail",
+        "p.t1.t2",
+        "p.--t1",
+        "p.--t1.--t2",
+        "p[title]",
+        "div[class=\"red\" i]",
+        "address[title=\"foo\"]",
+        "address[title=\"FoOIgnoRECaSe\" i]",
+        "address[title!=\"foo\"]",
+        "address[title!=\"foo\" i]",
+        "p[title!=\"FooBarUFoo\" i]",
+        "[  \t title        ~=       foo    ]",
+        "p[title~=\"FOO\" i]",
+        "p[title~=toofoo i]",
+        "[title~=\"hello world\"]",
+        "[title~=\"hello\" i]",
+        "[title~=\"hello\"          I]",
+        "[lang|=\"en\"]",
+        "[lang|=\"EN\" i]",
+        "[lang|=\"EN\"     i]",
+        "[title^=\"foo\"]",
+        "[title^=\"foo\" i]",
+        "[title$=\"bar\"]",
+        "[title$=\"BAR\" i]",
+        "[title*=\"bar\"]",
+        "[title*=\"BaRu\" i]",
+        "[title*=\"BaRu\" I]",
+        "p[class$=\" \"]",
+        "p[class$=\"\"]",
+        "p[class^=\" \"]",
+        "p[class^=\"\"]",
+        "p[class*=\" \"]",
+        "p[class*=\"\"]",
+        "input[name=Sex][value=F]",
+        "table[border=\"0\"][cellpadding=\"0\"][cellspacing=\"0\"]",
+        ".t1:not(.t2)",
+        "div:not(.t1)",
+        "div:not([class=\"t2\"])",
+        "li:nth-child(odd)",
+        "li:nth-child(even)",
+        "li:nth-child(-n+2)",
+        "li:nth-child(3n+1)",
+        "li:nth-last-child(odd)",
+        "li:nth-last-child(even)",
+        "li:nth-last-child(-n+2)",
+        "li:nth-last-child(3n+1)",
+        "span:first-child",
+        "span:last-child",
+        "p:nth-of-type(2)",
+        "p:nth-last-of-type(2)",
+        "p:last-of-type",
+        "p:first-of-type",
+        "p:only-child",
+        "p:only-of-type",
+        ":empty",
+        "div p",
+        "div table p",
+        "div > p",
+        "p ~ p",
+        "p + p",
+        "li, p",
+        "p +/*This is a comment*/ p",
+        "p:contains(\"that wraps\")",
+        "p:containsOwn(\"that wraps\")",
+        ":containsOwn(\"inner\")",
+        "p:containsOwn(\"block\")",
+        "div:has(#p1)",
+        "div:has(:containsOwn(\"2\"))",
+        "body :has(:containsOwn(\"2\"))",
+        "body :haschild(:containsOwn(\"2\"))",
+        "p:matches([\\d])",
+        "p:matches([a-z])",
+        "p:matches([a-zA-Z])",
+        "p:matches([^\\d])",
+        "p:matches(^(0|a))",
+        "p:matches(^\\d+$)",
+        "p:not(:matches(^\\d+$))",
+        "div :matchesOwn(^\\d+$)",
+        "[href#=(fina)]:not([href#=(\\/\\/[^\\/]+untrusted)])",
+        "[href#=(^https:\\/\\/[^\\/]*\\/?news)]",
+        ":input",
+        ":root",
+        "*:root",
+        "html:nth-child(1)",
+        "*:root:first-child",
+        "*:root:nth-child(1)",
+        "a:not(:root)",
+        "body > *:nth-child(3n+2)",
+        "input:disabled",
+        ":disabled",
+        ":enabled",
+        "div.class1, div.class2",
+    };
+
+    for (testcases) |tc| {
+        const s = parse(alloc, tc, .{}) catch |e| {
+            std.debug.print("query {s}", .{tc});
+            return e;
+        };
+        defer s.deinit(alloc);
+    }
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: CSS" {
+    try testing.htmlRunner("css.html");
+}

src/browser/css/libdom.zig

@@ -0,0 +1,427 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+const css = @import("css.zig");
+const Allocator = std.mem.Allocator;
+
+// Node implementation with Netsurf Libdom C lib.
+pub const Node = struct {
+    node: *parser.Node,
+
+    pub fn firstChild(n: Node) !?Node {
+        const c = try parser.nodeFirstChild(n.node);
+        if (c) |cc| return .{ .node = cc };
+
+        return null;
+    }
+
+    pub fn lastChild(n: Node) !?Node {
+        const c = try parser.nodeLastChild(n.node);
+        if (c) |cc| return .{ .node = cc };
+
+        return null;
+    }
+
+    pub fn nextSibling(n: Node) !?Node {
+        const c = try parser.nodeNextSibling(n.node);
+        if (c) |cc| return .{ .node = cc };
+
+        return null;
+    }
+
+    pub fn prevSibling(n: Node) !?Node {
+        const c = try parser.nodePreviousSibling(n.node);
+        if (c) |cc| return .{ .node = cc };
+
+        return null;
+    }
+
+    pub fn parent(n: Node) !?Node {
+        const c = try parser.nodeParentNode(n.node);
+        if (c) |cc| return .{ .node = cc };
+
+        return null;
+    }
+
+    pub fn isElement(n: Node) bool {
+        const t = parser.nodeType(n.node) catch return false;
+        return t == .element;
+    }
+
+    pub fn isDocument(n: Node) bool {
+        const t = parser.nodeType(n.node) catch return false;
+        return t == .document;
+    }
+
+    pub fn isComment(n: Node) bool {
+        const t = parser.nodeType(n.node) catch return false;
+        return t == .comment;
+    }
+
+    pub fn isText(n: Node) bool {
+        const t = parser.nodeType(n.node) catch return false;
+        return t == .text;
+    }
+
+    pub fn text(n: Node) !?[]const u8 {
+        const data = try parser.nodeTextContent(n.node);
+        if (data == null) return null;
+        if (data.?.len == 0) return null;
+
+        return std.mem.trim(u8, data.?, &std.ascii.whitespace);
+    }
+
+    pub fn isEmptyText(n: Node) !bool {
+        const data = try parser.nodeTextContent(n.node);
+        if (data == null) return true;
+        if (data.?.len == 0) return true;
+
+        return std.mem.trim(u8, data.?, &std.ascii.whitespace).len == 0;
+    }
+
+    pub fn tag(n: Node) ![]const u8 {
+        return try parser.nodeName(n.node);
+    }
+
+    pub fn attr(n: Node, key: []const u8) !?[]const u8 {
+        if (!n.isElement()) return null;
+        return try parser.elementGetAttribute(parser.nodeToElement(n.node), key);
+    }
+
+    pub fn eql(a: Node, b: Node) bool {
+        return a.node == b.node;
+    }
+};
+
+const MatcherTest = struct {
+    const Nodes = std.ArrayListUnmanaged(Node);
+
+    nodes: Nodes,
+    allocator: Allocator,
+
+    fn init(allocator: Allocator) MatcherTest {
+        return .{
+            .nodes = .empty,
+            .allocator = allocator,
+        };
+    }
+
+    fn deinit(m: *MatcherTest) void {
+        m.nodes.deinit(m.allocator);
+    }
+
+    fn reset(m: *MatcherTest) void {
+        m.nodes.clearRetainingCapacity();
+    }
+
+    pub fn match(m: *MatcherTest, n: Node) !void {
+        try m.nodes.append(m.allocator, n);
+    }
+};
+
+test "Browser.CSS.Libdom: matchFirst" {
+    const alloc = std.testing.allocator;
+
+    try parser.init();
+    defer parser.deinit();
+
+    var matcher = MatcherTest.init(alloc);
+    defer matcher.deinit();
+
+    const testcases = [_]struct {
+        q: []const u8,
+        html: []const u8,
+        exp: usize,
+    }{
+        .{ .q = "address", .html = "<body><address>This address...</address></body>", .exp = 1 },
+        .{ .q = "*", .html = "<!-- comment --><html><head></head><body>text</body></html>", .exp = 1 },
+        .{ .q = "*", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "#foo", .html = "<p id=\"foo\"><p id=\"bar\">", .exp = 1 },
+        .{ .q = "li#t1", .html = "<ul><li id=\"t1\"><p id=\"t1\">", .exp = 1 },
+        .{ .q = ".t3", .html = "<ul><li class=\"t1\"><li class=\"t2 t3\">", .exp = 1 },
+        .{ .q = "*#t4", .html = "<ol><li id=\"t4\"><li id=\"t44\">", .exp = 1 },
+        .{ .q = ".t1", .html = "<ul><li class=\"t1\"><li class=\"t2\">", .exp = 1 },
+        .{ .q = "p.t1", .html = "<p class=\"t1 t2\">", .exp = 1 },
+        .{ .q = "div.teST", .html = "<div class=\"test\">", .exp = 0 },
+        .{ .q = ".t1.fail", .html = "<p class=\"t1 t2\">", .exp = 0 },
+        .{ .q = "p.t1.t2", .html = "<p class=\"t1 t2\">", .exp = 1 },
+        .{ .q = "p.--t1", .html = "<p class=\"--t1 --t2\">", .exp = 1 },
+        .{ .q = "p.--t1.--t2", .html = "<p class=\"--t1 --t2\">", .exp = 1 },
+        .{ .q = "p[title]", .html = "<p><p title=\"title\">", .exp = 1 },
+        .{ .q = "div[class=\"red\" i]", .html = "<div><div class=\"Red\">", .exp = 1 },
+        .{ .q = "address[title=\"foo\"]", .html = "<address><address title=\"foo\"><address title=\"bar\">", .exp = 1 },
+        .{ .q = "address[title=\"FoOIgnoRECaSe\" i]", .html = "<address><address title=\"fooIgnoreCase\"><address title=\"bar\">", .exp = 1 },
+        .{ .q = "address[title!=\"foo\"]", .html = "<address><address title=\"foo\"><address title=\"bar\">", .exp = 1 },
+        .{ .q = "address[title!=\"foo\" i]", .html = "<address><address title=\"FOO\"><address title=\"bar\">", .exp = 1 },
+        .{ .q = "p[title!=\"FooBarUFoo\" i]", .html = "<p title=\"fooBARuFOO\"><p title=\"varfoo\">", .exp = 1 },
+        .{ .q = "[   title        ~=       foo    ]", .html = "<p title=\"tot foo bar\">", .exp = 1 },
+        .{ .q = "p[title~=\"FOO\" i]", .html = "<p title=\"tot foo bar\">", .exp = 1 },
+        .{ .q = "p[title~=toofoo i]", .html = "<p title=\"tot foo bar\">", .exp = 0 },
+        .{ .q = "[title~=\"hello world\"]", .html = "<p title=\"hello world\">", .exp = 0 },
+        .{ .q = "[title~=\"hello\" i]", .html = "<p title=\"HELLO world\">", .exp = 1 },
+        .{ .q = "[title~=\"hello\"          I]", .html = "<p title=\"HELLO world\">", .exp = 1 },
+        .{ .q = "[lang|=\"en\"]", .html = "<p lang=\"en\"><p lang=\"en-gb\"><p lang=\"enough\"><p lang=\"fr-en\">", .exp = 1 },
+        .{ .q = "[lang|=\"EN\" i]", .html = "<p lang=\"en\"><p lang=\"En-gb\"><p lang=\"enough\"><p lang=\"fr-en\">", .exp = 1 },
+        .{ .q = "[lang|=\"EN\"     i]", .html = "<p lang=\"en\"><p lang=\"En-gb\"><p lang=\"enough\"><p lang=\"fr-en\">", .exp = 1 },
+        .{ .q = "[title^=\"foo\"]", .html = "<p title=\"foobar\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title^=\"foo\" i]", .html = "<p title=\"FooBAR\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title$=\"bar\"]", .html = "<p title=\"foobar\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title$=\"BAR\" i]", .html = "<p title=\"foobar\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title*=\"bar\"]", .html = "<p title=\"foobarufoo\">", .exp = 1 },
+        .{ .q = "[title*=\"BaRu\" i]", .html = "<p title=\"foobarufoo\">", .exp = 1 },
+        .{ .q = "[title*=\"BaRu\" I]", .html = "<p title=\"foobarufoo\">", .exp = 1 },
+        .{ .q = "p[class$=\" \"]", .html = "<p class=\" \">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class$=\"\"]", .html = "<p class=\"\">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class^=\" \"]", .html = "<p class=\" \">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class^=\"\"]", .html = "<p class=\"\">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class*=\" \"]", .html = "<p class=\" \">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class*=\"\"]", .html = "<p class=\"\">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "input[name=Sex][value=F]", .html = "<input type=\"radio\" name=\"Sex\" value=\"F\"/>", .exp = 1 },
+        .{ .q = "table[border=\"0\"][cellpadding=\"0\"][cellspacing=\"0\"]", .html = "<table border=\"0\" cellpadding=\"0\" cellspacing=\"0\" style=\"table-layout: fixed; width: 100%; border: 0 dashed; border-color: #FFFFFF\"><tr style=\"height:64px\">aaa</tr></table>", .exp = 1 },
+        .{ .q = ".t1:not(.t2)", .html = "<p class=\"t1 t2\">", .exp = 0 },
+        .{ .q = "div:not(.t1)", .html = "<div class=\"t3\">", .exp = 1 },
+        .{ .q = "div:not([class=\"t2\"])", .html = "<div><div class=\"t2\"><div class=\"t3\">", .exp = 1 },
+        .{ .q = "li:nth-child(odd)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 1 },
+        .{ .q = "li:nth-child(even)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 1 },
+        .{ .q = "li:nth-child(-n+2)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 1 },
+        .{ .q = "li:nth-child(3n+1)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 1 },
+        .{ .q = "li:nth-last-child(odd)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 1 },
+        .{ .q = "li:nth-last-child(even)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 1 },
+        .{ .q = "li:nth-last-child(-n+2)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 1 },
+        .{ .q = "li:nth-last-child(3n+1)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 1 },
+        .{ .q = "span:first-child", .html = "<p>some text <span id=\"1\">and a span</span><span id=\"2\"> and another</span></p>", .exp = 1 },
+        .{ .q = "span:last-child", .html = "<span>a span</span> and some text", .exp = 1 },
+        .{ .q = "p:nth-of-type(2)", .html = "<address></address><p id=1><p id=2>", .exp = 1 },
+        .{ .q = "p:nth-last-of-type(2)", .html = "<address></address><p id=1><p id=2></p><a>", .exp = 1 },
+        .{ .q = "p:last-of-type", .html = "<address></address><p id=1><p id=2></p><a>", .exp = 1 },
+        .{ .q = "p:first-of-type", .html = "<address></address><p id=1><p id=2></p><a>", .exp = 1 },
+        .{ .q = "p:only-child", .html = "<div><p id=\"1\"></p><a></a></div><div><p id=\"2\"></p></div>", .exp = 1 },
+        .{ .q = "p:only-of-type", .html = "<div><p id=\"1\"></p><a></a></div><div><p id=\"2\"></p><p id=\"3\"></p></div>", .exp = 1 },
+        .{ .q = ":empty", .html = "<p id=\"1\"><!-- --><p id=\"2\">Hello<p id=\"3\"><span>", .exp = 1 },
+        .{ .q = "div p", .html = "<div><p id=\"1\"><table><tr><td><p id=\"2\"></table></div><p id=\"3\">", .exp = 1 },
+        .{ .q = "div table p", .html = "<div><p id=\"1\"><table><tr><td><p id=\"2\"></table></div><p id=\"3\">", .exp = 1 },
+        .{ .q = "div > p", .html = "<div><p id=\"1\"><div><p id=\"2\"></div><table><tr><td><p id=\"3\"></table></div>", .exp = 1 },
+        .{ .q = "p ~ p", .html = "<p id=\"1\"><p id=\"2\"></p><address></address><p id=\"3\">", .exp = 1 },
+        .{ .q = "p + p", .html = "<p id=\"1\"></p> <!--comment--> <p id=\"2\"></p><address></address><p id=\"3\">", .exp = 1 },
+        .{ .q = "li, p", .html = "<ul><li></li><li></li></ul><p>", .exp = 1 },
+        .{ .q = "p +/*This is a comment*/ p", .html = "<p id=\"1\"><p id=\"2\"></p><address></address><p id=\"3\">", .exp = 1 },
+        // .{ .q = "p:contains(\"that wraps\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 1 },
+        .{ .q = "p:containsOwn(\"that wraps\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 0 },
+        .{ .q = ":containsOwn(\"inner\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 1 },
+        .{ .q = ":containsOwn(\"Inner\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 0 },
+        .{ .q = "p:containsOwn(\"block\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 1 },
+        // .{ .q = "div:has(#p1)", .html = "<div id=\"d1\"><p id=\"p1\"><span>text content</span></p></div><div id=\"d2\"/>", .exp = 1 },
+        .{ .q = "div:has(:containsOwn(\"2\"))", .html = "<div id=\"d1\"><p id=\"p1\"><span>contents 1</span></p></div> <div id=\"d2\"><p>contents <em>2</em></p></div>", .exp = 1 },
+        .{ .q = "body :has(:containsOwn(\"2\"))", .html = "<body><div id=\"d1\"><p id=\"p1\"><span>contents 1</span></p></div> <div id=\"d2\"><p id=\"p2\">contents <em>2</em></p></div></body>", .exp = 1 },
+        .{ .q = "body :haschild(:containsOwn(\"2\"))", .html = "<body><div id=\"d1\"><p id=\"p1\"><span>contents 1</span></p></div> <div id=\"d2\"><p id=\"p2\">contents <em>2</em></p></div></body>", .exp = 1 },
+        // .{ .q = "p:matches([\\d])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:matches([a-z])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:matches([a-zA-Z])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:matches([^\\d])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:matches(^(0|a))", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:matches(^\\d+$)", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:not(:matches(^\\d+$))", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "div :matchesOwn(^\\d+$)", .html = "<div><p id=\"p1\">01234<em>567</em>89</p><div>", .exp = 1 },
+        // .{ .q = "[href#=(fina)]:not([href#=(\\/\\/[^\\/]+untrusted)])", .html = "<ul> <li><a id=\"a1\" href=\"http://www.google.com/finance\"></a> <li><a id=\"a2\" href=\"http://finance.yahoo.com/\"></a> <li><a id=\"a2\" href=\"http://finance.untrusted.com/\"/> <li><a id=\"a3\" href=\"https://www.google.com/news\"/> <li><a id=\"a4\" href=\"http://news.yahoo.com\"/> </ul>", .exp = 1 },
+        // .{ .q = "[href#=(^https:\\/\\/[^\\/]*\\/?news)]", .html = "<ul> <li><a id=\"a1\" href=\"http://www.google.com/finance\"/> <li><a id=\"a2\" href=\"http://finance.yahoo.com/\"/> <li><a id=\"a3\" href=\"https://www.google.com/news\"></a> <li><a id=\"a4\" href=\"http://news.yahoo.com\"/> </ul>", .exp = 1 },
+        .{ .q = ":input", .html = "<form> <label>Username <input type=\"text\" name=\"username\" /></label> <label>Password <input type=\"password\" name=\"password\" /></label> <label>Country <select name=\"country\"> <option value=\"ca\">Canada</option> <option value=\"us\">United States</option> </select> </label> <label>Bio <textarea name=\"bio\"></textarea></label> <button>Sign up</button> </form>", .exp = 1 },
+        .{ .q = ":root", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "*:root", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "html:nth-child(1)", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "*:root:first-child", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "*:root:nth-child(1)", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "a:not(:root)", .html = "<html><head></head><body><a href=\"http://www.foo.com\"></a></body></html>", .exp = 1 },
+        .{ .q = "body > *:nth-child(3n+2)", .html = "<html><head></head><body><p></p><div></div><span></span><a></a><form></form></body></html>", .exp = 1 },
+        .{ .q = "input:disabled", .html = "<html><head></head><body><fieldset disabled><legend id=\"1\"><input id=\"i1\"/></legend><legend id=\"2\"><input id=\"i2\"/></legend></fieldset></body></html>", .exp = 1 },
+        .{ .q = ":disabled", .html = "<html><head></head><body><fieldset disabled></fieldset></body></html>", .exp = 1 },
+        .{ .q = ":enabled", .html = "<html><head></head><body><fieldset></fieldset></body></html>", .exp = 1 },
+        .{ .q = "div.class1, div.class2", .html = "<div class=class1></div><div class=class2></div><div class=class3></div>", .exp = 1 },
+    };
+
+    for (testcases) |tc| {
+        matcher.reset();
+
+        const doc = try parser.documentHTMLParseFromStr(tc.html);
+        defer parser.documentHTMLClose(doc) catch {};
+
+        const s = css.parse(alloc, tc.q, .{}) catch |e| {
+            std.debug.print("parse, query: {s}\n", .{tc.q});
+            return e;
+        };
+
+        defer s.deinit(alloc);
+
+        const node = Node{ .node = parser.documentHTMLToNode(doc) };
+
+        _ = css.matchFirst(&s, node, &matcher) catch |e| {
+            std.debug.print("match, query: {s}\n", .{tc.q});
+            return e;
+        };
+        std.testing.expectEqual(tc.exp, matcher.nodes.items.len) catch |e| {
+            std.debug.print("expectation, query: {s}\n", .{tc.q});
+            return e;
+        };
+    }
+}
+
+test "Browser.CSS.Libdom: matchAll" {
+    const alloc = std.testing.allocator;
+
+    try parser.init();
+    defer parser.deinit();
+
+    var matcher = MatcherTest.init(alloc);
+    defer matcher.deinit();
+
+    const testcases = [_]struct {
+        q: []const u8,
+        html: []const u8,
+        exp: usize,
+    }{
+        .{ .q = "address", .html = "<body><address>This address...</address></body>", .exp = 1 },
+        .{ .q = "*", .html = "<!-- comment --><html><head></head><body>text</body></html>", .exp = 3 },
+        .{ .q = "*", .html = "<html><head></head><body></body></html>", .exp = 3 },
+        .{ .q = "#foo", .html = "<p id=\"foo\"><p id=\"bar\">", .exp = 1 },
+        .{ .q = "li#t1", .html = "<ul><li id=\"t1\"><p id=\"t1\">", .exp = 1 },
+        .{ .q = ".t3", .html = "<ul><li class=\"t1\"><li class=\"t2 t3\">", .exp = 1 },
+        .{ .q = "*#t4", .html = "<ol><li id=\"t4\"><li id=\"t44\">", .exp = 1 },
+        .{ .q = ".t1", .html = "<ul><li class=\"t1\"><li class=\"t2\">", .exp = 1 },
+        .{ .q = "p.t1", .html = "<p class=\"t1 t2\">", .exp = 1 },
+        .{ .q = "div.teST", .html = "<div class=\"test\">", .exp = 0 },
+        .{ .q = ".t1.fail", .html = "<p class=\"t1 t2\">", .exp = 0 },
+        .{ .q = "p.t1.t2", .html = "<p class=\"t1 t2\">", .exp = 1 },
+        .{ .q = "p.--t1", .html = "<p class=\"--t1 --t2\">", .exp = 1 },
+        .{ .q = "p.--t1.--t2", .html = "<p class=\"--t1 --t2\">", .exp = 1 },
+        .{ .q = "p[title]", .html = "<p><p title=\"title\">", .exp = 1 },
+        .{ .q = "div[class=\"red\" i]", .html = "<div><div class=\"Red\">", .exp = 1 },
+        .{ .q = "address[title=\"foo\"]", .html = "<address><address title=\"foo\"><address title=\"bar\">", .exp = 1 },
+        .{ .q = "address[title=\"FoOIgnoRECaSe\" i]", .html = "<address><address title=\"fooIgnoreCase\"><address title=\"bar\">", .exp = 1 },
+        .{ .q = "address[title!=\"foo\"]", .html = "<address><address title=\"foo\"><address title=\"bar\">", .exp = 2 },
+        .{ .q = "address[title!=\"foo\" i]", .html = "<address><address title=\"FOO\"><address title=\"bar\">", .exp = 2 },
+        .{ .q = "p[title!=\"FooBarUFoo\" i]", .html = "<p title=\"fooBARuFOO\"><p title=\"varfoo\">", .exp = 1 },
+        .{ .q = "[   title        ~=       foo    ]", .html = "<p title=\"tot foo bar\">", .exp = 1 },
+        .{ .q = "p[title~=\"FOO\" i]", .html = "<p title=\"tot foo bar\">", .exp = 1 },
+        .{ .q = "p[title~=toofoo i]", .html = "<p title=\"tot foo bar\">", .exp = 0 },
+        .{ .q = "[title~=\"hello world\"]", .html = "<p title=\"hello world\">", .exp = 0 },
+        .{ .q = "[title~=\"hello\" i]", .html = "<p title=\"HELLO world\">", .exp = 1 },
+        .{ .q = "[title~=\"hello\"          I]", .html = "<p title=\"HELLO world\">", .exp = 1 },
+        .{ .q = "[lang|=\"en\"]", .html = "<p lang=\"en\"><p lang=\"en-gb\"><p lang=\"enough\"><p lang=\"fr-en\">", .exp = 2 },
+        .{ .q = "[lang|=\"EN\" i]", .html = "<p lang=\"en\"><p lang=\"En-gb\"><p lang=\"enough\"><p lang=\"fr-en\">", .exp = 2 },
+        .{ .q = "[lang|=\"EN\"     i]", .html = "<p lang=\"en\"><p lang=\"En-gb\"><p lang=\"enough\"><p lang=\"fr-en\">", .exp = 2 },
+        .{ .q = "[title^=\"foo\"]", .html = "<p title=\"foobar\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title^=\"foo\" i]", .html = "<p title=\"FooBAR\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title$=\"bar\"]", .html = "<p title=\"foobar\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title$=\"BAR\" i]", .html = "<p title=\"foobar\"><p title=\"barfoo\">", .exp = 1 },
+        .{ .q = "[title*=\"bar\"]", .html = "<p title=\"foobarufoo\">", .exp = 1 },
+        .{ .q = "[title*=\"BaRu\" i]", .html = "<p title=\"foobarufoo\">", .exp = 1 },
+        .{ .q = "[title*=\"BaRu\" I]", .html = "<p title=\"foobarufoo\">", .exp = 1 },
+        .{ .q = "p[class$=\" \"]", .html = "<p class=\" \">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class$=\"\"]", .html = "<p class=\"\">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class^=\" \"]", .html = "<p class=\" \">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class^=\"\"]", .html = "<p class=\"\">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class*=\" \"]", .html = "<p class=\" \">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "p[class*=\"\"]", .html = "<p class=\"\">This text should be green.</p><p>This text should be green.</p>", .exp = 0 },
+        .{ .q = "input[name=Sex][value=F]", .html = "<input type=\"radio\" name=\"Sex\" value=\"F\"/>", .exp = 1 },
+        .{ .q = "table[border=\"0\"][cellpadding=\"0\"][cellspacing=\"0\"]", .html = "<table border=\"0\" cellpadding=\"0\" cellspacing=\"0\" style=\"table-layout: fixed; width: 100%; border: 0 dashed; border-color: #FFFFFF\"><tr style=\"height:64px\">aaa</tr></table>", .exp = 1 },
+        .{ .q = ".t1:not(.t2)", .html = "<p class=\"t1 t2\">", .exp = 0 },
+        .{ .q = "div:not(.t1)", .html = "<div class=\"t3\">", .exp = 1 },
+        .{ .q = "div:not([class=\"t2\"])", .html = "<div><div class=\"t2\"><div class=\"t3\">", .exp = 2 },
+        .{ .q = "li:nth-child(odd)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 2 },
+        .{ .q = "li:nth-child(even)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 1 },
+        .{ .q = "li:nth-child(-n+2)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 2 },
+        .{ .q = "li:nth-child(3n+1)", .html = "<ol><li id=1><li id=2><li id=3></ol>", .exp = 1 },
+        .{ .q = "li:nth-last-child(odd)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 2 },
+        .{ .q = "li:nth-last-child(even)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 2 },
+        .{ .q = "li:nth-last-child(-n+2)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 2 },
+        .{ .q = "li:nth-last-child(3n+1)", .html = "<ol><li id=1><li id=2><li id=3><li id=4></ol>", .exp = 2 },
+        .{ .q = "span:first-child", .html = "<p>some text <span id=\"1\">and a span</span><span id=\"2\"> and another</span></p>", .exp = 1 },
+        .{ .q = "span:last-child", .html = "<span>a span</span> and some text", .exp = 1 },
+        .{ .q = "p:nth-of-type(2)", .html = "<address></address><p id=1><p id=2>", .exp = 1 },
+        .{ .q = "p:nth-last-of-type(2)", .html = "<address></address><p id=1><p id=2></p><a>", .exp = 1 },
+        .{ .q = "p:last-of-type", .html = "<address></address><p id=1><p id=2></p><a>", .exp = 1 },
+        .{ .q = "p:first-of-type", .html = "<address></address><p id=1><p id=2></p><a>", .exp = 1 },
+        .{ .q = "p:only-child", .html = "<div><p id=\"1\"></p><a></a></div><div><p id=\"2\"></p></div>", .exp = 1 },
+        .{ .q = "p:only-of-type", .html = "<div><p id=\"1\"></p><a></a></div><div><p id=\"2\"></p><p id=\"3\"></p></div>", .exp = 1 },
+        .{ .q = ":empty", .html = "<p id=\"1\"><!-- --><p id=\"2\">Hello<p id=\"3\"><span>", .exp = 3 },
+        .{ .q = "div p", .html = "<div><p id=\"1\"><table><tr><td><p id=\"2\"></table></div><p id=\"3\">", .exp = 2 },
+        .{ .q = "div table p", .html = "<div><p id=\"1\"><table><tr><td><p id=\"2\"></table></div><p id=\"3\">", .exp = 1 },
+        .{ .q = "div > p", .html = "<div><p id=\"1\"><div><p id=\"2\"></div><table><tr><td><p id=\"3\"></table></div>", .exp = 2 },
+        .{ .q = "p ~ p", .html = "<p id=\"1\"><p id=\"2\"></p><address></address><p id=\"3\">", .exp = 2 },
+        .{ .q = "p + p", .html = "<p id=\"1\"></p> <!--comment--> <p id=\"2\"></p><address></address><p id=\"3\">", .exp = 1 },
+        .{ .q = "li, p", .html = "<ul><li></li><li></li></ul><p>", .exp = 3 },
+        .{ .q = "p +/*This is a comment*/ p", .html = "<p id=\"1\"><p id=\"2\"></p><address></address><p id=\"3\">", .exp = 1 },
+        // .{ .q = "p:contains(\"that wraps\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 1 },
+        .{ .q = "p:containsOwn(\"that wraps\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 0 },
+        .{ .q = ":containsOwn(\"inner\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 1 },
+        .{ .q = ":containsOwn(\"Inner\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 0 },
+        .{ .q = "p:containsOwn(\"block\")", .html = "<p>Text block that <span>wraps inner text</span> and continues</p>", .exp = 1 },
+        .{ .q = "div:has(#p1)", .html = "<div id=\"d1\"><p id=\"p1\"><span>text content</span></p></div><div id=\"d2\"/>", .exp = 1 },
+        .{ .q = "div:has(:containsOwn(\"2\"))", .html = "<div id=\"d1\"><p id=\"p1\"><span>contents 1</span></p></div> <div id=\"d2\"><p>contents <em>2</em></p></div>", .exp = 1 },
+        .{ .q = "body :has(:containsOwn(\"2\"))", .html = "<body><div id=\"d1\"><p id=\"p1\"><span>contents 1</span></p></div> <div id=\"d2\"><p id=\"p2\">contents <em>2</em></p></div></body>", .exp = 2 },
+        .{ .q = "body :haschild(:containsOwn(\"2\"))", .html = "<body><div id=\"d1\"><p id=\"p1\"><span>contents 1</span></p></div> <div id=\"d2\"><p id=\"p2\">contents <em>2</em></p></div></body>", .exp = 1 },
+        // .{ .q = "p:matches([\\d])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 2 },
+        // .{ .q = "p:matches([a-z])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:matches([a-zA-Z])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 2 },
+        // .{ .q = "p:matches([^\\d])", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 2 },
+        // .{ .q = "p:matches(^(0|a))", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 3 },
+        // .{ .q = "p:matches(^\\d+$)", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 1 },
+        // .{ .q = "p:not(:matches(^\\d+$))", .html = "<p id=\"p1\">0123456789</p><p id=\"p2\">abcdef</p><p id=\"p3\">0123ABCD</p>", .exp = 2 },
+        // .{ .q = "div :matchesOwn(^\\d+$)", .html = "<div><p id=\"p1\">01234<em>567</em>89</p><div>", .exp = 2 },
+        // .{ .q = "[href#=(fina)]:not([href#=(\\/\\/[^\\/]+untrusted)])", .html = "<ul> <li><a id=\"a1\" href=\"http://www.google.com/finance\"></a> <li><a id=\"a2\" href=\"http://finance.yahoo.com/\"></a> <li><a id=\"a2\" href=\"http://finance.untrusted.com/\"/> <li><a id=\"a3\" href=\"https://www.google.com/news\"/> <li><a id=\"a4\" href=\"http://news.yahoo.com\"/> </ul>", .exp = 2 },
+        // .{ .q = "[href#=(^https:\\/\\/[^\\/]*\\/?news)]", .html = "<ul> <li><a id=\"a1\" href=\"http://www.google.com/finance\"/> <li><a id=\"a2\" href=\"http://finance.yahoo.com/\"/> <li><a id=\"a3\" href=\"https://www.google.com/news\"></a> <li><a id=\"a4\" href=\"http://news.yahoo.com\"/> </ul>", .exp = 1 },
+        .{ .q = ":input", .html = "<form> <label>Username <input type=\"text\" name=\"username\" /></label> <label>Password <input type=\"password\" name=\"password\" /></label> <label>Country <select name=\"country\"> <option value=\"ca\">Canada</option> <option value=\"us\">United States</option> </select> </label> <label>Bio <textarea name=\"bio\"></textarea></label> <button>Sign up</button> </form>", .exp = 5 },
+        .{ .q = ":root", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "*:root", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "html:nth-child(1)", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "*:root:first-child", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "*:root:nth-child(1)", .html = "<html><head></head><body></body></html>", .exp = 1 },
+        .{ .q = "a:not(:root)", .html = "<html><head></head><body><a href=\"http://www.foo.com\"></a></body></html>", .exp = 1 },
+        .{ .q = "body > *:nth-child(3n+2)", .html = "<html><head></head><body><p></p><div></div><span></span><a></a><form></form></body></html>", .exp = 2 },
+        .{ .q = "input:disabled", .html = "<html><head></head><body><fieldset disabled><legend id=\"1\"><input id=\"i1\"/></legend><legend id=\"2\"><input id=\"i2\"/></legend></fieldset></body></html>", .exp = 1 },
+        .{ .q = ":disabled", .html = "<html><head></head><body><fieldset disabled></fieldset></body></html>", .exp = 1 },
+        .{ .q = ":enabled", .html = "<html><head></head><body><fieldset></fieldset></body></html>", .exp = 1 },
+        .{ .q = "div.class1, div.class2", .html = "<div class=class1></div><div class=class2></div><div class=class3></div>", .exp = 2 },
+    };
+
+    for (testcases) |tc| {
+        matcher.reset();
+
+        const doc = try parser.documentHTMLParseFromStr(tc.html);
+        defer parser.documentHTMLClose(doc) catch {};
+
+        const s = css.parse(alloc, tc.q, .{}) catch |e| {
+            std.debug.print("parse, query: {s}\n", .{tc.q});
+            return e;
+        };
+        defer s.deinit(alloc);
+
+        const node = Node{ .node = parser.documentHTMLToNode(doc) };
+
+        _ = css.matchAll(&s, node, &matcher) catch |e| {
+            std.debug.print("match, query: {s}\n", .{tc.q});
+            return e;
+        };
+        std.testing.expectEqual(tc.exp, matcher.nodes.items.len) catch |e| {
+            std.debug.print("expectation, query: {s}\n", .{tc.q});
+            return e;
+        };
+    }
+}

src/browser/css/parser.zig

@@ -0,0 +1,994 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+// CSS Selector parser
+// This file is a rewrite in Zig of Cascadia CSS Selector parser.
+// see https://github.com/andybalholm/cascadia
+// see https://github.com/andybalholm/cascadia/blob/master/parser.go
+const std = @import("std");
+const ascii = std.ascii;
+const Allocator = std.mem.Allocator;
+
+const selector = @import("selector.zig");
+const Selector = selector.Selector;
+const PseudoClass = selector.PseudoClass;
+const AttributeOP = selector.AttributeOP;
+const Combinator = selector.Combinator;
+
+const REPLACEMENT_CHARACTER = &.{ 239, 191, 189 };
+
+pub const ParseError = error{
+    ExpectedSelector,
+    ExpectedIdentifier,
+    ExpectedName,
+    ExpectedIDSelector,
+    ExpectedClassSelector,
+    ExpectedAttributeSelector,
+    ExpectedString,
+    ExpectedRegexp,
+    ExpectedPseudoClassSelector,
+    ExpectedParenthesis,
+    ExpectedParenthesisClose,
+    ExpectedNthExpression,
+    ExpectedInteger,
+    InvalidEscape,
+    EscapeLineEndingOutsideString,
+    InvalidUnicode,
+    UnicodeIsNotHandled,
+    WriteError,
+    PseudoElementNotAtSelectorEnd,
+    PseudoElementNotUnique,
+    PseudoElementDisabled,
+    InvalidAttributeOperator,
+    InvalidAttributeSelector,
+    InvalidString,
+    InvalidRegexp,
+    InvalidPseudoClassSelector,
+    EmptyPseudoClassSelector,
+    InvalidPseudoClass,
+    InvalidPseudoElement,
+    UnmatchParenthesis,
+    NotHandled,
+    UnknownPseudoSelector,
+    InvalidNthExpression,
+} || PseudoClass.Error || Combinator.Error || std.mem.Allocator.Error;
+
+pub const ParseOptions = struct {
+    accept_pseudo_elts: bool = true,
+};
+
+pub const Parser = struct {
+    s: []const u8, // string to parse
+    i: usize = 0, // current position
+
+    opts: ParseOptions,
+
+    pub fn parse(p: *Parser, allocator: Allocator) ParseError!Selector {
+        return p.parseSelectorGroup(allocator);
+    }
+
+    // skipWhitespace consumes whitespace characters and comments.
+    // It returns true if there was actually anything to skip.
+    fn skipWhitespace(p: *Parser) bool {
+        var i = p.i;
+        while (i < p.s.len) {
+            const c = p.s[i];
+            // Whitespaces.
+            if (ascii.isWhitespace(c)) {
+                i += 1;
+                continue;
+            }
+
+            // Comments.
+            if (c == '/') {
+                if (std.mem.startsWith(u8, p.s[i..], "/*")) {
+                    if (std.mem.indexOf(u8, p.s[i..], "*/")) |end| {
+                        i += end + "*/".len;
+                        continue;
+                    }
+                }
+            }
+            break;
+        }
+
+        if (i > p.i) {
+            p.i = i;
+            return true;
+        }
+
+        return false;
+    }
+
+    // parseSimpleSelectorSequence parses a selector sequence that applies to
+    // a single element.
+    fn parseSimpleSelectorSequence(p: *Parser, allocator: Allocator) ParseError!Selector {
+        if (p.i >= p.s.len) {
+            return ParseError.ExpectedSelector;
+        }
+
+        var buf: std.ArrayListUnmanaged(Selector) = .empty;
+        defer buf.deinit(allocator);
+
+        switch (p.s[p.i]) {
+            '*' => {
+                // It's the universal selector. Just skip over it, since it
+                // doesn't affect the meaning.
+                p.i += 1;
+
+                // other version of universal selector
+                if (p.i + 2 < p.s.len and std.mem.eql(u8, "|*", p.s[p.i .. p.i + 2])) {
+                    p.i += 2;
+                }
+            },
+            '#', '.', '[', ':' => {
+                // There's no type selector. Wait to process the other till the
+                // main loop.
+            },
+            else => try buf.append(allocator, try p.parseTypeSelector(allocator)),
+        }
+
+        var pseudo_elt: ?PseudoClass = null;
+
+        loop: while (p.i < p.s.len) {
+            var ns: Selector = switch (p.s[p.i]) {
+                '#' => try p.parseIDSelector(allocator),
+                '.' => try p.parseClassSelector(allocator),
+                '[' => try p.parseAttributeSelector(allocator),
+                ':' => try p.parsePseudoclassSelector(allocator),
+                else => break :loop,
+            };
+            errdefer ns.deinit(allocator);
+
+            // From https://drafts.csswg.org/selectors-3/#pseudo-elements :
+            // "Only one pseudo-element may appear per selector, and if present
+            // it must appear after the sequence of simple selectors that
+            // represents the subjects of the selector.""
+            switch (ns) {
+                .pseudo_element => |e| {
+                    //  We found a pseudo-element.
+                    //  Only one pseudo-element is accepted per selector.
+                    if (pseudo_elt != null) return ParseError.PseudoElementNotUnique;
+                    if (!p.opts.accept_pseudo_elts) return ParseError.PseudoElementDisabled;
+
+                    pseudo_elt = e;
+                    ns.deinit(allocator);
+                },
+                else => {
+                    if (pseudo_elt != null) return ParseError.PseudoElementNotAtSelectorEnd;
+                    try buf.append(allocator, ns);
+                },
+            }
+        }
+
+        // no need wrap the selectors in compoundSelector
+        if (buf.items.len == 1 and pseudo_elt == null) {
+            return buf.items[0];
+        }
+
+        return .{
+            .compound = .{ .selectors = try buf.toOwnedSlice(allocator), .pseudo_elt = pseudo_elt },
+        };
+    }
+
+    // parseTypeSelector parses a type selector (one that matches by tag name).
+    fn parseTypeSelector(p: *Parser, allocator: Allocator) ParseError!Selector {
+        var buf: std.ArrayListUnmanaged(u8) = .empty;
+        defer buf.deinit(allocator);
+        try p.parseIdentifier(buf.writer(allocator));
+
+        return .{ .tag = try buf.toOwnedSlice(allocator) };
+    }
+
+    // parseIdentifier parses an identifier.
+    fn parseIdentifier(p: *Parser, w: anytype) ParseError!void {
+        const prefix = '-';
+        var numPrefix: usize = 0;
+
+        while (p.s.len > p.i and p.s[p.i] == prefix) {
+            p.i += 1;
+            numPrefix += 1;
+        }
+
+        if (p.s.len <= p.i) {
+            return ParseError.ExpectedSelector;
+        }
+
+        const c = p.s[p.i];
+        if (!(nameStart(c) or c == '\\')) {
+            return ParseError.ExpectedSelector;
+        }
+
+        var ii: usize = 0;
+        while (ii < numPrefix) {
+            w.writeByte(prefix) catch return ParseError.WriteError;
+            ii += 1;
+        }
+        try parseName(p, w);
+    }
+
+    // parseName parses a name (which is like an identifier, but doesn't have
+    // extra restrictions on the first character).
+    fn parseName(p: *Parser, w: anytype) ParseError!void {
+        const sel = p.s;
+        const sel_len = sel.len;
+
+        var i = p.i;
+        var ok = false;
+
+        while (i < sel_len) {
+            const c = sel[i];
+
+            if (nameChar(c)) {
+                const start = i;
+                while (i < sel_len and nameChar(sel[i])) i += 1;
+                w.writeAll(sel[start..i]) catch return ParseError.WriteError;
+                ok = true;
+            } else if (c == '\\') {
+                p.i = i;
+                try p.parseEscape(w);
+                i = p.i;
+                ok = true;
+            } else if (c == 0) {
+                w.writeAll(REPLACEMENT_CHARACTER) catch return ParseError.WriteError;
+                i += 1;
+                if (i == sel_len) {
+                    ok = true;
+                }
+            } else {
+                // default:
+                break;
+            }
+        }
+
+        if (!ok) return ParseError.ExpectedName;
+        p.i = i;
+    }
+
+    // parseEscape parses a backslash escape.
+    // The returned string is owned by the caller.
+    fn parseEscape(p: *Parser, w: anytype) ParseError!void {
+        const sel = p.s;
+        const sel_len = sel.len;
+
+        if (sel_len < p.i + 2 or sel[p.i] != '\\') {
+            p.i += 1;
+            w.writeAll(REPLACEMENT_CHARACTER) catch return ParseError.WriteError;
+            return;
+        }
+
+        const start = p.i + 1;
+        const c = sel[start];
+
+        // unicode escape (hex)
+        if (ascii.isHex(c)) {
+            var i: usize = start;
+            while (i < start + 6 and i < sel_len and ascii.isHex(sel[i])) {
+                i += 1;
+            }
+
+            const v = std.fmt.parseUnsigned(u21, sel[start..i], 16) catch {
+                p.i = i;
+                w.writeAll(REPLACEMENT_CHARACTER) catch return ParseError.WriteError;
+                return;
+            };
+
+            if (sel_len >= i) {
+                if (sel_len > i) {
+                    switch (sel[i]) {
+                        '\r' => {
+                            i += 1;
+                            if (sel_len > i and sel[i] == '\n') i += 1;
+                        },
+                        ' ', '\t', '\n', std.ascii.control_code.ff => i += 1,
+                        else => {},
+                    }
+                }
+                p.i = i;
+                if (v == 0) {
+                    w.writeAll(REPLACEMENT_CHARACTER) catch return ParseError.WriteError;
+                    return;
+                }
+                var buf: [4]u8 = undefined;
+                const ln = std.unicode.utf8Encode(v, &buf) catch {
+                    w.writeAll(REPLACEMENT_CHARACTER) catch return ParseError.WriteError;
+                    return;
+                };
+                w.writeAll(buf[0..ln]) catch return ParseError.WriteError;
+                return;
+            }
+        }
+
+        // Return the literal character after the backslash.
+        p.i += 2;
+        w.writeByte(sel[start]) catch return ParseError.WriteError;
+    }
+
+    // parseIDSelector parses a selector that matches by id attribute.
+    fn parseIDSelector(p: *Parser, allocator: Allocator) ParseError!Selector {
+        if (p.i >= p.s.len) return ParseError.ExpectedIDSelector;
+        if (p.s[p.i] != '#') return ParseError.ExpectedIDSelector;
+
+        p.i += 1;
+
+        var buf: std.ArrayListUnmanaged(u8) = .empty;
+        defer buf.deinit(allocator);
+
+        try p.parseName(buf.writer(allocator));
+        return .{ .id = try buf.toOwnedSlice(allocator) };
+    }
+
+    // parseClassSelector parses a selector that matches by class attribute.
+    fn parseClassSelector(p: *Parser, allocator: Allocator) ParseError!Selector {
+        if (p.i >= p.s.len) return ParseError.ExpectedClassSelector;
+        if (p.s[p.i] != '.') return ParseError.ExpectedClassSelector;
+
+        p.i += 1;
+
+        var buf: std.ArrayListUnmanaged(u8) = .empty;
+        defer buf.deinit(allocator);
+
+        try p.parseIdentifier(buf.writer(allocator));
+        return .{ .class = try buf.toOwnedSlice(allocator) };
+    }
+
+    // parseAttributeSelector parses a selector that matches by attribute value.
+    fn parseAttributeSelector(p: *Parser, allocator: Allocator) ParseError!Selector {
+        if (p.i >= p.s.len) return ParseError.ExpectedAttributeSelector;
+        if (p.s[p.i] != '[') return ParseError.ExpectedAttributeSelector;
+
+        p.i += 1;
+        _ = p.skipWhitespace();
+
+        var buf: std.ArrayListUnmanaged(u8) = .empty;
+        defer buf.deinit(allocator);
+
+        try p.parseIdentifier(buf.writer(allocator));
+        const key = try buf.toOwnedSlice(allocator);
+        errdefer allocator.free(key);
+
+        lowerstr(key);
+
+        _ = p.skipWhitespace();
+        if (p.i >= p.s.len) return ParseError.ExpectedAttributeSelector;
+        if (p.s[p.i] == ']') {
+            p.i += 1;
+            return .{ .attribute = .{ .key = key } };
+        }
+
+        if (p.i + 2 >= p.s.len) return ParseError.ExpectedAttributeSelector;
+
+        const op = try parseAttributeOP(p.s[p.i .. p.i + 2]);
+        p.i += op.len();
+
+        _ = p.skipWhitespace();
+        if (p.i >= p.s.len) return ParseError.ExpectedAttributeSelector;
+
+        buf.clearRetainingCapacity();
+        var is_val: bool = undefined;
+        if (op == .regexp) {
+            is_val = false;
+            try p.parseRegex(buf.writer(allocator));
+        } else {
+            is_val = true;
+            switch (p.s[p.i]) {
+                '\'', '"' => try p.parseString(buf.writer(allocator)),
+                else => try p.parseIdentifier(buf.writer(allocator)),
+            }
+        }
+
+        _ = p.skipWhitespace();
+        if (p.i >= p.s.len) return ParseError.ExpectedAttributeSelector;
+
+        // check if the attribute contains an ignore case flag
+        var ci = false;
+        if (p.s[p.i] == 'i' or p.s[p.i] == 'I') {
+            ci = true;
+            p.i += 1;
+        }
+
+        _ = p.skipWhitespace();
+        if (p.i >= p.s.len) return ParseError.ExpectedAttributeSelector;
+
+        if (p.s[p.i] != ']') return ParseError.InvalidAttributeSelector;
+        p.i += 1;
+
+        return .{ .attribute = .{
+            .key = key,
+            .val = if (is_val) try buf.toOwnedSlice(allocator) else null,
+            .regexp = if (!is_val) try buf.toOwnedSlice(allocator) else null,
+            .op = op,
+            .ci = ci,
+        } };
+    }
+
+    // parseString parses a single- or double-quoted string.
+    fn parseString(p: *Parser, writer: anytype) ParseError!void {
+        const sel = p.s;
+        const sel_len = sel.len;
+
+        var i = p.i;
+        if (sel_len < i + 2) return ParseError.ExpectedString;
+
+        const quote = sel[i];
+        i += 1;
+
+        loop: while (i < sel_len) {
+            switch (sel[i]) {
+                '\\' => {
+                    if (sel_len > i + 1) {
+                        const c = sel[i + 1];
+                        switch (c) {
+                            '\r' => {
+                                if (sel_len > i + 2 and sel[i + 2] == '\n') {
+                                    i += 3;
+                                    continue :loop;
+                                }
+                                i += 2;
+                                continue :loop;
+                            },
+                            '\n', std.ascii.control_code.ff => {
+                                i += 2;
+                                continue :loop;
+                            },
+                            else => {},
+                        }
+                    }
+                    p.i = i;
+                    try p.parseEscape(writer);
+                    i = p.i;
+                },
+                '\r', '\n', std.ascii.control_code.ff => return ParseError.InvalidString,
+                else => |c| {
+                    if (c == quote) break :loop;
+                    const start = i;
+                    while (i < sel_len) {
+                        const cc = sel[i];
+                        if (cc == quote or cc == '\\' or c == '\r' or c == '\n' or c == std.ascii.control_code.ff) break;
+                        i += 1;
+                    }
+                    writer.writeAll(sel[start..i]) catch return ParseError.WriteError;
+                },
+            }
+        }
+
+        if (i >= sel_len) return ParseError.InvalidString;
+
+        // Consume the final quote.
+        i += 1;
+        p.i = i;
+    }
+
+    // parseRegex parses a regular expression; the end is defined by encountering an
+    // unmatched closing ')' or ']' which is not consumed
+    fn parseRegex(p: *Parser, writer: anytype) ParseError!void {
+        var i = p.i;
+        if (p.s.len < i + 2) return ParseError.ExpectedRegexp;
+
+        // number of open parens or brackets;
+        // when it becomes negative, finished parsing regex
+        var open: isize = 0;
+
+        loop: while (i < p.s.len) {
+            switch (p.s[i]) {
+                '(', '[' => open += 1,
+                ')', ']' => {
+                    open -= 1;
+                    if (open < 0) break :loop;
+                },
+                else => {},
+            }
+            i += 1;
+        }
+
+        if (i >= p.s.len) return ParseError.InvalidRegexp;
+        writer.writeAll(p.s[p.i..i]) catch return ParseError.WriteError;
+        p.i = i;
+    }
+
+    // parsePseudoclassSelector parses a pseudoclass selector like :not(p) or a pseudo-element
+    // For backwards compatibility, both ':' and '::' prefix are allowed for pseudo-elements.
+    // https://drafts.csswg.org/selectors-3/#pseudo-elements
+    fn parsePseudoclassSelector(p: *Parser, allocator: Allocator) ParseError!Selector {
+        if (p.i >= p.s.len) return ParseError.ExpectedPseudoClassSelector;
+        if (p.s[p.i] != ':') return ParseError.ExpectedPseudoClassSelector;
+
+        p.i += 1;
+
+        var must_pseudo_elt: bool = false;
+        if (p.i >= p.s.len) return ParseError.EmptyPseudoClassSelector;
+        if (p.s[p.i] == ':') { // we found a pseudo-element
+            must_pseudo_elt = true;
+            p.i += 1;
+        }
+
+        var buf: std.ArrayListUnmanaged(u8) = .empty;
+        defer buf.deinit(allocator);
+
+        try p.parseIdentifier(buf.writer(allocator));
+
+        const pseudo_class = try PseudoClass.parse(buf.items);
+
+        // reset the buffer to reuse it.
+        buf.clearRetainingCapacity();
+
+        if (must_pseudo_elt and !pseudo_class.isPseudoElement()) return ParseError.InvalidPseudoElement;
+
+        switch (pseudo_class) {
+            .not, .has, .haschild => {
+                if (!p.consumeParenthesis()) return ParseError.ExpectedParenthesis;
+
+                const sel = try p.parseSelectorGroup(allocator);
+                if (!p.consumeClosingParenthesis()) return ParseError.ExpectedParenthesisClose;
+
+                const s = try allocator.create(Selector);
+                errdefer allocator.destroy(s);
+                s.* = sel;
+
+                return .{ .pseudo_class_relative = .{ .pseudo_class = pseudo_class, .match = s } };
+            },
+            .contains, .containsown => {
+                if (!p.consumeParenthesis()) return ParseError.ExpectedParenthesis;
+                if (p.i == p.s.len) return ParseError.UnmatchParenthesis;
+
+                switch (p.s[p.i]) {
+                    '\'', '"' => try p.parseString(buf.writer(allocator)),
+                    else => try p.parseString(buf.writer(allocator)),
+                }
+
+                _ = p.skipWhitespace();
+                if (p.i >= p.s.len) return ParseError.InvalidPseudoClass;
+                if (!p.consumeClosingParenthesis()) return ParseError.ExpectedParenthesisClose;
+
+                const val = try buf.toOwnedSlice(allocator);
+                errdefer allocator.free(val);
+
+                return .{ .pseudo_class_contains = .{ .own = pseudo_class == .containsown, .val = val } };
+            },
+            .matches, .matchesown => {
+                if (!p.consumeParenthesis()) return ParseError.ExpectedParenthesis;
+
+                try p.parseRegex(buf.writer(allocator));
+                if (p.i >= p.s.len) return ParseError.InvalidPseudoClassSelector;
+                if (!p.consumeClosingParenthesis()) return ParseError.ExpectedParenthesisClose;
+
+                return .{ .pseudo_class_regexp = .{ .own = pseudo_class == .matchesown, .regexp = try buf.toOwnedSlice(allocator) } };
+            },
+            .nth_child, .nth_last_child, .nth_of_type, .nth_last_of_type => {
+                if (!p.consumeParenthesis()) return ParseError.ExpectedParenthesis;
+                const nth = try p.parseNth(allocator);
+                if (!p.consumeClosingParenthesis()) return ParseError.ExpectedParenthesisClose;
+
+                const last = pseudo_class == .nth_last_child or pseudo_class == .nth_last_of_type;
+                const of_type = pseudo_class == .nth_of_type or pseudo_class == .nth_last_of_type;
+                return .{ .pseudo_class_nth = .{ .a = nth[0], .b = nth[1], .of_type = of_type, .last = last } };
+            },
+            .first_child => return .{ .pseudo_class_nth = .{ .a = 0, .b = 1, .of_type = false, .last = false } },
+            .last_child => return .{ .pseudo_class_nth = .{ .a = 0, .b = 1, .of_type = false, .last = true } },
+            .first_of_type => return .{ .pseudo_class_nth = .{ .a = 0, .b = 1, .of_type = true, .last = false } },
+            .last_of_type => return .{ .pseudo_class_nth = .{ .a = 0, .b = 1, .of_type = true, .last = true } },
+            .only_child => return .{ .pseudo_class_only_child = false },
+            .only_of_type => return .{ .pseudo_class_only_child = true },
+            .input, .empty, .root, .link => return .{ .pseudo_class = pseudo_class },
+            .enabled, .disabled, .checked => return .{ .pseudo_class = pseudo_class },
+            .visible => return .{ .pseudo_class = pseudo_class },
+            .lang => {
+                if (!p.consumeParenthesis()) return ParseError.ExpectedParenthesis;
+                if (p.i == p.s.len) return ParseError.UnmatchParenthesis;
+
+                try p.parseIdentifier(buf.writer(allocator));
+
+                _ = p.skipWhitespace();
+                if (p.i >= p.s.len) return ParseError.InvalidPseudoClass;
+                if (!p.consumeClosingParenthesis()) return ParseError.ExpectedParenthesisClose;
+
+                const val = try buf.toOwnedSlice(allocator);
+                errdefer allocator.free(val);
+                lowerstr(val);
+
+                return .{ .pseudo_class_lang = val };
+            },
+            .visited, .hover, .active, .focus, .target => {
+                // Not applicable in a static context: never match.
+                return .{ .never_match = pseudo_class };
+            },
+            .after, .backdrop, .before, .cue, .first_letter => return .{ .pseudo_element = pseudo_class },
+            .first_line, .grammar_error, .marker, .placeholder => return .{ .pseudo_element = pseudo_class },
+            .selection, .spelling_error => return .{ .pseudo_element = pseudo_class },
+            .modal, .popover_open => return .{ .pseudo_element = pseudo_class },
+        }
+    }
+
+    // consumeParenthesis consumes an opening parenthesis and any following
+    // whitespace. It returns true if there was actually a parenthesis to skip.
+    fn consumeParenthesis(p: *Parser) bool {
+        if (p.i < p.s.len and p.s[p.i] == '(') {
+            p.i += 1;
+            _ = p.skipWhitespace();
+            return true;
+        }
+        return false;
+    }
+
+    // parseSelectorGroup parses a group of selectors, separated by commas.
+    fn parseSelectorGroup(p: *Parser, allocator: Allocator) ParseError!Selector {
+        const s = try p.parseSelector(allocator);
+
+        var buf: std.ArrayListUnmanaged(Selector) = .empty;
+        defer buf.deinit(allocator);
+
+        try buf.append(allocator, s);
+
+        while (p.i < p.s.len) {
+            if (p.s[p.i] != ',') break;
+            p.i += 1;
+            const ss = try p.parseSelector(allocator);
+            try buf.append(allocator, ss);
+        }
+
+        if (buf.items.len == 1) {
+            return buf.items[0];
+        }
+
+        return .{ .group = try buf.toOwnedSlice(allocator) };
+    }
+
+    // parseSelector parses a selector that may include combinators.
+    fn parseSelector(p: *Parser, allocator: Allocator) ParseError!Selector {
+        _ = p.skipWhitespace();
+        var s = try p.parseSimpleSelectorSequence(allocator);
+
+        while (true) {
+            var combinator: Combinator = .empty;
+            if (p.skipWhitespace()) {
+                combinator = .descendant;
+            }
+            if (p.i >= p.s.len) {
+                return s;
+            }
+
+            switch (p.s[p.i]) {
+                '+', '>', '~' => {
+                    combinator = try Combinator.parse(p.s[p.i]);
+                    p.i += 1;
+                    _ = p.skipWhitespace();
+                },
+                // These characters can't begin a selector, but they can legally occur after one.
+                ',', ')' => {
+                    return s;
+                },
+                else => {},
+            }
+
+            if (combinator == .empty) {
+                return s;
+            }
+
+            const c = try p.parseSimpleSelectorSequence(allocator);
+
+            const first = try allocator.create(Selector);
+            errdefer allocator.destroy(first);
+            first.* = s;
+
+            const second = try allocator.create(Selector);
+            errdefer allocator.destroy(second);
+            second.* = c;
+
+            s = Selector{ .combined = .{
+                .first = first,
+                .second = second,
+                .combinator = combinator,
+            } };
+        }
+
+        return s;
+    }
+
+    // consumeClosingParenthesis consumes a closing parenthesis and any preceding
+    // whitespace. It returns true if there was actually a parenthesis to skip.
+    fn consumeClosingParenthesis(p: *Parser) bool {
+        const i = p.i;
+        _ = p.skipWhitespace();
+        if (p.i < p.s.len and p.s[p.i] == ')') {
+            p.i += 1;
+            return true;
+        }
+        p.i = i;
+        return false;
+    }
+
+    // parseInteger parses a  decimal integer.
+    fn parseInteger(p: *Parser) ParseError!isize {
+        var i = p.i;
+        const start = i;
+        while (i < p.s.len and '0' <= p.s[i] and p.s[i] <= '9') i += 1;
+        if (i == start) return ParseError.ExpectedInteger;
+        p.i = i;
+
+        return std.fmt.parseUnsigned(isize, p.s[start..i], 10) catch ParseError.ExpectedInteger;
+    }
+
+    fn parseNthReadN(p: *Parser, a: isize) ParseError![2]isize {
+        _ = p.skipWhitespace();
+        if (p.i >= p.s.len) return ParseError.ExpectedNthExpression;
+
+        return switch (p.s[p.i]) {
+            '+' => {
+                p.i += 1;
+                _ = p.skipWhitespace();
+                const b = try p.parseInteger();
+                return .{ a, b };
+            },
+            '-' => {
+                p.i += 1;
+                _ = p.skipWhitespace();
+                const b = try p.parseInteger();
+                return .{ a, -b };
+            },
+            else => .{ a, 0 },
+        };
+    }
+
+    fn parseNthReadA(p: *Parser, a: isize) ParseError![2]isize {
+        if (p.i >= p.s.len) return ParseError.ExpectedNthExpression;
+        return switch (p.s[p.i]) {
+            'n', 'N' => {
+                p.i += 1;
+                return p.parseNthReadN(a);
+            },
+            else => .{ 0, a },
+        };
+    }
+
+    fn parseNthNegativeA(p: *Parser) ParseError![2]isize {
+        if (p.i >= p.s.len) return ParseError.ExpectedNthExpression;
+        const c = p.s[p.i];
+        if (std.ascii.isDigit(c)) {
+            const a = try p.parseInteger() * -1;
+            return p.parseNthReadA(a);
+        }
+        if (c == 'n' or c == 'N') {
+            p.i += 1;
+            return p.parseNthReadN(-1);
+        }
+
+        return ParseError.InvalidNthExpression;
+    }
+
+    fn parseNthPositiveA(p: *Parser) ParseError![2]isize {
+        if (p.i >= p.s.len) return ParseError.ExpectedNthExpression;
+        const c = p.s[p.i];
+        if (std.ascii.isDigit(c)) {
+            const a = try p.parseInteger();
+            return p.parseNthReadA(a);
+        }
+        if (c == 'n' or c == 'N') {
+            p.i += 1;
+            return p.parseNthReadN(1);
+        }
+
+        return ParseError.InvalidNthExpression;
+    }
+
+    // parseNth parses the argument for :nth-child (normally of the form an+b).
+    fn parseNth(p: *Parser, allocator: Allocator) ParseError![2]isize {
+        // initial state
+        if (p.i >= p.s.len) return ParseError.ExpectedNthExpression;
+        return switch (p.s[p.i]) {
+            '-' => {
+                p.i += 1;
+                return p.parseNthNegativeA();
+            },
+            '+' => {
+                p.i += 1;
+                return p.parseNthPositiveA();
+            },
+            '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' => p.parseNthPositiveA(),
+            'n', 'N' => {
+                p.i += 1;
+                return p.parseNthReadN(1);
+            },
+            'o', 'O', 'e', 'E' => {
+                var buf: std.ArrayListUnmanaged(u8) = .empty;
+                defer buf.deinit(allocator);
+
+                try p.parseName(buf.writer(allocator));
+
+                if (std.ascii.eqlIgnoreCase("odd", buf.items)) return .{ 2, 1 };
+                if (std.ascii.eqlIgnoreCase("even", buf.items)) return .{ 2, 0 };
+
+                return ParseError.InvalidNthExpression;
+            },
+            else => ParseError.InvalidNthExpression,
+        };
+    }
+};
+
+// nameStart returns whether c can be the first character of an identifier
+// (not counting an initial hyphen, or an escape sequence).
+fn nameStart(c: u8) bool {
+    return 'a' <= c and c <= 'z' or 'A' <= c and c <= 'Z' or c == '_' or c > 127;
+}
+
+// nameChar returns whether c can be a character within an identifier
+// (not counting an escape sequence).
+fn nameChar(c: u8) bool {
+    return 'a' <= c and c <= 'z' or 'A' <= c and c <= 'Z' or c == '_' or c > 127 or
+        c == '-' or '0' <= c and c <= '9';
+}
+
+fn lowerstr(str: []u8) void {
+    for (str, 0..) |c, i| {
+        str[i] = std.ascii.toLower(c);
+    }
+}
+
+// parseAttributeOP parses an AttributeOP from a string of 1 or 2 bytes.
+fn parseAttributeOP(s: []const u8) ParseError!AttributeOP {
+    if (s.len < 1 or s.len > 2) return ParseError.InvalidAttributeOperator;
+
+    // if the first sign is equal, we don't check anything else.
+    if (s[0] == '=') return .eql;
+
+    if (s.len != 2 or s[1] != '=') return ParseError.InvalidAttributeOperator;
+
+    return switch (s[0]) {
+        '=' => .eql,
+        '!' => .not_eql,
+        '~' => .one_of,
+        '|' => .prefix_hyphen,
+        '^' => .prefix,
+        '$' => .suffix,
+        '*' => .contains,
+        '#' => .regexp,
+        else => ParseError.InvalidAttributeOperator,
+    };
+}
+
+test "parser.skipWhitespace" {
+    const testcases = [_]struct {
+        s: []const u8,
+        i: usize,
+        r: bool,
+    }{
+        .{ .s = "", .i = 0, .r = false },
+        .{ .s = "foo", .i = 0, .r = false },
+        .{ .s = " ", .i = 1, .r = true },
+        .{ .s = " foo", .i = 1, .r = true },
+        .{ .s = "/* foo */ bar", .i = 10, .r = true },
+        .{ .s = "/* foo", .i = 0, .r = false },
+    };
+
+    for (testcases) |tc| {
+        var p = Parser{ .s = tc.s, .opts = .{} };
+        const res = p.skipWhitespace();
+        try std.testing.expectEqual(tc.r, res);
+        try std.testing.expectEqual(tc.i, p.i);
+    }
+}
+
+test "parser.parseIdentifier" {
+    const allocator = std.testing.allocator;
+
+    const testcases = [_]struct {
+        s: []const u8, // given value
+        exp: []const u8, // expected value
+        err: bool = false,
+    }{
+        .{ .s = "x", .exp = "x" },
+        .{ .s = "96", .exp = "", .err = true },
+        .{ .s = "-x", .exp = "-x" },
+        .{ .s = "r\\e9 sumé", .exp = "résumé" },
+        .{ .s = "r\\0000e9 sumé", .exp = "résumé" },
+        .{ .s = "r\\0000e9sumé", .exp = "résumé" },
+        .{ .s = "a\\\"b", .exp = "a\"b" },
+    };
+
+    var buf: std.ArrayListUnmanaged(u8) = .empty;
+    defer buf.deinit(allocator);
+
+    for (testcases) |tc| {
+        buf.clearRetainingCapacity();
+
+        var p = Parser{ .s = tc.s, .opts = .{} };
+        p.parseIdentifier(buf.writer(allocator)) catch |e| {
+            // if error was expected, continue.
+            if (tc.err) continue;
+
+            std.debug.print("test case {s}\n", .{tc.s});
+            return e;
+        };
+        std.testing.expectEqualDeep(tc.exp, buf.items) catch |e| {
+            std.debug.print("test case {s} : {s}\n", .{ tc.s, buf.items });
+            return e;
+        };
+    }
+}
+
+test "parser.parseString" {
+    const allocator = std.testing.allocator;
+
+    const testcases = [_]struct {
+        s: []const u8, // given value
+        exp: []const u8, // expected value
+        err: bool = false,
+    }{
+        .{ .s = "\"x\"", .exp = "x" },
+        .{ .s = "'x'", .exp = "x" },
+        .{ .s = "'x", .exp = "", .err = true },
+        .{ .s = "'x\\\r\nx'", .exp = "xx" },
+        .{ .s = "\"r\\e9 sumé\"", .exp = "résumé" },
+        .{ .s = "\"r\\0000e9 sumé\"", .exp = "résumé" },
+        .{ .s = "\"r\\0000e9sumé\"", .exp = "résumé" },
+        .{ .s = "\"a\\\"b\"", .exp = "a\"b" },
+        .{ .s = "\"\\\n\"", .exp = "" },
+        .{ .s = "\"hello world\"", .exp = "hello world" },
+    };
+
+    var buf: std.ArrayListUnmanaged(u8) = .empty;
+    defer buf.deinit(allocator);
+
+    for (testcases) |tc| {
+        buf.clearRetainingCapacity();
+
+        var p = Parser{ .s = tc.s, .opts = .{} };
+        p.parseString(buf.writer(allocator)) catch |e| {
+            // if error was expected, continue.
+            if (tc.err) continue;
+
+            std.debug.print("test case {s}\n", .{tc.s});
+            return e;
+        };
+        std.testing.expectEqualDeep(tc.exp, buf.items) catch |e| {
+            std.debug.print("test case {s} : {s}\n", .{ tc.s, buf.items });
+            return e;
+        };
+    }
+}
+
+test "parser.parse" {
+    var arena = std.heap.ArenaAllocator.init(std.testing.allocator);
+    defer arena.deinit();
+    const allocator = arena.allocator();
+
+    const testcases = [_]struct {
+        s: []const u8, // given value
+        exp: Selector, // expected value
+        err: bool = false,
+    }{
+        .{ .s = "root", .exp = .{ .tag = "root" } },
+        .{ .s = ".root", .exp = .{ .class = "root" } },
+        .{ .s = ":root", .exp = .{ .pseudo_class = .root } },
+        .{ .s = ".\\:bar", .exp = .{ .class = ":bar" } },
+        .{ .s = ".foo\\:bar", .exp = .{ .class = "foo:bar" } },
+    };
+
+    for (testcases) |tc| {
+        var p = Parser{ .s = tc.s, .opts = .{} };
+        const sel = p.parse(allocator) catch |e| {
+            // if error was expected, continue.
+            if (tc.err) continue;
+
+            std.debug.print("test case {s}\n", .{tc.s});
+            return e;
+        };
+        std.testing.expectEqualDeep(tc.exp, sel) catch |e| {
+            std.debug.print("test case {s} : {}\n", .{ tc.s, sel });
+            return e;
+        };
+    }
+}

src/browser/cssom/CSSParser.zig

@@ -0,0 +1,289 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const Allocator = std.mem.Allocator;
+
+const CSSConstants = struct {
+    const IMPORTANT = "!important";
+    const URL_PREFIX = "url(";
+};
+
+const CSSParserState = enum {
+    seek_name,
+    in_name,
+    seek_colon,
+    seek_value,
+    in_value,
+    in_quoted_value,
+    in_single_quoted_value,
+    in_url,
+    in_important,
+};
+
+const CSSDeclaration = struct {
+    name: []const u8,
+    value: []const u8,
+    is_important: bool,
+};
+
+const CSSParser = @This();
+state: CSSParserState,
+name_start: usize,
+name_end: usize,
+value_start: usize,
+position: usize,
+paren_depth: usize,
+escape_next: bool,
+
+pub fn init() CSSParser {
+    return .{
+        .state = .seek_name,
+        .name_start = 0,
+        .name_end = 0,
+        .value_start = 0,
+        .position = 0,
+        .paren_depth = 0,
+        .escape_next = false,
+    };
+}
+
+pub fn parseDeclarations(arena: Allocator, text: []const u8) ![]CSSDeclaration {
+    var parser = init();
+    var declarations: std.ArrayListUnmanaged(CSSDeclaration) = .empty;
+
+    while (parser.position < text.len) {
+        const c = text[parser.position];
+
+        switch (parser.state) {
+            .seek_name => {
+                if (!std.ascii.isWhitespace(c)) {
+                    parser.name_start = parser.position;
+                    parser.state = .in_name;
+                    continue;
+                }
+            },
+            .in_name => {
+                if (c == ':') {
+                    parser.name_end = parser.position;
+                    parser.state = .seek_value;
+                } else if (std.ascii.isWhitespace(c)) {
+                    parser.name_end = parser.position;
+                    parser.state = .seek_colon;
+                }
+            },
+            .seek_colon => {
+                if (c == ':') {
+                    parser.state = .seek_value;
+                } else if (!std.ascii.isWhitespace(c)) {
+                    parser.state = .seek_name;
+                    continue;
+                }
+            },
+            .seek_value => {
+                if (!std.ascii.isWhitespace(c)) {
+                    parser.value_start = parser.position;
+                    if (c == '"') {
+                        parser.state = .in_quoted_value;
+                    } else if (c == '\'') {
+                        parser.state = .in_single_quoted_value;
+                    } else if (c == 'u' and parser.position + CSSConstants.URL_PREFIX.len <= text.len and std.mem.startsWith(u8, text[parser.position..], CSSConstants.URL_PREFIX)) {
+                        parser.state = .in_url;
+                        parser.paren_depth = 1;
+                        parser.position += 3;
+                    } else {
+                        parser.state = .in_value;
+                        continue;
+                    }
+                }
+            },
+            .in_value => {
+                if (parser.escape_next) {
+                    parser.escape_next = false;
+                } else if (c == '\\') {
+                    parser.escape_next = true;
+                } else if (c == '(') {
+                    parser.paren_depth += 1;
+                } else if (c == ')' and parser.paren_depth > 0) {
+                    parser.paren_depth -= 1;
+                } else if (c == ';' and parser.paren_depth == 0) {
+                    try parser.finishDeclaration(arena, &declarations, text);
+                    parser.state = .seek_name;
+                }
+            },
+            .in_quoted_value => {
+                if (parser.escape_next) {
+                    parser.escape_next = false;
+                } else if (c == '\\') {
+                    parser.escape_next = true;
+                } else if (c == '"') {
+                    parser.state = .in_value;
+                }
+            },
+            .in_single_quoted_value => {
+                if (parser.escape_next) {
+                    parser.escape_next = false;
+                } else if (c == '\\') {
+                    parser.escape_next = true;
+                } else if (c == '\'') {
+                    parser.state = .in_value;
+                }
+            },
+            .in_url => {
+                if (parser.escape_next) {
+                    parser.escape_next = false;
+                } else if (c == '\\') {
+                    parser.escape_next = true;
+                } else if (c == '(') {
+                    parser.paren_depth += 1;
+                } else if (c == ')') {
+                    parser.paren_depth -= 1;
+                    if (parser.paren_depth == 0) {
+                        parser.state = .in_value;
+                    }
+                }
+            },
+            .in_important => {},
+        }
+
+        parser.position += 1;
+    }
+
+    try parser.finalize(arena, &declarations, text);
+
+    return declarations.items;
+}
+
+fn finishDeclaration(self: *CSSParser, arena: Allocator, declarations: *std.ArrayListUnmanaged(CSSDeclaration), text: []const u8) !void {
+    const name = std.mem.trim(u8, text[self.name_start..self.name_end], &std.ascii.whitespace);
+    if (name.len == 0) return;
+
+    const raw_value = text[self.value_start..self.position];
+    const value = std.mem.trim(u8, raw_value, &std.ascii.whitespace);
+
+    var final_value = value;
+    var is_important = false;
+
+    if (std.mem.endsWith(u8, value, CSSConstants.IMPORTANT)) {
+        is_important = true;
+        final_value = std.mem.trimRight(u8, value[0 .. value.len - CSSConstants.IMPORTANT.len], &std.ascii.whitespace);
+    }
+
+    try declarations.append(arena, .{
+        .name = name,
+        .value = final_value,
+        .is_important = is_important,
+    });
+}
+
+fn finalize(self: *CSSParser, arena: Allocator, declarations: *std.ArrayListUnmanaged(CSSDeclaration), text: []const u8) !void {
+    if (self.state != .in_value) {
+        return;
+    }
+    return self.finishDeclaration(arena, declarations, text);
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: CSS.Parser - Simple property" {
+    defer testing.reset();
+
+    const text = "color: red;";
+    const allocator = testing.arena_allocator;
+
+    const declarations = try CSSParser.parseDeclarations(allocator, text);
+
+    try testing.expectEqual(1, declarations.len);
+    try testing.expectEqual("color", declarations[0].name);
+    try testing.expectEqual("red", declarations[0].value);
+    try testing.expectEqual(false, declarations[0].is_important);
+}
+
+test "Browser: CSS.Parser - Property with !important" {
+    defer testing.reset();
+    const text = "margin: 10px !important;";
+    const allocator = testing.arena_allocator;
+
+    const declarations = try CSSParser.parseDeclarations(allocator, text);
+
+    try testing.expectEqual(1, declarations.len);
+    try testing.expectEqual("margin", declarations[0].name);
+    try testing.expectEqual("10px", declarations[0].value);
+    try testing.expectEqual(true, declarations[0].is_important);
+}
+
+test "Browser: CSS.Parser - Multiple properties" {
+    defer testing.reset();
+    const text = "color: red; font-size: 12px; margin: 5px !important;";
+    const allocator = testing.arena_allocator;
+
+    const declarations = try CSSParser.parseDeclarations(allocator, text);
+
+    try testing.expect(declarations.len == 3);
+
+    try testing.expectEqual("color", declarations[0].name);
+    try testing.expectEqual("red", declarations[0].value);
+    try testing.expectEqual(false, declarations[0].is_important);
+
+    try testing.expectEqual("font-size", declarations[1].name);
+    try testing.expectEqual("12px", declarations[1].value);
+    try testing.expectEqual(false, declarations[1].is_important);
+
+    try testing.expectEqual("margin", declarations[2].name);
+    try testing.expectEqual("5px", declarations[2].value);
+    try testing.expectEqual(true, declarations[2].is_important);
+}
+
+test "Browser: CSS.Parser - Quoted value with semicolon" {
+    defer testing.reset();
+    const text = "content: \"Hello; world!\";";
+    const allocator = testing.arena_allocator;
+
+    const declarations = try CSSParser.parseDeclarations(allocator, text);
+
+    try testing.expectEqual(1, declarations.len);
+    try testing.expectEqual("content", declarations[0].name);
+    try testing.expectEqual("\"Hello; world!\"", declarations[0].value);
+    try testing.expectEqual(false, declarations[0].is_important);
+}
+
+test "Browser: CSS.Parser - URL value" {
+    defer testing.reset();
+    const text = "background-image: url(\"test.png\");";
+    const allocator = testing.arena_allocator;
+
+    const declarations = try CSSParser.parseDeclarations(allocator, text);
+
+    try testing.expectEqual(1, declarations.len);
+    try testing.expectEqual("background-image", declarations[0].name);
+    try testing.expectEqual("url(\"test.png\")", declarations[0].value);
+    try testing.expectEqual(false, declarations[0].is_important);
+}
+
+test "Browser: CSS.Parser - Whitespace handling" {
+    defer testing.reset();
+    const text = "  color  :  purple  ;  margin  :  10px  ;  ";
+    const allocator = testing.arena_allocator;
+
+    const declarations = try CSSParser.parseDeclarations(allocator, text);
+
+    try testing.expectEqual(2, declarations.len);
+    try testing.expectEqual("color", declarations[0].name);
+    try testing.expectEqual("purple", declarations[0].value);
+    try testing.expectEqual("margin", declarations[1].name);
+    try testing.expectEqual("10px", declarations[1].value);
+}

src/browser/cssom/CSSRule.zig

@@ -0,0 +1,41 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const CSSStyleSheet = @import("CSSStyleSheet.zig");
+
+pub const Interfaces = .{
+    CSSRule,
+    CSSImportRule,
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/CSSRule
+const CSSRule = @This();
+css_text: []const u8,
+parent_rule: ?*CSSRule = null,
+parent_stylesheet: ?*CSSStyleSheet = null,
+
+pub const CSSImportRule = struct {
+    pub const prototype = *CSSRule;
+    href: []const u8,
+    layer_name: ?[]const u8,
+    media: void,
+    style_sheet: CSSStyleSheet,
+    supports_text: ?[]const u8,
+};

src/browser/cssom/CSSRuleList.zig

@@ -0,0 +1,52 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const CSSRule = @import("CSSRule.zig");
+const StyleSheet = @import("StyleSheet.zig").StyleSheet;
+
+const CSSImportRule = CSSRule.CSSImportRule;
+
+const CSSRuleList = @This();
+list: std.ArrayListUnmanaged([]const u8),
+
+pub fn constructor() CSSRuleList {
+    return .{ .list = .empty };
+}
+
+pub fn _item(self: *CSSRuleList, _index: u32) ?CSSRule {
+    const index: usize = @intCast(_index);
+
+    if (index > self.list.items.len) {
+        return null;
+    }
+
+    // todo: for now, just return null.
+    // this depends on properly parsing CSSRule
+    return null;
+}
+
+pub fn get_length(self: *CSSRuleList) u32 {
+    return @intCast(self.list.items.len);
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: CSS.CSSRuleList" {
+    try testing.htmlRunner("cssom/css_rule_list.html");
+}

src/browser/cssom/CSSStyleDeclaration.zig

@@ -0,0 +1,958 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const Page = @import("../page.zig").Page;
+const CSSRule = @import("CSSRule.zig");
+const CSSParser = @import("CSSParser.zig");
+
+const Property = struct {
+    value: []const u8,
+    priority: bool,
+};
+
+const CSSStyleDeclaration = @This();
+
+properties: std.StringArrayHashMapUnmanaged(Property),
+
+pub const empty: CSSStyleDeclaration = .{
+    .properties = .empty,
+};
+
+pub fn get_cssFloat(self: *const CSSStyleDeclaration) []const u8 {
+    return self._getPropertyValue("float");
+}
+
+pub fn set_cssFloat(self: *CSSStyleDeclaration, value: ?[]const u8, page: *Page) !void {
+    const final_value = value orelse "";
+    return self._setProperty("float", final_value, null, page);
+}
+
+pub fn get_cssText(self: *const CSSStyleDeclaration, page: *Page) ![]const u8 {
+    var buffer: std.ArrayListUnmanaged(u8) = .empty;
+    const writer = buffer.writer(page.call_arena);
+    var it = self.properties.iterator();
+    while (it.next()) |entry| {
+        const name = entry.key_ptr.*;
+        const property = entry.value_ptr;
+        const escaped = try escapeCSSValue(page.call_arena, property.value);
+        try writer.print("{s}: {s}", .{ name, escaped });
+        if (property.priority) {
+            try writer.writeAll(" !important; ");
+        } else {
+            try writer.writeAll("; ");
+        }
+    }
+    return buffer.items;
+}
+
+// TODO Propagate also upward to parent node
+pub fn set_cssText(self: *CSSStyleDeclaration, text: []const u8, page: *Page) !void {
+    self.properties.clearRetainingCapacity();
+
+    // call_arena is safe here, because _setProperty will dupe the name
+    // using the page's longer-living arena.
+    const declarations = try CSSParser.parseDeclarations(page.call_arena, text);
+
+    for (declarations) |decl| {
+        if (!isValidPropertyName(decl.name)) {
+            continue;
+        }
+        const priority: ?[]const u8 = if (decl.is_important) "important" else null;
+        try self._setProperty(decl.name, decl.value, priority, page);
+    }
+}
+
+pub fn get_length(self: *const CSSStyleDeclaration) usize {
+    return self.properties.count();
+}
+
+pub fn get_parentRule(_: *const CSSStyleDeclaration) ?CSSRule {
+    return null;
+}
+
+pub fn _getPropertyPriority(self: *const CSSStyleDeclaration, name: []const u8) []const u8 {
+    const property = self.properties.getPtr(name) orelse return "";
+    return if (property.priority) "important" else "";
+}
+
+// TODO should handle properly shorthand properties and canonical forms
+pub fn _getPropertyValue(self: *const CSSStyleDeclaration, name: []const u8) []const u8 {
+    if (self.properties.getPtr(name)) |property| {
+        return property.value;
+    }
+
+    // default to everything being visible (unless it's been explicitly set)
+    if (std.mem.eql(u8, name, "visibility")) {
+        return "visible";
+    }
+
+    return "";
+}
+
+pub fn _item(self: *const CSSStyleDeclaration, index: usize) []const u8 {
+    const values = self.properties.entries.items(.key);
+    if (index >= values.len) {
+        return "";
+    }
+    return values[index];
+}
+
+pub fn _removeProperty(self: *CSSStyleDeclaration, name: []const u8) ![]const u8 {
+    const property = self.properties.fetchOrderedRemove(name) orelse return "";
+    return property.value.value;
+}
+
+pub fn _setProperty(self: *CSSStyleDeclaration, name: []const u8, value: []const u8, priority: ?[]const u8, page: *Page) !void {
+    const gop = try self.properties.getOrPut(page.arena, name);
+    if (!gop.found_existing) {
+        const owned_name = try page.arena.dupe(u8, name);
+        gop.key_ptr.* = owned_name;
+    }
+
+    const owned_value = try page.arena.dupe(u8, value);
+    const is_important = priority != null and std.ascii.eqlIgnoreCase(priority.?, "important");
+    gop.value_ptr.* = .{ .value = owned_value, .priority = is_important };
+}
+
+pub fn named_get(self: *const CSSStyleDeclaration, name: []const u8, _: *bool) []const u8 {
+    return self._getPropertyValue(name);
+}
+
+pub fn named_set(self: *CSSStyleDeclaration, name: []const u8, value: []const u8, _: *bool, page: *Page) !void {
+    return self._setProperty(name, value, null, page);
+}
+
+fn isNumericWithUnit(value: []const u8) bool {
+    if (value.len == 0) {
+        return false;
+    }
+
+    const first = value[0];
+
+    if (!std.ascii.isDigit(first) and first != '+' and first != '-' and first != '.') {
+        return false;
+    }
+
+    var i: usize = 0;
+    var has_digit = false;
+    var decimal_point = false;
+
+    while (i < value.len) : (i += 1) {
+        const c = value[i];
+        if (std.ascii.isDigit(c)) {
+            has_digit = true;
+        } else if (c == '.' and !decimal_point) {
+            decimal_point = true;
+        } else if ((c == 'e' or c == 'E') and has_digit) {
+            if (i + 1 >= value.len) return false;
+            if (value[i + 1] != '+' and value[i + 1] != '-' and !std.ascii.isDigit(value[i + 1])) break;
+            i += 1;
+            if (value[i] == '+' or value[i] == '-') {
+                i += 1;
+            }
+            var has_exp_digits = false;
+            while (i < value.len and std.ascii.isDigit(value[i])) : (i += 1) {
+                has_exp_digits = true;
+            }
+            if (!has_exp_digits) return false;
+            break;
+        } else if (c != '-' and c != '+') {
+            break;
+        }
+    }
+
+    if (!has_digit) {
+        return false;
+    }
+
+    if (i == value.len) {
+        return true;
+    }
+
+    const unit = value[i..];
+    return CSSKeywords.isValidUnit(unit);
+}
+
+fn isHexColor(value: []const u8) bool {
+    if (value.len == 0) {
+        return false;
+    }
+    if (value[0] != '#') {
+        return false;
+    }
+
+    const hex_part = value[1..];
+    if (hex_part.len != 3 and hex_part.len != 6 and hex_part.len != 8) {
+        return false;
+    }
+
+    for (hex_part) |c| {
+        if (!std.ascii.isHex(c)) {
+            return false;
+        }
+    }
+
+    return true;
+}
+
+fn isMultiValueProperty(value: []const u8) bool {
+    var parts = std.mem.splitAny(u8, value, " ");
+    var multi_value_parts: usize = 0;
+    var all_parts_valid = true;
+
+    while (parts.next()) |part| {
+        if (part.len == 0) continue;
+        multi_value_parts += 1;
+
+        if (isNumericWithUnit(part)) {
+            continue;
+        }
+        if (isHexColor(part)) {
+            continue;
+        }
+        if (CSSKeywords.isKnownKeyword(part)) {
+            continue;
+        }
+        if (CSSKeywords.startsWithFunction(part)) {
+            continue;
+        }
+
+        all_parts_valid = false;
+        break;
+    }
+
+    return multi_value_parts >= 2 and all_parts_valid;
+}
+
+fn isAlreadyQuoted(value: []const u8) bool {
+    return value.len >= 2 and ((value[0] == '"' and value[value.len - 1] == '"') or
+        (value[0] == '\'' and value[value.len - 1] == '\''));
+}
+
+fn isValidPropertyName(name: []const u8) bool {
+    if (name.len == 0) return false;
+
+    if (std.mem.startsWith(u8, name, "--")) {
+        if (name.len == 2) return false;
+        for (name[2..]) |c| {
+            if (!std.ascii.isAlphanumeric(c) and c != '-' and c != '_') {
+                return false;
+            }
+        }
+        return true;
+    }
+
+    const first_char = name[0];
+    if (!std.ascii.isAlphabetic(first_char) and first_char != '-') {
+        return false;
+    }
+
+    if (first_char == '-') {
+        if (name.len < 2) return false;
+
+        if (!std.ascii.isAlphabetic(name[1])) {
+            return false;
+        }
+
+        for (name[2..]) |c| {
+            if (!std.ascii.isAlphanumeric(c) and c != '-') {
+                return false;
+            }
+        }
+    } else {
+        for (name[1..]) |c| {
+            if (!std.ascii.isAlphanumeric(c) and c != '-') {
+                return false;
+            }
+        }
+    }
+
+    return true;
+}
+
+fn extractImportant(value: []const u8) struct { value: []const u8, is_important: bool } {
+    const trimmed = std.mem.trim(u8, value, &std.ascii.whitespace);
+
+    if (std.mem.endsWith(u8, trimmed, "!important")) {
+        const clean_value = std.mem.trimRight(u8, trimmed[0 .. trimmed.len - 10], &std.ascii.whitespace);
+        return .{ .value = clean_value, .is_important = true };
+    }
+
+    return .{ .value = trimmed, .is_important = false };
+}
+
+fn needsQuotes(value: []const u8) bool {
+    if (value.len == 0) return true;
+    if (isAlreadyQuoted(value)) return false;
+
+    if (CSSKeywords.containsSpecialChar(value)) {
+        return true;
+    }
+
+    if (std.mem.indexOfScalar(u8, value, ' ') == null) {
+        return false;
+    }
+
+    const is_url = std.mem.startsWith(u8, value, "url(");
+    const is_function = CSSKeywords.startsWithFunction(value);
+
+    return !isMultiValueProperty(value) and
+        !is_url and
+        !is_function;
+}
+
+fn escapeCSSValue(arena: std.mem.Allocator, value: []const u8) ![]const u8 {
+    if (!needsQuotes(value)) {
+        return value;
+    }
+    var out: std.ArrayListUnmanaged(u8) = .empty;
+
+    // We'll need at least this much space, +2 for the quotes
+    try out.ensureTotalCapacity(arena, value.len + 2);
+    const writer = out.writer(arena);
+
+    try writer.writeByte('"');
+
+    for (value, 0..) |c, i| {
+        switch (c) {
+            '"' => try writer.writeAll("\\\""),
+            '\\' => try writer.writeAll("\\\\"),
+            '\n' => try writer.writeAll("\\A "),
+            '\r' => try writer.writeAll("\\D "),
+            '\t' => try writer.writeAll("\\9 "),
+            0...8, 11, 12, 14...31, 127 => {
+                try writer.print("\\{x}", .{c});
+                if (i + 1 < value.len and std.ascii.isHex(value[i + 1])) {
+                    try writer.writeByte(' ');
+                }
+            },
+            else => try writer.writeByte(c),
+        }
+    }
+
+    try writer.writeByte('"');
+    return out.items;
+}
+
+fn isKnownKeyword(value: []const u8) bool {
+    return CSSKeywords.isKnownKeyword(value);
+}
+
+fn containsSpecialChar(value: []const u8) bool {
+    return CSSKeywords.containsSpecialChar(value);
+}
+
+const CSSKeywords = struct {
+    const BORDER_STYLES = [_][]const u8{
+        "none", "solid", "dotted", "dashed", "double", "groove", "ridge", "inset", "outset",
+    };
+
+    const COLOR_NAMES = [_][]const u8{
+        "black",        "white",   "red", "green", "blue", "yellow", "purple", "gray", "transparent",
+        "currentColor", "inherit",
+    };
+
+    const POSITION_KEYWORDS = [_][]const u8{
+        "auto", "center", "left", "right", "top", "bottom",
+    };
+
+    const BACKGROUND_REPEAT = [_][]const u8{
+        "repeat", "no-repeat", "repeat-x", "repeat-y", "space", "round",
+    };
+
+    const FONT_STYLES = [_][]const u8{
+        "normal", "italic", "oblique", "bold", "bolder", "lighter",
+    };
+
+    const FONT_SIZES = [_][]const u8{
+        "xx-small", "x-small", "small", "medium", "large", "x-large", "xx-large",
+        "smaller",  "larger",
+    };
+
+    const FONT_FAMILIES = [_][]const u8{
+        "serif", "sans-serif", "monospace", "cursive", "fantasy", "system-ui",
+    };
+
+    const CSS_GLOBAL = [_][]const u8{
+        "initial", "inherit", "unset", "revert",
+    };
+
+    const DISPLAY_VALUES = [_][]const u8{
+        "block", "inline", "inline-block", "flex", "grid", "none",
+    };
+
+    const UNITS = [_][]const u8{
+        // LENGTH
+        "px",   "em", "rem", "vw",  "vh",  "vmin", "vmax", "%", "pt", "pc", "in",  "cm",  "mm",
+        "ex",   "ch", "fr",
+
+        // ANGLE
+         "deg", "rad", "grad", "turn",
+
+        // TIME
+        "s", "ms",
+
+        // FREQUENCY
+        "hz", "khz",
+
+        // RESOLUTION
+        "dpi", "dpcm",
+        "dppx",
+    };
+
+    const SPECIAL_CHARS = [_]u8{
+        '"', '\'', ';', '{', '}', '\\', '<', '>', '/', '\n', '\t', '\r', '\x00', '\x7F',
+    };
+
+    const FUNCTIONS = [_][]const u8{
+        "rgb(",             "rgba(",            "hsl(",            "hsla(",      "url(",    "calc(",  "var(",  "attr(",
+        "linear-gradient(", "radial-gradient(", "conic-gradient(", "translate(", "rotate(", "scale(", "skew(", "matrix(",
+    };
+
+    const KEYWORDS = BORDER_STYLES ++ COLOR_NAMES ++ POSITION_KEYWORDS ++
+        BACKGROUND_REPEAT ++ FONT_STYLES ++ FONT_SIZES ++ FONT_FAMILIES ++
+        CSS_GLOBAL ++ DISPLAY_VALUES;
+
+    const MAX_KEYWORD_LEN = lengthOfLongestValue(&KEYWORDS);
+
+    pub fn isKnownKeyword(value: []const u8) bool {
+        if (value.len > MAX_KEYWORD_LEN) {
+            return false;
+        }
+        var buf: [MAX_KEYWORD_LEN]u8 = undefined;
+        const normalized = std.ascii.lowerString(&buf, value);
+
+        for (KEYWORDS) |keyword| {
+            if (std.ascii.eqlIgnoreCase(normalized, keyword)) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+
+    pub fn containsSpecialChar(value: []const u8) bool {
+        return std.mem.indexOfAny(u8, value, &SPECIAL_CHARS) != null;
+    }
+
+    const MAX_UNIT_LEN = lengthOfLongestValue(&UNITS);
+
+    pub fn isValidUnit(unit: []const u8) bool {
+        if (unit.len > MAX_UNIT_LEN) {
+            return false;
+        }
+        var buf: [MAX_UNIT_LEN]u8 = undefined;
+        const normalized = std.ascii.lowerString(&buf, unit);
+
+        for (UNITS) |u| {
+            if (std.mem.eql(u8, normalized, u)) {
+                return true;
+            }
+        }
+        return false;
+    }
+
+    pub fn startsWithFunction(value: []const u8) bool {
+        const pos = std.mem.indexOfScalar(u8, value, '(') orelse return false;
+        if (pos == 0) return false;
+
+        if (std.mem.indexOfScalarPos(u8, value, pos, ')') == null) {
+            return false;
+        }
+        const function_name = value[0..pos];
+        return isValidFunctionName(function_name);
+    }
+
+    fn isValidFunctionName(name: []const u8) bool {
+        if (name.len == 0) return false;
+
+        const first = name[0];
+        if (!std.ascii.isAlphabetic(first) and first != '_' and first != '-') {
+            return false;
+        }
+
+        for (name[1..]) |c| {
+            if (!std.ascii.isAlphanumeric(c) and c != '_' and c != '-') {
+                return false;
+            }
+        }
+
+        return true;
+    }
+};
+
+fn lengthOfLongestValue(values: []const []const u8) usize {
+    var max: usize = 0;
+    for (values) |v| {
+        max = @max(v.len, max);
+    }
+    return max;
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: CSS.StyleDeclaration" {
+    try testing.htmlRunner("cssom/css_style_declaration.html");
+}
+
+test "Browser: CSS.StyleDeclaration: isNumericWithUnit - valid numbers with units" {
+    try testing.expect(isNumericWithUnit("10px"));
+    try testing.expect(isNumericWithUnit("3.14em"));
+    try testing.expect(isNumericWithUnit("-5rem"));
+    try testing.expect(isNumericWithUnit("+12.5%"));
+    try testing.expect(isNumericWithUnit("0vh"));
+    try testing.expect(isNumericWithUnit(".5vw"));
+}
+
+test "Browser: CSS.StyleDeclaration: isNumericWithUnit - scientific notation" {
+    try testing.expect(isNumericWithUnit("1e5px"));
+    try testing.expect(isNumericWithUnit("2.5E-3em"));
+    try testing.expect(isNumericWithUnit("1e+2rem"));
+    try testing.expect(isNumericWithUnit("-3.14e10px"));
+}
+
+test "Browser: CSS.StyleDeclaration: isNumericWithUnit - edge cases and invalid inputs" {
+    try testing.expect(!isNumericWithUnit(""));
+
+    try testing.expect(!isNumericWithUnit("px"));
+    try testing.expect(!isNumericWithUnit("--px"));
+    try testing.expect(!isNumericWithUnit(".px"));
+
+    try testing.expect(!isNumericWithUnit("1e"));
+    try testing.expect(!isNumericWithUnit("1epx"));
+    try testing.expect(!isNumericWithUnit("1e+"));
+    try testing.expect(!isNumericWithUnit("1e+px"));
+
+    try testing.expect(!isNumericWithUnit("1.2.3px"));
+
+    try testing.expect(!isNumericWithUnit("10xyz"));
+    try testing.expect(!isNumericWithUnit("5invalid"));
+
+    try testing.expect(isNumericWithUnit("10"));
+    try testing.expect(isNumericWithUnit("3.14"));
+    try testing.expect(isNumericWithUnit("-5"));
+}
+
+test "Browser: CSS.StyleDeclaration: isHexColor - valid hex colors" {
+    try testing.expect(isHexColor("#000"));
+    try testing.expect(isHexColor("#fff"));
+    try testing.expect(isHexColor("#123456"));
+    try testing.expect(isHexColor("#abcdef"));
+    try testing.expect(isHexColor("#ABCDEF"));
+    try testing.expect(isHexColor("#12345678"));
+}
+
+test "Browser: CSS.StyleDeclaration: isHexColor - invalid hex colors" {
+    try testing.expect(!isHexColor(""));
+    try testing.expect(!isHexColor("#"));
+    try testing.expect(!isHexColor("000"));
+    try testing.expect(!isHexColor("#00"));
+    try testing.expect(!isHexColor("#0000"));
+    try testing.expect(!isHexColor("#00000"));
+    try testing.expect(!isHexColor("#0000000"));
+    try testing.expect(!isHexColor("#000000000"));
+    try testing.expect(!isHexColor("#gggggg"));
+    try testing.expect(!isHexColor("#123xyz"));
+}
+
+test "Browser: CSS.StyleDeclaration: isMultiValueProperty - valid multi-value properties" {
+    try testing.expect(isMultiValueProperty("10px 20px"));
+    try testing.expect(isMultiValueProperty("solid red"));
+    try testing.expect(isMultiValueProperty("#fff black"));
+    try testing.expect(isMultiValueProperty("1em 2em 3em 4em"));
+    try testing.expect(isMultiValueProperty("rgb(255,0,0) solid"));
+}
+
+test "Browser: CSS.StyleDeclaration: isMultiValueProperty - invalid multi-value properties" {
+    try testing.expect(!isMultiValueProperty(""));
+    try testing.expect(!isMultiValueProperty("10px"));
+    try testing.expect(!isMultiValueProperty("invalid unknown"));
+    try testing.expect(!isMultiValueProperty("10px invalid"));
+    try testing.expect(!isMultiValueProperty("   "));
+}
+
+test "Browser: CSS.StyleDeclaration: isAlreadyQuoted - various quoting scenarios" {
+    try testing.expect(isAlreadyQuoted("\"hello\""));
+    try testing.expect(isAlreadyQuoted("'world'"));
+    try testing.expect(isAlreadyQuoted("\"\""));
+    try testing.expect(isAlreadyQuoted("''"));
+
+    try testing.expect(!isAlreadyQuoted(""));
+    try testing.expect(!isAlreadyQuoted("hello"));
+    try testing.expect(!isAlreadyQuoted("\""));
+    try testing.expect(!isAlreadyQuoted("'"));
+    try testing.expect(!isAlreadyQuoted("\"hello'"));
+    try testing.expect(!isAlreadyQuoted("'hello\""));
+    try testing.expect(!isAlreadyQuoted("\"hello"));
+    try testing.expect(!isAlreadyQuoted("hello\""));
+}
+
+test "Browser: CSS.StyleDeclaration: isValidPropertyName - valid property names" {
+    try testing.expect(isValidPropertyName("color"));
+    try testing.expect(isValidPropertyName("background-color"));
+    try testing.expect(isValidPropertyName("-webkit-transform"));
+    try testing.expect(isValidPropertyName("font-size"));
+    try testing.expect(isValidPropertyName("margin-top"));
+    try testing.expect(isValidPropertyName("z-index"));
+    try testing.expect(isValidPropertyName("line-height"));
+}
+
+test "Browser: CSS.StyleDeclaration: isValidPropertyName - invalid property names" {
+    try testing.expect(!isValidPropertyName(""));
+    try testing.expect(!isValidPropertyName("123color"));
+    try testing.expect(!isValidPropertyName("color!"));
+    try testing.expect(!isValidPropertyName("color space"));
+    try testing.expect(!isValidPropertyName("@color"));
+    try testing.expect(!isValidPropertyName("color.test"));
+    try testing.expect(!isValidPropertyName("color_test"));
+}
+
+test "Browser: CSS.StyleDeclaration: extractImportant - with and without !important" {
+    var result = extractImportant("red !important");
+    try testing.expect(result.is_important);
+    try testing.expectEqual("red", result.value);
+
+    result = extractImportant("blue");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("blue", result.value);
+
+    result = extractImportant("  green  !important  ");
+    try testing.expect(result.is_important);
+    try testing.expectEqual("green", result.value);
+
+    result = extractImportant("!important");
+    try testing.expect(result.is_important);
+    try testing.expectEqual("", result.value);
+
+    result = extractImportant("important");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("important", result.value);
+}
+
+test "Browser: CSS.StyleDeclaration: needsQuotes - various scenarios" {
+    try testing.expect(needsQuotes(""));
+    try testing.expect(needsQuotes("hello world"));
+    try testing.expect(needsQuotes("test;"));
+    try testing.expect(needsQuotes("a{b}"));
+    try testing.expect(needsQuotes("test\"quote"));
+
+    try testing.expect(!needsQuotes("\"already quoted\""));
+    try testing.expect(!needsQuotes("'already quoted'"));
+    try testing.expect(!needsQuotes("url(image.png)"));
+    try testing.expect(!needsQuotes("rgb(255, 0, 0)"));
+    try testing.expect(!needsQuotes("10px 20px"));
+    try testing.expect(!needsQuotes("simple"));
+}
+
+test "Browser: CSS.StyleDeclaration: escapeCSSValue - escaping various characters" {
+    const allocator = testing.arena_allocator;
+
+    var result = try escapeCSSValue(allocator, "simple");
+    try testing.expectEqual("simple", result);
+
+    result = try escapeCSSValue(allocator, "\"already quoted\"");
+    try testing.expectEqual("\"already quoted\"", result);
+
+    result = try escapeCSSValue(allocator, "test\"quote");
+    try testing.expectEqual("\"test\\\"quote\"", result);
+
+    result = try escapeCSSValue(allocator, "test\nline");
+    try testing.expectEqual("\"test\\A line\"", result);
+
+    result = try escapeCSSValue(allocator, "test\\back");
+    try testing.expectEqual("\"test\\\\back\"", result);
+}
+
+test "Browser: CSS.StyleDeclaration: CSSKeywords.isKnownKeyword - case sensitivity" {
+    try testing.expect(CSSKeywords.isKnownKeyword("red"));
+    try testing.expect(CSSKeywords.isKnownKeyword("solid"));
+    try testing.expect(CSSKeywords.isKnownKeyword("center"));
+    try testing.expect(CSSKeywords.isKnownKeyword("inherit"));
+
+    try testing.expect(CSSKeywords.isKnownKeyword("RED"));
+    try testing.expect(CSSKeywords.isKnownKeyword("Red"));
+    try testing.expect(CSSKeywords.isKnownKeyword("SOLID"));
+    try testing.expect(CSSKeywords.isKnownKeyword("Center"));
+
+    try testing.expect(!CSSKeywords.isKnownKeyword("invalid"));
+    try testing.expect(!CSSKeywords.isKnownKeyword("unknown"));
+    try testing.expect(!CSSKeywords.isKnownKeyword(""));
+}
+
+test "Browser: CSS.StyleDeclaration: CSSKeywords.containsSpecialChar - various special characters" {
+    try testing.expect(CSSKeywords.containsSpecialChar("test\"quote"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test'quote"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test;end"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test{brace"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test}brace"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test\\back"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test<angle"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test>angle"));
+    try testing.expect(CSSKeywords.containsSpecialChar("test/slash"));
+
+    try testing.expect(!CSSKeywords.containsSpecialChar("normal-text"));
+    try testing.expect(!CSSKeywords.containsSpecialChar("text123"));
+    try testing.expect(!CSSKeywords.containsSpecialChar(""));
+}
+
+test "Browser: CSS.StyleDeclaration: CSSKeywords.isValidUnit - various units" {
+    try testing.expect(CSSKeywords.isValidUnit("px"));
+    try testing.expect(CSSKeywords.isValidUnit("em"));
+    try testing.expect(CSSKeywords.isValidUnit("rem"));
+    try testing.expect(CSSKeywords.isValidUnit("%"));
+
+    try testing.expect(CSSKeywords.isValidUnit("deg"));
+    try testing.expect(CSSKeywords.isValidUnit("rad"));
+
+    try testing.expect(CSSKeywords.isValidUnit("s"));
+    try testing.expect(CSSKeywords.isValidUnit("ms"));
+
+    try testing.expect(CSSKeywords.isValidUnit("PX"));
+
+    try testing.expect(!CSSKeywords.isValidUnit("invalid"));
+    try testing.expect(!CSSKeywords.isValidUnit(""));
+}
+
+test "Browser: CSS.StyleDeclaration: CSSKeywords.startsWithFunction - function detection" {
+    try testing.expect(CSSKeywords.startsWithFunction("rgb(255, 0, 0)"));
+    try testing.expect(CSSKeywords.startsWithFunction("rgba(255, 0, 0, 0.5)"));
+    try testing.expect(CSSKeywords.startsWithFunction("url(image.png)"));
+    try testing.expect(CSSKeywords.startsWithFunction("calc(100% - 20px)"));
+    try testing.expect(CSSKeywords.startsWithFunction("var(--custom-property)"));
+    try testing.expect(CSSKeywords.startsWithFunction("linear-gradient(to right, red, blue)"));
+
+    try testing.expect(CSSKeywords.startsWithFunction("custom-function(args)"));
+    try testing.expect(CSSKeywords.startsWithFunction("unknown(test)"));
+
+    try testing.expect(!CSSKeywords.startsWithFunction("not-a-function"));
+    try testing.expect(!CSSKeywords.startsWithFunction("missing-paren)"));
+    try testing.expect(!CSSKeywords.startsWithFunction("missing-close("));
+    try testing.expect(!CSSKeywords.startsWithFunction(""));
+    try testing.expect(!CSSKeywords.startsWithFunction("rgb"));
+}
+
+test "Browser: CSS.StyleDeclaration: isNumericWithUnit - whitespace handling" {
+    try testing.expect(!isNumericWithUnit(" 10px"));
+    try testing.expect(!isNumericWithUnit("10 px"));
+    try testing.expect(!isNumericWithUnit("10px "));
+    try testing.expect(!isNumericWithUnit(" 10 px "));
+}
+
+test "Browser: CSS.StyleDeclaration: extractImportant - whitespace edge cases" {
+    var result = extractImportant("   ");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("", result.value);
+
+    result = extractImportant("\t\n\r !important\t\n");
+    try testing.expect(result.is_important);
+    try testing.expectEqual("", result.value);
+
+    result = extractImportant("red\t!important");
+    try testing.expect(result.is_important);
+    try testing.expectEqual("red", result.value);
+}
+
+test "Browser: CSS.StyleDeclaration: isHexColor - mixed case handling" {
+    try testing.expect(isHexColor("#AbC"));
+    try testing.expect(isHexColor("#123aBc"));
+    try testing.expect(isHexColor("#FFffFF"));
+    try testing.expect(isHexColor("#000FFF"));
+}
+
+test "Browser: CSS.StyleDeclaration: edge case - very long inputs" {
+    const long_valid = "a" ** 1000 ++ "px";
+    try testing.expect(!isNumericWithUnit(long_valid)); // not numeric
+
+    const long_property = "a-" ** 100 ++ "property";
+    try testing.expect(isValidPropertyName(long_property));
+
+    const long_hex = "#" ++ "a" ** 20;
+    try testing.expect(!isHexColor(long_hex));
+}
+
+test "Browser: CSS.StyleDeclaration: boundary conditions - numeric parsing" {
+    try testing.expect(isNumericWithUnit("0px"));
+    try testing.expect(isNumericWithUnit("0.0px"));
+    try testing.expect(isNumericWithUnit(".0px"));
+    try testing.expect(isNumericWithUnit("0.px"));
+
+    try testing.expect(isNumericWithUnit("999999999px"));
+    try testing.expect(isNumericWithUnit("1.7976931348623157e+308px"));
+
+    try testing.expect(isNumericWithUnit("0.000000001px"));
+    try testing.expect(isNumericWithUnit("1e-100px"));
+}
+
+test "Browser: CSS.StyleDeclaration: extractImportant - malformed important declarations" {
+    var result = extractImportant("red ! important");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("red ! important", result.value);
+
+    result = extractImportant("red !Important");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("red !Important", result.value);
+
+    result = extractImportant("red !IMPORTANT");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("red !IMPORTANT", result.value);
+
+    result = extractImportant("!importantred");
+    try testing.expect(!result.is_important);
+    try testing.expectEqual("!importantred", result.value);
+
+    result = extractImportant("red !important !important");
+    try testing.expect(result.is_important);
+    try testing.expectEqual("red !important", result.value);
+}
+
+test "Browser: CSS.StyleDeclaration: isMultiValueProperty - complex spacing scenarios" {
+    try testing.expect(isMultiValueProperty("10px    20px"));
+    try testing.expect(isMultiValueProperty("solid     red"));
+
+    try testing.expect(isMultiValueProperty("  10px 20px  "));
+
+    try testing.expect(!isMultiValueProperty("10px\t20px"));
+    try testing.expect(!isMultiValueProperty("10px\n20px"));
+
+    try testing.expect(isMultiValueProperty("10px   20px   30px"));
+}
+
+test "Browser: CSS.StyleDeclaration: isAlreadyQuoted - edge cases with quotes" {
+    try testing.expect(isAlreadyQuoted("\"'hello'\""));
+    try testing.expect(isAlreadyQuoted("'\"hello\"'"));
+
+    try testing.expect(isAlreadyQuoted("\"hello\\\"world\""));
+    try testing.expect(isAlreadyQuoted("'hello\\'world'"));
+
+    try testing.expect(!isAlreadyQuoted("\"hello"));
+    try testing.expect(!isAlreadyQuoted("hello\""));
+    try testing.expect(!isAlreadyQuoted("'hello"));
+    try testing.expect(!isAlreadyQuoted("hello'"));
+
+    try testing.expect(isAlreadyQuoted("\"a\""));
+    try testing.expect(isAlreadyQuoted("'b'"));
+}
+
+test "Browser: CSS.StyleDeclaration: needsQuotes - function and URL edge cases" {
+    try testing.expect(!needsQuotes("rgb(255, 0, 0)"));
+    try testing.expect(!needsQuotes("calc(100% - 20px)"));
+
+    try testing.expect(!needsQuotes("url(path with spaces.jpg)"));
+
+    try testing.expect(!needsQuotes("linear-gradient(to right, red, blue)"));
+
+    try testing.expect(needsQuotes("rgb(255, 0, 0"));
+}
+
+test "Browser: CSS.StyleDeclaration: escapeCSSValue - control characters and Unicode" {
+    const allocator = testing.arena_allocator;
+
+    var result = try escapeCSSValue(allocator, "test\ttab");
+    try testing.expectEqual("\"test\\9 tab\"", result);
+
+    result = try escapeCSSValue(allocator, "test\rreturn");
+    try testing.expectEqual("\"test\\D return\"", result);
+
+    result = try escapeCSSValue(allocator, "test\x00null");
+    try testing.expectEqual("\"test\\0null\"", result);
+
+    result = try escapeCSSValue(allocator, "test\x7Fdel");
+    try testing.expectEqual("\"test\\7f del\"", result);
+
+    result = try escapeCSSValue(allocator, "test\"quote\nline\\back");
+    try testing.expectEqual("\"test\\\"quote\\A line\\\\back\"", result);
+}
+
+test "Browser: CSS.StyleDeclaration: isValidPropertyName - CSS custom properties and vendor prefixes" {
+    try testing.expect(isValidPropertyName("--custom-color"));
+    try testing.expect(isValidPropertyName("--my-variable"));
+    try testing.expect(isValidPropertyName("--123"));
+
+    try testing.expect(isValidPropertyName("-webkit-transform"));
+    try testing.expect(isValidPropertyName("-moz-border-radius"));
+    try testing.expect(isValidPropertyName("-ms-filter"));
+    try testing.expect(isValidPropertyName("-o-transition"));
+
+    try testing.expect(!isValidPropertyName("-123invalid"));
+    try testing.expect(!isValidPropertyName("--"));
+    try testing.expect(!isValidPropertyName("-"));
+}
+
+test "Browser: CSS.StyleDeclaration: startsWithFunction - case sensitivity and partial matches" {
+    try testing.expect(CSSKeywords.startsWithFunction("RGB(255, 0, 0)"));
+    try testing.expect(CSSKeywords.startsWithFunction("Rgb(255, 0, 0)"));
+    try testing.expect(CSSKeywords.startsWithFunction("URL(image.png)"));
+
+    try testing.expect(CSSKeywords.startsWithFunction("rg(something)"));
+    try testing.expect(CSSKeywords.startsWithFunction("ur(something)"));
+
+    try testing.expect(CSSKeywords.startsWithFunction("rgb(1,2,3)"));
+    try testing.expect(CSSKeywords.startsWithFunction("rgba(1,2,3,4)"));
+
+    try testing.expect(CSSKeywords.startsWithFunction("my-custom-function(args)"));
+    try testing.expect(CSSKeywords.startsWithFunction("function-with-dashes(test)"));
+
+    try testing.expect(!CSSKeywords.startsWithFunction("123function(test)"));
+}
+
+test "Browser: CSS.StyleDeclaration: isHexColor - Unicode and invalid characters" {
+    try testing.expect(!isHexColor("#ghijkl"));
+    try testing.expect(!isHexColor("#12345g"));
+    try testing.expect(!isHexColor("#xyz"));
+
+    try testing.expect(!isHexColor("#АВС"));
+
+    try testing.expect(!isHexColor("#1234567g"));
+    try testing.expect(!isHexColor("#g2345678"));
+}
+
+test "Browser: CSS.StyleDeclaration: complex integration scenarios" {
+    const allocator = testing.arena_allocator;
+
+    try testing.expect(isMultiValueProperty("rgb(255,0,0) url(bg.jpg)"));
+
+    try testing.expect(!needsQuotes("calc(100% - 20px)"));
+
+    const result = try escapeCSSValue(allocator, "fake(function with spaces");
+    try testing.expectEqual("\"fake(function with spaces\"", result);
+
+    const important_result = extractImportant("rgb(255,0,0) !important");
+    try testing.expect(important_result.is_important);
+    try testing.expectEqual("rgb(255,0,0)", important_result.value);
+}
+
+test "Browser: CSS.StyleDeclaration: performance edge cases - empty and minimal inputs" {
+    try testing.expect(!isNumericWithUnit(""));
+    try testing.expect(!isHexColor(""));
+    try testing.expect(!isMultiValueProperty(""));
+    try testing.expect(!isAlreadyQuoted(""));
+    try testing.expect(!isValidPropertyName(""));
+    try testing.expect(needsQuotes(""));
+    try testing.expect(!CSSKeywords.isKnownKeyword(""));
+    try testing.expect(!CSSKeywords.containsSpecialChar(""));
+    try testing.expect(!CSSKeywords.isValidUnit(""));
+    try testing.expect(!CSSKeywords.startsWithFunction(""));
+
+    try testing.expect(!isNumericWithUnit("a"));
+    try testing.expect(!isHexColor("a"));
+    try testing.expect(!isMultiValueProperty("a"));
+    try testing.expect(!isAlreadyQuoted("a"));
+    try testing.expect(isValidPropertyName("a"));
+    try testing.expect(!needsQuotes("a"));
+}

src/browser/cssom/CSSStyleSheet.zig

@@ -0,0 +1,97 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+const StyleSheet = @import("StyleSheet.zig");
+const CSSRuleList = @import("CSSRuleList.zig");
+const CSSImportRule = @import("CSSRule.zig").CSSImportRule;
+
+const CSSStyleSheet = @This();
+pub const prototype = *StyleSheet;
+
+proto: StyleSheet,
+css_rules: CSSRuleList,
+owner_rule: ?*CSSImportRule,
+
+const CSSStyleSheetOpts = struct {
+    base_url: ?[]const u8 = null,
+    // TODO: Suupport media
+    disabled: bool = false,
+};
+
+pub fn constructor(_opts: ?CSSStyleSheetOpts) !CSSStyleSheet {
+    const opts = _opts orelse CSSStyleSheetOpts{};
+    return .{
+        .proto = .{ .disabled = opts.disabled },
+        .css_rules = .constructor(),
+        .owner_rule = null,
+    };
+}
+
+pub fn get_ownerRule(_: *CSSStyleSheet) ?*CSSImportRule {
+    return null;
+}
+
+pub fn get_cssRules(self: *CSSStyleSheet) *CSSRuleList {
+    return &self.css_rules;
+}
+
+pub fn _insertRule(self: *CSSStyleSheet, rule: []const u8, _index: ?usize, page: *Page) !usize {
+    const index = _index orelse 0;
+    if (index > self.css_rules.list.items.len) {
+        return error.IndexSize;
+    }
+
+    const arena = page.arena;
+    try self.css_rules.list.insert(arena, index, try arena.dupe(u8, rule));
+    return index;
+}
+
+pub fn _deleteRule(self: *CSSStyleSheet, index: usize) !void {
+    if (index > self.css_rules.list.items.len) {
+        return error.IndexSize;
+    }
+
+    _ = self.css_rules.list.orderedRemove(index);
+}
+
+pub fn _replace(self: *CSSStyleSheet, text: []const u8, page: *Page) !Env.Promise {
+    _ = self;
+    _ = text;
+    // TODO: clear self.css_rules
+    // parse text and re-populate self.css_rules
+
+    const resolver = page.main_context.createPromiseResolver();
+    try resolver.resolve({});
+    return resolver.promise();
+}
+
+pub fn _replaceSync(self: *CSSStyleSheet, text: []const u8) !void {
+    _ = self;
+    _ = text;
+    // TODO: clear self.css_rules
+    // parse text and re-populate self.css_rules
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: CSS.StyleSheet" {
+    try testing.htmlRunner("cssom/css_stylesheet.html");
+}

src/browser/cssom/StyleSheet.zig

@@ -0,0 +1,55 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+
+// https://developer.mozilla.org/en-US/docs/Web/API/StyleSheet#specifications
+const StyleSheet = @This();
+
+disabled: bool = false,
+href: []const u8 = "",
+owner_node: ?*parser.Node = null,
+parent_stylesheet: ?*StyleSheet = null,
+title: []const u8 = "",
+type: []const u8 = "text/css",
+
+pub fn get_disabled(self: *const StyleSheet) bool {
+    return self.disabled;
+}
+
+pub fn get_href(self: *const StyleSheet) []const u8 {
+    return self.href;
+}
+
+// TODO: media
+
+pub fn get_ownerNode(self: *const StyleSheet) ?*parser.Node {
+    return self.owner_node;
+}
+
+pub fn get_parentStyleSheet(self: *const StyleSheet) ?*StyleSheet {
+    return self.parent_stylesheet;
+}
+
+pub fn get_title(self: *const StyleSheet) []const u8 {
+    return self.title;
+}
+
+pub fn get_type(self: *const StyleSheet) []const u8 {
+    return self.type;
+}

src/browser/cssom/cssom.zig

@@ -0,0 +1,25 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+pub const Interfaces = .{
+    @import("StyleSheet.zig"),
+    @import("CSSStyleSheet.zig"),
+    @import("CSSStyleDeclaration.zig"),
+    @import("CSSRuleList.zig"),
+    @import("CSSRule.zig").Interfaces,
+};

src/browser/dom/Animation.zig

@@ -0,0 +1,109 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const Page = @import("../page.zig").Page;
+const JsObject = @import("../env.zig").JsObject;
+const Promise = @import("../env.zig").Promise;
+const PromiseResolver = @import("../env.zig").PromiseResolver;
+
+const Animation = @This();
+
+effect: ?JsObject,
+timeline: ?JsObject,
+ready_resolver: ?PromiseResolver,
+finished_resolver: ?PromiseResolver,
+
+pub fn constructor(effect: ?JsObject, timeline: ?JsObject) !Animation {
+    return .{
+        .effect = if (effect) |eo| try eo.persist() else null,
+        .timeline = if (timeline) |to| try to.persist() else null,
+        .ready_resolver = null,
+        .finished_resolver = null,
+    };
+}
+
+pub fn get_playState(self: *const Animation) []const u8 {
+    _ = self;
+    return "finished";
+}
+
+pub fn get_pending(self: *const Animation) bool {
+    _ = self;
+    return false;
+}
+
+pub fn get_finished(self: *Animation, page: *Page) !Promise {
+    if (self.finished_resolver == null) {
+        const resolver = page.main_context.createPromiseResolver();
+        try resolver.resolve(self);
+        self.finished_resolver = resolver;
+    }
+    return self.finished_resolver.?.promise();
+}
+
+pub fn get_ready(self: *Animation, page: *Page) !Promise {
+    // never resolved, because we're always "finished"
+    if (self.ready_resolver == null) {
+        const resolver = page.main_context.createPromiseResolver();
+        self.ready_resolver = resolver;
+    }
+    return self.ready_resolver.?.promise();
+}
+
+pub fn get_effect(self: *const Animation) ?JsObject {
+    return self.effect;
+}
+
+pub fn set_effect(self: *Animation, effect: JsObject) !void {
+    self.effect = try effect.persist();
+}
+
+pub fn get_timeline(self: *const Animation) ?JsObject {
+    return self.timeline;
+}
+
+pub fn set_timeline(self: *Animation, timeline: JsObject) !void {
+    self.timeline = try timeline.persist();
+}
+
+pub fn _play(self: *const Animation) void {
+    _ = self;
+}
+
+pub fn _pause(self: *const Animation) void {
+    _ = self;
+}
+
+pub fn _cancel(self: *const Animation) void {
+    _ = self;
+}
+
+pub fn _finish(self: *const Animation) void {
+    _ = self;
+}
+
+pub fn _reverse(self: *const Animation) void {
+    _ = self;
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Animation" {
+    try testing.htmlRunner("dom/animation.html");
+}

src/browser/dom/MessageChannel.zig

@@ -0,0 +1,291 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+const EventTarget = @import("../dom/event_target.zig").EventTarget;
+const EventHandler = @import("../events/event.zig").EventHandler;
+
+const JsObject = Env.JsObject;
+const Function = Env.Function;
+const Allocator = std.mem.Allocator;
+
+const MAX_QUEUE_SIZE = 10;
+
+pub const Interfaces = .{ MessageChannel, MessagePort };
+
+const MessageChannel = @This();
+
+port1: *MessagePort,
+port2: *MessagePort,
+
+pub fn constructor(page: *Page) !MessageChannel {
+    // Why do we allocate this rather than storing directly in the struct?
+    // https://github.com/lightpanda-io/project/discussions/165
+    const port1 = try page.arena.create(MessagePort);
+    const port2 = try page.arena.create(MessagePort);
+    port1.* = .{
+        .pair = port2,
+    };
+    port2.* = .{
+        .pair = port1,
+    };
+
+    return .{
+        .port1 = port1,
+        .port2 = port2,
+    };
+}
+
+pub fn get_port1(self: *const MessageChannel) *MessagePort {
+    return self.port1;
+}
+
+pub fn get_port2(self: *const MessageChannel) *MessagePort {
+    return self.port2;
+}
+
+pub const MessagePort = struct {
+    pub const prototype = *EventTarget;
+
+    proto: parser.EventTargetTBase = .{ .internal_target_type = .message_port },
+
+    pair: *MessagePort,
+    closed: bool = false,
+    started: bool = false,
+    onmessage_cbk: ?Function = null,
+    onmessageerror_cbk: ?Function = null,
+    // This is the queue of messages to dispatch to THIS MessagePort when the
+    // MessagePort is started.
+    queue: std.ArrayListUnmanaged(JsObject) = .empty,
+
+    pub const PostMessageOption = union(enum) {
+        transfer: JsObject,
+        options: Opts,
+
+        pub const Opts = struct {
+            transfer: JsObject,
+        };
+    };
+
+    pub fn _postMessage(self: *MessagePort, obj: JsObject, opts_: ?PostMessageOption, page: *Page) !void {
+        if (self.closed) {
+            return;
+        }
+
+        if (opts_ != null) {
+            log.warn(.web_api, "not implemented", .{ .feature = "MessagePort postMessage options" });
+            return error.NotImplemented;
+        }
+
+        try self.pair.dispatchOrQueue(obj, page.arena);
+    }
+
+    // Start impacts the ability to receive a message.
+    // Given pair1 (started) and pair2 (not started), then:
+    //    pair2.postMessage('x'); //will be dispatched to pair1.onmessage
+    //    pair1.postMessage('x'); // will be queued until pair2 is started
+    pub fn _start(self: *MessagePort) !void {
+        if (self.started) {
+            return;
+        }
+        self.started = true;
+        for (self.queue.items) |data| {
+            try self.dispatch(data);
+        }
+        // we'll never use this queue again, but it's allocated with an arena
+        // we don't even need to clear it, but it seems a bit safer to do at
+        // least that
+        self.queue.clearRetainingCapacity();
+    }
+
+    // Closing seems to stop both the publishing and receiving of messages,
+    // effectively rendering the channel useless. It cannot be reversed.
+    pub fn _close(self: *MessagePort) void {
+        self.closed = true;
+        self.pair.closed = true;
+    }
+
+    pub fn get_onmessage(self: *MessagePort) ?Function {
+        return self.onmessage_cbk;
+    }
+    pub fn get_onmessageerror(self: *MessagePort) ?Function {
+        return self.onmessageerror_cbk;
+    }
+
+    pub fn set_onmessage(self: *MessagePort, listener: EventHandler.Listener, page: *Page) !void {
+        if (self.onmessage_cbk) |cbk| {
+            try self.unregister("message", cbk.id);
+        }
+        self.onmessage_cbk = try self.register(page.arena, "message", listener);
+
+        // When onmessage is set directly, then it's like start() was called.
+        // If addEventListener('message') is used, the app has to call start()
+        // explicitly.
+        try self._start();
+    }
+
+    pub fn set_onmessageerror(self: *MessagePort, listener: EventHandler.Listener, page: *Page) !void {
+        if (self.onmessageerror_cbk) |cbk| {
+            try self.unregister("messageerror", cbk.id);
+        }
+        self.onmessageerror_cbk = try self.register(page.arena, "messageerror", listener);
+    }
+
+    // called from our pair. If port1.postMessage("x") is called, then this
+    // will be called on port2.
+    fn dispatchOrQueue(self: *MessagePort, obj: JsObject, arena: Allocator) !void {
+        // our pair should have checked this already
+        std.debug.assert(self.closed == false);
+
+        if (self.started) {
+            return self.dispatch(try obj.persist());
+        }
+
+        if (self.queue.items.len > MAX_QUEUE_SIZE) {
+            // This isn't part of the spec, but not putting a limit is reckless
+            return error.MessageQueueLimit;
+        }
+        return self.queue.append(arena, try obj.persist());
+    }
+
+    fn dispatch(self: *MessagePort, obj: JsObject) !void {
+        // obj is already persisted, don't use `MessageEvent.constructor`, but
+        // go directly to `init`, which assumes persisted objects.
+        var evt = try MessageEvent.init(.{ .data = obj });
+        _ = try parser.eventTargetDispatchEvent(
+            parser.toEventTarget(MessagePort, self),
+            @as(*parser.Event, @ptrCast(&evt)),
+        );
+    }
+
+    fn register(
+        self: *MessagePort,
+        alloc: Allocator,
+        typ: []const u8,
+        listener: EventHandler.Listener,
+    ) !?Function {
+        const target = @as(*parser.EventTarget, @ptrCast(self));
+        const eh = (try EventHandler.register(alloc, target, typ, listener, null)) orelse unreachable;
+        return eh.callback;
+    }
+
+    fn unregister(self: *MessagePort, typ: []const u8, cbk_id: usize) !void {
+        const et = @as(*parser.EventTarget, @ptrCast(self));
+        const lst = try parser.eventTargetHasListener(et, typ, false, cbk_id);
+        if (lst == null) {
+            return;
+        }
+        try parser.eventTargetRemoveEventListener(et, typ, lst.?, false);
+    }
+};
+
+pub const MessageEvent = struct {
+    const Event = @import("../events/event.zig").Event;
+    const DOMException = @import("exceptions.zig").DOMException;
+
+    pub const prototype = *Event;
+    pub const Exception = DOMException;
+    pub const union_make_copy = true;
+
+    proto: parser.Event,
+    data: ?JsObject,
+
+    // You would think if port1 sends to port2, the source would be port2
+    // (which is how I read the documentation), but it appears to always be
+    // null. It can always be set explicitly via the constructor;
+    source: ?JsObject,
+
+    origin: []const u8,
+
+    // This is used for Server-Sent events. Appears to always be an empty
+    // string for MessagePort messages.
+    last_event_id: []const u8,
+
+    // This might be related to the "transfer" option of postMessage which
+    // we don't yet support. For "normal" message, it's always an empty array.
+    // Though it could be set explicitly via the constructor
+    ports: []*MessagePort,
+
+    const Options = struct {
+        data: ?JsObject = null,
+        source: ?JsObject = null,
+        origin: []const u8 = "",
+        lastEventId: []const u8 = "",
+        ports: []*MessagePort = &.{},
+    };
+
+    pub fn constructor(opts: Options) !MessageEvent {
+        return init(.{
+            .data = if (opts.data) |obj| try obj.persist() else null,
+            .source = if (opts.source) |obj| try obj.persist() else null,
+            .ports = opts.ports,
+            .origin = opts.origin,
+            .lastEventId = opts.lastEventId,
+        });
+    }
+
+    // This is like "constructor", but it assumes JsObjects have already been
+    // persisted. Necessary because this `new MessageEvent()` can be called
+    // directly from JS OR from a port.postMessage. In the latter case, data
+    // may have already been persisted (as it might need to be queued);
+    fn init(opts: Options) !MessageEvent {
+        const event = try parser.eventCreate();
+        defer parser.eventDestroy(event);
+        try parser.eventInit(event, "message", .{});
+        parser.eventSetInternalType(event, .message_event);
+
+        return .{
+            .proto = event.*,
+            .data = opts.data,
+            .source = opts.source,
+            .ports = opts.ports,
+            .origin = opts.origin,
+            .last_event_id = opts.lastEventId,
+        };
+    }
+
+    pub fn get_data(self: *const MessageEvent) !?JsObject {
+        return self.data;
+    }
+
+    pub fn get_origin(self: *const MessageEvent) []const u8 {
+        return self.origin;
+    }
+
+    pub fn get_source(self: *const MessageEvent) ?JsObject {
+        return self.source;
+    }
+
+    pub fn get_ports(self: *const MessageEvent) []*MessagePort {
+        return self.ports;
+    }
+
+    pub fn get_lastEventId(self: *const MessageEvent) []const u8 {
+        return self.last_event_id;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.MessageChannel" {
+    try testing.htmlRunner("dom/message_channel.html");
+}

src/browser/dom/attribute.zig

@@ -0,0 +1,75 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+const parser = @import("../netsurf.zig");
+
+const Node = @import("node.zig").Node;
+
+// WEB IDL https://dom.spec.whatwg.org/#attr
+pub const Attr = struct {
+    pub const Self = parser.Attribute;
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    pub fn get_namespaceURI(self: *parser.Attribute) !?[]const u8 {
+        return try parser.nodeGetNamespace(parser.attributeToNode(self));
+    }
+
+    pub fn get_prefix(self: *parser.Attribute) !?[]const u8 {
+        return try parser.nodeGetPrefix(parser.attributeToNode(self));
+    }
+
+    pub fn get_localName(self: *parser.Attribute) ![]const u8 {
+        return try parser.nodeLocalName(parser.attributeToNode(self));
+    }
+
+    pub fn get_name(self: *parser.Attribute) ![]const u8 {
+        return try parser.attributeGetName(self);
+    }
+
+    pub fn get_value(self: *parser.Attribute) !?[]const u8 {
+        return try parser.attributeGetValue(self);
+    }
+
+    pub fn set_value(self: *parser.Attribute, v: []const u8) !?[]const u8 {
+        if (try parser.attributeGetOwnerElement(self)) |el| {
+            // if possible, go through the element, as that triggers a
+            // DOMAttrModified event (which MutationObserver cares about)
+            const name = try parser.attributeGetName(self);
+            try parser.elementSetAttribute(el, name, v);
+        } else {
+            try parser.attributeSetValue(self, v);
+        }
+        return v;
+    }
+
+    pub fn get_ownerElement(self: *parser.Attribute) !?*parser.Element {
+        return try parser.attributeGetOwnerElement(self);
+    }
+
+    pub fn get_specified(_: *parser.Attribute) bool {
+        return true;
+    }
+};
+
+// Tests
+// -----
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Attribute" {
+    try testing.htmlRunner("dom/attribute.html");
+}

src/browser/dom/cdata_section.zig

@@ -0,0 +1,28 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+
+const Text = @import("text.zig").Text;
+
+// https://dom.spec.whatwg.org/#cdatasection
+pub const CDATASection = struct {
+    pub const Self = parser.CDATASection;
+    pub const prototype = *Text;
+    pub const subtype = .node;
+};

src/browser/dom/character_data.zig

@@ -0,0 +1,134 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+
+const Node = @import("node.zig").Node;
+const Comment = @import("comment.zig").Comment;
+const Text = @import("text.zig");
+const ProcessingInstruction = @import("processing_instruction.zig").ProcessingInstruction;
+const Element = @import("element.zig").Element;
+const ElementUnion = @import("element.zig").Union;
+
+// CharacterData interfaces
+pub const Interfaces = .{
+    Comment,
+    Text.Text,
+    Text.Interfaces,
+    ProcessingInstruction,
+};
+
+// CharacterData implementation
+pub const CharacterData = struct {
+    pub const Self = parser.CharacterData;
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    // JS funcs
+    // --------
+
+    // Read attributes
+
+    pub fn get_length(self: *parser.CharacterData) !u32 {
+        return try parser.characterDataLength(self);
+    }
+
+    pub fn get_nextElementSibling(self: *parser.CharacterData) !?ElementUnion {
+        const res = try parser.nodeNextElementSibling(parser.characterDataToNode(self));
+        if (res == null) {
+            return null;
+        }
+        return try Element.toInterface(res.?);
+    }
+
+    pub fn get_previousElementSibling(self: *parser.CharacterData) !?ElementUnion {
+        const res = try parser.nodePreviousElementSibling(parser.characterDataToNode(self));
+        if (res == null) {
+            return null;
+        }
+        return try Element.toInterface(res.?);
+    }
+
+    // Read/Write attributes
+
+    pub fn get_data(self: *parser.CharacterData) ![]const u8 {
+        return try parser.characterDataData(self);
+    }
+
+    pub fn set_data(self: *parser.CharacterData, data: []const u8) !void {
+        return try parser.characterDataSetData(self, data);
+    }
+
+    // JS methods
+    // ----------
+
+    pub fn _appendData(self: *parser.CharacterData, data: []const u8) !void {
+        return try parser.characterDataAppendData(self, data);
+    }
+
+    pub fn _deleteData(self: *parser.CharacterData, offset: u32, count: u32) !void {
+        return try parser.characterDataDeleteData(self, offset, count);
+    }
+
+    pub fn _insertData(self: *parser.CharacterData, offset: u32, data: []const u8) !void {
+        return try parser.characterDataInsertData(self, offset, data);
+    }
+
+    pub fn _replaceData(self: *parser.CharacterData, offset: u32, count: u32, data: []const u8) !void {
+        return try parser.characterDataReplaceData(self, offset, count, data);
+    }
+
+    pub fn _substringData(self: *parser.CharacterData, offset: u32, count: u32) ![]const u8 {
+        return try parser.characterDataSubstringData(self, offset, count);
+    }
+
+    // netsurf's CharacterData (text, comment) doesn't implement the
+    // dom_node_get_attributes and thus will crash if we try to call nodeIsEqualNode.
+    pub fn _isEqualNode(self: *parser.CharacterData, other_node: *parser.Node) !bool {
+        if (try parser.nodeType(@ptrCast(@alignCast(self))) != try parser.nodeType(other_node)) {
+            return false;
+        }
+
+        const other: *parser.CharacterData = @ptrCast(other_node);
+        if (std.mem.eql(u8, try get_data(self), try get_data(other)) == false) {
+            return false;
+        }
+
+        return true;
+    }
+
+    pub fn _before(self: *parser.CharacterData, nodes: []const Node.NodeOrText) !void {
+        const ref_node = parser.characterDataToNode(self);
+        return Node.before(ref_node, nodes);
+    }
+
+    pub fn _after(self: *parser.CharacterData, nodes: []const Node.NodeOrText) !void {
+        const ref_node = parser.characterDataToNode(self);
+        return Node.after(ref_node, nodes);
+    }
+};
+
+// Tests
+// -----
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.CharacterData" {
+    try testing.htmlRunner("dom/character_data.html");
+}

src/browser/dom/comment.zig

@@ -0,0 +1,45 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+
+const CharacterData = @import("character_data.zig").CharacterData;
+
+const Page = @import("../page.zig").Page;
+
+// https://dom.spec.whatwg.org/#interface-comment
+pub const Comment = struct {
+    pub const Self = parser.Comment;
+    pub const prototype = *CharacterData;
+    pub const subtype = .node;
+
+    pub fn constructor(data: ?[]const u8, page: *const Page) !*parser.Comment {
+        return parser.documentCreateComment(
+            parser.documentHTMLToDocument(page.window.document),
+            data orelse "",
+        );
+    }
+};
+
+// Tests
+// -----
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Comment" {
+    try testing.htmlRunner("dom/comment.html");
+}

src/browser/dom/css.zig

@@ -0,0 +1,80 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+
+const css = @import("../css/css.zig");
+const Node = @import("../css/libdom.zig").Node;
+const NodeList = @import("nodelist.zig").NodeList;
+
+const MatchFirst = struct {
+    n: ?*parser.Node = null,
+
+    pub fn match(m: *MatchFirst, n: Node) !void {
+        m.n = n.node;
+    }
+};
+
+pub fn querySelector(alloc: std.mem.Allocator, n: *parser.Node, selector: []const u8) !?*parser.Node {
+    const ps = try css.parse(alloc, selector, .{ .accept_pseudo_elts = true });
+    defer ps.deinit(alloc);
+
+    var m = MatchFirst{};
+
+    _ = try css.matchFirst(&ps, Node{ .node = n }, &m);
+    return m.n;
+}
+
+const MatchAll = struct {
+    alloc: std.mem.Allocator,
+    nl: NodeList,
+
+    fn init(alloc: std.mem.Allocator) MatchAll {
+        return .{
+            .alloc = alloc,
+            .nl = .{},
+        };
+    }
+
+    fn deinit(m: *MatchAll) void {
+        m.nl.deinit(m.alloc);
+    }
+
+    pub fn match(m: *MatchAll, n: Node) !void {
+        try m.nl.append(m.alloc, n.node);
+    }
+
+    fn toOwnedList(m: *MatchAll) NodeList {
+        // reset it.
+        defer m.nl = .{};
+        return m.nl;
+    }
+};
+
+pub fn querySelectorAll(alloc: std.mem.Allocator, n: *parser.Node, selector: []const u8) !NodeList {
+    const ps = try css.parse(alloc, selector, .{ .accept_pseudo_elts = true });
+    defer ps.deinit(alloc);
+
+    var m = MatchAll.init(alloc);
+    defer m.deinit();
+
+    try css.matchAll(&ps, Node{ .node = n }, &m);
+    return m.toOwnedList();
+}

src/browser/dom/document.zig

@@ -0,0 +1,331 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const Node = @import("node.zig").Node;
+const NodeList = @import("nodelist.zig").NodeList;
+const NodeUnion = @import("node.zig").Union;
+
+const collection = @import("html_collection.zig");
+const css = @import("css.zig");
+
+const Element = @import("element.zig").Element;
+const ElementUnion = @import("element.zig").Union;
+const TreeWalker = @import("tree_walker.zig").TreeWalker;
+const CSSStyleSheet = @import("../cssom/CSSStyleSheet.zig");
+const NodeIterator = @import("node_iterator.zig").NodeIterator;
+const Range = @import("range.zig").Range;
+
+const CustomEvent = @import("../events/custom_event.zig").CustomEvent;
+
+const Env = @import("../env.zig").Env;
+
+const DOMImplementation = @import("implementation.zig").DOMImplementation;
+
+// WEB IDL https://dom.spec.whatwg.org/#document
+pub const Document = struct {
+    pub const Self = parser.Document;
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    pub fn constructor(page: *const Page) !*parser.DocumentHTML {
+        const doc = try parser.documentCreateDocument(
+            try parser.documentHTMLGetTitle(page.window.document),
+        );
+
+        // we have to work w/ document instead of html document.
+        const ddoc = parser.documentHTMLToDocument(doc);
+        const ccur = parser.documentHTMLToDocument(page.window.document);
+        try parser.documentSetDocumentURI(ddoc, try parser.documentGetDocumentURI(ccur));
+        try parser.documentSetInputEncoding(ddoc, try parser.documentGetInputEncoding(ccur));
+
+        return doc;
+    }
+
+    // JS funcs
+    // --------
+    pub fn get_implementation(_: *parser.Document) DOMImplementation {
+        return DOMImplementation{};
+    }
+
+    pub fn get_documentElement(self: *parser.Document) !?ElementUnion {
+        const e = try parser.documentGetDocumentElement(self);
+        if (e == null) return null;
+        return try Element.toInterface(e.?);
+    }
+
+    pub fn get_documentURI(self: *parser.Document) ![]const u8 {
+        return try parser.documentGetDocumentURI(self);
+    }
+
+    pub fn get_URL(self: *parser.Document) ![]const u8 {
+        return try get_documentURI(self);
+    }
+
+    // TODO implement contentType
+    pub fn get_contentType(self: *parser.Document) []const u8 {
+        _ = self;
+        return "text/html";
+    }
+
+    // TODO implement compactMode
+    pub fn get_compatMode(self: *parser.Document) []const u8 {
+        _ = self;
+        return "CSS1Compat";
+    }
+
+    pub fn get_characterSet(self: *parser.Document) ![]const u8 {
+        return try parser.documentGetInputEncoding(self);
+    }
+
+    // alias of get_characterSet
+    pub fn get_charset(self: *parser.Document) ![]const u8 {
+        return try get_characterSet(self);
+    }
+
+    // alias of get_characterSet
+    pub fn get_inputEncoding(self: *parser.Document) ![]const u8 {
+        return try get_characterSet(self);
+    }
+
+    pub fn get_doctype(self: *parser.Document) !?*parser.DocumentType {
+        return try parser.documentGetDoctype(self);
+    }
+
+    pub fn _createEvent(_: *parser.Document, eventCstr: []const u8) !union(enum) {
+        base: *parser.Event,
+        custom: CustomEvent,
+    } {
+        const eqlIgnoreCase = std.ascii.eqlIgnoreCase;
+
+        if (eqlIgnoreCase(eventCstr, "Event") or eqlIgnoreCase(eventCstr, "Events") or eqlIgnoreCase(eventCstr, "HTMLEvents")) {
+            return .{ .base = try parser.eventCreate() };
+        }
+
+        // Not documented in MDN but supported in Chrome.
+        // This is actually both instance of `Event` and `CustomEvent`.
+        if (std.ascii.eqlIgnoreCase(eventCstr, "CustomEvent")) {
+            return .{ .custom = try CustomEvent.constructor(eventCstr, null) };
+        }
+
+        return error.NotSupported;
+    }
+
+    pub fn _getElementById(self: *parser.Document, id: []const u8) !?ElementUnion {
+        const e = try parser.documentGetElementById(self, id) orelse return null;
+        return try Element.toInterface(e);
+    }
+
+    pub fn _createElement(self: *parser.Document, tag_name: []const u8) !ElementUnion {
+        // The element’s namespace is the HTML namespace when document is an HTML document
+        // https://dom.spec.whatwg.org/#ref-for-dom-document-createelement%E2%91%A0
+        const e = try parser.documentCreateElementNS(self, "http://www.w3.org/1999/xhtml", tag_name);
+        return Element.toInterface(e);
+    }
+
+    pub fn _createElementNS(self: *parser.Document, ns: []const u8, tag_name: []const u8) !ElementUnion {
+        const e = try parser.documentCreateElementNS(self, ns, tag_name);
+        return try Element.toInterface(e);
+    }
+
+    // We can't simply use libdom dom_document_get_elements_by_tag_name here.
+    // Indeed, netsurf implemented a previous dom spec when
+    // getElementsByTagName returned a NodeList.
+    // But since
+    // https://github.com/whatwg/dom/commit/190700b7c12ecfd3b5ebdb359ab1d6ea9cbf7749
+    // the spec changed to return an HTMLCollection instead.
+    // That's why we reimplemented getElementsByTagName by using an
+    // HTMLCollection in zig here.
+    pub fn _getElementsByTagName(
+        self: *parser.Document,
+        tag_name: []const u8,
+        page: *Page,
+    ) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByTagName(page.arena, parser.documentToNode(self), tag_name, .{
+            .include_root = true,
+        });
+    }
+
+    pub fn _getElementsByClassName(
+        self: *parser.Document,
+        classNames: []const u8,
+        page: *Page,
+    ) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByClassName(page.arena, parser.documentToNode(self), classNames, .{
+            .include_root = true,
+        });
+    }
+
+    pub fn _createDocumentFragment(self: *parser.Document) !*parser.DocumentFragment {
+        return try parser.documentCreateDocumentFragment(self);
+    }
+
+    pub fn _createTextNode(self: *parser.Document, data: []const u8) !*parser.Text {
+        return try parser.documentCreateTextNode(self, data);
+    }
+
+    pub fn _createCDATASection(self: *parser.Document, data: []const u8) !*parser.CDATASection {
+        return try parser.documentCreateCDATASection(self, data);
+    }
+
+    pub fn _createComment(self: *parser.Document, data: []const u8) !*parser.Comment {
+        return try parser.documentCreateComment(self, data);
+    }
+
+    pub fn _createProcessingInstruction(self: *parser.Document, target: []const u8, data: []const u8) !*parser.ProcessingInstruction {
+        return try parser.documentCreateProcessingInstruction(self, target, data);
+    }
+
+    pub fn _importNode(self: *parser.Document, node: *parser.Node, deep: ?bool) !NodeUnion {
+        const n = try parser.documentImportNode(self, node, deep orelse false);
+        return try Node.toInterface(n);
+    }
+
+    pub fn _adoptNode(self: *parser.Document, node: *parser.Node) !NodeUnion {
+        const n = try parser.documentAdoptNode(self, node);
+        return try Node.toInterface(n);
+    }
+
+    pub fn _createAttribute(self: *parser.Document, name: []const u8) !*parser.Attribute {
+        return try parser.documentCreateAttribute(self, name);
+    }
+
+    pub fn _createAttributeNS(self: *parser.Document, ns: []const u8, qname: []const u8) !*parser.Attribute {
+        return try parser.documentCreateAttributeNS(self, ns, qname);
+    }
+
+    // ParentNode
+    // https://dom.spec.whatwg.org/#parentnode
+    pub fn get_children(self: *parser.Document) !collection.HTMLCollection {
+        return collection.HTMLCollectionChildren(parser.documentToNode(self), .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_firstElementChild(self: *parser.Document) !?ElementUnion {
+        const elt = try parser.documentGetDocumentElement(self) orelse return null;
+        return try Element.toInterface(elt);
+    }
+
+    pub fn get_lastElementChild(self: *parser.Document) !?ElementUnion {
+        const elt = try parser.documentGetDocumentElement(self) orelse return null;
+        return try Element.toInterface(elt);
+    }
+
+    pub fn get_childElementCount(self: *parser.Document) !u32 {
+        _ = try parser.documentGetDocumentElement(self) orelse return 0;
+        return 1;
+    }
+
+    pub fn _querySelector(self: *parser.Document, selector: []const u8, page: *Page) !?ElementUnion {
+        if (selector.len == 0) return null;
+
+        const n = try css.querySelector(page.call_arena, parser.documentToNode(self), selector);
+
+        if (n == null) return null;
+
+        return try Element.toInterface(parser.nodeToElement(n.?));
+    }
+
+    pub fn _querySelectorAll(self: *parser.Document, selector: []const u8, page: *Page) !NodeList {
+        return css.querySelectorAll(page.arena, parser.documentToNode(self), selector);
+    }
+
+    pub fn _prepend(self: *parser.Document, nodes: []const Node.NodeOrText) !void {
+        return Node.prepend(parser.documentToNode(self), nodes);
+    }
+
+    pub fn _append(self: *parser.Document, nodes: []const Node.NodeOrText) !void {
+        return Node.append(parser.documentToNode(self), nodes);
+    }
+
+    pub fn _replaceChildren(self: *parser.Document, nodes: []const Node.NodeOrText) !void {
+        return Node.replaceChildren(parser.documentToNode(self), nodes);
+    }
+
+    pub fn _createTreeWalker(_: *parser.Document, root: *parser.Node, what_to_show: ?TreeWalker.WhatToShow, filter: ?TreeWalker.TreeWalkerOpts) !TreeWalker {
+        return TreeWalker.init(root, what_to_show, filter);
+    }
+
+    pub fn _createNodeIterator(_: *parser.Document, root: *parser.Node, what_to_show: ?NodeIterator.WhatToShow, filter: ?NodeIterator.NodeIteratorOpts) !NodeIterator {
+        return NodeIterator.init(root, what_to_show, filter);
+    }
+
+    pub fn getActiveElement(self: *parser.Document, page: *Page) !?*parser.Element {
+        if (page.getNodeState(@ptrCast(@alignCast(self)))) |state| {
+            if (state.active_element) |ae| {
+                return ae;
+            }
+        }
+
+        if (try parser.documentHTMLBody(page.window.document)) |body| {
+            return @ptrCast(@alignCast(body));
+        }
+
+        return try parser.documentGetDocumentElement(self);
+    }
+
+    pub fn get_activeElement(self: *parser.Document, page: *Page) !?ElementUnion {
+        const ae = (try getActiveElement(self, page)) orelse return null;
+        return try Element.toInterface(ae);
+    }
+
+    // TODO: some elements can't be focused, like if they're disabled
+    // but there doesn't seem to be a generic way to check this. For example
+    // we could look for the "disabled" attribute, but that's only meaningful
+    // on certain types, and libdom's vtable doesn't seem to expose this.
+    pub fn setFocus(self: *parser.Document, e: *parser.ElementHTML, page: *Page) !void {
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        state.active_element = @ptrCast(e);
+    }
+
+    pub fn _createRange(_: *parser.Document, page: *Page) Range {
+        return Range.constructor(page);
+    }
+
+    // TODO: dummy implementation
+    pub fn get_styleSheets(_: *parser.Document) []CSSStyleSheet {
+        return &.{};
+    }
+
+    pub fn get_adoptedStyleSheets(self: *parser.Document, page: *Page) !Env.JsObject {
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        if (state.adopted_style_sheets) |obj| {
+            return obj;
+        }
+
+        const obj = try page.main_context.newArray(0).persist();
+        state.adopted_style_sheets = obj;
+        return obj;
+    }
+
+    pub fn set_adoptedStyleSheets(self: *parser.Document, sheets: Env.JsObject, page: *Page) !void {
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        state.adopted_style_sheets = try sheets.persist();
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Document" {
+    try testing.htmlRunner("dom/document.html");
+}

src/browser/dom/document_fragment.zig

@@ -0,0 +1,96 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const css = @import("css.zig");
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+const NodeList = @import("nodelist.zig").NodeList;
+const Element = @import("element.zig").Element;
+const ElementUnion = @import("element.zig").Union;
+const collection = @import("html_collection.zig");
+
+const Node = @import("node.zig").Node;
+
+// WEB IDL https://dom.spec.whatwg.org/#documentfragment
+pub const DocumentFragment = struct {
+    pub const Self = parser.DocumentFragment;
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    pub fn constructor(page: *const Page) !*parser.DocumentFragment {
+        return parser.documentCreateDocumentFragment(
+            parser.documentHTMLToDocument(page.window.document),
+        );
+    }
+
+    pub fn _isEqualNode(self: *parser.DocumentFragment, other_node: *parser.Node) !bool {
+        const other_type = try parser.nodeType(other_node);
+        if (other_type != .document_fragment) {
+            return false;
+        }
+        _ = self;
+        return true;
+    }
+
+    pub fn _prepend(self: *parser.DocumentFragment, nodes: []const Node.NodeOrText) !void {
+        return Node.prepend(parser.documentFragmentToNode(self), nodes);
+    }
+
+    pub fn _append(self: *parser.DocumentFragment, nodes: []const Node.NodeOrText) !void {
+        return Node.append(parser.documentFragmentToNode(self), nodes);
+    }
+
+    pub fn _replaceChildren(self: *parser.DocumentFragment, nodes: []const Node.NodeOrText) !void {
+        return Node.replaceChildren(parser.documentFragmentToNode(self), nodes);
+    }
+
+    pub fn _querySelector(self: *parser.DocumentFragment, selector: []const u8, page: *Page) !?ElementUnion {
+        if (selector.len == 0) return null;
+
+        const n = try css.querySelector(page.call_arena, parser.documentFragmentToNode(self), selector);
+
+        if (n == null) return null;
+
+        return try Element.toInterface(parser.nodeToElement(n.?));
+    }
+
+    pub fn _querySelectorAll(self: *parser.DocumentFragment, selector: []const u8, page: *Page) !NodeList {
+        return css.querySelectorAll(page.arena, parser.documentFragmentToNode(self), selector);
+    }
+
+    pub fn get_childElementCount(self: *parser.DocumentFragment) !u32 {
+        var children = try get_children(self);
+        return children.get_length();
+    }
+
+    pub fn get_children(self: *parser.DocumentFragment) !collection.HTMLCollection {
+        return collection.HTMLCollectionChildren(parser.documentFragmentToNode(self), .{
+            .include_root = false,
+        });
+    }
+
+    pub fn _getElementById(self: *parser.DocumentFragment, id: []const u8) !?ElementUnion {
+        const e = try parser.nodeGetElementById(@ptrCast(@alignCast(self)), id) orelse return null;
+        return try Element.toInterface(e);
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.DocumentFragment" {
+    try testing.htmlRunner("dom/document_fragment.html");
+}

src/browser/dom/document_type.zig

@@ -0,0 +1,67 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+
+const Node = @import("node.zig").Node;
+
+// WEB IDL https://dom.spec.whatwg.org/#documenttype
+pub const DocumentType = struct {
+    pub const Self = parser.DocumentType;
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    pub fn get_name(self: *parser.DocumentType) ![]const u8 {
+        return try parser.documentTypeGetName(self);
+    }
+
+    pub fn get_publicId(self: *parser.DocumentType) ![]const u8 {
+        return try parser.documentTypeGetPublicId(self);
+    }
+
+    pub fn get_systemId(self: *parser.DocumentType) ![]const u8 {
+        return try parser.documentTypeGetSystemId(self);
+    }
+
+    // netsurf's DocumentType doesn't implement the dom_node_get_attributes
+    // and thus will crash if we try to call nodeIsEqualNode.
+    pub fn _isEqualNode(self: *parser.DocumentType, other_node: *parser.Node) !bool {
+        if (try parser.nodeType(other_node) != .document_type) {
+            return false;
+        }
+
+        const other: *parser.DocumentType = @ptrCast(other_node);
+        if (std.mem.eql(u8, try get_name(self), try get_name(other)) == false) {
+            return false;
+        }
+        if (std.mem.eql(u8, try get_publicId(self), try get_publicId(other)) == false) {
+            return false;
+        }
+        if (std.mem.eql(u8, try get_systemId(self), try get_systemId(other)) == false) {
+            return false;
+        }
+        return true;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.DocumentType" {
+    try testing.htmlRunner("dom/document_type.html");
+}

src/browser/dom/dom.zig

@@ -0,0 +1,57 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const DOMException = @import("exceptions.zig").DOMException;
+const EventTarget = @import("event_target.zig").EventTarget;
+const DOMImplementation = @import("implementation.zig").DOMImplementation;
+const NamedNodeMap = @import("namednodemap.zig").NamedNodeMap;
+const DOMTokenList = @import("token_list.zig");
+const NodeList = @import("nodelist.zig");
+const Node = @import("node.zig");
+const ResizeObserver = @import("resize_observer.zig");
+const MutationObserver = @import("mutation_observer.zig");
+const IntersectionObserver = @import("intersection_observer.zig");
+const DOMParser = @import("dom_parser.zig").DOMParser;
+const TreeWalker = @import("tree_walker.zig").TreeWalker;
+const NodeIterator = @import("node_iterator.zig").NodeIterator;
+const NodeFilter = @import("node_filter.zig").NodeFilter;
+const PerformanceObserver = @import("performance_observer.zig").PerformanceObserver;
+
+pub const Interfaces = .{
+    DOMException,
+    EventTarget,
+    DOMImplementation,
+    NamedNodeMap,
+    NamedNodeMap.Iterator,
+    DOMTokenList.Interfaces,
+    NodeList.Interfaces,
+    Node.Node,
+    Node.Interfaces,
+    ResizeObserver.Interfaces,
+    MutationObserver.Interfaces,
+    IntersectionObserver.Interfaces,
+    DOMParser,
+    TreeWalker,
+    NodeIterator,
+    NodeFilter,
+    @import("performance.zig").Interfaces,
+    PerformanceObserver,
+    @import("range.zig").Interfaces,
+    @import("Animation.zig"),
+    @import("MessageChannel.zig").Interfaces,
+};

src/browser/dom/dom_parser.zig

@@ -0,0 +1,41 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const parser = @import("../netsurf.zig");
+
+// https://developer.mozilla.org/en-US/docs/Web/API/DOMParser
+pub const DOMParser = struct {
+    pub fn constructor() !DOMParser {
+        return .{};
+    }
+
+    pub fn _parseFromString(_: *DOMParser, string: []const u8, mime_type: []const u8) !*parser.DocumentHTML {
+        if (!std.mem.eql(u8, mime_type, "text/html")) {
+            // TODO: Support XML
+            return error.TypeError;
+        }
+
+        return try parser.documentHTMLParseFromStr(string);
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Parser" {
+    try testing.htmlRunner("dom/dom_parser.html");
+}

src/browser/dom/element.zig

@@ -0,0 +1,609 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const css = @import("css.zig");
+const log = @import("../../log.zig");
+const dump = @import("../dump.zig");
+const collection = @import("html_collection.zig");
+
+const Node = @import("node.zig").Node;
+const Walker = @import("walker.zig").WalkerDepthFirst;
+const NodeList = @import("nodelist.zig").NodeList;
+const HTMLElem = @import("../html/elements.zig");
+const ShadowRoot = @import("../dom/shadow_root.zig").ShadowRoot;
+
+const Animation = @import("Animation.zig");
+const JsObject = @import("../env.zig").JsObject;
+
+pub const Union = @import("../html/elements.zig").Union;
+
+// WEB IDL https://dom.spec.whatwg.org/#element
+pub const Element = struct {
+    pub const Self = parser.Element;
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    pub const DOMRect = struct {
+        x: f64,
+        y: f64,
+        width: f64,
+        height: f64,
+        bottom: f64,
+        right: f64,
+        top: f64,
+        left: f64,
+    };
+
+    pub fn toInterface(e: *parser.Element) !Union {
+        return toInterfaceT(Union, e);
+    }
+
+    pub fn toInterfaceT(comptime T: type, e: *parser.Element) !T {
+        const tagname = try parser.elementGetTagName(e) orelse {
+            // If the owner's document is HTML, assume we have an HTMLElement.
+            const doc = try parser.nodeOwnerDocument(parser.elementToNode(e));
+            if (doc != null and !doc.?.is_html) {
+                return .{ .HTMLElement = @as(*parser.ElementHTML, @ptrCast(e)) };
+            }
+
+            return .{ .Element = e };
+        };
+
+        // TODO SVGElement and MathML are not supported yet.
+
+        const tag = parser.Tag.fromString(tagname) catch {
+            // If the owner's document is HTML, assume we have an HTMLElement.
+            const doc = try parser.nodeOwnerDocument(parser.elementToNode(e));
+            if (doc != null and doc.?.is_html) {
+                return .{ .HTMLElement = @as(*parser.ElementHTML, @ptrCast(e)) };
+            }
+
+            return .{ .Element = e };
+        };
+
+        return HTMLElem.toInterfaceFromTag(T, e, tag);
+    }
+
+    // JS funcs
+    // --------
+
+    pub fn get_namespaceURI(self: *parser.Element) !?[]const u8 {
+        return try parser.nodeGetNamespace(parser.elementToNode(self));
+    }
+
+    pub fn get_prefix(self: *parser.Element) !?[]const u8 {
+        return try parser.nodeGetPrefix(parser.elementToNode(self));
+    }
+
+    pub fn get_localName(self: *parser.Element) ![]const u8 {
+        return try parser.nodeLocalName(parser.elementToNode(self));
+    }
+
+    pub fn get_tagName(self: *parser.Element) ![]const u8 {
+        return try parser.nodeName(parser.elementToNode(self));
+    }
+
+    pub fn get_id(self: *parser.Element) ![]const u8 {
+        return try parser.elementGetAttribute(self, "id") orelse "";
+    }
+
+    pub fn set_id(self: *parser.Element, id: []const u8) !void {
+        return try parser.elementSetAttribute(self, "id", id);
+    }
+
+    pub fn get_className(self: *parser.Element) ![]const u8 {
+        return try parser.elementGetAttribute(self, "class") orelse "";
+    }
+
+    pub fn set_className(self: *parser.Element, class: []const u8) !void {
+        return try parser.elementSetAttribute(self, "class", class);
+    }
+
+    pub fn get_slot(self: *parser.Element) ![]const u8 {
+        return try parser.elementGetAttribute(self, "slot") orelse "";
+    }
+
+    pub fn set_slot(self: *parser.Element, slot: []const u8) !void {
+        return try parser.elementSetAttribute(self, "slot", slot);
+    }
+
+    pub fn get_classList(self: *parser.Element) !*parser.TokenList {
+        return try parser.tokenListCreate(self, "class");
+    }
+
+    pub fn get_attributes(self: *parser.Element) !*parser.NamedNodeMap {
+        // An element must have non-nil attributes.
+        return try parser.nodeGetAttributes(parser.elementToNode(self)) orelse unreachable;
+    }
+
+    pub fn get_innerHTML(self: *parser.Element, page: *Page) ![]const u8 {
+        var aw = std.Io.Writer.Allocating.init(page.call_arena);
+        try dump.writeChildren(parser.elementToNode(self), .{}, &aw.writer);
+        return aw.written();
+    }
+
+    pub fn get_outerHTML(self: *parser.Element, page: *Page) ![]const u8 {
+        var aw = std.Io.Writer.Allocating.init(page.call_arena);
+        try dump.writeNode(parser.elementToNode(self), .{}, &aw.writer);
+        return aw.written();
+    }
+
+    pub fn set_innerHTML(self: *parser.Element, str: []const u8, page: *Page) !void {
+        const node = parser.elementToNode(self);
+        const doc = try parser.nodeOwnerDocument(node) orelse return parser.DOMError.WrongDocument;
+        // parse the fragment
+        const fragment = try parser.documentParseFragmentFromStr(doc, str);
+
+        // remove existing children
+        try Node.removeChildren(node);
+
+        const fragment_node = parser.documentFragmentToNode(fragment);
+
+        // I'm not sure what the exact behavior is supposed to be. Initially,
+        // we were only copying the body of the document fragment. But it seems
+        // like head elements should be copied too. Specifically, some sites
+        // create script tags via innerHTML, which we need to capture.
+        // If you play with this in a browser, you should notice that the
+        // behavior is different depending on whether you're in a blank page
+        // or an actual document. In a blank page, something like:
+        //    x.innerHTML = '<script></script>';
+        // does _not_ create an empty script, but in a real page, it does. Weird.
+        const html = try parser.nodeFirstChild(fragment_node) orelse return;
+        const head = try parser.nodeFirstChild(html) orelse return;
+        const body = try parser.nodeNextSibling(head) orelse return;
+
+        if (try parser.elementTag(self) == .template) {
+            // HTMLElementTemplate is special. We don't append these as children
+            // of the template, but instead set its content as the body of the
+            // fragment. Simpler to do this by copying the body children into
+            // a new fragment
+            const clean = try parser.documentCreateDocumentFragment(doc);
+            const children = try parser.nodeGetChildNodes(body);
+            const ln = try parser.nodeListLength(children);
+            for (0..ln) |_| {
+                // always index 0, because nodeAppendChild moves the node out of
+                // the nodeList and into the new tree
+                const child = try parser.nodeListItem(children, 0) orelse continue;
+                _ = try parser.nodeAppendChild(@ptrCast(@alignCast(clean)), child);
+            }
+
+            const state = try page.getOrCreateNodeState(node);
+            state.template_content = clean;
+            return;
+        }
+
+        // For any node other than a template, we copy the head and body elements
+        // as child nodes of the element
+        {
+            // First, copy some of the head element
+            const children = try parser.nodeGetChildNodes(head);
+            const ln = try parser.nodeListLength(children);
+            for (0..ln) |_| {
+                // always index 0, because nodeAppendChild moves the node out of
+                // the nodeList and into the new tree
+                const child = try parser.nodeListItem(children, 0) orelse continue;
+                _ = try parser.nodeAppendChild(node, child);
+            }
+        }
+
+        {
+            const children = try parser.nodeGetChildNodes(body);
+            const ln = try parser.nodeListLength(children);
+            for (0..ln) |_| {
+                // always index 0, because nodeAppendChild moves the node out of
+                // the nodeList and into the new tree
+                const child = try parser.nodeListItem(children, 0) orelse continue;
+                _ = try parser.nodeAppendChild(node, child);
+            }
+        }
+    }
+
+    // The closest() method of the Element interface traverses the element and its parents (heading toward the document root) until it finds a node that matches the specified CSS selector.
+    // Returns the closest ancestor Element or itself, which matches the selectors. If there are no such element, null.
+    pub fn _closest(self: *parser.Element, selector: []const u8, page: *Page) !?*parser.Element {
+        const cssParse = @import("../css/css.zig").parse;
+        const CssNodeWrap = @import("../css/libdom.zig").Node;
+        const select = try cssParse(page.call_arena, selector, .{});
+
+        var current: CssNodeWrap = .{ .node = parser.elementToNode(self) };
+        while (true) {
+            if (try select.match(current)) {
+                if (!current.isElement()) {
+                    log.err(.browser, "closest invalid type", .{ .type = try current.tag() });
+                    return null;
+                }
+                return parser.nodeToElement(current.node);
+            }
+            current = try current.parent() orelse return null;
+        }
+    }
+
+    // don't use parser.nodeHasAttributes(...) because that returns true/false
+    // based on the type, e.g. a node never as attributes, an element always has
+    // attributes. But, Element.hasAttributes is supposed to return true only
+    // if the element has at least 1 attribute.
+    pub fn _hasAttributes(self: *parser.Element) !bool {
+        // an element _must_ have at least an empty attribute
+        const node_map = try parser.nodeGetAttributes(parser.elementToNode(self)) orelse unreachable;
+        return try parser.namedNodeMapGetLength(node_map) > 0;
+    }
+
+    pub fn _getAttribute(self: *parser.Element, qname: []const u8) !?[]const u8 {
+        return try parser.elementGetAttribute(self, qname);
+    }
+
+    pub fn _getAttributeNS(self: *parser.Element, ns: []const u8, qname: []const u8) !?[]const u8 {
+        return try parser.elementGetAttributeNS(self, ns, qname);
+    }
+
+    pub fn _setAttribute(self: *parser.Element, qname: []const u8, value: []const u8) !void {
+        return try parser.elementSetAttribute(self, qname, value);
+    }
+
+    pub fn _setAttributeNS(self: *parser.Element, ns: []const u8, qname: []const u8, value: []const u8) !void {
+        return try parser.elementSetAttributeNS(self, ns, qname, value);
+    }
+
+    pub fn _removeAttribute(self: *parser.Element, qname: []const u8) !void {
+        return try parser.elementRemoveAttribute(self, qname);
+    }
+
+    pub fn _removeAttributeNS(self: *parser.Element, ns: []const u8, qname: []const u8) !void {
+        return try parser.elementRemoveAttributeNS(self, ns, qname);
+    }
+
+    pub fn _hasAttribute(self: *parser.Element, qname: []const u8) !bool {
+        return try parser.elementHasAttribute(self, qname);
+    }
+
+    pub fn _hasAttributeNS(self: *parser.Element, ns: []const u8, qname: []const u8) !bool {
+        return try parser.elementHasAttributeNS(self, ns, qname);
+    }
+
+    // https://dom.spec.whatwg.org/#dom-element-toggleattribute
+    pub fn _toggleAttribute(self: *parser.Element, qname: []u8, force: ?bool) !bool {
+        _ = std.ascii.lowerString(qname, qname);
+        const exists = try parser.elementHasAttribute(self, qname);
+
+        // If attribute is null, then:
+        if (!exists) {
+            // If force is not given or is true, create an attribute whose
+            // local name is qualifiedName, value is the empty string and node
+            // document is this’s node document, then append this attribute to
+            // this, and then return true.
+            if (force == null or force.?) {
+                try parser.elementSetAttribute(self, qname, "");
+                return true;
+            }
+            if (try parser.validateName(qname) == false) {
+                return parser.DOMError.InvalidCharacter;
+            }
+
+            // Return false.
+            return false;
+        }
+
+        // Otherwise, if force is not given or is false, remove an attribute
+        // given qualifiedName and this, and then return false.
+        if (force == null or !force.?) {
+            try parser.elementRemoveAttribute(self, qname);
+            return false;
+        }
+
+        // Return true.
+        return true;
+    }
+
+    pub fn _getAttributeNames(self: *parser.Element, page: *Page) ![]const []const u8 {
+        const attributes = try parser.nodeGetAttributes(@ptrCast(self)) orelse return &.{};
+        const ln = try parser.namedNodeMapGetLength(attributes);
+
+        const names = try page.call_arena.alloc([]const u8, ln);
+        var at: usize = 0;
+
+        for (0..ln) |i| {
+            const attribute = try parser.namedNodeMapItem(attributes, @intCast(i)) orelse break;
+            names[at] = try parser.attributeGetName(attribute);
+            at += 1;
+        }
+
+        return names[0..at];
+    }
+
+    pub fn _getAttributeNode(self: *parser.Element, name: []const u8) !?*parser.Attribute {
+        return try parser.elementGetAttributeNode(self, name);
+    }
+
+    pub fn _getAttributeNodeNS(self: *parser.Element, ns: []const u8, name: []const u8) !?*parser.Attribute {
+        return try parser.elementGetAttributeNodeNS(self, ns, name);
+    }
+
+    pub fn _setAttributeNode(self: *parser.Element, attr: *parser.Attribute) !?*parser.Attribute {
+        return try parser.elementSetAttributeNode(self, attr);
+    }
+
+    pub fn _setAttributeNodeNS(self: *parser.Element, attr: *parser.Attribute) !?*parser.Attribute {
+        return try parser.elementSetAttributeNodeNS(self, attr);
+    }
+
+    pub fn _removeAttributeNode(self: *parser.Element, attr: *parser.Attribute) !*parser.Attribute {
+        return try parser.elementRemoveAttributeNode(self, attr);
+    }
+
+    pub fn _getElementsByTagName(
+        self: *parser.Element,
+        tag_name: []const u8,
+        page: *Page,
+    ) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByTagName(
+            page.arena,
+            parser.elementToNode(self),
+            tag_name,
+            .{ .include_root = false },
+        );
+    }
+
+    pub fn _getElementsByClassName(
+        self: *parser.Element,
+        classNames: []const u8,
+        page: *Page,
+    ) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByClassName(
+            page.arena,
+            parser.elementToNode(self),
+            classNames,
+            .{ .include_root = false },
+        );
+    }
+
+    // ParentNode
+    // https://dom.spec.whatwg.org/#parentnode
+    pub fn get_children(self: *parser.Element) !collection.HTMLCollection {
+        return collection.HTMLCollectionChildren(parser.elementToNode(self), .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_firstElementChild(self: *parser.Element) !?Union {
+        var children = try get_children(self);
+        return try children._item(0);
+    }
+
+    pub fn get_lastElementChild(self: *parser.Element) !?Union {
+        // TODO we could check the last child node first, if it's an element,
+        // we can return it directly instead of looping twice over the
+        // children.
+        var children = try get_children(self);
+        const ln = try children.get_length();
+        if (ln == 0) return null;
+        return try children._item(ln - 1);
+    }
+
+    pub fn get_childElementCount(self: *parser.Element) !u32 {
+        var children = try get_children(self);
+        return try children.get_length();
+    }
+
+    // NonDocumentTypeChildNode
+    // https://dom.spec.whatwg.org/#interface-nondocumenttypechildnode
+    pub fn get_previousElementSibling(self: *parser.Element) !?Union {
+        const res = try parser.nodePreviousElementSibling(parser.elementToNode(self));
+        if (res == null) return null;
+        return try toInterface(res.?);
+    }
+
+    pub fn get_nextElementSibling(self: *parser.Element) !?Union {
+        const res = try parser.nodeNextElementSibling(parser.elementToNode(self));
+        if (res == null) return null;
+        return try toInterface(res.?);
+    }
+
+    fn getElementById(self: *parser.Element, id: []const u8) !?*parser.Node {
+        // walk over the node tree fo find the node by id.
+        const root = parser.elementToNode(self);
+        const walker = Walker{};
+        var next: ?*parser.Node = null;
+        while (true) {
+            next = try walker.get_next(root, next) orelse return null;
+            // ignore non-element nodes.
+            if (try parser.nodeType(next.?) != .element) {
+                continue;
+            }
+            const e = parser.nodeToElement(next.?);
+            if (std.mem.eql(u8, id, try get_id(e))) return next;
+        }
+    }
+
+    pub fn _querySelector(self: *parser.Element, selector: []const u8, page: *Page) !?Union {
+        if (selector.len == 0) return null;
+
+        const n = try css.querySelector(page.call_arena, parser.elementToNode(self), selector);
+
+        if (n == null) return null;
+
+        return try toInterface(parser.nodeToElement(n.?));
+    }
+
+    pub fn _querySelectorAll(self: *parser.Element, selector: []const u8, page: *Page) !NodeList {
+        return css.querySelectorAll(page.arena, parser.elementToNode(self), selector);
+    }
+
+    pub fn _prepend(self: *parser.Element, nodes: []const Node.NodeOrText) !void {
+        return Node.prepend(parser.elementToNode(self), nodes);
+    }
+
+    pub fn _append(self: *parser.Element, nodes: []const Node.NodeOrText) !void {
+        return Node.append(parser.elementToNode(self), nodes);
+    }
+
+    pub fn _before(self: *parser.Element, nodes: []const Node.NodeOrText) !void {
+        const ref_node = parser.elementToNode(self);
+        return Node.before(ref_node, nodes);
+    }
+
+    pub fn _after(self: *parser.Element, nodes: []const Node.NodeOrText) !void {
+        const ref_node = parser.elementToNode(self);
+        return Node.after(ref_node, nodes);
+    }
+
+    pub fn _replaceChildren(self: *parser.Element, nodes: []const Node.NodeOrText) !void {
+        return Node.replaceChildren(parser.elementToNode(self), nodes);
+    }
+
+    // A DOMRect object providing information about the size of an element and its position relative to the viewport.
+    // Returns a 0 DOMRect object if the element is eventually detached from the main window
+    pub fn _getBoundingClientRect(self: *parser.Element, page: *Page) !DOMRect {
+        // Since we are lazy rendering we need to do this check. We could store the renderer in a viewport such that it could cache these, but it would require tracking changes.
+        if (!try page.isNodeAttached(parser.elementToNode(self))) {
+            return DOMRect{
+                .x = 0,
+                .y = 0,
+                .width = 0,
+                .height = 0,
+                .bottom = 0,
+                .right = 0,
+                .top = 0,
+                .left = 0,
+            };
+        }
+        return page.renderer.getRect(self);
+    }
+
+    // Returns a collection of DOMRect objects that indicate the bounding rectangles for each CSS border box in a client.
+    // We do not render so it only always return the element's bounding rect.
+    // Returns an empty array if the element is eventually detached from the main window
+    pub fn _getClientRects(self: *parser.Element, page: *Page) ![]DOMRect {
+        if (!try page.isNodeAttached(parser.elementToNode(self))) {
+            return &.{};
+        }
+        const heap_ptr = try page.call_arena.create(DOMRect);
+        heap_ptr.* = try page.renderer.getRect(self);
+        return heap_ptr[0..1];
+    }
+
+    pub fn get_clientWidth(_: *parser.Element, page: *Page) u32 {
+        return page.renderer.width();
+    }
+
+    pub fn get_clientHeight(_: *parser.Element, page: *Page) u32 {
+        return page.renderer.height();
+    }
+
+    pub fn _matches(self: *parser.Element, selectors: []const u8, page: *Page) !bool {
+        const cssParse = @import("../css/css.zig").parse;
+        const CssNodeWrap = @import("../css/libdom.zig").Node;
+        const s = try cssParse(page.call_arena, selectors, .{});
+        return s.match(CssNodeWrap{ .node = parser.elementToNode(self) });
+    }
+
+    pub fn _scrollIntoViewIfNeeded(_: *parser.Element, center_if_needed: ?bool) void {
+        _ = center_if_needed;
+    }
+
+    const CheckVisibilityOpts = struct {
+        contentVisibilityAuto: bool,
+        opacityProperty: bool,
+        visibilityProperty: bool,
+    };
+
+    pub fn _checkVisibility(self: *parser.Element, opts: ?CheckVisibilityOpts) bool {
+        _ = self;
+        _ = opts;
+        return true;
+    }
+
+    const AttachShadowOpts = struct {
+        mode: []const u8, // must be specified
+    };
+    pub fn _attachShadow(self: *parser.Element, opts: AttachShadowOpts, page: *Page) !*ShadowRoot {
+        const mode = std.meta.stringToEnum(ShadowRoot.Mode, opts.mode) orelse return error.InvalidArgument;
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        if (state.shadow_root) |sr| {
+            if (mode != sr.mode) {
+                // this is the behavior per the spec
+                return error.NotSupportedError;
+            }
+
+            try Node.removeChildren(@ptrCast(@alignCast(sr.proto)));
+            return sr;
+        }
+
+        // Not sure what to do if there is no owner document
+        const doc = try parser.nodeOwnerDocument(@ptrCast(self)) orelse return error.InvalidArgument;
+        const fragment = try parser.documentCreateDocumentFragment(doc);
+        const sr = try page.arena.create(ShadowRoot);
+        sr.* = .{
+            .host = self,
+            .mode = mode,
+            .proto = fragment,
+        };
+        state.shadow_root = sr;
+        parser.documentFragmentSetHost(sr.proto, @ptrCast(@alignCast(self)));
+
+        // Storing the ShadowRoot on the element makes sense, it's the ShadowRoot's
+        // parent. When we render, we go top-down, so we'll have the element, get
+        // its shadowroot, and go on. that's what the above code does.
+        // But we sometimes need to go bottom-up, e.g when we have a slot element
+        // and want to find the containing parent. Unforatunately , we don't have
+        // that link, so we need to create it. In the DOM, the ShadowRoot is
+        // represented by this DocumentFragment (it's the ShadowRoot's base prototype)
+        // So we can also store the ShadowRoot in the DocumentFragment's state.
+        const fragment_state = try page.getOrCreateNodeState(@ptrCast(@alignCast(fragment)));
+        fragment_state.shadow_root = sr;
+
+        return sr;
+    }
+
+    pub fn get_shadowRoot(self: *parser.Element, page: *Page) ?*ShadowRoot {
+        const state = page.getNodeState(@ptrCast(@alignCast(self))) orelse return null;
+        const sr = state.shadow_root orelse return null;
+        if (sr.mode == .closed) {
+            return null;
+        }
+        return sr;
+    }
+
+    pub fn _animate(self: *parser.Element, effect: JsObject, opts: JsObject) !Animation {
+        _ = self;
+        _ = opts;
+        return Animation.constructor(effect, null);
+    }
+
+    pub fn _remove(self: *parser.Element) !void {
+        // TODO: This hasn't been tested to make sure all references to this
+        // node are properly updated. A lot of libdom is lazy and will look
+        // for related elements JIT by walking the tree, but there could be
+        // cases in libdom or the Zig WebAPI where this reference is kept
+        const as_node: *parser.Node = @ptrCast(self);
+        const parent = try parser.nodeParentNode(as_node) orelse return;
+        _ = try Node._removeChild(parent, as_node);
+    }
+};
+
+// Tests
+// -----
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Element" {
+    try testing.htmlRunner("dom/element.html");
+}

src/browser/dom/event_target.zig

@@ -0,0 +1,166 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const Env = @import("../env.zig").Env;
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const EventHandler = @import("../events/event.zig").EventHandler;
+
+const DOMException = @import("exceptions.zig").DOMException;
+const nod = @import("node.zig");
+
+pub const Union = union(enum) {
+    node: nod.Union,
+    xhr: *@import("../xhr/xhr.zig").XMLHttpRequest,
+    plain: *parser.EventTarget,
+    message_port: *@import("MessageChannel.zig").MessagePort,
+    screen: *@import("../html/screen.zig").Screen,
+    screen_orientation: *@import("../html/screen.zig").ScreenOrientation,
+    performance: *@import("performance.zig").Performance,
+    media_query_list: *@import("../html/media_query_list.zig").MediaQueryList,
+};
+
+// EventTarget implementation
+pub const EventTarget = struct {
+    pub const Self = parser.EventTarget;
+    pub const Exception = DOMException;
+
+    // Extend libdom event target for pure zig struct.
+    base: parser.EventTargetTBase = parser.EventTargetTBase{ .internal_target_type = .plain },
+
+    pub fn toInterface(et: *parser.EventTarget, page: *Page) !Union {
+        // libdom assumes that all event targets are libdom nodes. They are not.
+
+        switch (try parser.eventTargetInternalType(et)) {
+            .libdom_node => {
+                return .{ .node = try nod.Node.toInterface(@as(*parser.Node, @ptrCast(et))) };
+            },
+            .plain => return .{ .plain = et },
+            .abort_signal => {
+                // AbortSignal is a special case, it has its own internal type.
+                // We return it as a node, but we need to handle it differently.
+                return .{ .node = .{ .AbortSignal = @fieldParentPtr("proto", @as(*parser.EventTargetTBase, @ptrCast(et))) } };
+            },
+            .window => {
+                // The window is a common non-node target, but it's easy to handle as its a singleton.
+                std.debug.assert(@intFromPtr(et) == @intFromPtr(&page.window.base));
+                return .{ .node = .{ .Window = &page.window } };
+            },
+            .xhr => {
+                const XMLHttpRequestEventTarget = @import("../xhr/event_target.zig").XMLHttpRequestEventTarget;
+                const base: *XMLHttpRequestEventTarget = @fieldParentPtr("base", @as(*parser.EventTargetTBase, @ptrCast(et)));
+                return .{ .xhr = @fieldParentPtr("proto", base) };
+            },
+            .message_port => {
+                return .{ .message_port = @fieldParentPtr("proto", @as(*parser.EventTargetTBase, @ptrCast(et))) };
+            },
+            .screen => {
+                return .{ .screen = @fieldParentPtr("proto", @as(*parser.EventTargetTBase, @ptrCast(et))) };
+            },
+            .screen_orientation => {
+                return .{ .screen_orientation = @fieldParentPtr("proto", @as(*parser.EventTargetTBase, @ptrCast(et))) };
+            },
+            .performance => {
+                return .{ .performance = @fieldParentPtr("base", @as(*parser.EventTargetTBase, @ptrCast(et))) };
+            },
+            .media_query_list => {
+                return .{ .media_query_list = @fieldParentPtr("base", @as(*parser.EventTargetTBase, @ptrCast(et))) };
+            },
+        }
+    }
+
+    // JS funcs
+    // --------
+    pub fn constructor(page: *Page) !*parser.EventTarget {
+        const et = try page.arena.create(EventTarget);
+        return @ptrCast(&et.base);
+    }
+
+    pub fn _addEventListener(
+        self: *parser.EventTarget,
+        typ: []const u8,
+        listener: EventHandler.Listener,
+        opts: ?EventHandler.Opts,
+        page: *Page,
+    ) !void {
+        _ = try EventHandler.register(page.arena, self, typ, listener, opts);
+    }
+
+    const RemoveEventListenerOpts = union(enum) {
+        opts: Opts,
+        capture: bool,
+
+        const Opts = struct {
+            capture: ?bool,
+        };
+    };
+
+    pub fn _removeEventListener(
+        self: *parser.EventTarget,
+        typ: []const u8,
+        listener: EventHandler.Listener,
+        opts_: ?RemoveEventListenerOpts,
+    ) !void {
+        var capture = false;
+        if (opts_) |opts| {
+            capture = switch (opts) {
+                .capture => |c| c,
+                .opts => |o| o.capture orelse false,
+            };
+        }
+
+        const cbk = (try listener.callback(self)) orelse return;
+
+        // check if event target has already this listener
+        const lst = try parser.eventTargetHasListener(
+            self,
+            typ,
+            capture,
+            cbk.id,
+        );
+        if (lst == null) {
+            return;
+        }
+
+        // remove listener
+        try parser.eventTargetRemoveEventListener(
+            self,
+            typ,
+            lst.?,
+            capture,
+        );
+    }
+
+    pub fn _dispatchEvent(self: *parser.EventTarget, event: *parser.Event, page: *Page) !bool {
+        const res = try parser.eventTargetDispatchEvent(self, event);
+
+        if (!parser.eventBubbles(event) or parser.eventIsStopped(event)) {
+            return res;
+        }
+
+        try page.window.dispatchForDocumentTarget(event);
+        return true;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.EventTarget" {
+    try testing.htmlRunner("dom/event_target.html");
+}

src/browser/dom/exceptions.zig

@@ -0,0 +1,224 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const allocPrint = std.fmt.allocPrint;
+
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+// https://webidl.spec.whatwg.org/#idl-DOMException
+pub const DOMException = struct {
+    err: ?parser.DOMError,
+    str: []const u8,
+
+    pub const ErrorSet = parser.DOMError;
+
+    // static attributes
+    pub const _INDEX_SIZE_ERR = 1;
+    pub const _DOMSTRING_SIZE_ERR = 2;
+    pub const _HIERARCHY_REQUEST_ERR = 3;
+    pub const _WRONG_DOCUMENT_ERR = 4;
+    pub const _INVALID_CHARACTER_ERR = 5;
+    pub const _NO_DATA_ALLOWED_ERR = 6;
+    pub const _NO_MODIFICATION_ALLOWED_ERR = 7;
+    pub const _NOT_FOUND_ERR = 8;
+    pub const _NOT_SUPPORTED_ERR = 9;
+    pub const _INUSE_ATTRIBUTE_ERR = 10;
+    pub const _INVALID_STATE_ERR = 11;
+    pub const _SYNTAX_ERR = 12;
+    pub const _INVALID_MODIFICATION_ERR = 13;
+    pub const _NAMESPACE_ERR = 14;
+    pub const _INVALID_ACCESS_ERR = 15;
+    pub const _VALIDATION_ERR = 16;
+    pub const _TYPE_MISMATCH_ERR = 17;
+    pub const _SECURITY_ERR = 18;
+    pub const _NETWORK_ERR = 19;
+    pub const _ABORT_ERR = 20;
+    pub const _URL_MISMATCH_ERR = 21;
+    pub const _QUOTA_EXCEEDED_ERR = 22;
+    pub const _TIMEOUT_ERR = 23;
+    pub const _INVALID_NODE_TYPE_ERR = 24;
+    pub const _DATA_CLONE_ERR = 25;
+
+    pub fn constructor(message_: ?[]const u8, name_: ?[]const u8, page: *const Page) !DOMException {
+        const message = message_ orelse "";
+        const err = if (name_) |n| error_from_str(n) else null;
+        const fixed_name = name(err);
+
+        if (message.len == 0) return .{ .err = err, .str = fixed_name };
+
+        const str = try allocPrint(page.arena, "{s}: {s}", .{ fixed_name, message });
+        return .{ .err = err, .str = str };
+    }
+
+    // TODO: deinit
+    pub fn init(alloc: std.mem.Allocator, err: anyerror, caller_name: []const u8) !DOMException {
+        const dom_error = @as(parser.DOMError, @errorCast(err));
+        const error_name = DOMException.name(dom_error);
+        const str = switch (dom_error) {
+            error.HierarchyRequest => try allocPrint(
+                alloc,
+                "{s}: Failed to execute '{s}' on 'Node': The new child element contains the parent.",
+                .{ error_name, caller_name },
+            ),
+            // todo add more custom error messages
+            else => try allocPrint(
+                alloc,
+                "{s}: Failed to execute '{s}' : {s}",
+                .{ error_name, caller_name, error_name },
+            ),
+            error.NoError => unreachable,
+        };
+        return .{ .err = dom_error, .str = str };
+    }
+
+    fn error_from_str(name_: []const u8) ?parser.DOMError {
+        // @speed: Consider length first, left as is for maintainability, awaiting switch on string support
+        if (std.mem.eql(u8, name_, "IndexSizeError")) return error.IndexSize;
+        if (std.mem.eql(u8, name_, "StringSizeError")) return error.StringSize;
+        if (std.mem.eql(u8, name_, "HierarchyRequestError")) return error.HierarchyRequest;
+        if (std.mem.eql(u8, name_, "WrongDocumentError")) return error.WrongDocument;
+        if (std.mem.eql(u8, name_, "InvalidCharacterError")) return error.InvalidCharacter;
+        if (std.mem.eql(u8, name_, "NoDataAllowedError")) return error.NoDataAllowed;
+        if (std.mem.eql(u8, name_, "NoModificationAllowedError")) return error.NoModificationAllowed;
+        if (std.mem.eql(u8, name_, "NotFoundError")) return error.NotFound;
+        if (std.mem.eql(u8, name_, "NotSupportedError")) return error.NotSupported;
+        if (std.mem.eql(u8, name_, "InuseAttributeError")) return error.InuseAttribute;
+        if (std.mem.eql(u8, name_, "InvalidStateError")) return error.InvalidState;
+        if (std.mem.eql(u8, name_, "SyntaxError")) return error.Syntax;
+        if (std.mem.eql(u8, name_, "InvalidModificationError")) return error.InvalidModification;
+        if (std.mem.eql(u8, name_, "NamespaceError")) return error.Namespace;
+        if (std.mem.eql(u8, name_, "InvalidAccessError")) return error.InvalidAccess;
+        if (std.mem.eql(u8, name_, "ValidationError")) return error.Validation;
+        if (std.mem.eql(u8, name_, "TypeMismatchError")) return error.TypeMismatch;
+        if (std.mem.eql(u8, name_, "SecurityError")) return error.Security;
+        if (std.mem.eql(u8, name_, "NetworkError")) return error.Network;
+        if (std.mem.eql(u8, name_, "AbortError")) return error.Abort;
+        if (std.mem.eql(u8, name_, "URLismatchError")) return error.URLismatch;
+        if (std.mem.eql(u8, name_, "QuotaExceededError")) return error.QuotaExceeded;
+        if (std.mem.eql(u8, name_, "TimeoutError")) return error.Timeout;
+        if (std.mem.eql(u8, name_, "InvalidNodeTypeError")) return error.InvalidNodeType;
+        if (std.mem.eql(u8, name_, "DataCloneError")) return error.DataClone;
+
+        // custom netsurf error
+        if (std.mem.eql(u8, name_, "UnspecifiedEventTypeError")) return error.UnspecifiedEventType;
+        if (std.mem.eql(u8, name_, "DispatchRequestError")) return error.DispatchRequest;
+        if (std.mem.eql(u8, name_, "NoMemoryError")) return error.NoMemory;
+        if (std.mem.eql(u8, name_, "AttributeWrongTypeError")) return error.AttributeWrongType;
+        return null;
+    }
+
+    fn name(err_: ?parser.DOMError) []const u8 {
+        const err = err_ orelse return "Error";
+
+        return switch (err) {
+            error.IndexSize => "IndexSizeError",
+            error.StringSize => "StringSizeError", // Legacy: DOMSTRING_SIZE_ERR
+            error.HierarchyRequest => "HierarchyRequestError",
+            error.WrongDocument => "WrongDocumentError",
+            error.InvalidCharacter => "InvalidCharacterError",
+            error.NoDataAllowed => "NoDataAllowedError", // Legacy: NO_DATA_ALLOWED_ERR
+            error.NoModificationAllowed => "NoModificationAllowedError",
+            error.NotFound => "NotFoundError",
+            error.NotSupported => "NotSupportedError",
+            error.InuseAttribute => "InuseAttributeError",
+            error.InvalidState => "InvalidStateError",
+            error.Syntax => "SyntaxError",
+            error.InvalidModification => "InvalidModificationError",
+            error.Namespace => "NamespaceError",
+            error.InvalidAccess => "InvalidAccessError",
+            error.Validation => "ValidationError", // Legacy: VALIDATION_ERR
+            error.TypeMismatch => "TypeMismatchError",
+            error.Security => "SecurityError",
+            error.Network => "NetworkError",
+            error.Abort => "AbortError",
+            error.URLismatch => "URLismatchError",
+            error.QuotaExceeded => "QuotaExceededError",
+            error.Timeout => "TimeoutError",
+            error.InvalidNodeType => "InvalidNodeTypeError",
+            error.DataClone => "DataCloneError",
+            error.NoError => unreachable,
+
+            // custom netsurf error
+            error.UnspecifiedEventType => "UnspecifiedEventTypeError",
+            error.DispatchRequest => "DispatchRequestError",
+            error.NoMemory => "NoMemoryError",
+            error.AttributeWrongType => "AttributeWrongTypeError",
+        };
+    }
+
+    // JS properties and methods
+
+    pub fn get_code(self: *const DOMException) u8 {
+        const err = self.err orelse return 0;
+        return switch (err) {
+            error.IndexSize => 1,
+            error.StringSize => 2,
+            error.HierarchyRequest => 3,
+            error.WrongDocument => 4,
+            error.InvalidCharacter => 5,
+            error.NoDataAllowed => 6,
+            error.NoModificationAllowed => 7,
+            error.NotFound => 8,
+            error.NotSupported => 9,
+            error.InuseAttribute => 10,
+            error.InvalidState => 11,
+            error.Syntax => 12,
+            error.InvalidModification => 13,
+            error.Namespace => 14,
+            error.InvalidAccess => 15,
+            error.Validation => 16,
+            error.TypeMismatch => 17,
+            error.Security => 18,
+            error.Network => 19,
+            error.Abort => 20,
+            error.URLismatch => 21,
+            error.QuotaExceeded => 22,
+            error.Timeout => 23,
+            error.InvalidNodeType => 24,
+            error.DataClone => 25,
+            error.NoError => unreachable,
+
+            // custom netsurf error
+            error.UnspecifiedEventType => 128,
+            error.DispatchRequest => 129,
+            error.NoMemory => 130,
+            error.AttributeWrongType => 131,
+        };
+    }
+
+    pub fn get_name(self: *const DOMException) []const u8 {
+        return DOMException.name(self.err);
+    }
+
+    pub fn get_message(self: *const DOMException) []const u8 {
+        const errName = DOMException.name(self.err);
+        if (self.str.len <= errName.len + 2) return "";
+        return self.str[errName.len + 2 ..]; // ! Requires str is formatted as "{name}: {message}"
+    }
+
+    pub fn _toString(self: *const DOMException) []const u8 {
+        return self.str;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Exceptions" {
+    try testing.htmlRunner("dom/exceptions.html");
+}

src/browser/dom/html_collection.zig

@@ -1,20 +1,30 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
 const std = @import("std");
+const Allocator = std.mem.Allocator;
 
 const parser = @import("../netsurf.zig");
 
-const jsruntime = @import("jsruntime");
-const Case = jsruntime.test_utils.Case;
-const checkCases = jsruntime.test_utils.checkCases;
-const generate = @import("../generate.zig");
-
-const utils = @import("utils.z");
 const Element = @import("element.zig").Element;
 const Union = @import("element.zig").Union;
-
+const JsThis = @import("../env.zig").JsThis;
 const Walker = @import("walker.zig").Walker;
-const WalkerDepthFirst = @import("walker.zig").WalkerDepthFirst;
-const WalkerChildren = @import("walker.zig").WalkerChildren;
-const WalkerNone = @import("walker.zig").WalkerNone;
 
 const Matcher = union(enum) {
     matchByName: MatchByName,
@@ -27,25 +37,11 @@ const Matcher = union(enum) {
 
     pub fn match(self: Matcher, node: *parser.Node) !bool {
         switch (self) {
-            inline .matchTrue => return true,
-            inline .matchFalse => return false,
-            inline .matchByTagName => |case| return case.match(node),
-            inline .matchByClassName => |case| return case.match(node),
-            inline .matchByName => |case| return case.match(node),
-            inline .matchByLinks => return MatchByLinks.match(node),
-            inline .matchByAnchors => return MatchByAnchors.match(node),
-        }
-    }
-
-    pub fn deinit(self: Matcher, alloc: std.mem.Allocator) void {
-        switch (self) {
-            inline .matchTrue => return,
-            inline .matchFalse => return,
-            inline .matchByTagName => |case| return case.deinit(alloc),
-            inline .matchByClassName => |case| return case.deinit(alloc),
-            inline .matchByName => |case| return case.deinit(alloc),
-            inline .matchByLinks => return,
-            inline .matchByAnchors => return,
+            .matchTrue => return true,
+            .matchFalse => return false,
+            .matchByLinks => return MatchByLinks.match(node),
+            .matchByAnchors => return MatchByAnchors.match(node),
+            inline else => |m| return m.match(node),
         }
     }
 };
@@ -56,54 +52,50 @@ pub const MatchByTagName = struct {
     tag: []const u8,
     is_wildcard: bool,
 
-    fn init(alloc: std.mem.Allocator, tag_name: []const u8) !MatchByTagName {
-        const tag_name_alloc = try alloc.alloc(u8, tag_name.len);
-        @memcpy(tag_name_alloc, tag_name);
-        return MatchByTagName{
-            .tag = tag_name_alloc,
-            .is_wildcard = std.mem.eql(u8, tag_name, "*"),
+    fn init(arena: Allocator, tag_name: []const u8) !MatchByTagName {
+        if (std.mem.eql(u8, tag_name, "*")) {
+            return .{ .tag = "*", .is_wildcard = true };
+        }
+
+        return .{
+            .tag = try arena.dupe(u8, tag_name),
+            .is_wildcard = false,
         };
     }
 
     pub fn match(self: MatchByTagName, node: *parser.Node) !bool {
         return self.is_wildcard or std.ascii.eqlIgnoreCase(self.tag, try parser.nodeName(node));
     }
-
-    fn deinit(self: MatchByTagName, alloc: std.mem.Allocator) void {
-        alloc.free(self.tag);
-    }
 };
 
 pub fn HTMLCollectionByTagName(
-    alloc: std.mem.Allocator,
+    arena: Allocator,
     root: ?*parser.Node,
     tag_name: []const u8,
-    include_root: bool,
+    opts: Opts,
 ) !HTMLCollection {
     return HTMLCollection{
         .root = root,
-        .walker = Walker{ .walkerDepthFirst = .{} },
-        .matcher = Matcher{
-            .matchByTagName = try MatchByTagName.init(alloc, tag_name),
-        },
-        .include_root = include_root,
+        .walker = .{ .walkerDepthFirst = .{} },
+        .matcher = .{ .matchByTagName = try MatchByTagName.init(arena, tag_name) },
+        .mutable = opts.mutable,
+        .include_root = opts.include_root,
     };
 }
 
 pub const MatchByClassName = struct {
-    classNames: []const u8,
+    class_names: []const u8,
 
-    fn init(alloc: std.mem.Allocator, classNames: []const u8) !MatchByClassName {
-        const class_names_alloc = try alloc.alloc(u8, classNames.len);
-        @memcpy(class_names_alloc, classNames);
-        return MatchByClassName{
-            .classNames = class_names_alloc,
+    fn init(arena: Allocator, class_names: []const u8) !MatchByClassName {
+        return .{
+            .class_names = try arena.dupe(u8, class_names),
         };
     }
 
     pub fn match(self: MatchByClassName, node: *parser.Node) !bool {
-        var it = std.mem.splitAny(u8, self.classNames, " ");
         const e = parser.nodeToElement(node);
+
+        var it = std.mem.splitScalar(u8, self.class_names, ' ');
         while (it.next()) |c| {
             if (!try parser.elementHasClass(e, c)) {
                 return false;
@@ -112,36 +104,29 @@ pub const MatchByClassName = struct {
 
         return true;
     }
-
-    fn deinit(self: MatchByClassName, alloc: std.mem.Allocator) void {
-        alloc.free(self.classNames);
-    }
 };
 
 pub fn HTMLCollectionByClassName(
-    alloc: std.mem.Allocator,
+    arena: Allocator,
     root: ?*parser.Node,
     classNames: []const u8,
-    include_root: bool,
+    opts: Opts,
 ) !HTMLCollection {
     return HTMLCollection{
         .root = root,
-        .walker = Walker{ .walkerDepthFirst = .{} },
-        .matcher = Matcher{
-            .matchByClassName = try MatchByClassName.init(alloc, classNames),
-        },
-        .include_root = include_root,
+        .walker = .{ .walkerDepthFirst = .{} },
+        .matcher = .{ .matchByClassName = try MatchByClassName.init(arena, classNames) },
+        .mutable = opts.mutable,
+        .include_root = opts.include_root,
     };
 }
 
 pub const MatchByName = struct {
     name: []const u8,
 
-    fn init(alloc: std.mem.Allocator, name: []const u8) !MatchByName {
-        const names_alloc = try alloc.alloc(u8, name.len);
-        @memcpy(names_alloc, name);
-        return MatchByName{
-            .name = names_alloc,
+    fn init(arena: Allocator, name: []const u8) !MatchByName {
+        return .{
+            .name = try arena.dupe(u8, name),
         };
     }
 
@@ -150,57 +135,79 @@ pub const MatchByName = struct {
         const nname = try parser.elementGetAttribute(e, "name") orelse return false;
         return std.mem.eql(u8, self.name, nname);
     }
-
-    fn deinit(self: MatchByName, alloc: std.mem.Allocator) void {
-        alloc.free(self.name);
-    }
 };
 
 pub fn HTMLCollectionByName(
-    alloc: std.mem.Allocator,
+    arena: Allocator,
     root: ?*parser.Node,
     name: []const u8,
-    include_root: bool,
+    opts: Opts,
 ) !HTMLCollection {
     return HTMLCollection{
         .root = root,
-        .walker = Walker{ .walkerDepthFirst = .{} },
-        .matcher = Matcher{
-            .matchByName = try MatchByName.init(alloc, name),
-        },
-        .include_root = include_root,
+        .walker = .{ .walkerDepthFirst = .{} },
+        .matcher = .{ .matchByName = try MatchByName.init(arena, name) },
+        .mutable = opts.mutable,
+        .include_root = opts.include_root,
     };
 }
 
-pub fn HTMLCollectionAll(
-    root: ?*parser.Node,
-    include_root: bool,
-) !HTMLCollection {
-    return HTMLCollection{
-        .root = root,
-        .walker = Walker{ .walkerDepthFirst = .{} },
-        .matcher = Matcher{ .matchTrue = .{} },
-        .include_root = include_root,
+// HTMLAllCollection is a special type: instances of it are falsy. It's the only
+// object in the WebAPI that behaves like this - in fact, it's even a special
+// case in the JavaScript spec.
+// This is important, because a lot of browser detection rely on this behavior
+// to determine what browser is running.
+
+// It's also possible to use an instance like a function:
+//   document.all(3)
+//   document.all('some_id')
+pub const HTMLAllCollection = struct {
+    pub const prototype = *HTMLCollection;
+
+    proto: HTMLCollection,
+
+    pub const mark_as_undetectable = true;
+
+    pub fn init(root: ?*parser.Node) HTMLAllCollection {
+        return .{ .proto = .{
+            .root = root,
+            .walker = .{ .walkerDepthFirst = .{} },
+            .matcher = .{ .matchTrue = .{} },
+            .include_root = true,
+        } };
+    }
+
+    const CAllAsFunctionArg = union(enum) {
+        index: u32,
+        id: []const u8,
     };
-}
+
+    pub fn jsCallAsFunction(self: *HTMLAllCollection, arg: CAllAsFunctionArg) !?Union {
+        return switch (arg) {
+            .index => |i| self.proto._item(i),
+            .id => |id| self.proto._namedItem(id),
+        };
+    }
+};
 
 pub fn HTMLCollectionChildren(
     root: ?*parser.Node,
-    include_root: bool,
-) !HTMLCollection {
+    opts: Opts,
+) HTMLCollection {
     return HTMLCollection{
         .root = root,
-        .walker = Walker{ .walkerChildren = .{} },
-        .matcher = Matcher{ .matchTrue = .{} },
-        .include_root = include_root,
+        .walker = .{ .walkerChildren = .{} },
+        .matcher = .{ .matchTrue = .{} },
+        .mutable = opts.mutable,
+        .include_root = opts.include_root,
     };
 }
 
 pub fn HTMLCollectionEmpty() !HTMLCollection {
     return HTMLCollection{
         .root = null,
-        .walker = Walker{ .walkerNone = .{} },
-        .matcher = Matcher{ .matchFalse = .{} },
+        .walker = .{ .walkerNone = .{} },
+        .matcher = .{ .matchFalse = .{} },
         .include_root = false,
     };
 }
@@ -221,15 +228,14 @@ pub const MatchByLinks = struct {
 
 pub fn HTMLCollectionByLinks(
     root: ?*parser.Node,
-    include_root: bool,
+    opts: Opts,
 ) !HTMLCollection {
     return HTMLCollection{
         .root = root,
-        .walker = Walker{ .walkerDepthFirst = .{} },
-        .matcher = Matcher{
-            .matchByLinks = MatchByLinks{},
-        },
-        .include_root = include_root,
+        .walker = .{ .walkerDepthFirst = .{} },
+        .matcher = .{ .matchByLinks = MatchByLinks{} },
+        .mutable = opts.mutable,
+        .include_root = opts.include_root,
     };
 }
 
@@ -248,21 +254,18 @@ pub const MatchByAnchors = struct {
 
 pub fn HTMLCollectionByAnchors(
     root: ?*parser.Node,
-    include_root: bool,
+    opts: Opts,
 ) !HTMLCollection {
     return HTMLCollection{
         .root = root,
-        .walker = Walker{ .walkerDepthFirst = .{} },
-        .matcher = Matcher{
-            .matchByAnchors = MatchByAnchors{},
-        },
-        .include_root = include_root,
+        .walker = .{ .walkerDepthFirst = .{} },
+        .matcher = .{ .matchByAnchors = MatchByAnchors{} },
+        .mutable = opts.mutable,
+        .include_root = opts.include_root,
     };
 }
 
 pub const HTMLCollectionIterator = struct {
-    pub const mem_guarantied = true;
-
     coll: *HTMLCollection,
     index: u32 = 0,
 
@@ -288,13 +291,16 @@ pub const HTMLCollectionIterator = struct {
     }
 };
 
+const Opts = struct {
+    include_root: bool,
+    mutable: bool = false,
+};
+
 // WEB IDL https://dom.spec.whatwg.org/#htmlcollection
 // HTMLCollection is re implemented in zig here because libdom
 // dom_html_collection expects a comparison function callback as arguement.
 // But we wanted a dynamically comparison here, according to the match tagname.
 pub const HTMLCollection = struct {
-    pub const mem_guarantied = true;
-
     matcher: Matcher,
     walker: Walker,
 
@@ -305,16 +311,14 @@ pub const HTMLCollection = struct {
     // itself.
     include_root: bool = false,
 
-    // save a state for the collection to improve the _item speed.
-    cur_idx: ?u32 = undefined,
-    cur_node: ?*parser.Node = undefined,
+    mutable: bool = false,
 
-    // array_like_keys is used to keep reference to array like interface implementation.
-    // the collection generates keys string which must be free on deinit.
-    array_like_keys: std.ArrayListUnmanaged([]u8) = .{},
+    // save a state for the collection to improve the _item speed.
+    cur_idx: ?u32 = null,
+    cur_node: ?*parser.Node = null,
 
     // start returns the first node to walk on.
-    fn start(self: HTMLCollection) !?*parser.Node {
+    fn start(self: *const HTMLCollection) !?*parser.Node {
         if (self.root == null) return null;
 
         if (self.include_root) {
@@ -359,7 +363,7 @@ pub const HTMLCollection = struct {
         var node: *parser.Node = undefined;
 
         // Use the current state to improve speed if possible.
-        if (self.cur_idx != null and index >= self.cur_idx.?) {
+        if (self.mutable == false and self.cur_idx != null and index >= self.cur_idx.?) {
             i = self.cur_idx.?;
             node = self.cur_node.?;
         } else {
@@ -394,7 +398,7 @@ pub const HTMLCollection = struct {
         return try Element.toInterface(e);
     }
 
-    pub fn _namedItem(self: *HTMLCollection, name: []const u8) !?Union {
+    pub fn _namedItem(self: *const HTMLCollection, name: []const u8) !?Union {
         if (self.root == null) return null;
         if (name.len == 0) return null;
 
@@ -436,81 +440,28 @@ pub const HTMLCollection = struct {
         return null;
     }
 
-    pub fn postAttach(self: *HTMLCollection, alloc: std.mem.Allocator, js_obj: jsruntime.JSObject) !void {
-        const ln = try self.get_length();
-        var i: u32 = 0;
-        while (i < ln) {
-            defer i += 1;
-            const k = try std.fmt.allocPrint(alloc, "{d}", .{i});
-            try self.array_like_keys.append(alloc, k);
-
-            const node = try self.item(i) orelse unreachable;
+    pub fn postAttach(self: *HTMLCollection, js_this: JsThis) !void {
+        const len = try self.get_length();
+        for (0..len) |i| {
+            const node = try self.item(@intCast(i)) orelse unreachable;
             const e = @as(*parser.Element, @ptrCast(node));
-            try js_obj.set(k, e);
+            const as_interface = try Element.toInterface(e);
+            try js_this.setIndex(@intCast(i), as_interface, .{});
 
             if (try item_name(e)) |name| {
-                try js_obj.set(name, e);
+                // Even though an entry might have an empty id, the spec says
+                // that namedItem("") should always return null
+                if (name.len > 0) {
+                    // Named fields should not be enumerable (it is defined with
+                    // the LegacyUnenumerableNamedProperties flag.)
+                    try js_this.set(name, as_interface, .{ .DONT_ENUM = true });
+                }
             }
         }
     }
-
-    pub fn deinit(self: *HTMLCollection, alloc: std.mem.Allocator) void {
-        for (self.array_like_keys_) |k| alloc.free(k);
-        self.array_like_keys.deinit(alloc);
-        self.matcher.deinit(alloc);
-    }
 };
 
-// Tests
-// -----
-
-pub fn testExecFn(
-    _: std.mem.Allocator,
-    js_env: *jsruntime.Env,
-) anyerror!void {
-    var getElementsByTagName = [_]Case{
-        .{ .src = "let getElementsByTagName = document.getElementsByTagName('p')", .ex = "undefined" },
-        .{ .src = "getElementsByTagName.length", .ex = "2" },
-        .{ .src = "let getElementsByTagNameCI = document.getElementsByTagName('P')", .ex = "undefined" },
-        .{ .src = "getElementsByTagNameCI.length", .ex = "2" },
-        .{ .src = "getElementsByTagName.item(0).localName", .ex = "p" },
-        .{ .src = "getElementsByTagName.item(1).localName", .ex = "p" },
-        .{ .src = "let getElementsByTagNameAll = document.getElementsByTagName('*')", .ex = "undefined" },
-        .{ .src = "getElementsByTagNameAll.length", .ex = "8" },
-        .{ .src = "getElementsByTagNameAll.item(0).localName", .ex = "html" },
-        .{ .src = "getElementsByTagNameAll.item(0).localName", .ex = "html" },
-        .{ .src = "getElementsByTagNameAll.item(1).localName", .ex = "head" },
-        .{ .src = "getElementsByTagNameAll.item(0).localName", .ex = "html" },
-        .{ .src = "getElementsByTagNameAll.item(2).localName", .ex = "body" },
-        .{ .src = "getElementsByTagNameAll.item(3).localName", .ex = "div" },
-        .{ .src = "getElementsByTagNameAll.item(7).localName", .ex = "p" },
-        .{ .src = "getElementsByTagNameAll.namedItem('para-empty-child').localName", .ex = "span" },
-
-        // array like
-        .{ .src = "getElementsByTagNameAll[0].localName", .ex = "html" },
-        .{ .src = "getElementsByTagNameAll[7].localName", .ex = "p" },
-        .{ .src = "getElementsByTagNameAll[8]", .ex = "undefined" },
-        .{ .src = "getElementsByTagNameAll['para-empty-child'].localName", .ex = "span" },
-        .{ .src = "getElementsByTagNameAll['foo']", .ex = "undefined" },
-
-        .{ .src = "document.getElementById('content').getElementsByTagName('*').length", .ex = "4" },
-        .{ .src = "document.getElementById('content').getElementsByTagName('p').length", .ex = "2" },
-        .{ .src = "document.getElementById('content').getElementsByTagName('div').length", .ex = "0" },
-
-        .{ .src = "document.children.length", .ex = "1" },
-        .{ .src = "document.getElementById('content').children.length", .ex = "3" },
-
-        // check liveness
-        .{ .src = "let content = document.getElementById('content')", .ex = "undefined" },
-        .{ .src = "let pe = document.getElementById('para-empty')", .ex = "undefined" },
-        .{ .src = "let p = document.createElement('p')", .ex = "undefined" },
-        .{ .src = "p.textContent = 'OK live'", .ex = "OK live" },
-        .{ .src = "getElementsByTagName.item(1).textContent", .ex = " And" },
-        .{ .src = "content.appendChild(p) != undefined", .ex = "true" },
-        .{ .src = "getElementsByTagName.length", .ex = "3" },
-        .{ .src = "getElementsByTagName.item(2).textContent", .ex = "OK live" },
-        .{ .src = "content.insertBefore(p, pe) != undefined", .ex = "true" },
-        .{ .src = "getElementsByTagName.item(0).textContent", .ex = "OK live" },
-    };
-    try checkCases(js_env, &getElementsByTagName);
+const testing = @import("../../testing.zig");
+test "Browser: DOM.HTMLCollection" {
+    try testing.htmlRunner("dom/html_collection.html");
 }

src/browser/dom/implementation.zig

@@ -0,0 +1,56 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+const DOMException = @import("exceptions.zig").DOMException;
+
+// WEB IDL https://dom.spec.whatwg.org/#domimplementation
+pub const DOMImplementation = struct {
+    pub const Exception = DOMException;
+
+    pub fn _createDocumentType(
+        _: *DOMImplementation,
+        qname: [:0]const u8,
+        publicId: [:0]const u8,
+        systemId: [:0]const u8,
+    ) !*parser.DocumentType {
+        return try parser.domImplementationCreateDocumentType(qname, publicId, systemId);
+    }
+
+    pub fn _createDocument(
+        _: *DOMImplementation,
+        namespace: ?[:0]const u8,
+        qname: ?[:0]const u8,
+        doctype: ?*parser.DocumentType,
+    ) !*parser.Document {
+        return try parser.domImplementationCreateDocument(namespace, qname, doctype);
+    }
+
+    pub fn _createHTMLDocument(_: *DOMImplementation, title: ?[]const u8) !*parser.DocumentHTML {
+        return try parser.domImplementationCreateHTMLDocument(title);
+    }
+
+    pub fn _hasFeature(_: *DOMImplementation) bool {
+        return true;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Implementation" {
+    try testing.htmlRunner("dom/implementation.html");
+}

src/browser/dom/intersection_observer.zig

@@ -0,0 +1,186 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const Env = @import("../env.zig").Env;
+const Element = @import("element.zig").Element;
+
+pub const Interfaces = .{
+    IntersectionObserver,
+    IntersectionObserverEntry,
+};
+
+// This is supposed to listen to change between the root and observation targets.
+// However, our rendered stores everything as 1 pixel sized boxes in a long row that never changes.
+// As such, there are no changes to intersections between the root and any target.
+// Instead we keep a list of all entries that are being observed.
+// The callback is called with all entries everytime a new entry is added(observed).
+// Potentially we should also call the callback at a regular interval.
+// The returned Entries are phony, they always indicate full intersection.
+// https://developer.mozilla.org/en-US/docs/Web/API/IntersectionObserver
+pub const IntersectionObserver = struct {
+    page: *Page,
+    callback: Env.Function,
+    options: IntersectionObserverOptions,
+
+    observed_entries: std.ArrayListUnmanaged(IntersectionObserverEntry),
+
+    // new IntersectionObserver(callback)
+    // new IntersectionObserver(callback, options) [not supported yet]
+    pub fn constructor(callback: Env.Function, options_: ?IntersectionObserverOptions, page: *Page) !IntersectionObserver {
+        var options = IntersectionObserverOptions{
+            .root = parser.documentToNode(parser.documentHTMLToDocument(page.window.document)),
+            .rootMargin = "0px 0px 0px 0px",
+            .threshold = .{ .single = 0.0 },
+        };
+        if (options_) |*o| {
+            if (o.root) |root| {
+                options.root = root;
+            } // Other properties are not used due to the way we render
+        }
+
+        return .{
+            .page = page,
+            .callback = callback,
+            .options = options,
+            .observed_entries = .{},
+        };
+    }
+
+    pub fn _disconnect(self: *IntersectionObserver) !void {
+        self.observed_entries = .{}; // We don't free as it is on an arena
+    }
+
+    pub fn _observe(self: *IntersectionObserver, target_element: *parser.Element) !void {
+        for (self.observed_entries.items) |*observer| {
+            if (observer.target == target_element) {
+                return; // Already observed
+            }
+        }
+
+        try self.observed_entries.append(self.page.arena, .{
+            .page = self.page,
+            .target = target_element,
+            .options = &self.options,
+        });
+
+        var result: Env.Function.Result = undefined;
+        self.callback.tryCall(void, .{self.observed_entries.items}, &result) catch {
+            log.debug(.user_script, "callback error", .{
+                .err = result.exception,
+                .stack = result.stack,
+                .source = "intersection observer",
+            });
+        };
+    }
+
+    pub fn _unobserve(self: *IntersectionObserver, target: *parser.Element) !void {
+        for (self.observed_entries.items, 0..) |*observer, index| {
+            if (observer.target == target) {
+                _ = self.observed_entries.swapRemove(index);
+                break;
+            }
+        }
+    }
+
+    pub fn _takeRecords(self: *IntersectionObserver) []IntersectionObserverEntry {
+        return self.observed_entries.items;
+    }
+};
+
+const IntersectionObserverOptions = struct {
+    root: ?*parser.Node, // Element or Document
+    rootMargin: ?[]const u8,
+    threshold: ?Threshold,
+
+    const Threshold = union(enum) {
+        single: f32,
+        list: []const f32,
+    };
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/IntersectionObserverEntry
+// https://w3c.github.io/IntersectionObserver/#intersection-observer-entry
+pub const IntersectionObserverEntry = struct {
+    page: *Page,
+    target: *parser.Element,
+    options: *IntersectionObserverOptions,
+
+    // Returns the bounds rectangle of the target element as a DOMRectReadOnly. The bounds are computed as described in the documentation for Element.getBoundingClientRect().
+    pub fn get_boundingClientRect(self: *const IntersectionObserverEntry) !Element.DOMRect {
+        return Element._getBoundingClientRect(self.target, self.page);
+    }
+
+    // Returns the ratio of the intersectionRect to the boundingClientRect.
+    pub fn get_intersectionRatio(_: *const IntersectionObserverEntry) f32 {
+        return 1.0;
+    }
+
+    // Returns a DOMRectReadOnly representing the target's visible area.
+    pub fn get_intersectionRect(self: *const IntersectionObserverEntry) !Element.DOMRect {
+        return Element._getBoundingClientRect(self.target, self.page);
+    }
+
+    // A Boolean value which is true if the target element intersects with the
+    // intersection observer's root. If this is true, then, the
+    // IntersectionObserverEntry describes a transition into a state of
+    // intersection; if it's false, then you know the transition is from
+    // intersecting to not-intersecting.
+    pub fn get_isIntersecting(_: *const IntersectionObserverEntry) bool {
+        return true;
+    }
+
+    // Returns a DOMRectReadOnly for the intersection observer's root.
+    pub fn get_rootBounds(self: *const IntersectionObserverEntry) !Element.DOMRect {
+        const root = self.options.root.?;
+        if (@intFromPtr(root) == @intFromPtr(self.page.window.document)) {
+            return self.page.renderer.boundingRect();
+        }
+
+        const root_type = try parser.nodeType(root);
+
+        var element: *parser.Element = undefined;
+        switch (root_type) {
+            .element => element = parser.nodeToElement(root),
+            .document => {
+                const doc = parser.nodeToDocument(root);
+                element = (try parser.documentGetDocumentElement(doc)).?;
+            },
+            else => return error.InvalidState,
+        }
+
+        return Element._getBoundingClientRect(element, self.page);
+    }
+
+    // The Element whose intersection with the root changed.
+    pub fn get_target(self: *const IntersectionObserverEntry) *parser.Element {
+        return self.target;
+    }
+
+    // TODO: pub fn get_time(self: *const IntersectionObserverEntry)
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.IntersectionObserver" {
+    try testing.htmlRunner("dom/intersection_observer.html");
+}

src/browser/dom/mutation_observer.zig

@@ -0,0 +1,358 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const Allocator = std.mem.Allocator;
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const Env = @import("../env.zig").Env;
+const NodeList = @import("nodelist.zig").NodeList;
+
+pub const Interfaces = .{
+    MutationObserver,
+    MutationRecord,
+};
+
+const Walker = @import("../dom/walker.zig").WalkerChildren;
+
+// WEB IDL https://dom.spec.whatwg.org/#interface-mutationobserver
+pub const MutationObserver = struct {
+    page: *Page,
+    cbk: Env.Function,
+    connected: bool,
+    scheduled: bool,
+
+    // List of records which were observed. When the call scope ends, we need to
+    // execute our callback with it.
+    observed: std.ArrayListUnmanaged(MutationRecord),
+
+    pub fn constructor(cbk: Env.Function, page: *Page) !MutationObserver {
+        return .{
+            .cbk = cbk,
+            .page = page,
+            .observed = .{},
+            .connected = true,
+            .scheduled = false,
+        };
+    }
+
+    pub fn _observe(self: *MutationObserver, node: *parser.Node, options_: ?Options) !void {
+        const arena = self.page.arena;
+        var options = options_ orelse Options{};
+        if (options.attributeFilter.len > 0) {
+            options.attributeFilter = try arena.dupe([]const u8, options.attributeFilter);
+        }
+
+        const observer = try arena.create(Observer);
+        observer.* = .{
+            .node = node,
+            .options = options,
+            .mutation_observer = self,
+            .event_node = .{ .id = self.cbk.id, .func = Observer.handle },
+        };
+
+        // register node's events
+        if (options.childList or options.subtree) {
+            _ = try parser.eventTargetAddEventListener(
+                parser.toEventTarget(parser.Node, node),
+                "DOMNodeInserted",
+                &observer.event_node,
+                false,
+            );
+            _ = try parser.eventTargetAddEventListener(
+                parser.toEventTarget(parser.Node, node),
+                "DOMNodeRemoved",
+                &observer.event_node,
+                false,
+            );
+        }
+        if (options.attr()) {
+            _ = try parser.eventTargetAddEventListener(
+                parser.toEventTarget(parser.Node, node),
+                "DOMAttrModified",
+                &observer.event_node,
+                false,
+            );
+        }
+        if (options.cdata()) {
+            _ = try parser.eventTargetAddEventListener(
+                parser.toEventTarget(parser.Node, node),
+                "DOMCharacterDataModified",
+                &observer.event_node,
+                false,
+            );
+        }
+        if (options.subtree) {
+            _ = try parser.eventTargetAddEventListener(
+                parser.toEventTarget(parser.Node, node),
+                "DOMSubtreeModified",
+                &observer.event_node,
+                false,
+            );
+        }
+    }
+
+    fn callback(ctx: *anyopaque) ?u32 {
+        const self: *MutationObserver = @ptrCast(@alignCast(ctx));
+        if (self.connected == false) {
+            self.scheduled = true;
+            return null;
+        }
+        self.scheduled = false;
+
+        const records = self.observed.items;
+        if (records.len == 0) {
+            return null;
+        }
+
+        defer self.observed.clearRetainingCapacity();
+
+        var result: Env.Function.Result = undefined;
+        self.cbk.tryCall(void, .{records}, &result) catch {
+            log.debug(.user_script, "callback error", .{
+                .err = result.exception,
+                .stack = result.stack,
+                .source = "mutation observer",
+            });
+        };
+        return null;
+    }
+
+    // TODO
+    pub fn _disconnect(self: *MutationObserver) !void {
+        self.connected = false;
+    }
+
+    // TODO
+    pub fn _takeRecords(_: *const MutationObserver) ?[]const u8 {
+        return &[_]u8{};
+    }
+};
+
+pub const MutationRecord = struct {
+    type: []const u8,
+    target: *parser.Node,
+    added_nodes: NodeList = .{},
+    removed_nodes: NodeList = .{},
+    previous_sibling: ?*parser.Node = null,
+    next_sibling: ?*parser.Node = null,
+    attribute_name: ?[]const u8 = null,
+    attribute_namespace: ?[]const u8 = null,
+    old_value: ?[]const u8 = null,
+
+    pub fn get_type(self: *const MutationRecord) []const u8 {
+        return self.type;
+    }
+
+    pub fn get_addedNodes(self: *MutationRecord) *NodeList {
+        return &self.added_nodes;
+    }
+
+    pub fn get_removedNodes(self: *MutationRecord) *NodeList {
+        return &self.removed_nodes;
+    }
+
+    pub fn get_target(self: *const MutationRecord) *parser.Node {
+        return self.target;
+    }
+
+    pub fn get_attributeName(self: *const MutationRecord) ?[]const u8 {
+        return self.attribute_name;
+    }
+
+    pub fn get_attributeNamespace(self: *const MutationRecord) ?[]const u8 {
+        return self.attribute_namespace;
+    }
+
+    pub fn get_previousSibling(self: *const MutationRecord) ?*parser.Node {
+        return self.previous_sibling;
+    }
+
+    pub fn get_nextSibling(self: *const MutationRecord) ?*parser.Node {
+        return self.next_sibling;
+    }
+
+    pub fn get_oldValue(self: *const MutationRecord) ?[]const u8 {
+        return self.old_value;
+    }
+};
+
+const Options = struct {
+    childList: bool = false,
+    attributes: bool = false,
+    characterData: bool = false,
+    subtree: bool = false,
+    attributeOldValue: bool = false,
+    characterDataOldValue: bool = false,
+    attributeFilter: [][]const u8 = &.{},
+
+    fn attr(self: Options) bool {
+        return self.attributes or self.attributeOldValue or self.attributeFilter.len > 0;
+    }
+
+    fn cdata(self: Options) bool {
+        return self.characterData or self.characterDataOldValue;
+    }
+};
+
+const Observer = struct {
+    node: *parser.Node,
+    options: Options,
+
+    // reference back to the MutationObserver so that we can access the arena
+    // and batch the mutation records.
+    mutation_observer: *MutationObserver,
+
+    event_node: parser.EventNode,
+
+    fn appliesTo(
+        self: *const Observer,
+        target: *parser.Node,
+        event_type: MutationEventType,
+        event: *parser.MutationEvent,
+    ) !bool {
+        if (event_type == .DOMAttrModified and self.options.attributeFilter.len > 0) {
+            const attribute_name = try parser.mutationEventAttributeName(event);
+            for (self.options.attributeFilter) |needle| blk: {
+                if (std.mem.eql(u8, attribute_name, needle)) {
+                    break :blk;
+                }
+            }
+            return false;
+        }
+
+        // mutation on any target is always ok.
+        if (self.options.subtree) {
+            return true;
+        }
+
+        // if target equals node, alway ok.
+        if (target == self.node) {
+            return true;
+        }
+
+        // no subtree, no same target and no childlist, always noky.
+        if (!self.options.childList) {
+            return false;
+        }
+
+        // target must be a child of o.node
+        const walker = Walker{};
+        var next: ?*parser.Node = null;
+        while (true) {
+            next = walker.get_next(self.node, next) catch break orelse break;
+            if (next.? == target) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+
+    fn handle(en: *parser.EventNode, event: *parser.Event) void {
+        const self: *Observer = @fieldParentPtr("event_node", en);
+        self._handle(event) catch |err| {
+            log.err(.web_api, "handle error", .{ .err = err, .source = "mutation observer" });
+        };
+    }
+
+    fn _handle(self: *Observer, event: *parser.Event) !void {
+        var mutation_observer = self.mutation_observer;
+
+        const node = blk: {
+            const event_target = parser.eventTarget(event) orelse return;
+            break :blk parser.eventTargetToNode(event_target);
+        };
+
+        const mutation_event = parser.eventToMutationEvent(event);
+        const event_type = blk: {
+            const t = try parser.eventType(event);
+            break :blk std.meta.stringToEnum(MutationEventType, t) orelse return;
+        };
+
+        if (try self.appliesTo(node, event_type, mutation_event) == false) {
+            return;
+        }
+
+        var record = MutationRecord{
+            .target = self.node,
+            .type = event_type.recordType(),
+        };
+
+        const arena = mutation_observer.page.arena;
+        switch (event_type) {
+            .DOMAttrModified => {
+                record.attribute_name = parser.mutationEventAttributeName(mutation_event) catch null;
+                if (self.options.attributeOldValue) {
+                    record.old_value = parser.mutationEventPrevValue(mutation_event) catch null;
+                }
+            },
+            .DOMCharacterDataModified => {
+                if (self.options.characterDataOldValue) {
+                    record.old_value = parser.mutationEventPrevValue(mutation_event) catch null;
+                }
+            },
+            .DOMNodeInserted => {
+                if (parser.mutationEventRelatedNode(mutation_event) catch null) |related_node| {
+                    try record.added_nodes.append(arena, related_node);
+                }
+            },
+            .DOMNodeRemoved => {
+                if (parser.mutationEventRelatedNode(mutation_event) catch null) |related_node| {
+                    try record.removed_nodes.append(arena, related_node);
+                }
+            },
+        }
+
+        try mutation_observer.observed.append(arena, record);
+
+        if (mutation_observer.scheduled == false) {
+            mutation_observer.scheduled = true;
+            try mutation_observer.page.scheduler.add(
+                mutation_observer,
+                MutationObserver.callback,
+                0,
+                .{ .name = "mutation_observer" },
+            );
+        }
+    }
+};
+
+const MutationEventType = enum {
+    DOMAttrModified,
+    DOMCharacterDataModified,
+    DOMNodeInserted,
+    DOMNodeRemoved,
+
+    fn recordType(self: MutationEventType) []const u8 {
+        return switch (self) {
+            .DOMAttrModified => "attributes",
+            .DOMCharacterDataModified => "characterData",
+            .DOMNodeInserted => "childList",
+            .DOMNodeRemoved => "childList",
+        };
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.MutationObserver" {
+    try testing.htmlRunner("dom/mutation_observer.html");
+}

src/browser/dom/namednodemap.zig

@@ -0,0 +1,121 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+
+const DOMException = @import("exceptions.zig").DOMException;
+
+// WEB IDL https://dom.spec.whatwg.org/#namednodemap
+pub const NamedNodeMap = struct {
+    pub const Self = parser.NamedNodeMap;
+
+    pub const Exception = DOMException;
+    pub const Iterator = NamedNodeMapIterator;
+
+    // TODO implement LegacyUnenumerableNamedProperties.
+    // https://webidl.spec.whatwg.org/#LegacyUnenumerableNamedProperties
+
+    pub fn get_length(self: *parser.NamedNodeMap) !u32 {
+        return try parser.namedNodeMapGetLength(self);
+    }
+
+    pub fn _item(self: *parser.NamedNodeMap, index: u32) !?*parser.Attribute {
+        return try parser.namedNodeMapItem(self, index);
+    }
+
+    pub fn _getNamedItem(self: *parser.NamedNodeMap, qname: []const u8) !?*parser.Attribute {
+        return try parser.namedNodeMapGetNamedItem(self, qname);
+    }
+
+    pub fn _getNamedItemNS(
+        self: *parser.NamedNodeMap,
+        namespace: []const u8,
+        localname: []const u8,
+    ) !?*parser.Attribute {
+        return try parser.namedNodeMapGetNamedItemNS(self, namespace, localname);
+    }
+
+    pub fn _setNamedItem(self: *parser.NamedNodeMap, attr: *parser.Attribute) !?*parser.Attribute {
+        return try parser.namedNodeMapSetNamedItem(self, attr);
+    }
+
+    pub fn _setNamedItemNS(self: *parser.NamedNodeMap, attr: *parser.Attribute) !?*parser.Attribute {
+        return try parser.namedNodeMapSetNamedItemNS(self, attr);
+    }
+
+    pub fn _removeNamedItem(self: *parser.NamedNodeMap, qname: []const u8) !*parser.Attribute {
+        return try parser.namedNodeMapRemoveNamedItem(self, qname);
+    }
+
+    pub fn _removeNamedItemNS(
+        self: *parser.NamedNodeMap,
+        namespace: []const u8,
+        localname: []const u8,
+    ) !*parser.Attribute {
+        return try parser.namedNodeMapRemoveNamedItemNS(self, namespace, localname);
+    }
+
+    pub fn indexed_get(self: *parser.NamedNodeMap, index: u32, has_value: *bool) !*parser.Attribute {
+        return (try _item(self, index)) orelse {
+            has_value.* = false;
+            return undefined;
+        };
+    }
+
+    pub fn named_get(self: *parser.NamedNodeMap, name: []const u8, has_value: *bool) !*parser.Attribute {
+        return (try _getNamedItem(self, name)) orelse {
+            has_value.* = false;
+            return undefined;
+        };
+    }
+
+    pub fn _symbol_iterator(self: *parser.NamedNodeMap) NamedNodeMapIterator {
+        return .{ .map = self };
+    }
+};
+
+pub const NamedNodeMapIterator = struct {
+    index: u32 = 0,
+    map: *parser.NamedNodeMap,
+
+    pub const Return = struct {
+        done: bool,
+        value: ?*parser.Attribute,
+    };
+
+    pub fn _next(self: *NamedNodeMapIterator) !Return {
+        const e = try NamedNodeMap._item(self.map, self.index);
+        if (e == null) {
+            return .{
+                .value = null,
+                .done = true,
+            };
+        }
+
+        self.index += 1;
+        return .{
+            .value = e,
+            .done = false,
+        };
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.NamedNodeMap" {
+    try testing.htmlRunner("dom/named_node_map.html");
+}

src/browser/dom/node.zig

@@ -0,0 +1,637 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const generate = @import("../../runtime/generate.zig");
+
+const Page = @import("../page.zig").Page;
+const EventTarget = @import("event_target.zig").EventTarget;
+
+// DOM
+const Attr = @import("attribute.zig").Attr;
+const CData = @import("character_data.zig");
+const Element = @import("element.zig").Element;
+const ElementUnion = @import("element.zig").Union;
+const NodeList = @import("nodelist.zig").NodeList;
+const Document = @import("document.zig").Document;
+const DocumentType = @import("document_type.zig").DocumentType;
+const DocumentFragment = @import("document_fragment.zig").DocumentFragment;
+const HTMLCollection = @import("html_collection.zig").HTMLCollection;
+const HTMLAllCollection = @import("html_collection.zig").HTMLAllCollection;
+const HTMLCollectionIterator = @import("html_collection.zig").HTMLCollectionIterator;
+const ShadowRoot = @import("shadow_root.zig").ShadowRoot;
+const Walker = @import("walker.zig").WalkerDepthFirst;
+
+// HTML
+const HTML = @import("../html/html.zig");
+
+// Node interfaces
+pub const Interfaces = .{
+    Attr,
+    CData.CharacterData,
+    CData.Interfaces,
+    Element,
+    Document,
+    DocumentType,
+    DocumentFragment,
+    HTMLCollection,
+    HTMLAllCollection,
+    HTMLCollectionIterator,
+    HTML.Interfaces,
+};
+
+pub const Union = generate.Union(Interfaces);
+
+// Node implementation
+pub const Node = struct {
+    pub const Self = parser.Node;
+    pub const prototype = *EventTarget;
+    pub const subtype = .node;
+
+    pub fn toInterface(node: *parser.Node) !Union {
+        return switch (try parser.nodeType(node)) {
+            .element => try Element.toInterfaceT(
+                Union,
+                @as(*parser.Element, @ptrCast(node)),
+            ),
+            .comment => .{ .Comment = @as(*parser.Comment, @ptrCast(node)) },
+            .text => .{ .Text = @as(*parser.Text, @ptrCast(node)) },
+            .cdata_section => .{ .CDATASection = @as(*parser.CDATASection, @ptrCast(node)) },
+            .processing_instruction => .{ .ProcessingInstruction = @as(*parser.ProcessingInstruction, @ptrCast(node)) },
+            .document => blk: {
+                const doc: *parser.Document = @ptrCast(node);
+                if (doc.is_html) {
+                    break :blk .{ .HTMLDocument = @as(*parser.DocumentHTML, @ptrCast(node)) };
+                }
+
+                break :blk .{ .Document = doc };
+            },
+            .document_type => .{ .DocumentType = @as(*parser.DocumentType, @ptrCast(node)) },
+            .attribute => .{ .Attr = @as(*parser.Attribute, @ptrCast(node)) },
+            .document_fragment => .{ .DocumentFragment = @as(*parser.DocumentFragment, @ptrCast(node)) },
+            else => @panic("node type not handled"), // TODO
+        };
+    }
+
+    // class attributes
+
+    pub const _ELEMENT_NODE = @intFromEnum(parser.NodeType.element);
+    pub const _ATTRIBUTE_NODE = @intFromEnum(parser.NodeType.attribute);
+    pub const _TEXT_NODE = @intFromEnum(parser.NodeType.text);
+    pub const _CDATA_SECTION_NODE = @intFromEnum(parser.NodeType.cdata_section);
+    pub const _PROCESSING_INSTRUCTION_NODE = @intFromEnum(parser.NodeType.processing_instruction);
+    pub const _COMMENT_NODE = @intFromEnum(parser.NodeType.comment);
+    pub const _DOCUMENT_NODE = @intFromEnum(parser.NodeType.document);
+    pub const _DOCUMENT_TYPE_NODE = @intFromEnum(parser.NodeType.document_type);
+    pub const _DOCUMENT_FRAGMENT_NODE = @intFromEnum(parser.NodeType.document_fragment);
+
+    // These 3 are deprecated, but both Chrome and Firefox still expose them
+    pub const _ENTITY_REFERENCE_NODE = @intFromEnum(parser.NodeType.entity_reference);
+    pub const _ENTITY_NODE = @intFromEnum(parser.NodeType.entity);
+    pub const _NOTATION_NODE = @intFromEnum(parser.NodeType.notation);
+
+    pub const _DOCUMENT_POSITION_DISCONNECTED = @intFromEnum(parser.DocumentPosition.disconnected);
+    pub const _DOCUMENT_POSITION_PRECEDING = @intFromEnum(parser.DocumentPosition.preceding);
+    pub const _DOCUMENT_POSITION_FOLLOWING = @intFromEnum(parser.DocumentPosition.following);
+    pub const _DOCUMENT_POSITION_CONTAINS = @intFromEnum(parser.DocumentPosition.contains);
+    pub const _DOCUMENT_POSITION_CONTAINED_BY = @intFromEnum(parser.DocumentPosition.contained_by);
+    pub const _DOCUMENT_POSITION_IMPLEMENTATION_SPECIFIC = @intFromEnum(parser.DocumentPosition.implementation_specific);
+
+    // JS funcs
+    // --------
+
+    // Read-only attributes
+    pub fn get_baseURI(_: *parser.Node, page: *Page) ![]const u8 {
+        return page.url.raw;
+    }
+
+    pub fn get_firstChild(self: *parser.Node) !?Union {
+        const res = try parser.nodeFirstChild(self);
+        if (res == null) {
+            return null;
+        }
+        return try Node.toInterface(res.?);
+    }
+
+    pub fn get_lastChild(self: *parser.Node) !?Union {
+        const res = try parser.nodeLastChild(self);
+        if (res == null) {
+            return null;
+        }
+        return try Node.toInterface(res.?);
+    }
+
+    pub fn get_nextSibling(self: *parser.Node) !?Union {
+        const res = try parser.nodeNextSibling(self);
+        if (res == null) {
+            return null;
+        }
+        return try Node.toInterface(res.?);
+    }
+
+    pub fn get_previousSibling(self: *parser.Node) !?Union {
+        const res = try parser.nodePreviousSibling(self);
+        if (res == null) {
+            return null;
+        }
+        return try Node.toInterface(res.?);
+    }
+
+    pub fn get_parentNode(self: *parser.Node) !?Union {
+        const res = try parser.nodeParentNode(self);
+        if (res == null) {
+            return null;
+        }
+        return try Node.toInterface(res.?);
+    }
+
+    pub fn get_parentElement(self: *parser.Node) !?ElementUnion {
+        const res = try parser.nodeParentElement(self);
+        if (res == null) {
+            return null;
+        }
+        return try Element.toInterface(res.?);
+    }
+
+    pub fn get_nodeName(self: *parser.Node) ![]const u8 {
+        return try parser.nodeName(self);
+    }
+
+    pub fn get_nodeType(self: *parser.Node) !u8 {
+        return @intFromEnum(try parser.nodeType(self));
+    }
+
+    pub fn get_ownerDocument(self: *parser.Node) !?*parser.DocumentHTML {
+        const res = try parser.nodeOwnerDocument(self);
+        if (res == null) {
+            return null;
+        }
+        return @as(*parser.DocumentHTML, @ptrCast(res.?));
+    }
+
+    pub fn get_isConnected(self: *parser.Node) !bool {
+        var node = self;
+        while (true) {
+            const node_type = try parser.nodeType(node);
+            if (node_type == .document) {
+                return true;
+            }
+
+            if (try parser.nodeParentNode(node)) |parent| {
+                // didn't find a document, but node has a parent, let's see
+                // if it's connected;
+                node = parent;
+                continue;
+            }
+
+            if (node_type != .document_fragment) {
+                // doesn't have a parent and isn't a document_fragment
+                // can't be connected
+                return false;
+            }
+
+            if (parser.documentFragmentGetHost(@ptrCast(node))) |host| {
+                // node doesn't have a parent, but it's a document fragment
+                // with a host. The host is like the parent, but we only want to
+                // traverse up (or down) to it in specific cases, like isConnected.
+                node = host;
+                continue;
+            }
+            return false;
+        }
+    }
+
+    // Read/Write attributes
+
+    pub fn get_nodeValue(self: *parser.Node) !?[]const u8 {
+        return try parser.nodeValue(self);
+    }
+
+    pub fn set_nodeValue(self: *parser.Node, data: []u8) !void {
+        try parser.nodeSetValue(self, data);
+    }
+
+    pub fn get_textContent(self: *parser.Node) !?[]const u8 {
+        return try parser.nodeTextContent(self);
+    }
+
+    pub fn set_textContent(self: *parser.Node, data: []u8) !void {
+        return try parser.nodeSetTextContent(self, data);
+    }
+
+    // Methods
+
+    pub fn _appendChild(self: *parser.Node, child: *parser.Node) !Union {
+        const self_owner = try parser.nodeOwnerDocument(self);
+        const child_owner = try parser.nodeOwnerDocument(child);
+
+        // If the node to be inserted has a different ownerDocument than the parent node,
+        // modern browsers automatically adopt the node and its descendants into
+        // the parent's ownerDocument.
+        // This process is known as adoption.
+        // (7.1) https://dom.spec.whatwg.org/#concept-node-insert
+        if (child_owner == null or (self_owner != null and child_owner.? != self_owner.?)) {
+            const w = Walker{};
+            var current = child;
+            while (true) {
+                current.owner = self_owner;
+                current = try w.get_next(child, current) orelse break;
+            }
+        }
+
+        // TODO: DocumentFragment special case
+        const res = try parser.nodeAppendChild(self, child);
+        return try Node.toInterface(res);
+    }
+
+    pub fn _cloneNode(self: *parser.Node, deep: ?bool) !Union {
+        const clone = try parser.nodeCloneNode(self, deep orelse false);
+        return try Node.toInterface(clone);
+    }
+
+    pub fn _compareDocumentPosition(self: *parser.Node, other: *parser.Node) !u32 {
+        if (self == other) {
+            return 0;
+        }
+
+        const docself = try parser.nodeOwnerDocument(self) orelse blk: {
+            if (try parser.nodeType(self) == .document) {
+                break :blk @as(*parser.Document, @ptrCast(self));
+            }
+            break :blk null;
+        };
+        const docother = try parser.nodeOwnerDocument(other) orelse blk: {
+            if (try parser.nodeType(other) == .document) {
+                break :blk @as(*parser.Document, @ptrCast(other));
+            }
+            break :blk null;
+        };
+
+        // Both are in different document.
+        if (docself == null or docother == null or docself.? != docother.?) {
+            return @intFromEnum(parser.DocumentPosition.disconnected) +
+                @intFromEnum(parser.DocumentPosition.implementation_specific) +
+                @intFromEnum(parser.DocumentPosition.preceding);
+        }
+
+        if (@intFromPtr(self) == @intFromPtr(docself.?)) {
+            // if self is the document, and we already know other is in the
+            // document, then other is contained by and following self.
+            return @intFromEnum(parser.DocumentPosition.following) +
+                @intFromEnum(parser.DocumentPosition.contained_by);
+        }
+
+        const rootself = try parser.nodeGetRootNode(self);
+        const rootother = try parser.nodeGetRootNode(other);
+        if (rootself != rootother) {
+            return @intFromEnum(parser.DocumentPosition.disconnected) +
+                @intFromEnum(parser.DocumentPosition.implementation_specific) +
+                @intFromEnum(parser.DocumentPosition.preceding);
+        }
+
+        // TODO Both are in a different trees in the same document.
+
+        const w = Walker{};
+        var next: ?*parser.Node = null;
+
+        // Is other a descendant of self?
+        while (true) {
+            next = try w.get_next(self, next) orelse break;
+            if (other == next) {
+                return @intFromEnum(parser.DocumentPosition.following) +
+                    @intFromEnum(parser.DocumentPosition.contained_by);
+            }
+        }
+
+        // Is self a descendant of other?
+        next = null;
+        while (true) {
+            next = try w.get_next(other, next) orelse break;
+            if (self == next) {
+                return @intFromEnum(parser.DocumentPosition.contains) +
+                    @intFromEnum(parser.DocumentPosition.preceding);
+            }
+        }
+
+        next = null;
+        while (true) {
+            next = try w.get_next(parser.documentToNode(docself.?), next) orelse break;
+            if (other == next) {
+                // other precedes self.
+                return @intFromEnum(parser.DocumentPosition.preceding);
+            }
+            if (self == next) {
+                // other follows self.
+                return @intFromEnum(parser.DocumentPosition.following);
+            }
+        }
+
+        return 0;
+    }
+
+    pub fn _contains(self: *parser.Node, other: *parser.Node) !bool {
+        return try parser.nodeContains(self, other);
+    }
+
+    // Returns itself or ancestor object inheriting from Node.
+    // - An Element inside a standard web page will return an HTMLDocument object representing the entire page (or <iframe>).
+    // - An Element inside a shadow DOM will return the associated ShadowRoot.
+    // - An Element that is not attached to a document or a shadow tree will return the root of the DOM tree it belongs to
+    const GetRootNodeResult = union(enum) {
+        shadow_root: *ShadowRoot,
+        node: Union,
+    };
+    pub fn _getRootNode(self: *parser.Node, options: ?struct { composed: bool = false }, page: *Page) !GetRootNodeResult {
+        if (options) |options_| if (options_.composed) {
+            log.warn(.web_api, "not implemented", .{ .feature = "getRootNode composed" });
+        };
+
+        const root = try parser.nodeGetRootNode(self);
+        if (page.getNodeState(root)) |state| {
+            if (state.shadow_root) |sr| {
+                return .{ .shadow_root = sr };
+            }
+        }
+
+        return .{ .node = try Node.toInterface(root) };
+    }
+
+    pub fn _hasChildNodes(self: *parser.Node) !bool {
+        return try parser.nodeHasChildNodes(self);
+    }
+
+    pub fn get_childNodes(self: *parser.Node, page: *Page) !NodeList {
+        const allocator = page.arena;
+        var list: NodeList = .{};
+
+        var n = try parser.nodeFirstChild(self) orelse return list;
+        while (true) {
+            try list.append(allocator, n);
+            n = try parser.nodeNextSibling(n) orelse return list;
+        }
+    }
+
+    pub fn _insertBefore(self: *parser.Node, new_node: *parser.Node, ref_node_: ?*parser.Node) !Union {
+        if (ref_node_ == null) {
+            return _appendChild(self, new_node);
+        }
+
+        const self_owner = try parser.nodeOwnerDocument(self);
+        const new_node_owner = try parser.nodeOwnerDocument(new_node);
+
+        // If the node to be inserted has a different ownerDocument than the parent node,
+        // modern browsers automatically adopt the node and its descendants into
+        // the parent's ownerDocument.
+        // This process is known as adoption.
+        // (7.1) https://dom.spec.whatwg.org/#concept-node-insert
+        if (new_node_owner == null or (self_owner != null and new_node_owner.? != self_owner.?)) {
+            const w = Walker{};
+            var current = new_node;
+            while (true) {
+                current.owner = self_owner;
+                current = try w.get_next(new_node, current) orelse break;
+            }
+        }
+
+        return Node.toInterface(try parser.nodeInsertBefore(self, new_node, ref_node_.?));
+    }
+
+    pub fn _isDefaultNamespace(self: *parser.Node, namespace: ?[]const u8) !bool {
+        return try parser.nodeIsDefaultNamespace(self, namespace);
+    }
+
+    pub fn _isEqualNode(self: *parser.Node, other: *parser.Node) !bool {
+        // TODO: other is not an optional parameter, but can be null.
+        return try parser.nodeIsEqualNode(self, other);
+    }
+
+    pub fn _isSameNode(self: *parser.Node, other: *parser.Node) !bool {
+        // TODO: other is not an optional parameter, but can be null.
+        // NOTE: there is no need to use isSameNode(); instead use the === strict equality operator
+        return try parser.nodeIsSameNode(self, other);
+    }
+
+    pub fn _lookupPrefix(self: *parser.Node, namespace: ?[]const u8) !?[]const u8 {
+        // TODO: other is not an optional parameter, but can be null.
+        if (namespace == null) {
+            return null;
+        }
+        if (std.mem.eql(u8, namespace.?, "")) {
+            return null;
+        }
+        return try parser.nodeLookupPrefix(self, namespace.?);
+    }
+
+    pub fn _lookupNamespaceURI(self: *parser.Node, prefix: ?[]const u8) !?[]const u8 {
+        // TODO: other is not an optional parameter, but can be null.
+        return try parser.nodeLookupNamespaceURI(self, prefix);
+    }
+
+    pub fn _normalize(self: *parser.Node) !void {
+        return try parser.nodeNormalize(self);
+    }
+
+    pub fn _removeChild(self: *parser.Node, child: *parser.Node) !Union {
+        const res = try parser.nodeRemoveChild(self, child);
+        return try Node.toInterface(res);
+    }
+
+    pub fn _replaceChild(self: *parser.Node, new_child: *parser.Node, old_child: *parser.Node) !Union {
+        const res = try parser.nodeReplaceChild(self, new_child, old_child);
+        return try Node.toInterface(res);
+    }
+
+    // Check if the hierarchy node tree constraints are respected.
+    // For now, it checks only if new nodes are not self.
+    // TODO implements the others contraints.
+    // see https://dom.spec.whatwg.org/#concept-node-tree
+    pub fn hierarchy(self: *parser.Node, nodes: []const NodeOrText) bool {
+        for (nodes) |n| {
+            if (n.is(self)) {
+                return false;
+            }
+        }
+        return true;
+    }
+
+    pub fn prepend(self: *parser.Node, nodes: []const NodeOrText) !void {
+        if (nodes.len == 0) {
+            return;
+        }
+
+        // check hierarchy
+        if (!hierarchy(self, nodes)) {
+            return parser.DOMError.HierarchyRequest;
+        }
+
+        const doc = (try parser.nodeOwnerDocument(self)) orelse return;
+
+        if (try parser.nodeFirstChild(self)) |first| {
+            for (nodes) |node| {
+                _ = try parser.nodeInsertBefore(self, try node.toNode(doc), first);
+            }
+            return;
+        }
+
+        for (nodes) |node| {
+            _ = try parser.nodeAppendChild(self, try node.toNode(doc));
+        }
+    }
+
+    pub fn append(self: *parser.Node, nodes: []const NodeOrText) !void {
+        if (nodes.len == 0) {
+            return;
+        }
+
+        // check hierarchy
+        if (!hierarchy(self, nodes)) {
+            return parser.DOMError.HierarchyRequest;
+        }
+
+        const doc = (try parser.nodeOwnerDocument(self)) orelse return;
+        for (nodes) |node| {
+            _ = try parser.nodeAppendChild(self, try node.toNode(doc));
+        }
+    }
+
+    pub fn replaceChildren(self: *parser.Node, nodes: []const NodeOrText) !void {
+        if (nodes.len == 0) {
+            return;
+        }
+
+        // check hierarchy
+        if (!hierarchy(self, nodes)) {
+            return parser.DOMError.HierarchyRequest;
+        }
+
+        // remove existing children
+        try removeChildren(self);
+
+        const doc = (try parser.nodeOwnerDocument(self)) orelse return;
+        // add new children
+        for (nodes) |node| {
+            _ = try parser.nodeAppendChild(self, try node.toNode(doc));
+        }
+    }
+
+    pub fn removeChildren(self: *parser.Node) !void {
+        if (!try parser.nodeHasChildNodes(self)) return;
+
+        const children = try parser.nodeGetChildNodes(self);
+        const ln = try parser.nodeListLength(children);
+        var i: u32 = 0;
+        while (i < ln) {
+            defer i += 1;
+            // we always retrieve the 0 index child on purpose: libdom nodelist
+            // are dynamic. So the next child to remove is always as pos 0.
+            const child = try parser.nodeListItem(children, 0) orelse continue;
+            _ = try parser.nodeRemoveChild(self, child);
+        }
+    }
+
+    pub fn before(self: *parser.Node, nodes: []const NodeOrText) !void {
+        const parent = try parser.nodeParentNode(self) orelse return;
+        const doc = (try parser.nodeOwnerDocument(parent)) orelse return;
+
+        var sibling: ?*parser.Node = self;
+        // have to find the first sibling that isn't in nodes
+        CHECK: while (sibling) |s| {
+            for (nodes) |n| {
+                if (n.is(s)) {
+                    sibling = try parser.nodePreviousSibling(s);
+                    continue :CHECK;
+                }
+            }
+            break;
+        }
+
+        if (sibling == null) {
+            sibling = try parser.nodeFirstChild(parent);
+        }
+
+        if (sibling) |ref_node| {
+            for (nodes) |node| {
+                _ = try parser.nodeInsertBefore(parent, try node.toNode(doc), ref_node);
+            }
+            return;
+        }
+
+        return Node.prepend(self, nodes);
+    }
+
+    pub fn after(self: *parser.Node, nodes: []const NodeOrText) !void {
+        const parent = try parser.nodeParentNode(self) orelse return;
+        const doc = (try parser.nodeOwnerDocument(parent)) orelse return;
+
+        // have to find the first sibling that isn't in nodes
+        var sibling = try parser.nodeNextSibling(self);
+        CHECK: while (sibling) |s| {
+            for (nodes) |n| {
+                if (n.is(s)) {
+                    sibling = try parser.nodeNextSibling(s);
+                    continue :CHECK;
+                }
+            }
+            break;
+        }
+
+        if (sibling) |ref_node| {
+            for (nodes) |node| {
+                _ = try parser.nodeInsertBefore(parent, try node.toNode(doc), ref_node);
+            }
+            return;
+        }
+
+        for (nodes) |node| {
+            _ = try parser.nodeAppendChild(parent, try node.toNode(doc));
+        }
+    }
+
+    // A lot of functions take either a node or text input.
+    // The text input is to be converted into a Text node.
+    pub const NodeOrText = union(enum) {
+        text: []const u8,
+        node: *parser.Node,
+
+        fn toNode(self: NodeOrText, doc: *parser.Document) !*parser.Node {
+            return switch (self) {
+                .node => |n| n,
+                .text => |txt| @ptrCast(@alignCast(try parser.documentCreateTextNode(doc, txt))),
+            };
+        }
+
+        // Whether the node represented by the NodeOrText is the same as the
+        // given Node. Always false for text values as these represent as-of-yet
+        // created Text nodes.
+        fn is(self: NodeOrText, other: *parser.Node) bool {
+            return switch (self) {
+                .text => false,
+                .node => |n| n == other,
+            };
+        }
+    };
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Node" {
+    try testing.htmlRunner("dom/node.html");
+    try testing.htmlRunner("dom/node_owner.html");
+}

src/browser/dom/node_filter.zig

@@ -0,0 +1,80 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const parser = @import("../netsurf.zig");
+const Env = @import("../env.zig").Env;
+const Node = @import("node.zig").Node;
+
+pub const NodeFilter = struct {
+    pub const _FILTER_ACCEPT: u16 = 1;
+    pub const _FILTER_REJECT: u16 = 2;
+    pub const _FILTER_SKIP: u16 = 3;
+
+    pub const _SHOW_ALL: u32 = std.math.maxInt(u32);
+    pub const _SHOW_ELEMENT: u32 = 0b1;
+    pub const _SHOW_ATTRIBUTE: u32 = 0b10;
+    pub const _SHOW_TEXT: u32 = 0b100;
+    pub const _SHOW_CDATA_SECTION: u32 = 0b1000;
+    pub const _SHOW_ENTITY_REFERENCE: u32 = 0b10000;
+    pub const _SHOW_ENTITY: u32 = 0b100000;
+    pub const _SHOW_PROCESSING_INSTRUCTION: u32 = 0b1000000;
+    pub const _SHOW_COMMENT: u32 = 0b10000000;
+    pub const _SHOW_DOCUMENT: u32 = 0b100000000;
+    pub const _SHOW_DOCUMENT_TYPE: u32 = 0b1000000000;
+    pub const _SHOW_DOCUMENT_FRAGMENT: u32 = 0b10000000000;
+    pub const _SHOW_NOTATION: u32 = 0b100000000000;
+};
+
+const VerifyResult = enum { accept, skip, reject };
+
+pub fn verify(what_to_show: u32, filter: ?Env.Function, node: *parser.Node) !VerifyResult {
+    const node_type = try parser.nodeType(node);
+
+    // Verify that we can show this node type.
+    if (!switch (node_type) {
+        .attribute => what_to_show & NodeFilter._SHOW_ATTRIBUTE != 0,
+        .cdata_section => what_to_show & NodeFilter._SHOW_CDATA_SECTION != 0,
+        .comment => what_to_show & NodeFilter._SHOW_COMMENT != 0,
+        .document => what_to_show & NodeFilter._SHOW_DOCUMENT != 0,
+        .document_fragment => what_to_show & NodeFilter._SHOW_DOCUMENT_FRAGMENT != 0,
+        .document_type => what_to_show & NodeFilter._SHOW_DOCUMENT_TYPE != 0,
+        .element => what_to_show & NodeFilter._SHOW_ELEMENT != 0,
+        .entity => what_to_show & NodeFilter._SHOW_ENTITY != 0,
+        .entity_reference => what_to_show & NodeFilter._SHOW_ENTITY_REFERENCE != 0,
+        .notation => what_to_show & NodeFilter._SHOW_NOTATION != 0,
+        .processing_instruction => what_to_show & NodeFilter._SHOW_PROCESSING_INSTRUCTION != 0,
+        .text => what_to_show & NodeFilter._SHOW_TEXT != 0,
+    }) return .reject;
+
+    // Verify that we aren't filtering it out.
+    if (filter) |f| {
+        const acceptance = try f.call(u16, .{try Node.toInterface(node)});
+        return switch (acceptance) {
+            NodeFilter._FILTER_ACCEPT => .accept,
+            NodeFilter._FILTER_REJECT => .reject,
+            NodeFilter._FILTER_SKIP => .skip,
+            else => .reject,
+        };
+    } else return .accept;
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.NodeFilter" {
+    try testing.htmlRunner("dom/node_filter.html");
+}

src/browser/dom/node_iterator.zig

@@ -0,0 +1,273 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+const Env = @import("../env.zig").Env;
+const NodeFilter = @import("node_filter.zig");
+const Node = @import("node.zig").Node;
+const NodeUnion = @import("node.zig").Union;
+const DOMException = @import("exceptions.zig").DOMException;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/NodeIterator
+// While this is similar to TreeWalker it has its own implementation as there are several subtle differences
+// For example:
+// - nextNode returns the reference node, whereas TreeWalker returns the next node
+// - Skip and reject are equivalent for NodeIterator, for TreeWalker they are different
+pub const NodeIterator = struct {
+    pub const Exception = DOMException;
+
+    root: *parser.Node,
+    reference_node: *parser.Node,
+    what_to_show: u32,
+    filter: ?NodeIteratorOpts,
+    filter_func: ?Env.Function,
+    pointer_before_current: bool = true,
+    // used to track / block recursive filters
+    is_in_callback: bool = false,
+
+    // One of the few cases where null and undefined resolve to different default.
+    // We need the raw JsObject so that we can probe the tri state:
+    // null, undefined or i32.
+    pub const WhatToShow = Env.JsObject;
+
+    pub const NodeIteratorOpts = union(enum) {
+        function: Env.Function,
+        object: struct { acceptNode: Env.Function },
+    };
+
+    pub fn init(node: *parser.Node, what_to_show_: ?WhatToShow, filter: ?NodeIteratorOpts) !NodeIterator {
+        var filter_func: ?Env.Function = null;
+        if (filter) |f| {
+            filter_func = switch (f) {
+                .function => |func| func,
+                .object => |o| o.acceptNode,
+            };
+        }
+
+        var what_to_show: u32 = undefined;
+        if (what_to_show_) |wts| {
+            switch (try wts.triState(NodeIterator, "what_to_show", u32)) {
+                .null => what_to_show = 0,
+                .undefined => what_to_show = NodeFilter.NodeFilter._SHOW_ALL,
+                .value => |v| what_to_show = v,
+            }
+        } else {
+            what_to_show = NodeFilter.NodeFilter._SHOW_ALL;
+        }
+
+        return .{
+            .root = node,
+            .reference_node = node,
+            .what_to_show = what_to_show,
+            .filter = filter,
+            .filter_func = filter_func,
+        };
+    }
+
+    pub fn get_filter(self: *const NodeIterator) ?NodeIteratorOpts {
+        return self.filter;
+    }
+
+    pub fn get_pointerBeforeReferenceNode(self: *const NodeIterator) bool {
+        return self.pointer_before_current;
+    }
+
+    pub fn get_referenceNode(self: *const NodeIterator) !NodeUnion {
+        return try Node.toInterface(self.reference_node);
+    }
+
+    pub fn get_root(self: *const NodeIterator) !NodeUnion {
+        return try Node.toInterface(self.root);
+    }
+
+    pub fn get_whatToShow(self: *const NodeIterator) u32 {
+        return self.what_to_show;
+    }
+
+    pub fn _nextNode(self: *NodeIterator) !?NodeUnion {
+        try self.callbackStart();
+        defer self.callbackEnd();
+
+        if (self.pointer_before_current) {
+            // Unlike TreeWalker, NodeIterator starts at the first node
+            if (.accept == try NodeFilter.verify(self.what_to_show, self.filter_func, self.reference_node)) {
+                self.pointer_before_current = false;
+                return try Node.toInterface(self.reference_node);
+            }
+        }
+
+        if (try self.firstChild(self.reference_node)) |child| {
+            self.reference_node = child;
+            return try Node.toInterface(child);
+        }
+
+        var current = self.reference_node;
+        while (current != self.root) {
+            if (try self.nextSibling(current)) |sibling| {
+                self.reference_node = sibling;
+                return try Node.toInterface(sibling);
+            }
+
+            current = (try parser.nodeParentNode(current)) orelse break;
+        }
+
+        return null;
+    }
+
+    pub fn _previousNode(self: *NodeIterator) !?NodeUnion {
+        try self.callbackStart();
+        defer self.callbackEnd();
+
+        if (!self.pointer_before_current) {
+            if (.accept == try NodeFilter.verify(self.what_to_show, self.filter_func, self.reference_node)) {
+                self.pointer_before_current = true;
+                // Still need to verify as last may be first as well
+                return try Node.toInterface(self.reference_node);
+            }
+        }
+        if (self.reference_node == self.root) {
+            return null;
+        }
+
+        var current = self.reference_node;
+        while (try parser.nodePreviousSibling(current)) |previous| {
+            current = previous;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => {
+                    // Get last child if it has one.
+                    if (try self.lastChild(current)) |child| {
+                        self.reference_node = child;
+                        return try Node.toInterface(child);
+                    }
+
+                    // Otherwise, this node is our previous one.
+                    self.reference_node = current;
+                    return try Node.toInterface(current);
+                },
+                .reject, .skip => {
+                    // Get last child if it has one.
+                    if (try self.lastChild(current)) |child| {
+                        self.reference_node = child;
+                        return try Node.toInterface(child);
+                    }
+                },
+            }
+        }
+
+        if (current != self.root) {
+            if (try self.parentNode(current)) |parent| {
+                self.reference_node = parent;
+                return try Node.toInterface(parent);
+            }
+        }
+
+        return null;
+    }
+
+    pub fn _detach(self: *const NodeIterator) void {
+        // no-op as per spec
+        _ = self;
+    }
+
+    fn firstChild(self: *const NodeIterator, node: *parser.Node) !?*parser.Node {
+        const children = try parser.nodeGetChildNodes(node);
+        const child_count = try parser.nodeListLength(children);
+
+        for (0..child_count) |i| {
+            const index: u32 = @intCast(i);
+            const child = (try parser.nodeListItem(children, index)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, child)) {
+                .accept => return child, // NOTE: Skip and reject are equivalent for NodeIterator, this is different from TreeWalker
+                .reject, .skip => if (try self.firstChild(child)) |gchild| return gchild,
+            }
+        }
+
+        return null;
+    }
+
+    fn lastChild(self: *const NodeIterator, node: *parser.Node) !?*parser.Node {
+        const children = try parser.nodeGetChildNodes(node);
+        const child_count = try parser.nodeListLength(children);
+
+        var index: u32 = child_count;
+        while (index > 0) {
+            index -= 1;
+            const child = (try parser.nodeListItem(children, index)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, child)) {
+                .accept => return child, // NOTE: Skip and reject are equivalent for NodeIterator, this is different from TreeWalker
+                .reject, .skip => if (try self.lastChild(child)) |gchild| return gchild,
+            }
+        }
+
+        return null;
+    }
+
+    // This implementation is actually the same as :TreeWalker
+    fn parentNode(self: *const NodeIterator, node: *parser.Node) !?*parser.Node {
+        if (self.root == node) return null;
+
+        var current = node;
+        while (true) {
+            if (current == self.root) return null;
+            current = (try parser.nodeParentNode(current)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => return current,
+                .reject, .skip => continue,
+            }
+        }
+    }
+
+    // This implementation is actually the same as :TreeWalker
+    fn nextSibling(self: *const NodeIterator, node: *parser.Node) !?*parser.Node {
+        var current = node;
+
+        while (true) {
+            current = (try parser.nodeNextSibling(current)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => return current,
+                .skip, .reject => continue,
+            }
+        }
+
+        return null;
+    }
+
+    fn callbackStart(self: *NodeIterator) !void {
+        if (self.is_in_callback) {
+            // this is the correct DOMExeption
+            return error.InvalidState;
+        }
+        self.is_in_callback = true;
+    }
+
+    fn callbackEnd(self: *NodeIterator) void {
+        self.is_in_callback = false;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.NodeIterator" {
+    try testing.htmlRunner("dom/node_iterator.html");
+}

src/browser/dom/nodelist.zig

@@ -0,0 +1,182 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+
+const JsThis = @import("../env.zig").JsThis;
+const Function = @import("../env.zig").Function;
+
+const NodeUnion = @import("node.zig").Union;
+const Node = @import("node.zig").Node;
+
+const U32Iterator = @import("../iterator/iterator.zig").U32Iterator;
+
+const DOMException = @import("exceptions.zig").DOMException;
+
+pub const Interfaces = .{
+    NodeListIterator,
+    NodeList,
+};
+
+pub const NodeListIterator = struct {
+    coll: *NodeList,
+    index: u32 = 0,
+
+    pub const Return = struct {
+        value: ?NodeUnion,
+        done: bool,
+    };
+
+    pub fn _next(self: *NodeListIterator) !Return {
+        const e = try self.coll._item(self.index);
+        if (e == null) {
+            return Return{
+                .value = null,
+                .done = true,
+            };
+        }
+
+        self.index += 1;
+        return Return{
+            .value = e,
+            .done = false,
+        };
+    }
+};
+
+pub const NodeListEntriesIterator = struct {
+    coll: *NodeList,
+    index: u32 = 0,
+
+    pub const Return = struct {
+        value: ?NodeUnion,
+        done: bool,
+    };
+
+    pub fn _next(self: *NodeListEntriesIterator) !Return {
+        const e = try self.coll._item(self.index);
+        if (e == null) {
+            return Return{
+                .value = null,
+                .done = true,
+            };
+        }
+
+        self.index += 1;
+        return Return{
+            .value = e,
+            .done = false,
+        };
+    }
+};
+
+// Nodelist is implemented in pure Zig b/c libdom's NodeList doesn't allow to
+// append nodes.
+// WEB IDL https://dom.spec.whatwg.org/#nodelist
+//
+// TODO: a Nodelist can be either static or live. But the current
+// implementation allows only static nodelist.
+// see https://dom.spec.whatwg.org/#old-style-collections
+pub const NodeList = struct {
+    pub const Exception = DOMException;
+    const NodesArrayList = std.ArrayListUnmanaged(*parser.Node);
+
+    nodes: NodesArrayList = .{},
+
+    pub fn deinit(self: *NodeList, alloc: std.mem.Allocator) void {
+        // TODO unref all nodes
+        self.nodes.deinit(alloc);
+    }
+
+    pub fn append(self: *NodeList, alloc: std.mem.Allocator, node: *parser.Node) !void {
+        try self.nodes.append(alloc, node);
+    }
+
+    pub fn get_length(self: *const NodeList) u32 {
+        return @intCast(self.nodes.items.len);
+    }
+
+    pub fn _item(self: *const NodeList, index: u32) !?NodeUnion {
+        if (index >= self.nodes.items.len) {
+            return null;
+        }
+
+        const n = self.nodes.items[index];
+        return try Node.toInterface(n);
+    }
+
+    // This code works, but it's _MUCH_ slower than using postAttach. The benefit
+    // of this version, is that it's "live"..but we're talking many orders of
+    // magnitude slower.
+    //
+    // You can test it by commenting out `postAttach`, uncommenting this and
+    // running:
+    //    zig build wpt --  tests/wpt/dom/nodes/NodeList-static-length-getter-tampered-indexOf-1.html
+    //
+    // I think this _is_ the right way to do it, but I must be doing something
+    // wrong to make it so slow.
+    // pub fn indexed_get(self: *const NodeList, index: u32, has_value: *bool) !?NodeUnion {
+    //     return (try self._item(index)) orelse {
+    //         has_value.* = false;
+    //         return null;
+    //     };
+    // }
+
+    pub fn _forEach(self: *NodeList, cbk: Function) !void { // TODO handle thisArg
+        for (self.nodes.items, 0..) |n, i| {
+            const ii: u32 = @intCast(i);
+            var result: Function.Result = undefined;
+            cbk.tryCall(void, .{ n, ii, self }, &result) catch {
+                log.debug(.user_script, "forEach callback", .{ .err = result.exception, .stack = result.stack });
+            };
+        }
+    }
+
+    pub fn _keys(self: *NodeList) U32Iterator {
+        return .{
+            .length = self.get_length(),
+        };
+    }
+
+    pub fn _values(self: *NodeList) NodeListIterator {
+        return .{
+            .coll = self,
+        };
+    }
+
+    pub fn _symbol_iterator(self: *NodeList) NodeListIterator {
+        return self._values();
+    }
+
+    // TODO entries() https://developer.mozilla.org/en-US/docs/Web/API/NodeList/entries
+    pub fn postAttach(self: *NodeList, js_this: JsThis) !void {
+        const len = self.get_length();
+        for (0..len) |i| {
+            const node = try self._item(@intCast(i)) orelse unreachable;
+            try js_this.setIndex(@intCast(i), node, .{});
+        }
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.NodeList" {
+    try testing.htmlRunner("dom/node_list.html");
+}

src/browser/dom/performance.zig

@@ -0,0 +1,208 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+const EventTarget = @import("../dom/event_target.zig").EventTarget;
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+
+const milliTimestamp = @import("../../datetime.zig").milliTimestamp;
+
+pub const Interfaces = .{
+    Performance,
+    PerformanceEntry,
+    PerformanceMark,
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/Performance
+pub const Performance = struct {
+    pub const prototype = *EventTarget;
+
+    // Extend libdom event target for pure zig struct.
+    base: parser.EventTargetTBase = parser.EventTargetTBase{ .internal_target_type = .performance },
+
+    time_origin: u64,
+    // if (Window.crossOriginIsolated) -> Resolution in isolated contexts:       5 microseconds
+    // else                            -> Resolution in non-isolated contexts: 100 microseconds
+    const ms_resolution = 100;
+
+    pub fn init() Performance {
+        return .{
+            .time_origin = milliTimestamp(),
+        };
+    }
+
+    pub fn get_timeOrigin(self: *const Performance) u64 {
+        return self.time_origin;
+    }
+
+    pub fn reset(self: *Performance) void {
+        self.time_origin = milliTimestamp();
+    }
+
+    pub fn _now(self: *const Performance) u64 {
+        return milliTimestamp() - self.time_origin;
+    }
+
+    pub fn _mark(_: *Performance, name: []const u8, _options: ?PerformanceMark.Options, page: *Page) !PerformanceMark {
+        const mark: PerformanceMark = try .constructor(name, _options, page);
+        // TODO: Should store this in an entries list
+        return mark;
+    }
+
+    // TODO: fn _mark should record the marks in a lookup
+    pub fn _clearMarks(_: *Performance, name: ?[]const u8) void {
+        _ = name;
+    }
+
+    // TODO: fn _measures should record the marks in a lookup
+    pub fn _clearMeasures(_: *Performance, name: ?[]const u8) void {
+        _ = name;
+    }
+
+    // TODO: fn _measures should record the marks in a lookup
+    pub fn _getEntriesByName(_: *Performance, name: []const u8, typ: ?[]const u8) []PerformanceEntry {
+        _ = name;
+        _ = typ;
+        return &.{};
+    }
+
+    // TODO: fn _measures should record the marks in a lookup
+    pub fn _getEntriesByType(_: *Performance, typ: []const u8) []PerformanceEntry {
+        _ = typ;
+        return &.{};
+    }
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/PerformanceEntry
+pub const PerformanceEntry = struct {
+    const PerformanceEntryType = enum {
+        element,
+        event,
+        first_input,
+        largest_contentful_paint,
+        layout_shift,
+        long_animation_frame,
+        longtask,
+        mark,
+        measure,
+        navigation,
+        paint,
+        resource,
+        taskattribution,
+        visibility_state,
+
+        pub fn toString(self: PerformanceEntryType) []const u8 {
+            return switch (self) {
+                .first_input => "first-input",
+                .largest_contentful_paint => "largest-contentful-paint",
+                .layout_shift => "layout-shift",
+                .long_animation_frame => "long-animation-frame",
+                .visibility_state => "visibility-state",
+                else => @tagName(self),
+            };
+        }
+    };
+
+    duration: f64 = 0.0,
+    entry_type: PerformanceEntryType,
+    name: []const u8,
+    start_time: f64 = 0.0,
+
+    pub fn get_duration(self: *const PerformanceEntry) f64 {
+        return self.duration;
+    }
+
+    pub fn get_entryType(self: *const PerformanceEntry) PerformanceEntryType {
+        return self.entry_type;
+    }
+
+    pub fn get_name(self: *const PerformanceEntry) []const u8 {
+        return self.name;
+    }
+
+    pub fn get_startTime(self: *const PerformanceEntry) f64 {
+        return self.start_time;
+    }
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/PerformanceMark
+pub const PerformanceMark = struct {
+    pub const prototype = *PerformanceEntry;
+
+    proto: PerformanceEntry,
+    detail: ?Env.JsObject,
+
+    const Options = struct {
+        detail: ?Env.JsObject = null,
+        startTime: ?f64 = null,
+    };
+
+    pub fn constructor(name: []const u8, _options: ?Options, page: *Page) !PerformanceMark {
+        const perf = &page.window.performance;
+
+        const options = _options orelse Options{};
+        const start_time = options.startTime orelse @as(f64, @floatFromInt(perf._now()));
+
+        if (start_time < 0.0) {
+            return error.TypeError;
+        }
+
+        const detail = if (options.detail) |d| try d.persist() else null;
+
+        const duped_name = try page.arena.dupe(u8, name);
+        const proto = PerformanceEntry{ .name = duped_name, .entry_type = .mark, .start_time = start_time };
+
+        return .{ .proto = proto, .detail = detail };
+    }
+
+    pub fn get_detail(self: *const PerformanceMark) ?Env.JsObject {
+        return self.detail;
+    }
+};
+
+const testing = @import("./../../testing.zig");
+
+test "Performance: get_timeOrigin" {
+    var perf = Performance.init();
+    const time_origin = perf.get_timeOrigin();
+    try testing.expect(time_origin >= 0);
+}
+
+test "Performance: now" {
+    var perf = Performance.init();
+
+    // Monotonically increasing
+    var now = perf._now();
+    while (now <= 0) { // Loop for now to not be 0
+        try testing.expectEqual(now, 0);
+        now = perf._now();
+    }
+
+    var after = perf._now();
+    while (after <= now) { // Loop untill after > now
+        try testing.expectEqual(after, now);
+        after = perf._now();
+    }
+}
+
+test "Browser: Performance.Mark" {
+    try testing.htmlRunner("dom/performance.html");
+}

src/browser/dom/performance_observer.zig

@@ -0,0 +1,58 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const Env = @import("../env.zig").Env;
+
+const PerformanceEntry = @import("performance.zig").PerformanceEntry;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/PerformanceObserver
+pub const PerformanceObserver = struct {
+    pub const _supportedEntryTypes = [0][]const u8{};
+
+    pub fn constructor(cbk: Env.Function) PerformanceObserver {
+        _ = cbk;
+        return .{};
+    }
+
+    pub fn _observe(self: *const PerformanceObserver, options_: ?Options) void {
+        _ = self;
+        _ = options_;
+        return;
+    }
+
+    pub fn _disconnect(self: *PerformanceObserver) void {
+        _ = self;
+    }
+
+    pub fn _takeRecords(_: *const PerformanceObserver) []PerformanceEntry {
+        return &[_]PerformanceEntry{};
+    }
+};
+
+const Options = struct {
+    buffered: ?bool = null,
+    durationThreshold: ?f64 = null,
+    entryTypes: ?[]const []const u8 = null,
+    type: ?[]const u8 = null,
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.PerformanceObserver" {
+    try testing.htmlRunner("dom/performance_observer.html");
+}

src/browser/dom/processing_instruction.zig

@@ -0,0 +1,92 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+const Node = @import("node.zig").Node;
+const Page = @import("../page.zig").Page;
+
+// https://dom.spec.whatwg.org/#processinginstruction
+pub const ProcessingInstruction = struct {
+    pub const Self = parser.ProcessingInstruction;
+
+    // TODO for libdom processing instruction inherit from node.
+    // But the spec says it must inherit from CDATA.
+    pub const prototype = *Node;
+    pub const subtype = .node;
+
+    pub fn get_target(self: *parser.ProcessingInstruction) ![]const u8 {
+        // libdom stores the ProcessingInstruction target in the node's name.
+        return try parser.nodeName(parser.processingInstructionToNode(self));
+    }
+
+    // There's something wrong when we try to clone a ProcessInstruction normally.
+    // The resulting object can't be cast back into a node (it crashes). This is
+    // a simple workaround.
+    pub fn _cloneNode(self: *parser.ProcessingInstruction, _: ?bool, page: *Page) !*parser.ProcessingInstruction {
+        return try parser.documentCreateProcessingInstruction(
+            @ptrCast(page.window.document),
+            try get_target(self),
+            (try get_data(self)) orelse "",
+        );
+    }
+
+    pub fn get_data(self: *parser.ProcessingInstruction) !?[]const u8 {
+        return try parser.nodeValue(parser.processingInstructionToNode(self));
+    }
+
+    pub fn set_data(self: *parser.ProcessingInstruction, data: []u8) !void {
+        try parser.nodeSetValue(parser.processingInstructionToNode(self), data);
+    }
+
+    // netsurf's ProcessInstruction doesn't implement the dom_node_get_attributes
+    // and thus will crash if we try to call nodeIsEqualNode.
+    pub fn _isEqualNode(self: *parser.ProcessingInstruction, other_node: *parser.Node) !bool {
+        if (try parser.nodeType(other_node) != .processing_instruction) {
+            return false;
+        }
+
+        const other: *parser.ProcessingInstruction = @ptrCast(other_node);
+
+        if (std.mem.eql(u8, try get_target(self), try get_target(other)) == false) {
+            return false;
+        }
+
+        {
+            const self_data = try get_data(self);
+            const other_data = try get_data(other);
+            if (self_data == null and other_data != null) {
+                return false;
+            }
+            if (self_data != null and other_data == null) {
+                return false;
+            }
+            if (std.mem.eql(u8, self_data.?, other_data.?) == false) {
+                return false;
+            }
+        }
+
+        return true;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.ProcessingInstruction" {
+    try testing.htmlRunner("dom/processing_instruction.html");
+}

src/browser/dom/range.zig

@@ -0,0 +1,390 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const Node = @import("node.zig").Node;
+const NodeUnion = @import("node.zig").Union;
+const DOMException = @import("exceptions.zig").DOMException;
+
+pub const Interfaces = .{
+    AbstractRange,
+    Range,
+};
+
+pub const AbstractRange = struct {
+    collapsed: bool,
+    end_node: *parser.Node,
+    end_offset: u32,
+    start_node: *parser.Node,
+    start_offset: u32,
+
+    pub fn updateCollapsed(self: *AbstractRange) void {
+        // TODO: Eventually, compare properly.
+        self.collapsed = false;
+    }
+
+    pub fn get_collapsed(self: *const AbstractRange) bool {
+        return self.collapsed;
+    }
+
+    pub fn get_endContainer(self: *const AbstractRange) !NodeUnion {
+        return Node.toInterface(self.end_node);
+    }
+
+    pub fn get_endOffset(self: *const AbstractRange) u32 {
+        return self.end_offset;
+    }
+
+    pub fn get_startContainer(self: *const AbstractRange) !NodeUnion {
+        return Node.toInterface(self.start_node);
+    }
+
+    pub fn get_startOffset(self: *const AbstractRange) u32 {
+        return self.start_offset;
+    }
+};
+
+pub const Range = struct {
+    pub const Exception = DOMException;
+    pub const prototype = *AbstractRange;
+
+    proto: AbstractRange,
+
+    pub const _START_TO_START = 0;
+    pub const _START_TO_END = 1;
+    pub const _END_TO_END = 2;
+    pub const _END_TO_START = 3;
+
+    // The Range() constructor returns a newly created Range object whose start
+    // and end is the global Document object.
+    // https://developer.mozilla.org/en-US/docs/Web/API/Range/Range
+    pub fn constructor(page: *Page) Range {
+        const proto: AbstractRange = .{
+            .collapsed = true,
+            .end_node = parser.documentHTMLToNode(page.window.document),
+            .end_offset = 0,
+            .start_node = parser.documentHTMLToNode(page.window.document),
+            .start_offset = 0,
+        };
+
+        return .{ .proto = proto };
+    }
+
+    pub fn _setStart(self: *Range, node: *parser.Node, offset_: i32) !void {
+        try ensureValidOffset(node, offset_);
+        const offset: u32 = @intCast(offset_);
+        const position = compare(node, offset, self.proto.start_node, self.proto.start_offset) catch |err| switch (err) {
+            error.WrongDocument => blk: {
+                // allow a node with a different root than the current, or
+                // a disconnected one. Treat it as if it's "after", so that
+                // we also update the end_offset and end_node.
+                break :blk 1;
+            },
+            else => return err,
+        };
+
+        if (position == 1) {
+            // if we're setting the node after the current start, the end must
+            // be set too.
+            self.proto.end_offset = offset;
+            self.proto.end_node = node;
+        }
+        self.proto.start_node = node;
+        self.proto.start_offset = offset;
+        self.proto.updateCollapsed();
+    }
+
+    pub fn _setStartBefore(self: *Range, node: *parser.Node) !void {
+        const parent, const index = try getParentAndIndex(node);
+        self.proto.start_node = parent;
+        self.proto.start_offset = index;
+    }
+
+    pub fn _setStartAfter(self: *Range, node: *parser.Node) !void {
+        const parent, const index = try getParentAndIndex(node);
+        self.proto.start_node = parent;
+        self.proto.start_offset = index + 1;
+    }
+
+    pub fn _setEnd(self: *Range, node: *parser.Node, offset_: i32) !void {
+        try ensureValidOffset(node, offset_);
+        const offset: u32 = @intCast(offset_);
+
+        const position = compare(node, offset, self.proto.start_node, self.proto.start_offset) catch |err| switch (err) {
+            error.WrongDocument => blk: {
+                // allow a node with a different root than the current, or
+                // a disconnected one. Treat it as if it's "before", so that
+                // we also update the end_offset and end_node.
+                break :blk -1;
+            },
+            else => return err,
+        };
+
+        if (position == -1) {
+            // if we're setting the node before the current start, the start
+            // must be set too.
+            self.proto.start_offset = offset;
+            self.proto.start_node = node;
+        }
+
+        self.proto.end_node = node;
+        self.proto.end_offset = offset;
+        self.proto.updateCollapsed();
+    }
+
+    pub fn _setEndBefore(self: *Range, node: *parser.Node) !void {
+        const parent, const index = try getParentAndIndex(node);
+        self.proto.end_node = parent;
+        self.proto.end_offset = index;
+    }
+
+    pub fn _setEndAfter(self: *Range, node: *parser.Node) !void {
+        const parent, const index = try getParentAndIndex(node);
+        self.proto.end_node = parent;
+        self.proto.end_offset = index + 1;
+    }
+
+    pub fn _createContextualFragment(_: *Range, fragment: []const u8, page: *Page) !*parser.DocumentFragment {
+        const document_html = page.window.document;
+        const document = parser.documentHTMLToDocument(document_html);
+        const doc_frag = try parser.documentParseFragmentFromStr(document, fragment);
+        return doc_frag;
+    }
+
+    pub fn _selectNodeContents(self: *Range, node: *parser.Node) !void {
+        self.proto.start_node = node;
+        self.proto.start_offset = 0;
+        self.proto.end_node = node;
+
+        // Set end_offset
+        switch (try parser.nodeType(node)) {
+            .text, .cdata_section, .comment, .processing_instruction => {
+                // For text-like nodes, end_offset should be the length of the text data
+                if (try parser.nodeValue(node)) |text_data| {
+                    self.proto.end_offset = @intCast(text_data.len);
+                } else {
+                    self.proto.end_offset = 0;
+                }
+            },
+            else => {
+                // For element and other nodes, end_offset is the number of children
+                const child_nodes = try parser.nodeGetChildNodes(node);
+                const child_count = try parser.nodeListLength(child_nodes);
+                self.proto.end_offset = @intCast(child_count);
+            },
+        }
+
+        self.proto.updateCollapsed();
+    }
+
+    // creates a copy
+    pub fn _cloneRange(self: *const Range) Range {
+        return .{
+            .proto = .{
+                .collapsed = self.proto.collapsed,
+                .end_node = self.proto.end_node,
+                .end_offset = self.proto.end_offset,
+                .start_node = self.proto.start_node,
+                .start_offset = self.proto.start_offset,
+            },
+        };
+    }
+
+    pub fn _comparePoint(self: *const Range, node: *parser.Node, offset_: i32) !i32 {
+        const start = self.proto.start_node;
+        if (try parser.nodeGetRootNode(start) != try parser.nodeGetRootNode(node)) {
+            // WPT really wants this error to be first. Later, when we check
+            // if the relative position is 'disconnected', it'll also catch this
+            // case, but WPT will complain because it sometimes also sends
+            // invalid offsets, and it wants WrongDocument to be raised.
+            return error.WrongDocument;
+        }
+
+        if (try parser.nodeType(node) == .document_type) {
+            return error.InvalidNodeType;
+        }
+
+        try ensureValidOffset(node, offset_);
+
+        const offset: u32 = @intCast(offset_);
+        if (try compare(node, offset, start, self.proto.start_offset) == -1) {
+            return -1;
+        }
+
+        if (try compare(node, offset, self.proto.end_node, self.proto.end_offset) == 1) {
+            return 1;
+        }
+
+        return 0;
+    }
+
+    pub fn _isPointInRange(self: *const Range, node: *parser.Node, offset_: i32) !bool {
+        return self._comparePoint(node, offset_) catch |err| switch (err) {
+            error.WrongDocument => return false,
+            else => return err,
+        } == 0;
+    }
+
+    pub fn _intersectsNode(self: *const Range, node: *parser.Node) !bool {
+        const start_root = try parser.nodeGetRootNode(self.proto.start_node);
+        const node_root = try parser.nodeGetRootNode(node);
+        if (start_root != node_root) {
+            return false;
+        }
+
+        const parent, const index = getParentAndIndex(node) catch |err| switch (err) {
+            error.InvalidNodeType => return true, // if node has no parent, we return true.
+            else => return err,
+        };
+
+        if (try compare(parent, index + 1, self.proto.start_node, self.proto.start_offset) != 1) {
+            // node isn't after start, can't intersect
+            return false;
+        }
+
+        if (try compare(parent, index, self.proto.end_node, self.proto.end_offset) != -1) {
+            // node isn't before end, can't intersect
+            return false;
+        }
+
+        return true;
+    }
+
+    pub fn _compareBoundaryPoints(self: *const Range, how: i32, other: *const Range) !i32 {
+        return switch (how) {
+            _START_TO_START => compare(self.proto.start_node, self.proto.start_offset, other.proto.start_node, other.proto.start_offset),
+            _START_TO_END => compare(self.proto.start_node, self.proto.start_offset, other.proto.end_node, other.proto.end_offset),
+            _END_TO_END => compare(self.proto.end_node, self.proto.end_offset, other.proto.end_node, other.proto.end_offset),
+            _END_TO_START => compare(self.proto.end_node, self.proto.end_offset, other.proto.start_node, other.proto.start_offset),
+            else => error.NotSupported, // this is the correct DOM Exception to return
+        };
+    }
+
+    // The Range.detach() method does nothing. It used to disable the Range
+    // object and enable the browser to release associated resources. The
+    // method has been kept for compatibility.
+    // https://developer.mozilla.org/en-US/docs/Web/API/Range/detach
+    pub fn _detach(_: *Range) void {}
+};
+
+fn ensureValidOffset(node: *parser.Node, offset: i32) !void {
+    if (offset < 0) {
+        return error.IndexSize;
+    }
+
+    // not >= because 0 seems to represent the node itself.
+    if (offset > try nodeLength(node)) {
+        return error.IndexSize;
+    }
+}
+
+fn nodeLength(node: *parser.Node) !usize {
+    switch (try isTextual(node)) {
+        true => return ((try parser.nodeTextContent(node)) orelse "").len,
+        false => {
+            const children = try parser.nodeGetChildNodes(node);
+            return @intCast(try parser.nodeListLength(children));
+        },
+    }
+}
+
+fn isTextual(node: *parser.Node) !bool {
+    return switch (try parser.nodeType(node)) {
+        .text, .comment, .cdata_section => true,
+        else => false,
+    };
+}
+
+fn getParentAndIndex(child: *parser.Node) !struct { *parser.Node, u32 } {
+    const parent = (try parser.nodeParentNode(child)) orelse return error.InvalidNodeType;
+    const children = try parser.nodeGetChildNodes(parent);
+    const ln = try parser.nodeListLength(children);
+    var i: u32 = 0;
+    while (i < ln) {
+        defer i += 1;
+        const c = try parser.nodeListItem(children, i) orelse continue;
+        if (c == child) {
+            return .{ parent, i };
+        }
+    }
+
+    // should not be possible to reach this point
+    return error.InvalidNodeType;
+}
+
+// implementation is largely copied from the WPT helper called getPosition in
+// the common.js of the dom folder.
+fn compare(node_a: *parser.Node, offset_a: u32, node_b: *parser.Node, offset_b: u32) !i32 {
+    if (node_a == node_b) {
+        // This is a simple and common case, where the two nodes are the same
+        // We just need to compare their offsets
+        if (offset_a == offset_b) {
+            return 0;
+        }
+        return if (offset_a < offset_b) -1 else 1;
+    }
+
+    // We're probably comparing two different nodes. "Probably", because the
+    // above case on considered the offset if the two nodes were the same
+    // as-is. They could still be the same here, if we first consider the
+    // offset.
+    const position = try Node._compareDocumentPosition(node_b, node_a);
+    if (position & @intFromEnum(parser.DocumentPosition.disconnected) == @intFromEnum(parser.DocumentPosition.disconnected)) {
+        return error.WrongDocument;
+    }
+
+    if (position & @intFromEnum(parser.DocumentPosition.following) == @intFromEnum(parser.DocumentPosition.following)) {
+        return switch (try compare(node_b, offset_b, node_a, offset_a)) {
+            -1 => 1,
+            1 => -1,
+            else => unreachable,
+        };
+    }
+
+    if (position & @intFromEnum(parser.DocumentPosition.contains) == @intFromEnum(parser.DocumentPosition.contains)) {
+        // node_a contains node_b
+        var child = node_b;
+        while (try parser.nodeParentNode(child)) |parent| {
+            if (parent == node_a) {
+                // child.parentNode == node_a
+                break;
+            }
+            child = parent;
+        } else {
+            // this should not happen, because  Node._compareDocumentPosition
+            // has told us that node_a contains node_b, so one of node_b's
+            // parent's MUST be node_a. But somehow we do end up here sometimes.
+            return -1;
+        }
+
+        const child_parent, const child_index = try getParentAndIndex(child);
+        std.debug.assert(node_a == child_parent);
+        return if (child_index < offset_a) -1 else 1;
+    }
+
+    return -1;
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: Range" {
+    try testing.htmlRunner("dom/range.html");
+}

src/browser/dom/resize_observer.zig

@@ -0,0 +1,54 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const Env = @import("../env.zig").Env;
+const parser = @import("../netsurf.zig");
+
+pub const Interfaces = .{
+    ResizeObserver,
+};
+
+// WEB IDL https://drafts.csswg.org/resize-observer/#resize-observer-interface
+pub const ResizeObserver = struct {
+    pub fn constructor(cbk: Env.Function) ResizeObserver {
+        _ = cbk;
+        return .{};
+    }
+
+    pub fn _observe(self: *const ResizeObserver, element: *parser.Element, options_: ?Options) void {
+        _ = self;
+        _ = element;
+        _ = options_;
+        return;
+    }
+
+    pub fn _unobserve(self: *const ResizeObserver, element: *parser.Element) void {
+        _ = self;
+        _ = element;
+        return;
+    }
+
+    // TODO
+    pub fn _disconnect(self: *ResizeObserver) void {
+        _ = self;
+    }
+};
+
+const Options = struct {
+    box: []const u8,
+};

src/browser/dom/shadow_root.zig

@@ -0,0 +1,101 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const dump = @import("../dump.zig");
+const parser = @import("../netsurf.zig");
+
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+const Node = @import("node.zig").Node;
+const Element = @import("element.zig").Element;
+const ElementUnion = @import("element.zig").Union;
+
+// WEB IDL https://dom.spec.whatwg.org/#interface-shadowroot
+pub const ShadowRoot = struct {
+    pub const prototype = *parser.DocumentFragment;
+    pub const subtype = .node;
+
+    mode: Mode,
+    host: *parser.Element,
+    proto: *parser.DocumentFragment,
+    adopted_style_sheets: ?Env.JsObject = null,
+
+    pub const Mode = enum {
+        open,
+        closed,
+    };
+
+    pub fn get_host(self: *const ShadowRoot) !ElementUnion {
+        return Element.toInterface(self.host);
+    }
+
+    pub fn get_adoptedStyleSheets(self: *ShadowRoot, page: *Page) !Env.JsObject {
+        if (self.adopted_style_sheets) |obj| {
+            return obj;
+        }
+
+        const obj = try page.main_context.newArray(0).persist();
+        self.adopted_style_sheets = obj;
+        return obj;
+    }
+
+    pub fn set_adoptedStyleSheets(self: *ShadowRoot, sheets: Env.JsObject) !void {
+        self.adopted_style_sheets = try sheets.persist();
+    }
+
+    pub fn get_innerHTML(self: *ShadowRoot, page: *Page) ![]const u8 {
+        var aw = std.Io.Writer.Allocating.init(page.call_arena);
+        try dump.writeChildren(parser.documentFragmentToNode(self.proto), .{}, &aw.writer);
+        return aw.written();
+    }
+
+    pub fn set_innerHTML(self: *ShadowRoot, str_: ?[]const u8) !void {
+        const sr_doc = parser.documentFragmentToNode(self.proto);
+        const doc = try parser.nodeOwnerDocument(sr_doc) orelse return parser.DOMError.WrongDocument;
+        try Node.removeChildren(sr_doc);
+        const str = str_ orelse return;
+
+        const fragment = try parser.documentParseFragmentFromStr(doc, str);
+        const fragment_node = parser.documentFragmentToNode(fragment);
+
+        // Element.set_innerHTML also has some weirdness here. It isn't clear
+        // what should and shouldn't be set. Whatever string you pass to libdom,
+        // it always creates a full HTML document, with an html, head and body
+        // element.
+        // For ShadowRoot, it appears the only the children within the body should
+        // be set.
+        const html = try parser.nodeFirstChild(fragment_node) orelse return;
+        const head = try parser.nodeFirstChild(html) orelse return;
+        const body = try parser.nodeNextSibling(head) orelse return;
+
+        const children = try parser.nodeGetChildNodes(body);
+        const ln = try parser.nodeListLength(children);
+        for (0..ln) |_| {
+            // always index 0, because nodeAppendChild moves the node out of
+            // the nodeList and into the new tree
+            const child = try parser.nodeListItem(children, 0) orelse continue;
+            _ = try parser.nodeAppendChild(sr_doc, child);
+        }
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.ShadowRoot" {
+    try testing.htmlRunner("dom/shadow_root.html");
+}

src/browser/dom/text.zig

@@ -0,0 +1,62 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const CharacterData = @import("character_data.zig").CharacterData;
+const CDATASection = @import("cdata_section.zig").CDATASection;
+
+// Text interfaces
+pub const Interfaces = .{
+    CDATASection,
+};
+
+pub const Text = struct {
+    pub const Self = parser.Text;
+    pub const prototype = *CharacterData;
+    pub const subtype = .node;
+
+    pub fn constructor(data: ?[]const u8, page: *const Page) !*parser.Text {
+        return parser.documentCreateTextNode(
+            parser.documentHTMLToDocument(page.window.document),
+            data orelse "",
+        );
+    }
+
+    // JS funcs
+    // --------
+
+    // Read attributes
+
+    pub fn get_wholeText(self: *parser.Text) ![]const u8 {
+        return try parser.textWholdeText(self);
+    }
+
+    // JS methods
+    // ----------
+
+    pub fn _splitText(self: *parser.Text, offset: u32) !*parser.Text {
+        return try parser.textSplitText(self, offset);
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.Text" {
+    try testing.htmlRunner("dom/text.html");
+}

src/browser/dom/token_list.zig

@@ -0,0 +1,175 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const iterator = @import("../iterator/iterator.zig");
+
+const Function = @import("../env.zig").Function;
+const JsObject = @import("../env.zig").JsObject;
+const DOMException = @import("exceptions.zig").DOMException;
+
+pub const Interfaces = .{
+    DOMTokenList,
+    DOMTokenListIterable,
+    TokenListEntriesIterator,
+    TokenListEntriesIterator.Iterable,
+};
+
+// https://dom.spec.whatwg.org/#domtokenlist
+pub const DOMTokenList = struct {
+    pub const Self = parser.TokenList;
+    pub const Exception = DOMException;
+
+    pub fn get_length(self: *parser.TokenList) !u32 {
+        return parser.tokenListGetLength(self);
+    }
+
+    pub fn _item(self: *parser.TokenList, index: u32) !?[]const u8 {
+        return parser.tokenListItem(self, index);
+    }
+
+    pub fn _contains(self: *parser.TokenList, token: []const u8) !bool {
+        return parser.tokenListContains(self, token);
+    }
+
+    pub fn _add(self: *parser.TokenList, tokens: []const []const u8) !void {
+        for (tokens) |token| {
+            try parser.tokenListAdd(self, token);
+        }
+    }
+
+    pub fn _remove(self: *parser.TokenList, tokens: []const []const u8) !void {
+        for (tokens) |token| {
+            try parser.tokenListRemove(self, token);
+        }
+    }
+
+    /// If token is the empty string, then throw a "SyntaxError" DOMException.
+    /// If token contains any ASCII whitespace, then throw an
+    /// "InvalidCharacterError" DOMException.
+    fn validateToken(token: []const u8) !void {
+        if (token.len == 0) {
+            return parser.DOMError.Syntax;
+        }
+        for (token) |c| {
+            if (std.ascii.isWhitespace(c)) return parser.DOMError.InvalidCharacter;
+        }
+    }
+
+    pub fn _toggle(self: *parser.TokenList, token: []const u8, force: ?bool) !bool {
+        try validateToken(token);
+        const exists = try parser.tokenListContains(self, token);
+        if (exists) {
+            if (force == null or force.? == false) {
+                try parser.tokenListRemove(self, token);
+                return false;
+            }
+            return true;
+        }
+
+        if (force == null or force.? == true) {
+            try parser.tokenListAdd(self, token);
+            return true;
+        }
+        return false;
+    }
+
+    pub fn _replace(self: *parser.TokenList, token: []const u8, new: []const u8) !bool {
+        try validateToken(token);
+        try validateToken(new);
+        const exists = try parser.tokenListContains(self, token);
+        if (!exists) return false;
+        try parser.tokenListRemove(self, token);
+        try parser.tokenListAdd(self, new);
+        return true;
+    }
+
+    // TODO to implement.
+    pub fn _supports(_: *parser.TokenList, token: []const u8) !bool {
+        try validateToken(token);
+        return error.TypeError;
+    }
+
+    pub fn get_value(self: *parser.TokenList) !?[]const u8 {
+        return (try parser.tokenListGetValue(self)) orelse "";
+    }
+
+    pub fn set_value(self: *parser.TokenList, value: []const u8) !void {
+        return parser.tokenListSetValue(self, value);
+    }
+
+    pub fn _toString(self: *parser.TokenList) ![]const u8 {
+        return (try get_value(self)) orelse "";
+    }
+
+    pub fn _keys(self: *parser.TokenList) !iterator.U32Iterator {
+        return .{ .length = try get_length(self) };
+    }
+
+    pub fn _values(self: *parser.TokenList) DOMTokenListIterable {
+        return DOMTokenListIterable.init(.{ .token_list = self });
+    }
+
+    pub fn _entries(self: *parser.TokenList) TokenListEntriesIterator {
+        return TokenListEntriesIterator.init(.{ .token_list = self });
+    }
+
+    pub fn _symbol_iterator(self: *parser.TokenList) DOMTokenListIterable {
+        return _values(self);
+    }
+
+    // TODO handle thisArg
+    pub fn _forEach(self: *parser.TokenList, cbk: Function, this_arg: JsObject) !void {
+        var entries = _entries(self);
+        while (try entries._next()) |entry| {
+            var result: Function.Result = undefined;
+            cbk.tryCallWithThis(void, this_arg, .{ entry.@"1", entry.@"0", self }, &result) catch {
+                log.debug(.user_script, "callback error", .{
+                    .err = result.exception,
+                    .stack = result.stack,
+                    .soure = "tokenList foreach",
+                });
+            };
+        }
+    }
+};
+
+const DOMTokenListIterable = iterator.Iterable(Iterator, "DOMTokenListIterable");
+const TokenListEntriesIterator = iterator.NumericEntries(Iterator, "TokenListEntriesIterator");
+
+pub const Iterator = struct {
+    index: u32 = 0,
+    token_list: *parser.TokenList,
+
+    // used when wrapped in an iterator.NumericEntries
+    pub const Error = parser.DOMError;
+
+    pub fn _next(self: *Iterator) !?[]const u8 {
+        const index = self.index;
+        self.index = index + 1;
+        return DOMTokenList._item(self.token_list, index);
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: DOM.TokenList" {
+    try testing.htmlRunner("dom/token_list.html");
+}

src/browser/dom/tree_walker.zig

@@ -0,0 +1,281 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const parser = @import("../netsurf.zig");
+
+const NodeFilter = @import("node_filter.zig");
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+const Node = @import("node.zig").Node;
+const NodeUnion = @import("node.zig").Union;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/TreeWalker
+pub const TreeWalker = struct {
+    root: *parser.Node,
+    current_node: *parser.Node,
+    what_to_show: u32,
+    filter: ?TreeWalkerOpts,
+    filter_func: ?Env.Function,
+
+    // One of the few cases where null and undefined resolve to different default.
+    // We need the raw JsObject so that we can probe the tri state:
+    // null, undefined or i32.
+    pub const WhatToShow = Env.JsObject;
+
+    pub const TreeWalkerOpts = union(enum) {
+        function: Env.Function,
+        object: struct { acceptNode: Env.Function },
+    };
+
+    pub fn init(node: *parser.Node, what_to_show_: ?WhatToShow, filter: ?TreeWalkerOpts) !TreeWalker {
+        var filter_func: ?Env.Function = null;
+
+        if (filter) |f| {
+            filter_func = switch (f) {
+                .function => |func| func,
+                .object => |o| o.acceptNode,
+            };
+        }
+
+        var what_to_show: u32 = undefined;
+        if (what_to_show_) |wts| {
+            switch (try wts.triState(TreeWalker, "what_to_show", u32)) {
+                .null => what_to_show = 0,
+                .undefined => what_to_show = NodeFilter.NodeFilter._SHOW_ALL,
+                .value => |v| what_to_show = v,
+            }
+        } else {
+            what_to_show = NodeFilter.NodeFilter._SHOW_ALL;
+        }
+
+        return .{
+            .root = node,
+            .current_node = node,
+            .what_to_show = what_to_show,
+            .filter = filter,
+            .filter_func = filter_func,
+        };
+    }
+
+    pub fn get_root(self: *TreeWalker) !NodeUnion {
+        return try Node.toInterface(self.root);
+    }
+
+    pub fn get_currentNode(self: *TreeWalker) !NodeUnion {
+        return try Node.toInterface(self.current_node);
+    }
+
+    pub fn get_whatToShow(self: *TreeWalker) u32 {
+        return self.what_to_show;
+    }
+
+    pub fn get_filter(self: *TreeWalker) ?TreeWalkerOpts {
+        return self.filter;
+    }
+
+    pub fn set_currentNode(self: *TreeWalker, node: *parser.Node) !void {
+        self.current_node = node;
+    }
+
+    fn firstChild(self: *const TreeWalker, node: *parser.Node) !?*parser.Node {
+        const children = try parser.nodeGetChildNodes(node);
+        const child_count = try parser.nodeListLength(children);
+
+        for (0..child_count) |i| {
+            const index: u32 = @intCast(i);
+            const child = (try parser.nodeListItem(children, index)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, child)) {
+                .accept => return child,
+                .reject => continue,
+                .skip => if (try self.firstChild(child)) |gchild| return gchild,
+            }
+        }
+
+        return null;
+    }
+
+    fn lastChild(self: *const TreeWalker, node: *parser.Node) !?*parser.Node {
+        const children = try parser.nodeGetChildNodes(node);
+        const child_count = try parser.nodeListLength(children);
+
+        var index: u32 = child_count;
+        while (index > 0) {
+            index -= 1;
+            const child = (try parser.nodeListItem(children, index)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, child)) {
+                .accept => return child,
+                .reject => continue,
+                .skip => if (try self.lastChild(child)) |gchild| return gchild,
+            }
+        }
+
+        return null;
+    }
+
+    fn nextSibling(self: *const TreeWalker, node: *parser.Node) !?*parser.Node {
+        var current = node;
+
+        while (true) {
+            current = (try parser.nodeNextSibling(current)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => return current,
+                .skip, .reject => continue,
+            }
+        }
+
+        return null;
+    }
+
+    fn previousSibling(self: *const TreeWalker, node: *parser.Node) !?*parser.Node {
+        var current = node;
+
+        while (true) {
+            current = (try parser.nodePreviousSibling(current)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => return current,
+                .skip, .reject => continue,
+            }
+        }
+
+        return null;
+    }
+
+    fn parentNode(self: *const TreeWalker, node: *parser.Node) !?*parser.Node {
+        if (self.root == node) return null;
+
+        var current = node;
+        while (true) {
+            if (current == self.root) return null;
+            current = (try parser.nodeParentNode(current)) orelse return null;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => return current,
+                .reject, .skip => continue,
+            }
+        }
+    }
+
+    pub fn _firstChild(self: *TreeWalker) !?NodeUnion {
+        if (try self.firstChild(self.current_node)) |child| {
+            self.current_node = child;
+            return try Node.toInterface(child);
+        }
+
+        return null;
+    }
+
+    pub fn _lastChild(self: *TreeWalker) !?NodeUnion {
+        if (try self.lastChild(self.current_node)) |child| {
+            self.current_node = child;
+            return try Node.toInterface(child);
+        }
+
+        return null;
+    }
+
+    pub fn _nextNode(self: *TreeWalker) !?NodeUnion {
+        if (try self.firstChild(self.current_node)) |child| {
+            self.current_node = child;
+            return try Node.toInterface(child);
+        }
+
+        var current = self.current_node;
+        while (current != self.root) {
+            if (try self.nextSibling(current)) |sibling| {
+                self.current_node = sibling;
+                return try Node.toInterface(sibling);
+            }
+
+            current = (try parser.nodeParentNode(current)) orelse break;
+        }
+
+        return null;
+    }
+
+    pub fn _nextSibling(self: *TreeWalker) !?NodeUnion {
+        if (try self.nextSibling(self.current_node)) |sibling| {
+            self.current_node = sibling;
+            return try Node.toInterface(sibling);
+        }
+
+        return null;
+    }
+
+    pub fn _parentNode(self: *TreeWalker) !?NodeUnion {
+        if (try self.parentNode(self.current_node)) |parent| {
+            self.current_node = parent;
+            return try Node.toInterface(parent);
+        }
+
+        return null;
+    }
+
+    pub fn _previousNode(self: *TreeWalker) !?NodeUnion {
+        if (self.current_node == self.root) return null;
+
+        var current = self.current_node;
+        while (try parser.nodePreviousSibling(current)) |previous| {
+            current = previous;
+
+            switch (try NodeFilter.verify(self.what_to_show, self.filter_func, current)) {
+                .accept => {
+                    // Get last child if it has one.
+                    if (try self.lastChild(current)) |child| {
+                        self.current_node = child;
+                        return try Node.toInterface(child);
+                    }
+
+                    // Otherwise, this node is our previous one.
+                    self.current_node = current;
+                    return try Node.toInterface(current);
+                },
+                .reject => continue,
+                .skip => {
+                    // Get last child if it has one.
+                    if (try self.lastChild(current)) |child| {
+                        self.current_node = child;
+                        return try Node.toInterface(child);
+                    }
+                },
+            }
+        }
+
+        if (current != self.root) {
+            if (try self.parentNode(current)) |parent| {
+                self.current_node = parent;
+                return try Node.toInterface(parent);
+            }
+        }
+
+        return null;
+    }
+
+    pub fn _previousSibling(self: *TreeWalker) !?NodeUnion {
+        if (try self.previousSibling(self.current_node)) |sibling| {
+            self.current_node = sibling;
+            return try Node.toInterface(sibling);
+        }
+
+        return null;
+    }
+};

src/browser/dom/walker.zig

@@ -1,4 +1,20 @@
-const std = @import("std");
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
 
 const parser = @import("../netsurf.zig");
 

src/browser/dump.zig

@@ -1,96 +1,283 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
 const std = @import("std");
-const File = std.fs.File;
 
-const parser = @import("../netsurf.zig");
-const Walker = @import("../dom/walker.zig").WalkerChildren;
+const parser = @import("netsurf.zig");
+const Page = @import("page.zig").Page;
+const Walker = @import("dom/walker.zig").WalkerChildren;
 
-pub fn htmlFile(doc: *parser.Document, out: File) !void {
-    try out.writeAll("<!DOCTYPE html>\n");
-    try nodeFile(parser.documentToNode(doc), out);
-    try out.writeAll("\n");
+pub const Opts = struct {
+    // set to include element shadowroots in the dump
+    page: ?*const Page = null,
+
+    exclude_scripts: bool = false,
+};
+
+// writer must be a std.io.Writer
+pub fn writeHTML(doc: *parser.Document, opts: Opts, writer: *std.Io.Writer) !void {
+    try writer.writeAll("<!DOCTYPE html>\n");
+    try writeChildren(parser.documentToNode(doc), opts, writer);
+    try writer.writeAll("\n");
 }
 
-fn nodeFile(root: *parser.Node, out: File) !void {
+// Spec: https://www.w3.org/TR/xml/#sec-prolog-dtd
+pub fn writeDocType(doc_type: *parser.DocumentType, writer: *std.Io.Writer) !void {
+    try writer.writeAll("<!DOCTYPE ");
+    try writer.writeAll(try parser.documentTypeGetName(doc_type));
+
+    const public_id = try parser.documentTypeGetPublicId(doc_type);
+    const system_id = try parser.documentTypeGetSystemId(doc_type);
+    if (public_id.len != 0 and system_id.len != 0) {
+        try writer.writeAll(" PUBLIC \"");
+        try writeEscapedAttributeValue(writer, public_id);
+        try writer.writeAll("\" \"");
+        try writeEscapedAttributeValue(writer, system_id);
+        try writer.writeAll("\"");
+    } else if (public_id.len != 0) {
+        try writer.writeAll(" PUBLIC \"");
+        try writeEscapedAttributeValue(writer, public_id);
+        try writer.writeAll("\"");
+    } else if (system_id.len != 0) {
+        try writer.writeAll(" SYSTEM \"");
+        try writeEscapedAttributeValue(writer, system_id);
+        try writer.writeAll("\"");
+    }
+    // Internal subset is not implemented
+    try writer.writeAll(">");
+}
+
+pub fn writeNode(node: *parser.Node, opts: Opts, writer: *std.Io.Writer) anyerror!void {
+    switch (try parser.nodeType(node)) {
+        .element => {
+            // open the tag
+            const tag_type = try parser.nodeHTMLGetTagType(node) orelse .undef;
+            if (opts.exclude_scripts and try isScriptOrRelated(tag_type, node)) {
+                return;
+            }
+
+            const tag = try parser.nodeLocalName(node);
+            try writer.writeAll("<");
+            try writer.writeAll(tag);
+
+            // write the attributes
+            const _map = try parser.nodeGetAttributes(node);
+            if (_map) |map| {
+                const ln = try parser.namedNodeMapGetLength(map);
+                for (0..ln) |i| {
+                    const attr = try parser.namedNodeMapItem(map, @intCast(i)) orelse break;
+                    try writer.writeAll(" ");
+                    try writer.writeAll(try parser.attributeGetName(attr));
+                    try writer.writeAll("=\"");
+                    const attribute_value = try parser.attributeGetValue(attr) orelse "";
+                    try writeEscapedAttributeValue(writer, attribute_value);
+                    try writer.writeAll("\"");
+                }
+            }
+
+            try writer.writeAll(">");
+
+            if (opts.page) |page| {
+                if (page.getNodeState(node)) |state| {
+                    if (state.shadow_root) |sr| {
+                        try writeChildren(@ptrCast(@alignCast(sr.proto)), opts, writer);
+                    }
+                }
+            }
+
+            // void elements can't have any content.
+            if (try isVoid(parser.nodeToElement(node))) return;
+
+            if (tag_type == .script) {
+                try writer.writeAll(try parser.nodeTextContent(node) orelse "");
+            } else {
+                // write the children
+                // TODO avoid recursion
+                try writeChildren(node, opts, writer);
+            }
+
+            // close the tag
+            try writer.writeAll("</");
+            try writer.writeAll(tag);
+            try writer.writeAll(">");
+        },
+        .text => {
+            const v = try parser.nodeValue(node) orelse return;
+            try writeEscapedTextNode(writer, v);
+        },
+        .cdata_section => {
+            const v = try parser.nodeValue(node) orelse return;
+            try writer.writeAll("<![CDATA[");
+            try writer.writeAll(v);
+            try writer.writeAll("]]>");
+        },
+        .comment => {
+            const v = try parser.nodeValue(node) orelse return;
+            try writer.writeAll("<!--");
+            try writer.writeAll(v);
+            try writer.writeAll("-->");
+        },
+        // TODO handle processing instruction dump
+        .processing_instruction => return,
+        // document fragment is outside of the main document DOM, so we
+        // don't output it.
+        .document_fragment => return,
+        // document will never be called, but required for completeness.
+        .document => return,
+        // done globally instead, but required for completeness. Only the outer DOCTYPE should be written
+        .document_type => return,
+        // deprecated
+        .attribute => return,
+        .entity_reference => return,
+        .entity => return,
+        .notation => return,
+    }
+}
+
+// writer must be a std.io.Writer
+pub fn writeChildren(root: *parser.Node, opts: Opts, writer: *std.Io.Writer) !void {
     const walker = Walker{};
     var next: ?*parser.Node = null;
     while (true) {
         next = try walker.get_next(root, next) orelse break;
-        switch (try parser.nodeType(next.?)) {
-            .element => {
-                // open the tag
-                const tag = try parser.nodeLocalName(next.?);
-                try out.writeAll("<");
-                try out.writeAll(tag);
-
-                // write the attributes
-                const map = try parser.nodeGetAttributes(next.?);
-                const ln = try parser.namedNodeMapGetLength(map);
-                var i: u32 = 0;
-                while (i < ln) {
-                    const attr = try parser.namedNodeMapItem(map, i) orelse break;
-                    try out.writeAll(" ");
-                    try out.writeAll(try parser.attributeGetName(attr));
-                    try out.writeAll("=\"");
-                    try out.writeAll(try parser.attributeGetValue(attr) orelse "");
-                    try out.writeAll("\"");
-                    i += 1;
-                }
-
-                try out.writeAll(">");
-
-                // write the children
-                // TODO avoid recursion
-                try nodeFile(next.?, out);
-
-                // close the tag
-                try out.writeAll("</");
-                try out.writeAll(tag);
-                try out.writeAll(">");
-            },
-            .text => {
-                const v = try parser.nodeValue(next.?) orelse continue;
-                try out.writeAll(v);
-            },
-            .cdata_section => {
-                const v = try parser.nodeValue(next.?) orelse continue;
-                try out.writeAll("<![CDATA[");
-                try out.writeAll(v);
-                try out.writeAll("]]>");
-            },
-            .comment => {
-                const v = try parser.nodeValue(next.?) orelse continue;
-                try out.writeAll("<!--");
-                try out.writeAll(v);
-                try out.writeAll("-->");
-            },
-            // TODO handle processing instruction dump
-            .processing_instruction => continue,
-            // document fragment is outside of the main document DOM, so we
-            // don't output it.
-            .document_fragment => continue,
-            // document will never be called, but required for completeness.
-            .document => continue,
-            // done globally instead, but required for completeness.
-            .document_type => continue,
-            // deprecated
-            .attribute => continue,
-            .entity_reference => continue,
-            .entity => continue,
-            .notation => continue,
-        }
+        try writeNode(next.?, opts, writer);
     }
 }
 
-// HTMLFileTestFn is run by run_tests.zig
-pub fn HTMLFileTestFn(out: File) !void {
-    const file = try std.fs.cwd().openFile("test.html", .{});
-    defer file.close();
+// When `exclude_scripts` is passed to dump, we don't include <script> tags.
+// We also want to omit <link rel=preload as=ascript>
+fn isScriptOrRelated(tag_type: parser.Tag, node: *parser.Node) !bool {
+    if (tag_type == .script) {
+        return true;
+    }
+    if (tag_type == .link) {
+        const el = parser.nodeToElement(node);
+        const as = try parser.elementGetAttribute(el, "as") orelse return false;
+        if (!std.ascii.eqlIgnoreCase(as, "script")) {
+            return false;
+        }
 
-    const doc_html = try parser.documentHTMLParse(file.reader(), "UTF-8");
-    // ignore close error
+        const rel = try parser.elementGetAttribute(el, "rel") orelse return false;
+        return std.ascii.eqlIgnoreCase(rel, "preload");
+    }
+    return false;
+}
+
+// area, base, br, col, embed, hr, img, input, link, meta, source, track, wbr
+// https://html.spec.whatwg.org/#void-elements
+fn isVoid(elem: *parser.Element) !bool {
+    const tag = try parser.elementTag(elem);
+    return switch (tag) {
+        .area, .base, .br, .col, .embed, .hr, .img, .input, .link => true,
+        .meta, .source, .track, .wbr => true,
+        else => false,
+    };
+}
+
+fn writeEscapedTextNode(writer: anytype, value: []const u8) !void {
+    var v = value;
+    while (v.len > 0) {
+        const index = std.mem.indexOfAnyPos(u8, v, 0, &.{ '&', '<', '>' }) orelse {
+            return writer.writeAll(v);
+        };
+        try writer.writeAll(v[0..index]);
+        switch (v[index]) {
+            '&' => try writer.writeAll("&amp;"),
+            '<' => try writer.writeAll("&lt;"),
+            '>' => try writer.writeAll("&gt;"),
+            else => unreachable,
+        }
+        v = v[index + 1 ..];
+    }
+}
+
+fn writeEscapedAttributeValue(writer: anytype, value: []const u8) !void {
+    var v = value;
+    while (v.len > 0) {
+        const index = std.mem.indexOfAnyPos(u8, v, 0, &.{ '&', '<', '>', '"' }) orelse {
+            return writer.writeAll(v);
+        };
+        try writer.writeAll(v[0..index]);
+        switch (v[index]) {
+            '&' => try writer.writeAll("&amp;"),
+            '<' => try writer.writeAll("&lt;"),
+            '>' => try writer.writeAll("&gt;"),
+            '"' => try writer.writeAll("&quot;"),
+            else => unreachable,
+        }
+        v = v[index + 1 ..];
+    }
+}
+
+const testing = std.testing;
+test "dump.writeHTML" {
+    try parser.init();
+    defer parser.deinit();
+
+    try testWriteHTML(
+        "<div id=\"content\">Over 9000!</div>",
+        "<div id=\"content\">Over 9000!</div>",
+    );
+
+    try testWriteHTML(
+        "<root><!-- a comment --></root>",
+        "<root><!-- a comment --></root>",
+    );
+
+    try testWriteHTML(
+        "<p>&lt; &gt; &amp;</p>",
+        "<p>&lt; &gt; &amp;</p>",
+    );
+
+    try testWriteHTML(
+        "<p id=\"&quot;&gt;&lt;&amp;&quot;''\">wat?</p>",
+        "<p id='\">&lt;&amp;&quot;&#39;&apos;'>wat?</p>",
+    );
+
+    try testWriteFullHTML(
+        \\<!DOCTYPE html>
+        \\<html><head><title>It's over what?</title><meta name="a" value="b">
+        \\</head><body>9000</body></html>
+        \\
+    , "<html><title>It's over what?</title><meta name=a value=\"b\">\n<body>9000");
+
+    try testWriteHTML(
+        "<p>hi</p><script>alert(power > 9000)</script>",
+        "<p>hi</p><script>alert(power > 9000)</script>",
+    );
+}
+
+fn testWriteHTML(comptime expected_body: []const u8, src: []const u8) !void {
+    const expected =
+        "<!DOCTYPE html>\n<html><head></head><body>" ++
+        expected_body ++
+        "</body></html>\n";
+    return testWriteFullHTML(expected, src);
+}
+
+fn testWriteFullHTML(comptime expected: []const u8, src: []const u8) !void {
+    var aw = std.Io.Writer.Allocating.init(testing.allocator);
+    defer aw.deinit();
+
+    const doc_html = try parser.documentHTMLParseFromStr(src);
     defer parser.documentHTMLClose(doc_html) catch {};
 
     const doc = parser.documentHTMLToDocument(doc_html);
-
-    try htmlFile(doc, out);
+    try writeHTML(doc, .{}, &aw.writer);
+    try testing.expectEqualStrings(expected, aw.written());
 }

src/browser/encoding/TextDecoder.zig

@@ -0,0 +1,103 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+
+// https://encoding.spec.whatwg.org/#interface-textdecoder
+const TextDecoder = @This();
+
+const SupportedLabels = enum {
+    utf8,
+    @"utf-8",
+    @"unicode-1-1-utf-8",
+};
+
+const Options = struct {
+    fatal: bool = false,
+    ignoreBOM: bool = false,
+};
+
+fatal: bool,
+ignore_bom: bool,
+stream: std.ArrayList(u8),
+
+pub fn constructor(label_: ?[]const u8, opts_: ?Options) !TextDecoder {
+    if (label_) |l| {
+        _ = std.meta.stringToEnum(SupportedLabels, l) orelse {
+            log.warn(.web_api, "not implemented", .{ .feature = "TextDecoder label", .label = l });
+            return error.NotImplemented;
+        };
+    }
+    const opts = opts_ orelse Options{};
+    return .{
+        .stream = .empty,
+        .fatal = opts.fatal,
+        .ignore_bom = opts.ignoreBOM,
+    };
+}
+
+pub fn get_encoding(_: *const TextDecoder) []const u8 {
+    return "utf-8";
+}
+
+pub fn get_ignoreBOM(self: *const TextDecoder) bool {
+    return self.ignore_bom;
+}
+
+pub fn get_fatal(self: *const TextDecoder) bool {
+    return self.fatal;
+}
+
+const DecodeOptions = struct {
+    stream: bool = false,
+};
+pub fn _decode(self: *TextDecoder, input_: ?[]const u8, opts_: ?DecodeOptions, page: *Page) ![]const u8 {
+    var str = input_ orelse return "";
+    const opts: DecodeOptions = opts_ orelse .{};
+
+    if (self.stream.items.len > 0) {
+        try self.stream.appendSlice(page.arena, str);
+        str = self.stream.items;
+    }
+
+    if (self.fatal and !std.unicode.utf8ValidateSlice(str)) {
+        if (opts.stream) {
+            if (self.stream.items.len == 0) {
+                try self.stream.appendSlice(page.arena, str);
+            }
+            return "";
+        }
+        return error.InvalidUtf8;
+    }
+
+    self.stream.clearRetainingCapacity();
+    if (self.ignore_bom == false and std.mem.startsWith(u8, str, &.{ 0xEF, 0xBB, 0xBF })) {
+        return str[3..];
+    }
+
+    return str;
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: Encoding.TextDecoder" {
+    try testing.htmlRunner("encoding/decoder.html");
+}

src/browser/encoding/TextEncoder.zig

@@ -0,0 +1,48 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const Env = @import("../env.zig").Env;
+
+// https://encoding.spec.whatwg.org/#interface-textencoder
+const TextEncoder = @This();
+
+pub fn constructor() !TextEncoder {
+    return .{};
+}
+
+pub fn get_encoding(_: *const TextEncoder) []const u8 {
+    return "utf-8";
+}
+
+pub fn _encode(_: *const TextEncoder, v: []const u8) !Env.TypedArray(u8) {
+    // Ensure the input is a valid utf-8
+    // It seems chrome accepts invalid utf-8 sequence.
+    //
+    if (!std.unicode.utf8ValidateSlice(v)) {
+        return error.InvalidUtf8;
+    }
+
+    return .{ .values = v };
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: Encoding.TextEncoder" {
+    try testing.htmlRunner("encoding/encoder.html");
+}

src/browser/encoding/encoding.zig

@@ -0,0 +1,22 @@
+// Copyright (C) 2023-2025  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+pub const Interfaces = .{
+    @import("TextDecoder.zig"),
+    @import("TextEncoder.zig"),
+};

src/browser/env.zig

@@ -0,0 +1,51 @@
+const std = @import("std");
+
+const Page = @import("page.zig").Page;
+const js = @import("../runtime/js.zig");
+const generate = @import("../runtime/generate.zig");
+
+const WebApis = struct {
+    // Wrapped like this for debug ergonomics.
+    // When we create our Env, a few lines down, we define it as:
+    //   pub const Env = js.Env(*Page, WebApis);
+    //
+    // If there's a compile time error witht he Env, it's type will be readable,
+    // i.e.: runtime.js.Env(*browser.env.Page, browser.env.WebApis)
+    //
+    // But if we didn't wrap it in the struct, like we once didn't, and defined
+    // env as:
+    //   pub const Env = js.Env(*Page, Interfaces);
+    //
+    // Because Interfaces is an anynoumous type, it doesn't have a friendly name
+    // and errors would be something like:
+    //   runtime.js.Env(*browser.Page, .{...A HUNDRED TYPES...})
+    pub const Interfaces = generate.Tuple(.{
+        @import("crypto/crypto.zig").Crypto,
+        @import("console/console.zig").Console,
+        @import("css/css.zig").Interfaces,
+        @import("cssom/cssom.zig").Interfaces,
+        @import("dom/dom.zig").Interfaces,
+        @import("dom/shadow_root.zig").ShadowRoot,
+        @import("encoding/encoding.zig").Interfaces,
+        @import("events/event.zig").Interfaces,
+        @import("html/html.zig").Interfaces,
+        @import("iterator/iterator.zig").Interfaces,
+        @import("storage/storage.zig").Interfaces,
+        @import("url/url.zig").Interfaces,
+        @import("xhr/xhr.zig").Interfaces,
+        @import("xhr/form_data.zig").Interfaces,
+        @import("xhr/File.zig"),
+        @import("xmlserializer/xmlserializer.zig").Interfaces,
+        @import("fetch/fetch.zig").Interfaces,
+        @import("streams/streams.zig").Interfaces,
+    });
+};
+
+pub const JsThis = Env.JsThis;
+pub const JsObject = Env.JsObject;
+pub const Function = Env.Function;
+pub const Promise = Env.Promise;
+pub const PromiseResolver = Env.PromiseResolver;
+
+pub const Env = js.Env(*Page, WebApis);
+pub const Global = @import("html/window.zig").Window;

src/browser/events/custom_event.zig

@@ -0,0 +1,85 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const parser = @import("../netsurf.zig");
+const Event = @import("event.zig").Event;
+const JsObject = @import("../env.zig").JsObject;
+const netsurf = @import("../netsurf.zig");
+
+// https://dom.spec.whatwg.org/#interface-customevent
+pub const CustomEvent = struct {
+    pub const prototype = *Event;
+    pub const union_make_copy = true;
+
+    proto: parser.Event,
+    detail: ?JsObject,
+
+    const CustomEventInit = struct {
+        bubbles: bool = false,
+        cancelable: bool = false,
+        composed: bool = false,
+        detail: ?JsObject = null,
+    };
+
+    pub fn constructor(event_type: []const u8, opts_: ?CustomEventInit) !CustomEvent {
+        const opts = opts_ orelse CustomEventInit{};
+
+        const event = try parser.eventCreate();
+        defer parser.eventDestroy(event);
+        try parser.eventInit(event, event_type, .{
+            .bubbles = opts.bubbles,
+            .cancelable = opts.cancelable,
+            .composed = opts.composed,
+        });
+
+        return .{
+            .proto = event.*,
+            .detail = if (opts.detail) |d| try d.persist() else null,
+        };
+    }
+
+    pub fn get_detail(self: *CustomEvent) ?JsObject {
+        return self.detail;
+    }
+
+    // Initializes an already created `CustomEvent`.
+    // https://developer.mozilla.org/en-US/docs/Web/API/CustomEvent/initCustomEvent
+    pub fn _initCustomEvent(
+        self: *CustomEvent,
+        event_type: []const u8,
+        can_bubble: bool,
+        cancelable: bool,
+        maybe_detail: ?JsObject,
+    ) !void {
+        // This function can only be called after the constructor has called.
+        // So we assume proto is initialized already by constructor.
+        self.proto.type = try netsurf.strFromData(event_type);
+        self.proto.bubble = can_bubble;
+        self.proto.cancelable = cancelable;
+        self.proto.is_initialised = true;
+        // Detail is stored separately.
+        if (maybe_detail) |detail| {
+            self.detail = try detail.persist();
+        }
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: Events.Custom" {
+    try testing.htmlRunner("events/custom.html");
+}

src/browser/events/event.zig

@@ -0,0 +1,400 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const Allocator = std.mem.Allocator;
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const generate = @import("../../runtime/generate.zig");
+
+const Page = @import("../page.zig").Page;
+const Node = @import("../dom/node.zig").Node;
+const DOMException = @import("../dom/exceptions.zig").DOMException;
+const EventTarget = @import("../dom/event_target.zig").EventTarget;
+const EventTargetUnion = @import("../dom/event_target.zig").Union;
+const AbortSignal = @import("../html/AbortController.zig").AbortSignal;
+
+const CustomEvent = @import("custom_event.zig").CustomEvent;
+const ProgressEvent = @import("../xhr/progress_event.zig").ProgressEvent;
+const MouseEvent = @import("mouse_event.zig").MouseEvent;
+const KeyboardEvent = @import("keyboard_event.zig").KeyboardEvent;
+const ErrorEvent = @import("../html/error_event.zig").ErrorEvent;
+const MessageEvent = @import("../dom/MessageChannel.zig").MessageEvent;
+
+// Event interfaces
+pub const Interfaces = .{
+    Event,
+    CustomEvent,
+    ProgressEvent,
+    MouseEvent,
+    KeyboardEvent,
+    ErrorEvent,
+    MessageEvent,
+};
+
+pub const Union = generate.Union(Interfaces);
+
+// https://dom.spec.whatwg.org/#event
+pub const Event = struct {
+    pub const Self = parser.Event;
+    pub const Exception = DOMException;
+
+    pub const EventInit = parser.EventInit;
+
+    // JS
+    // --
+
+    pub const _CAPTURING_PHASE = 1;
+    pub const _AT_TARGET = 2;
+    pub const _BUBBLING_PHASE = 3;
+
+    pub fn toInterface(evt: *parser.Event) Union {
+        return switch (parser.eventGetInternalType(evt)) {
+            .event, .abort_signal, .xhr_event => .{ .Event = evt },
+            .custom_event => .{ .CustomEvent = @as(*CustomEvent, @ptrCast(evt)).* },
+            .progress_event => .{ .ProgressEvent = @as(*ProgressEvent, @ptrCast(evt)).* },
+            .mouse_event => .{ .MouseEvent = @as(*parser.MouseEvent, @ptrCast(evt)) },
+            .error_event => .{ .ErrorEvent = @as(*ErrorEvent, @ptrCast(evt)).* },
+            .message_event => .{ .MessageEvent = @as(*MessageEvent, @ptrCast(evt)).* },
+            .keyboard_event => .{ .KeyboardEvent = @as(*parser.KeyboardEvent, @ptrCast(evt)) },
+        };
+    }
+
+    pub fn constructor(event_type: []const u8, opts: ?EventInit) !*parser.Event {
+        const event = try parser.eventCreate();
+        try parser.eventInit(event, event_type, opts orelse EventInit{});
+        return event;
+    }
+
+    // Getters
+
+    pub fn get_type(self: *parser.Event) ![]const u8 {
+        return try parser.eventType(self);
+    }
+
+    pub fn get_target(self: *parser.Event, page: *Page) !?EventTargetUnion {
+        const et = parser.eventTarget(self);
+        if (et == null) return null;
+        return try EventTarget.toInterface(et.?, page);
+    }
+
+    pub fn get_currentTarget(self: *parser.Event, page: *Page) !?EventTargetUnion {
+        const et = parser.eventCurrentTarget(self);
+        if (et == null) return null;
+        return try EventTarget.toInterface(et.?, page);
+    }
+
+    pub fn get_eventPhase(self: *parser.Event) u8 {
+        return parser.eventPhase(self);
+    }
+
+    pub fn get_bubbles(self: *parser.Event) bool {
+        return parser.eventBubbles(self);
+    }
+
+    pub fn get_cancelable(self: *parser.Event) bool {
+        return parser.eventCancelable(self);
+    }
+
+    pub fn get_defaultPrevented(self: *parser.Event) bool {
+        return parser.eventDefaultPrevented(self);
+    }
+
+    pub fn get_isTrusted(self: *parser.Event) bool {
+        return parser.eventIsTrusted(self);
+    }
+
+    // Even though this is supposed to to provide microsecond resolution, browser
+    // return coarser values to protect against fingerprinting. libdom returns
+    // seconds, which is good enough.
+    pub fn get_timeStamp(self: *parser.Event) u64 {
+        return parser.eventTimestamp(self);
+    }
+
+    // Methods
+
+    pub fn _initEvent(
+        self: *parser.Event,
+        eventType: []const u8,
+        bubbles: ?bool,
+        cancelable: ?bool,
+    ) !void {
+        const opts = EventInit{
+            .bubbles = bubbles orelse false,
+            .cancelable = cancelable orelse false,
+        };
+        return try parser.eventInit(self, eventType, opts);
+    }
+
+    pub fn _stopPropagation(self: *parser.Event) !void {
+        return parser.eventStopPropagation(self);
+    }
+
+    pub fn _stopImmediatePropagation(self: *parser.Event) !void {
+        return parser.eventStopImmediatePropagation(self);
+    }
+
+    pub fn _preventDefault(self: *parser.Event) !void {
+        return parser.eventPreventDefault(self);
+    }
+
+    pub fn _composedPath(self: *parser.Event, page: *Page) ![]const EventTargetUnion {
+        const et_ = parser.eventTarget(self);
+        const et = et_ orelse return &.{};
+
+        var node: ?*parser.Node = switch (try parser.eventTargetInternalType(et)) {
+            .libdom_node => @as(*parser.Node, @ptrCast(et)),
+            .plain => parser.eventTargetToNode(et),
+            else => {
+                // Window, XHR, MessagePort, etc...no path beyond the event itself
+                return &.{try EventTarget.toInterface(et, page)};
+            },
+        };
+
+        const arena = page.call_arena;
+        var path: std.ArrayListUnmanaged(EventTargetUnion) = .empty;
+        while (node) |n| {
+            try path.append(arena, .{
+                .node = try Node.toInterface(n),
+            });
+
+            node = try parser.nodeParentNode(n);
+            if (node == null and try parser.nodeType(n) == .document_fragment) {
+                // we have a non-continuous hook from a shadowroot to its host (
+                // it's parent element). libdom doesn't really support ShdowRoots
+                // and, for the most part, that works out well since it naturally
+                // provides isolation. But events don't follow the same
+                // shadowroot isolation as most other things, so, if this is
+                // a parent-less document fragment, we need to check if it has
+                // a host.
+                if (parser.documentFragmentGetHost(@ptrCast(n))) |host| {
+                    node = host;
+
+                    // If a document fragment has a host, then that host
+                    // _has_ to have a state and that state _has_ to have
+                    // a shadow_root field. All of this is set in Element._attachShadow
+                    if (page.getNodeState(host).?.shadow_root.?.mode == .closed) {
+                        // if the shadow root is closed, then the composedPath
+                        // starts at the host element.
+                        path.clearRetainingCapacity();
+                    }
+                } else {
+                    // Our document fragement has no parent and no host, we
+                    // can break out of the loop.
+                    break;
+                }
+            }
+        }
+
+        if (path.getLastOrNull()) |last| {
+            // the Window isn't part of the DOM hierarchy, but for events, it
+            // is, so we need to glue it on.
+            if (last.node == .HTMLDocument and last.node.HTMLDocument == page.window.document) {
+                try path.append(arena, .{ .node = .{ .Window = &page.window } });
+            }
+        }
+        return path.items;
+    }
+};
+
+pub const EventHandler = struct {
+    once: bool,
+    capture: bool,
+    callback: Function,
+    node: parser.EventNode,
+    listener: *parser.EventListener,
+
+    const Env = @import("../env.zig").Env;
+    const Function = Env.Function;
+
+    pub const Listener = union(enum) {
+        function: Function,
+        object: Env.JsObject,
+
+        pub fn callback(self: Listener, target: *parser.EventTarget) !?Function {
+            return switch (self) {
+                .function => |func| try func.withThis(target),
+                .object => |obj| blk: {
+                    const func = (try obj.getFunction("handleEvent")) orelse return null;
+                    break :blk try func.withThis(try obj.persist());
+                },
+            };
+        }
+    };
+
+    pub const Opts = union(enum) {
+        flags: Flags,
+        capture: bool,
+
+        const Flags = struct {
+            once: ?bool,
+            capture: ?bool,
+            // We ignore this property. It seems to be largely used to help the
+            // browser make certain performance tweaks (i.e. the browser knows
+            // that the listener won't call preventDefault() and thus can safely
+            // run the default as needed).
+            passive: ?bool,
+            signal: ?*AbortSignal, // currently does nothing
+        };
+    };
+
+    pub fn register(
+        allocator: Allocator,
+        target: *parser.EventTarget,
+        typ: []const u8,
+        listener: Listener,
+        opts_: ?Opts,
+    ) !?*EventHandler {
+        var once = false;
+        var capture = false;
+        var signal: ?*AbortSignal = null;
+
+        if (opts_) |opts| {
+            switch (opts) {
+                .capture => |c| capture = c,
+                .flags => |f| {
+                    once = f.once orelse false;
+                    signal = f.signal orelse null;
+                    capture = f.capture orelse false;
+                },
+            }
+        }
+        const callback = (try listener.callback(target)) orelse return null;
+
+        if (signal) |s| {
+            const signal_target = parser.toEventTarget(AbortSignal, s);
+
+            const scb = try allocator.create(SignalCallback);
+            scb.* = .{
+                .target = target,
+                .capture = capture,
+                .callback_id = callback.id,
+                .typ = try allocator.dupe(u8, typ),
+                .signal_target = signal_target,
+                .signal_listener = undefined,
+                .node = .{ .func = SignalCallback.handle },
+            };
+
+            scb.signal_listener = try parser.eventTargetAddEventListener(
+                signal_target,
+                "abort",
+                &scb.node,
+                false,
+            );
+        }
+
+        // check if event target has already this listener
+        if (try parser.eventTargetHasListener(target, typ, capture, callback.id) != null) {
+            return null;
+        }
+
+        const eh = try allocator.create(EventHandler);
+        eh.* = .{
+            .once = once,
+            .capture = capture,
+            .callback = callback,
+            .node = .{
+                .id = callback.id,
+                .func = handle,
+            },
+            .listener = undefined,
+        };
+
+        eh.listener = try parser.eventTargetAddEventListener(
+            target,
+            typ,
+            &eh.node,
+            capture,
+        );
+        return eh;
+    }
+
+    fn handle(node: *parser.EventNode, event: *parser.Event) void {
+        const ievent = Event.toInterface(event);
+        const self: *EventHandler = @fieldParentPtr("node", node);
+        var result: Function.Result = undefined;
+        self.callback.tryCall(void, .{ievent}, &result) catch {
+            log.debug(.user_script, "callback error", .{
+                .err = result.exception,
+                .stack = result.stack,
+                .source = "event handler",
+            });
+        };
+
+        if (self.once) {
+            const target = parser.eventTarget(event).?;
+            const typ = parser.eventType(event) catch return;
+            parser.eventTargetRemoveEventListener(
+                target,
+                typ,
+                self.listener,
+                self.capture,
+            ) catch {};
+        }
+    }
+};
+
+const SignalCallback = struct {
+    typ: []const u8,
+    capture: bool,
+    callback_id: usize,
+    node: parser.EventNode,
+    target: *parser.EventTarget,
+    signal_target: *parser.EventTarget,
+    signal_listener: *parser.EventListener,
+
+    fn handle(node: *parser.EventNode, _: *parser.Event) void {
+        const self: *SignalCallback = @fieldParentPtr("node", node);
+        self._handle() catch |err| {
+            log.err(.app, "event signal handler", .{ .err = err });
+        };
+    }
+
+    fn _handle(self: *SignalCallback) !void {
+        const lst = try parser.eventTargetHasListener(
+            self.target,
+            self.typ,
+            self.capture,
+            self.callback_id,
+        );
+        if (lst == null) {
+            return;
+        }
+
+        try parser.eventTargetRemoveEventListener(
+            self.target,
+            self.typ,
+            lst.?,
+            self.capture,
+        );
+
+        // remove the abort signal listener itself
+        try parser.eventTargetRemoveEventListener(
+            self.signal_target,
+            "abort",
+            self.signal_listener,
+            false,
+        );
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: Event" {
+    try testing.htmlRunner("events/event.html");
+}

src/browser/events/keyboard_event.zig

@@ -0,0 +1,160 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+const builtin = @import("builtin");
+
+const parser = @import("../netsurf.zig");
+const Event = @import("event.zig").Event;
+const JsObject = @import("../env.zig").JsObject;
+
+// TODO: We currently don't have a UIEvent interface so we skip it in the prototype chain.
+// https://developer.mozilla.org/en-US/docs/Web/API/UIEvent
+const UIEvent = Event;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/KeyboardEvent
+pub const KeyboardEvent = struct {
+    pub const Self = parser.KeyboardEvent;
+    pub const prototype = *UIEvent;
+
+    pub const ConstructorOptions = struct {
+        key: []const u8 = "",
+        code: []const u8 = "",
+        location: parser.KeyboardEventOpts.LocationCode = .standard,
+        repeat: bool = false,
+        isComposing: bool = false,
+        // Currently not supported but we take as argument.
+        charCode: u32 = 0,
+        // Currently not supported but we take as argument.
+        keyCode: u32 = 0,
+        // Currently not supported but we take as argument.
+        which: u32 = 0,
+        ctrlKey: bool = false,
+        shiftKey: bool = false,
+        altKey: bool = false,
+        metaKey: bool = false,
+    };
+
+    pub fn constructor(event_type: []const u8, maybe_options: ?ConstructorOptions) !*parser.KeyboardEvent {
+        const options: ConstructorOptions = maybe_options orelse .{};
+
+        var event = try parser.keyboardEventCreate();
+        parser.eventSetInternalType(@ptrCast(&event), .keyboard_event);
+
+        try parser.keyboardEventInit(
+            event,
+            event_type,
+            .{
+                .key = options.key,
+                .code = options.code,
+                .location = options.location,
+                .repeat = options.repeat,
+                .is_composing = options.isComposing,
+                .ctrl_key = options.ctrlKey,
+                .shift_key = options.shiftKey,
+                .alt_key = options.altKey,
+                .meta_key = options.metaKey,
+            },
+        );
+
+        return event;
+    }
+
+    // Returns the modifier state for given modifier key.
+    pub fn _getModifierState(self: *Self, key: []const u8) bool {
+        // Chrome and Firefox do case-sensitive match, here we prefer the same.
+        if (std.mem.eql(u8, key, "Alt")) {
+            return get_altKey(self);
+        }
+
+        if (std.mem.eql(u8, key, "AltGraph")) {
+            return (get_altKey(self) and get_ctrlKey(self));
+        }
+
+        if (std.mem.eql(u8, key, "Control")) {
+            return get_ctrlKey(self);
+        }
+
+        if (std.mem.eql(u8, key, "Shift")) {
+            return get_shiftKey(self);
+        }
+
+        if (std.mem.eql(u8, key, "Meta") or std.mem.eql(u8, key, "OS")) {
+            return get_metaKey(self);
+        }
+
+        // Special case for IE.
+        if (comptime builtin.os.tag == .windows) {
+            if (std.mem.eql(u8, key, "Win")) {
+                return get_metaKey(self);
+            }
+        }
+
+        // getModifierState() also accepts a deprecated virtual modifier named "Accel".
+        // event.getModifierState("Accel") returns true when at least one of
+        // KeyboardEvent.ctrlKey or KeyboardEvent.metaKey is true.
+        //
+        // https://developer.mozilla.org/en-US/docs/Web/API/KeyboardEvent/getModifierState#accel_virtual_modifier
+        if (std.mem.eql(u8, key, "Accel")) {
+            return (get_ctrlKey(self) or get_metaKey(self));
+        }
+
+        // TODO: Add support for "CapsLock", "ScrollLock".
+        return false;
+    }
+
+    // Getters.
+
+    pub fn get_altKey(self: *Self) bool {
+        return parser.keyboardEventKeyIsSet(self, .alt);
+    }
+
+    pub fn get_ctrlKey(self: *Self) bool {
+        return parser.keyboardEventKeyIsSet(self, .ctrl);
+    }
+
+    pub fn get_metaKey(self: *Self) bool {
+        return parser.keyboardEventKeyIsSet(self, .meta);
+    }
+
+    pub fn get_shiftKey(self: *Self) bool {
+        return parser.keyboardEventKeyIsSet(self, .shift);
+    }
+
+    pub fn get_isComposing(self: *Self) bool {
+        return self.is_composing;
+    }
+
+    pub fn get_location(self: *Self) u32 {
+        return self.location;
+    }
+
+    pub fn get_key(self: *Self) ![]const u8 {
+        return parser.keyboardEventGetKey(self);
+    }
+
+    pub fn get_repeat(self: *Self) bool {
+        return self.repeat;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: Events.Keyboard" {
+    try testing.htmlRunner("events/keyboard.html");
+}

src/browser/events/mouse_event.zig

@@ -0,0 +1,112 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+
+const parser = @import("../netsurf.zig");
+const Event = @import("event.zig").Event;
+const JsObject = @import("../env.zig").JsObject;
+
+// TODO: We currently don't have a UIEvent interface so we skip it in the prototype chain.
+// https://developer.mozilla.org/en-US/docs/Web/API/UIEvent
+const UIEvent = Event;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/MouseEvent
+pub const MouseEvent = struct {
+    pub const Self = parser.MouseEvent;
+    pub const prototype = *UIEvent;
+
+    const MouseButton = enum(u16) {
+        main_button = 0,
+        auxillary_button = 1,
+        secondary_button = 2,
+        fourth_button = 3,
+        fifth_button = 4,
+    };
+
+    const MouseEventInit = struct {
+        screenX: i32 = 0,
+        screenY: i32 = 0,
+        clientX: i32 = 0,
+        clientY: i32 = 0,
+        ctrlKey: bool = false,
+        shiftKey: bool = false,
+        altKey: bool = false,
+        metaKey: bool = false,
+        button: MouseButton = .main_button,
+    };
+
+    pub fn constructor(event_type: []const u8, opts_: ?MouseEventInit) !*parser.MouseEvent {
+        const opts = opts_ orelse MouseEventInit{};
+
+        var mouse_event = try parser.mouseEventCreate();
+        parser.eventSetInternalType(@ptrCast(&mouse_event), .mouse_event);
+
+        try parser.mouseEventInit(mouse_event, event_type, .{
+            .x = opts.clientX,
+            .y = opts.clientY,
+            .ctrl = opts.ctrlKey,
+            .shift = opts.shiftKey,
+            .alt = opts.altKey,
+            .meta = opts.metaKey,
+            .button = @intFromEnum(opts.button),
+        });
+
+        if (!std.mem.eql(u8, event_type, "click")) {
+            log.warn(.mouse_event, "unsupported mouse event", .{ .event = event_type });
+        }
+
+        return mouse_event;
+    }
+
+    pub fn get_button(self: *parser.MouseEvent) u16 {
+        return self.button;
+    }
+
+    // These is just an alias for clientX.
+    pub fn get_x(self: *parser.MouseEvent) i32 {
+        return self.cx;
+    }
+
+    // These is just an alias for clientY.
+    pub fn get_y(self: *parser.MouseEvent) i32 {
+        return self.cy;
+    }
+
+    pub fn get_clientX(self: *parser.MouseEvent) i32 {
+        return self.cx;
+    }
+
+    pub fn get_clientY(self: *parser.MouseEvent) i32 {
+        return self.cy;
+    }
+
+    pub fn get_screenX(self: *parser.MouseEvent) i32 {
+        return self.sx;
+    }
+
+    pub fn get_screenY(self: *parser.MouseEvent) i32 {
+        return self.sy;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: Events.Mouse" {
+    try testing.htmlRunner("events/mouse.html");
+}

src/browser/fetch/Headers.zig

@@ -0,0 +1,227 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+const URL = @import("../../url.zig").URL;
+const Page = @import("../page.zig").Page;
+
+const iterator = @import("../iterator/iterator.zig");
+
+const v8 = @import("v8");
+const Env = @import("../env.zig").Env;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/Headers
+const Headers = @This();
+
+// Case-Insensitive String HashMap.
+// This allows us to avoid having to allocate lowercase keys all the time.
+const HeaderHashMap = std.HashMapUnmanaged([]const u8, []const u8, struct {
+    pub fn hash(_: @This(), s: []const u8) u64 {
+        var buf: [64]u8 = undefined;
+        var hasher = std.hash.Wyhash.init(s.len);
+
+        var key = s;
+        while (key.len >= 64) {
+            const lower = std.ascii.lowerString(buf[0..], key[0..64]);
+            hasher.update(lower);
+            key = key[64..];
+        }
+
+        if (key.len > 0) {
+            const lower = std.ascii.lowerString(buf[0..key.len], key);
+            hasher.update(lower);
+        }
+
+        return hasher.final();
+    }
+
+    pub fn eql(_: @This(), a: []const u8, b: []const u8) bool {
+        return std.ascii.eqlIgnoreCase(a, b);
+    }
+}, 80);
+
+headers: HeaderHashMap = .empty,
+
+// They can either be:
+//
+// 1. An array of string pairs.
+// 2. An object with string keys to string values.
+// 3. Another Headers object.
+pub const HeadersInit = union(enum) {
+    // List of Pairs of []const u8
+    strings: []const [2][]const u8,
+    // Headers
+    headers: *Headers,
+    // Mappings
+    object: Env.JsObject,
+};
+
+pub fn constructor(_init: ?HeadersInit, page: *Page) !Headers {
+    const arena = page.arena;
+    var headers: HeaderHashMap = .empty;
+
+    if (_init) |init| {
+        switch (init) {
+            .strings => |kvs| {
+                for (kvs) |pair| {
+                    const key = try arena.dupe(u8, pair[0]);
+                    const value = try arena.dupe(u8, pair[1]);
+
+                    try headers.put(arena, key, value);
+                }
+            },
+            .headers => |hdrs| {
+                var iter = hdrs.headers.iterator();
+                while (iter.next()) |entry| {
+                    try headers.put(arena, entry.key_ptr.*, entry.value_ptr.*);
+                }
+            },
+            .object => |obj| {
+                var iter = obj.nameIterator();
+                while (try iter.next()) |name_value| {
+                    const name = try name_value.toString(arena);
+                    const value = try obj.get(name);
+                    const value_string = try value.toString(arena);
+
+                    try headers.put(arena, name, value_string);
+                }
+            },
+        }
+    }
+
+    return .{
+        .headers = headers,
+    };
+}
+
+pub fn append(self: *Headers, name: []const u8, value: []const u8, allocator: std.mem.Allocator) !void {
+    const key = try allocator.dupe(u8, name);
+    const gop = try self.headers.getOrPut(allocator, key);
+
+    if (gop.found_existing) {
+        // If we found it, append the value.
+        const new_value = try std.fmt.allocPrint(allocator, "{s}, {s}", .{ gop.value_ptr.*, value });
+        gop.value_ptr.* = new_value;
+    } else {
+        // Otherwise, we should just put it in.
+        gop.value_ptr.* = try allocator.dupe(u8, value);
+    }
+}
+
+pub fn _append(self: *Headers, name: []const u8, value: []const u8, page: *Page) !void {
+    const arena = page.arena;
+    try self.append(name, value, arena);
+}
+
+pub fn _delete(self: *Headers, name: []const u8) void {
+    _ = self.headers.remove(name);
+}
+
+pub const HeadersEntryIterator = struct {
+    slot: [2][]const u8,
+    iter: HeaderHashMap.Iterator,
+
+    // TODO: these SHOULD be in lexigraphical order but I'm not sure how actually
+    // important that is.
+    pub fn _next(self: *HeadersEntryIterator) ?[2][]const u8 {
+        if (self.iter.next()) |entry| {
+            self.slot[0] = entry.key_ptr.*;
+            self.slot[1] = entry.value_ptr.*;
+            return self.slot;
+        } else {
+            return null;
+        }
+    }
+};
+
+pub fn _entries(self: *const Headers) HeadersEntryIterable {
+    return .{
+        .inner = .{
+            .slot = undefined,
+            .iter = self.headers.iterator(),
+        },
+    };
+}
+
+pub fn _forEach(self: *Headers, callback_fn: Env.Function, this_arg: ?Env.JsObject) !void {
+    var iter = self.headers.iterator();
+
+    const cb = if (this_arg) |this| try callback_fn.withThis(this) else callback_fn;
+
+    while (iter.next()) |entry| {
+        try cb.call(void, .{ entry.key_ptr.*, entry.value_ptr.*, self });
+    }
+}
+
+pub fn _get(self: *const Headers, name: []const u8) ?[]const u8 {
+    return self.headers.get(name);
+}
+
+pub fn _has(self: *const Headers, name: []const u8) bool {
+    return self.headers.contains(name);
+}
+
+pub const HeadersKeyIterator = struct {
+    iter: HeaderHashMap.KeyIterator,
+
+    pub fn _next(self: *HeadersKeyIterator) ?[]const u8 {
+        if (self.iter.next()) |key| {
+            return key.*;
+        } else {
+            return null;
+        }
+    }
+};
+
+pub fn _keys(self: *const Headers) HeadersKeyIterable {
+    return .{ .inner = .{ .iter = self.headers.keyIterator() } };
+}
+
+pub fn _set(self: *Headers, name: []const u8, value: []const u8, page: *Page) !void {
+    const arena = page.arena;
+
+    const key = try arena.dupe(u8, name);
+    const gop = try self.headers.getOrPut(arena, key);
+    gop.value_ptr.* = try arena.dupe(u8, value);
+}
+
+pub const HeadersValueIterator = struct {
+    iter: HeaderHashMap.ValueIterator,
+
+    pub fn _next(self: *HeadersValueIterator) ?[]const u8 {
+        if (self.iter.next()) |value| {
+            return value.*;
+        } else {
+            return null;
+        }
+    }
+};
+
+pub fn _values(self: *const Headers) HeadersValueIterable {
+    return .{ .inner = .{ .iter = self.headers.valueIterator() } };
+}
+
+pub const HeadersKeyIterable = iterator.Iterable(HeadersKeyIterator, "HeadersKeyIterator");
+pub const HeadersValueIterable = iterator.Iterable(HeadersValueIterator, "HeadersValueIterator");
+pub const HeadersEntryIterable = iterator.Iterable(HeadersEntryIterator, "HeadersEntryIterator");
+
+const testing = @import("../../testing.zig");
+test "fetch: Headers" {
+    try testing.htmlRunner("fetch/headers.html");
+}

src/browser/fetch/Request.zig

@@ -0,0 +1,266 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+
+const URL = @import("../../url.zig").URL;
+const Page = @import("../page.zig").Page;
+
+const Response = @import("./Response.zig");
+const Http = @import("../../http/Http.zig");
+const ReadableStream = @import("../streams/ReadableStream.zig");
+
+const v8 = @import("v8");
+const Env = @import("../env.zig").Env;
+
+const Headers = @import("Headers.zig");
+const HeadersInit = @import("Headers.zig").HeadersInit;
+
+pub const RequestInput = union(enum) {
+    string: []const u8,
+    request: *Request,
+};
+
+pub const RequestCache = enum {
+    default,
+    @"no-store",
+    reload,
+    @"no-cache",
+    @"force-cache",
+    @"only-if-cached",
+
+    pub fn fromString(str: []const u8) ?RequestCache {
+        for (std.enums.values(RequestCache)) |cache| {
+            if (std.ascii.eqlIgnoreCase(str, @tagName(cache))) {
+                return cache;
+            }
+        } else {
+            return null;
+        }
+    }
+
+    pub fn toString(self: RequestCache) []const u8 {
+        return @tagName(self);
+    }
+};
+
+pub const RequestCredentials = enum {
+    omit,
+    @"same-origin",
+    include,
+
+    pub fn fromString(str: []const u8) ?RequestCredentials {
+        for (std.enums.values(RequestCredentials)) |cache| {
+            if (std.ascii.eqlIgnoreCase(str, @tagName(cache))) {
+                return cache;
+            }
+        } else {
+            return null;
+        }
+    }
+
+    pub fn toString(self: RequestCredentials) []const u8 {
+        return @tagName(self);
+    }
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/RequestInit
+pub const RequestInit = struct {
+    body: ?[]const u8 = null,
+    cache: ?[]const u8 = null,
+    credentials: ?[]const u8 = null,
+    headers: ?HeadersInit = null,
+    integrity: ?[]const u8 = null,
+    method: ?[]const u8 = null,
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/Request/Request
+const Request = @This();
+
+method: Http.Method,
+url: [:0]const u8,
+cache: RequestCache,
+credentials: RequestCredentials,
+headers: Headers,
+body: ?[]const u8,
+body_used: bool = false,
+integrity: []const u8,
+
+pub fn constructor(input: RequestInput, _options: ?RequestInit, page: *Page) !Request {
+    const arena = page.arena;
+    const options: RequestInit = _options orelse .{};
+
+    const url: [:0]const u8 = blk: switch (input) {
+        .string => |str| {
+            break :blk try URL.stitch(arena, str, page.url.raw, .{ .null_terminated = true });
+        },
+        .request => |req| {
+            break :blk try arena.dupeZ(u8, req.url);
+        },
+    };
+
+    const body = if (options.body) |body| try arena.dupe(u8, body) else null;
+    const cache = (if (options.cache) |cache| RequestCache.fromString(cache) else null) orelse RequestCache.default;
+    const credentials = (if (options.credentials) |creds| RequestCredentials.fromString(creds) else null) orelse RequestCredentials.@"same-origin";
+    const integrity = if (options.integrity) |integ| try arena.dupe(u8, integ) else "";
+    const headers: Headers = if (options.headers) |hdrs| try Headers.constructor(hdrs, page) else .{};
+
+    const method: Http.Method = blk: {
+        if (options.method) |given_method| {
+            for (std.enums.values(Http.Method)) |method| {
+                if (std.ascii.eqlIgnoreCase(given_method, @tagName(method))) {
+                    break :blk method;
+                }
+            } else {
+                return error.TypeError;
+            }
+        } else {
+            break :blk Http.Method.GET;
+        }
+    };
+
+    return .{
+        .method = method,
+        .url = url,
+        .cache = cache,
+        .credentials = credentials,
+        .headers = headers,
+        .body = body,
+        .integrity = integrity,
+    };
+}
+
+pub fn get_body(self: *const Request, page: *Page) !?*ReadableStream {
+    if (self.body) |body| {
+        const stream = try ReadableStream.constructor(null, null, page);
+        try stream.queue.append(page.arena, body);
+        return stream;
+    } else return null;
+}
+
+pub fn get_bodyUsed(self: *const Request) bool {
+    return self.body_used;
+}
+
+pub fn get_cache(self: *const Request) RequestCache {
+    return self.cache;
+}
+
+pub fn get_credentials(self: *const Request) RequestCredentials {
+    return self.credentials;
+}
+
+pub fn get_headers(self: *Request) *Headers {
+    return &self.headers;
+}
+
+pub fn get_integrity(self: *const Request) []const u8 {
+    return self.integrity;
+}
+
+// TODO: If we ever support the Navigation API, we need isHistoryNavigation
+// https://developer.mozilla.org/en-US/docs/Web/API/Request/isHistoryNavigation
+
+pub fn get_method(self: *const Request) []const u8 {
+    return @tagName(self.method);
+}
+
+pub fn get_url(self: *const Request) []const u8 {
+    return self.url;
+}
+
+pub fn _clone(self: *Request) !Request {
+    // Not allowed to clone if the body was used.
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    // OK to just return the same fields BECAUSE
+    // all of these fields are read-only and can't be modified.
+    return Request{
+        .body = self.body,
+        .body_used = self.body_used,
+        .cache = self.cache,
+        .credentials = self.credentials,
+        .headers = self.headers,
+        .method = self.method,
+        .integrity = self.integrity,
+        .url = self.url,
+    };
+}
+
+pub fn _bytes(self: *Response, page: *Page) !Env.Promise {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    const resolver = Env.PromiseResolver{
+        .js_context = page.main_context,
+        .resolver = v8.PromiseResolver.init(page.main_context.v8_context),
+    };
+
+    try resolver.resolve(self.body);
+    self.body_used = true;
+    return resolver.promise();
+}
+
+pub fn _json(self: *Response, page: *Page) !Env.Promise {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    const resolver = Env.PromiseResolver{
+        .js_context = page.main_context,
+        .resolver = v8.PromiseResolver.init(page.main_context.v8_context),
+    };
+
+    const p = std.json.parseFromSliceLeaky(
+        std.json.Value,
+        page.call_arena,
+        self.body,
+        .{},
+    ) catch |e| {
+        log.info(.browser, "invalid json", .{ .err = e, .source = "Request" });
+        return error.SyntaxError;
+    };
+
+    try resolver.resolve(p);
+    self.body_used = true;
+    return resolver.promise();
+}
+
+pub fn _text(self: *Response, page: *Page) !Env.Promise {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    const resolver = Env.PromiseResolver{
+        .js_context = page.main_context,
+        .resolver = v8.PromiseResolver.init(page.main_context.v8_context),
+    };
+
+    try resolver.resolve(self.body);
+    self.body_used = true;
+    return resolver.promise();
+}
+
+const testing = @import("../../testing.zig");
+test "fetch: Request" {
+    try testing.htmlRunner("fetch/request.html");
+}

src/browser/fetch/Response.zig

@@ -0,0 +1,196 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+
+const v8 = @import("v8");
+
+const HttpClient = @import("../../http/Client.zig");
+const Http = @import("../../http/Http.zig");
+const URL = @import("../../url.zig").URL;
+
+const ReadableStream = @import("../streams/ReadableStream.zig");
+const Headers = @import("Headers.zig");
+const HeadersInit = @import("Headers.zig").HeadersInit;
+
+const Env = @import("../env.zig").Env;
+const Mime = @import("../mime.zig").Mime;
+const Page = @import("../page.zig").Page;
+
+// https://developer.mozilla.org/en-US/docs/Web/API/Response
+const Response = @This();
+
+status: u16 = 200,
+status_text: []const u8 = "",
+headers: Headers,
+mime: ?Mime = null,
+url: []const u8 = "",
+body: []const u8 = "",
+body_used: bool = false,
+redirected: bool = false,
+
+const ResponseBody = union(enum) {
+    string: []const u8,
+};
+
+const ResponseOptions = struct {
+    status: u16 = 200,
+    statusText: ?[]const u8 = null,
+    headers: ?HeadersInit = null,
+};
+
+pub fn constructor(_input: ?ResponseBody, _options: ?ResponseOptions, page: *Page) !Response {
+    const arena = page.arena;
+
+    const options: ResponseOptions = _options orelse .{};
+
+    const body = blk: {
+        if (_input) |input| {
+            switch (input) {
+                .string => |str| {
+                    break :blk try arena.dupe(u8, str);
+                },
+            }
+        } else {
+            break :blk "";
+        }
+    };
+
+    const headers: Headers = if (options.headers) |hdrs| try Headers.constructor(hdrs, page) else .{};
+    const status_text = if (options.statusText) |st| try arena.dupe(u8, st) else "";
+
+    return .{
+        .body = body,
+        .headers = headers,
+        .status = options.status,
+        .status_text = status_text,
+    };
+}
+
+pub fn get_body(self: *const Response, page: *Page) !*ReadableStream {
+    const stream = try ReadableStream.constructor(null, null, page);
+    try stream.queue.append(page.arena, self.body);
+    return stream;
+}
+
+pub fn get_bodyUsed(self: *const Response) bool {
+    return self.body_used;
+}
+
+pub fn get_headers(self: *Response) *Headers {
+    return &self.headers;
+}
+
+pub fn get_ok(self: *const Response) bool {
+    return self.status >= 200 and self.status <= 299;
+}
+
+pub fn get_redirected(self: *const Response) bool {
+    return self.redirected;
+}
+
+pub fn get_status(self: *const Response) u16 {
+    return self.status;
+}
+
+pub fn get_statusText(self: *const Response) []const u8 {
+    return self.status_text;
+}
+
+pub fn get_url(self: *const Response) []const u8 {
+    return self.url;
+}
+
+pub fn _clone(self: *const Response) !Response {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    // OK to just return the same fields BECAUSE
+    // all of these fields are read-only and can't be modified.
+    return Response{
+        .body = self.body,
+        .body_used = self.body_used,
+        .mime = self.mime,
+        .headers = self.headers,
+        .redirected = self.redirected,
+        .status = self.status,
+        .url = self.url,
+    };
+}
+
+pub fn _bytes(self: *Response, page: *Page) !Env.Promise {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    const resolver = Env.PromiseResolver{
+        .js_context = page.main_context,
+        .resolver = v8.PromiseResolver.init(page.main_context.v8_context),
+    };
+
+    try resolver.resolve(self.body);
+    self.body_used = true;
+    return resolver.promise();
+}
+
+pub fn _json(self: *Response, page: *Page) !Env.Promise {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    const resolver = Env.PromiseResolver{
+        .js_context = page.main_context,
+        .resolver = v8.PromiseResolver.init(page.main_context.v8_context),
+    };
+
+    const p = std.json.parseFromSliceLeaky(
+        std.json.Value,
+        page.call_arena,
+        self.body,
+        .{},
+    ) catch |e| {
+        log.info(.browser, "invalid json", .{ .err = e, .source = "Response" });
+        return error.SyntaxError;
+    };
+
+    try resolver.resolve(p);
+    self.body_used = true;
+    return resolver.promise();
+}
+
+pub fn _text(self: *Response, page: *Page) !Env.Promise {
+    if (self.body_used) {
+        return error.TypeError;
+    }
+
+    const resolver = Env.PromiseResolver{
+        .js_context = page.main_context,
+        .resolver = v8.PromiseResolver.init(page.main_context.v8_context),
+    };
+
+    try resolver.resolve(self.body);
+    self.body_used = true;
+    return resolver.promise();
+}
+
+const testing = @import("../../testing.zig");
+test "fetch: Response" {
+    try testing.htmlRunner("fetch/response.html");
+}

src/browser/fetch/fetch.zig

@@ -0,0 +1,212 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+
+const Http = @import("../../http/Http.zig");
+const HttpClient = @import("../../http/Client.zig");
+const Mime = @import("../mime.zig").Mime;
+
+const Headers = @import("Headers.zig");
+
+const RequestInput = @import("Request.zig").RequestInput;
+const RequestInit = @import("Request.zig").RequestInit;
+const Request = @import("Request.zig");
+const Response = @import("Response.zig");
+
+pub const Interfaces = .{
+    @import("Headers.zig"),
+    @import("Headers.zig").HeadersEntryIterable,
+    @import("Headers.zig").HeadersKeyIterable,
+    @import("Headers.zig").HeadersValueIterable,
+    @import("Request.zig"),
+    @import("Response.zig"),
+};
+
+pub const FetchContext = struct {
+    arena: std.mem.Allocator,
+    js_ctx: *Env.JsContext,
+    promise_resolver: Env.PersistentPromiseResolver,
+
+    method: Http.Method,
+    url: []const u8,
+    body: std.ArrayListUnmanaged(u8) = .empty,
+    headers: std.ArrayListUnmanaged([]const u8) = .empty,
+    status: u16 = 0,
+    mime: ?Mime = null,
+    transfer: ?*HttpClient.Transfer = null,
+
+    /// This effectively takes ownership of the FetchContext.
+    ///
+    /// We just return the underlying slices used for `headers`
+    /// and for `body` here to avoid an allocation.
+    pub fn toResponse(self: *const FetchContext) !Response {
+        var headers: Headers = .{};
+
+        // convert into Headers
+        for (self.headers.items) |hdr| {
+            var iter = std.mem.splitScalar(u8, hdr, ':');
+            const name = iter.next() orelse "";
+            const value = iter.next() orelse "";
+            try headers.append(name, value, self.arena);
+        }
+
+        return Response{
+            .status = self.status,
+            .headers = headers,
+            .mime = self.mime,
+            .body = self.body.items,
+            .url = self.url,
+        };
+    }
+};
+
+// https://developer.mozilla.org/en-US/docs/Web/API/Window/fetch
+pub fn fetch(input: RequestInput, options: ?RequestInit, page: *Page) !Env.Promise {
+    const arena = page.arena;
+
+    const req = try Request.constructor(input, options, page);
+    var headers = try Http.Headers.init();
+
+    // Copy our headers into the HTTP headers.
+    var header_iter = req.headers.headers.iterator();
+    while (header_iter.next()) |entry| {
+        const combined = try std.fmt.allocPrintSentinel(
+            page.arena,
+            "{s}: {s}",
+            .{ entry.key_ptr.*, entry.value_ptr.* },
+            0,
+        );
+        try headers.add(combined.ptr);
+    }
+
+    try page.requestCookie(.{}).headersForRequest(arena, req.url, &headers);
+
+    const resolver = try page.main_context.createPersistentPromiseResolver();
+
+    const fetch_ctx = try arena.create(FetchContext);
+    fetch_ctx.* = .{
+        .arena = arena,
+        .js_ctx = page.main_context,
+        .promise_resolver = resolver,
+        .method = req.method,
+        .url = req.url,
+    };
+
+    try page.http_client.request(.{
+        .ctx = @ptrCast(fetch_ctx),
+        .url = req.url,
+        .method = req.method,
+        .headers = headers,
+        .body = req.body,
+        .cookie_jar = page.cookie_jar,
+        .resource_type = .fetch,
+
+        .start_callback = struct {
+            fn startCallback(transfer: *HttpClient.Transfer) !void {
+                const self: *FetchContext = @ptrCast(@alignCast(transfer.ctx));
+                log.debug(.fetch, "request start", .{ .method = self.method, .url = self.url, .source = "fetch" });
+
+                self.transfer = transfer;
+            }
+        }.startCallback,
+        .header_callback = struct {
+            fn headerCallback(transfer: *HttpClient.Transfer) !void {
+                const self: *FetchContext = @ptrCast(@alignCast(transfer.ctx));
+
+                const header = &transfer.response_header.?;
+
+                log.debug(.fetch, "request header", .{
+                    .source = "fetch",
+                    .method = self.method,
+                    .url = self.url,
+                    .status = header.status,
+                });
+
+                if (header.contentType()) |ct| {
+                    self.mime = Mime.parse(ct) catch {
+                        return error.MimeParsing;
+                    };
+                }
+
+                if (transfer.getContentLength()) |cl| {
+                    try self.body.ensureTotalCapacity(self.arena, cl);
+                }
+
+                var it = transfer.responseHeaderIterator();
+                while (it.next()) |hdr| {
+                    const joined = try std.fmt.allocPrint(self.arena, "{s}: {s}", .{ hdr.name, hdr.value });
+                    try self.headers.append(self.arena, joined);
+                }
+
+                self.status = header.status;
+            }
+        }.headerCallback,
+        .data_callback = struct {
+            fn dataCallback(transfer: *HttpClient.Transfer, data: []const u8) !void {
+                const self: *FetchContext = @ptrCast(@alignCast(transfer.ctx));
+                try self.body.appendSlice(self.arena, data);
+            }
+        }.dataCallback,
+        .done_callback = struct {
+            fn doneCallback(ctx: *anyopaque) !void {
+                const self: *FetchContext = @ptrCast(@alignCast(ctx));
+                self.transfer = null;
+
+                log.info(.fetch, "request complete", .{
+                    .source = "fetch",
+                    .method = self.method,
+                    .url = self.url,
+                    .status = self.status,
+                });
+
+                const response = try self.toResponse();
+                try self.promise_resolver.resolve(response);
+            }
+        }.doneCallback,
+        .error_callback = struct {
+            fn errorCallback(ctx: *anyopaque, err: anyerror) void {
+                const self: *FetchContext = @ptrCast(@alignCast(ctx));
+                self.transfer = null;
+
+                log.err(.fetch, "error", .{
+                    .url = self.url,
+                    .err = err,
+                    .source = "fetch error",
+                });
+
+                // We throw an Abort error when the page is getting closed so,
+                // in this case, we don't need to reject the promise.
+                if (err != error.Abort) {
+                    self.promise_resolver.reject(@errorName(err)) catch unreachable;
+                }
+            }
+        }.errorCallback,
+    });
+
+    return resolver.promise();
+}
+
+const testing = @import("../../testing.zig");
+test "fetch: fetch" {
+    try testing.htmlRunner("fetch/fetch.html");
+}

src/browser/html/AbortController.zig

@@ -0,0 +1,143 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+const EventTarget = @import("../dom/event_target.zig").EventTarget;
+
+pub const Interfaces = .{
+    AbortController,
+    AbortSignal,
+};
+
+const AbortController = @This();
+
+signal: *AbortSignal,
+
+pub fn constructor(page: *Page) !AbortController {
+    // Why do we allocate this rather than storing directly in the struct?
+    // https://github.com/lightpanda-io/project/discussions/165
+    const signal = try page.arena.create(AbortSignal);
+    signal.* = .init;
+
+    return .{
+        .signal = signal,
+    };
+}
+
+pub fn get_signal(self: *AbortController) *AbortSignal {
+    return self.signal;
+}
+
+pub fn _abort(self: *AbortController, reason_: ?[]const u8) !void {
+    return self.signal.abort(reason_);
+}
+
+pub const AbortSignal = struct {
+    const DEFAULT_REASON = "AbortError";
+
+    pub const prototype = *EventTarget;
+    proto: parser.EventTargetTBase = .{ .internal_target_type = .abort_signal },
+
+    aborted: bool,
+    reason: ?[]const u8,
+
+    pub const init: AbortSignal = .{
+        .reason = null,
+        .aborted = false,
+    };
+
+    pub fn static_abort(reason_: ?[]const u8) AbortSignal {
+        return .{
+            .aborted = true,
+            .reason = reason_ orelse DEFAULT_REASON,
+        };
+    }
+
+    pub fn static_timeout(delay: u32, page: *Page) !*AbortSignal {
+        const callback = try page.arena.create(TimeoutCallback);
+        callback.* = .{
+            .signal = .init,
+        };
+
+        try page.scheduler.add(callback, TimeoutCallback.run, delay, .{ .name = "abort_signal" });
+        return &callback.signal;
+    }
+
+    pub fn get_aborted(self: *const AbortSignal) bool {
+        return self.aborted;
+    }
+
+    fn abort(self: *AbortSignal, reason_: ?[]const u8) !void {
+        self.aborted = true;
+        self.reason = reason_ orelse DEFAULT_REASON;
+
+        const abort_event = try parser.eventCreate();
+        parser.eventSetInternalType(abort_event, .abort_signal);
+
+        defer parser.eventDestroy(abort_event);
+        try parser.eventInit(abort_event, "abort", .{});
+        _ = try parser.eventTargetDispatchEvent(
+            parser.toEventTarget(AbortSignal, self),
+            abort_event,
+        );
+    }
+
+    const Reason = union(enum) {
+        reason: []const u8,
+        undefined: void,
+    };
+    pub fn get_reason(self: *const AbortSignal) Reason {
+        if (self.reason) |r| {
+            return .{ .reason = r };
+        }
+        return .{ .undefined = {} };
+    }
+
+    const ThrowIfAborted = union(enum) {
+        exception: Env.Exception,
+        undefined: void,
+    };
+    pub fn _throwIfAborted(self: *const AbortSignal, page: *Page) ThrowIfAborted {
+        if (self.aborted) {
+            const ex = page.main_context.throw(self.reason orelse DEFAULT_REASON);
+            return .{ .exception = ex };
+        }
+        return .{ .undefined = {} };
+    }
+};
+
+const TimeoutCallback = struct {
+    signal: AbortSignal,
+
+    fn run(ctx: *anyopaque) ?u32 {
+        const self: *TimeoutCallback = @ptrCast(@alignCast(ctx));
+        self.signal.abort("TimeoutError") catch |err| {
+            log.warn(.app, "abort signal timeout", .{ .err = err });
+        };
+        return null;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: HTML.AbortController" {
+    try testing.htmlRunner("html/abort_controller.html");
+}

src/browser/html/DataSet.zig

@@ -0,0 +1,81 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+const std = @import("std");
+const parser = @import("../netsurf.zig");
+const Env = @import("../env.zig").Env;
+const Page = @import("../page.zig").Page;
+
+const Allocator = std.mem.Allocator;
+
+const DataSet = @This();
+
+element: *parser.Element,
+
+pub fn named_get(self: *const DataSet, name: []const u8, _: *bool, page: *Page) !Env.UndefinedOr([]const u8) {
+    const normalized_name = try normalize(page.call_arena, name);
+    if (try parser.elementGetAttribute(self.element, normalized_name)) |value| {
+        return .{ .value = value };
+    }
+    return .undefined;
+}
+
+pub fn named_set(self: *DataSet, name: []const u8, value: []const u8, _: *bool, page: *Page) !void {
+    const normalized_name = try normalize(page.call_arena, name);
+    try parser.elementSetAttribute(self.element, normalized_name, value);
+}
+
+pub fn named_delete(self: *DataSet, name: []const u8, _: *bool, page: *Page) !void {
+    const normalized_name = try normalize(page.call_arena, name);
+    try parser.elementRemoveAttribute(self.element, normalized_name);
+}
+
+fn normalize(allocator: Allocator, name: []const u8) ![]const u8 {
+    var upper_count: usize = 0;
+    for (name) |c| {
+        if (std.ascii.isUpper(c)) {
+            upper_count += 1;
+        }
+    }
+    // for every upper-case letter, we'll probably need a dash before it
+    // and we need the 'data-' prefix
+    var normalized = try allocator.alloc(u8, name.len + upper_count + 5);
+
+    @memcpy(normalized[0..5], "data-");
+    if (upper_count == 0) {
+        @memcpy(normalized[5..], name);
+        return normalized;
+    }
+
+    var pos: usize = 5;
+    for (name) |c| {
+        if (std.ascii.isUpper(c)) {
+            normalized[pos] = '-';
+            pos += 1;
+            normalized[pos] = c + 32;
+        } else {
+            normalized[pos] = c;
+        }
+        pos += 1;
+    }
+    return normalized;
+}
+
+const testing = @import("../../testing.zig");
+test "Browser: HTML.DataSet" {
+    try testing.htmlRunner("html/dataset.html");
+}

src/browser/html/document.zig

@@ -0,0 +1,320 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+
+const std = @import("std");
+
+const log = @import("../../log.zig");
+const parser = @import("../netsurf.zig");
+const Page = @import("../page.zig").Page;
+
+const Window = @import("window.zig").Window;
+const Element = @import("../dom/element.zig").Element;
+const ElementUnion = @import("../dom/element.zig").Union;
+const Document = @import("../dom/document.zig").Document;
+const NodeList = @import("../dom/nodelist.zig").NodeList;
+const Location = @import("location.zig").Location;
+
+const collection = @import("../dom/html_collection.zig");
+const Walker = @import("../dom/walker.zig").WalkerDepthFirst;
+const Cookie = @import("../storage/cookie.zig").Cookie;
+
+// WEB IDL https://html.spec.whatwg.org/#the-document-object
+pub const HTMLDocument = struct {
+    pub const Self = parser.DocumentHTML;
+    pub const prototype = *Document;
+    pub const subtype = .node;
+
+    // JS funcs
+    // --------
+
+    pub fn get_domain(self: *parser.DocumentHTML, page: *Page) ![]const u8 {
+        // libdom's document_html get_domain always returns null, this is
+        // the way MDN recommends getting the domain anyways, since document.domain
+        // is deprecated.
+        const location = try parser.documentHTMLGetLocation(Location, self) orelse return "";
+        return location.get_host(page);
+    }
+
+    pub fn set_domain(_: *parser.DocumentHTML, _: []const u8) ![]const u8 {
+        return error.NotImplemented;
+    }
+
+    pub fn get_referrer(self: *parser.DocumentHTML) ![]const u8 {
+        return try parser.documentHTMLGetReferrer(self);
+    }
+
+    pub fn set_referrer(_: *parser.DocumentHTML, _: []const u8) ![]const u8 {
+        return error.NotImplemented;
+    }
+
+    pub fn get_body(self: *parser.DocumentHTML) !?*parser.Body {
+        return try parser.documentHTMLBody(self);
+    }
+
+    pub fn set_body(self: *parser.DocumentHTML, elt: ?*parser.ElementHTML) !?*parser.Body {
+        try parser.documentHTMLSetBody(self, elt);
+        return try get_body(self);
+    }
+
+    pub fn get_head(self: *parser.DocumentHTML) !?*parser.Head {
+        const root = parser.documentHTMLToNode(self);
+        const walker = Walker{};
+        var next: ?*parser.Node = null;
+        while (true) {
+            next = try walker.get_next(root, next) orelse return null;
+            if (std.ascii.eqlIgnoreCase("head", try parser.nodeName(next.?))) {
+                return @as(*parser.Head, @ptrCast(next.?));
+            }
+        }
+    }
+
+    pub fn get_cookie(_: *parser.DocumentHTML, page: *Page) ![]const u8 {
+        var buf: std.ArrayListUnmanaged(u8) = .{};
+        try page.cookie_jar.forRequest(&page.url.uri, buf.writer(page.arena), .{
+            .is_http = false,
+            .is_navigation = true,
+        });
+        return buf.items;
+    }
+
+    pub fn set_cookie(_: *parser.DocumentHTML, cookie_str: []const u8, page: *Page) ![]const u8 {
+        // we use the cookie jar's allocator to parse the cookie because it
+        // outlives the page's arena.
+        const c = try Cookie.parse(page.cookie_jar.allocator, &page.url.uri, cookie_str);
+        errdefer c.deinit();
+        if (c.http_only) {
+            c.deinit();
+            return ""; // HttpOnly cookies cannot be set from JS
+        }
+        try page.cookie_jar.add(c, std.time.timestamp());
+        return cookie_str;
+    }
+
+    pub fn get_title(self: *parser.DocumentHTML) ![]const u8 {
+        return try parser.documentHTMLGetTitle(self);
+    }
+
+    pub fn set_title(self: *parser.DocumentHTML, v: []const u8) ![]const u8 {
+        try parser.documentHTMLSetTitle(self, v);
+        return v;
+    }
+
+    pub fn _getElementsByName(self: *parser.DocumentHTML, name: []const u8, page: *Page) !NodeList {
+        const arena = page.arena;
+        var list: NodeList = .{};
+
+        if (name.len == 0) return list;
+
+        const root = parser.documentHTMLToNode(self);
+        var c = try collection.HTMLCollectionByName(arena, root, name, .{
+            .include_root = false,
+        });
+
+        const ln = try c.get_length();
+        var i: u32 = 0;
+        while (i < ln) {
+            const n = try c.item(i) orelse break;
+            try list.append(arena, n);
+            i += 1;
+        }
+
+        return list;
+    }
+
+    pub fn get_images(self: *parser.DocumentHTML, page: *Page) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByTagName(page.arena, parser.documentHTMLToNode(self), "img", .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_embeds(self: *parser.DocumentHTML, page: *Page) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByTagName(page.arena, parser.documentHTMLToNode(self), "embed", .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_plugins(self: *parser.DocumentHTML, page: *Page) !collection.HTMLCollection {
+        return get_embeds(self, page);
+    }
+
+    pub fn get_forms(self: *parser.DocumentHTML, page: *Page) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByTagName(page.arena, parser.documentHTMLToNode(self), "form", .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_scripts(self: *parser.DocumentHTML, page: *Page) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByTagName(page.arena, parser.documentHTMLToNode(self), "script", .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_applets(_: *parser.DocumentHTML) !collection.HTMLCollection {
+        return try collection.HTMLCollectionEmpty();
+    }
+
+    pub fn get_links(self: *parser.DocumentHTML) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByLinks(parser.documentHTMLToNode(self), .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_anchors(self: *parser.DocumentHTML) !collection.HTMLCollection {
+        return try collection.HTMLCollectionByAnchors(parser.documentHTMLToNode(self), .{
+            .include_root = false,
+        });
+    }
+
+    pub fn get_all(self: *parser.DocumentHTML) collection.HTMLAllCollection {
+        return collection.HTMLAllCollection.init(parser.documentHTMLToNode(self));
+    }
+
+    pub fn get_currentScript(self: *parser.DocumentHTML) !?*parser.Script {
+        return try parser.documentHTMLGetCurrentScript(self);
+    }
+
+    pub fn get_location(self: *parser.DocumentHTML) !?*Location {
+        return try parser.documentHTMLGetLocation(Location, self);
+    }
+
+    pub fn set_location(_: *const parser.DocumentHTML, url: []const u8, page: *Page) !void {
+        return page.navigateFromWebAPI(url, .{ .reason = .script });
+    }
+
+    pub fn get_designMode(_: *parser.DocumentHTML) []const u8 {
+        return "off";
+    }
+
+    pub fn set_designMode(_: *parser.DocumentHTML, _: []const u8) []const u8 {
+        return "off";
+    }
+
+    pub fn get_defaultView(_: *parser.DocumentHTML, page: *Page) *Window {
+        return &page.window;
+    }
+
+    pub fn get_readyState(self: *parser.DocumentHTML, page: *Page) ![]const u8 {
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        return @tagName(state.ready_state);
+    }
+
+    // noop legacy functions
+    // https://html.spec.whatwg.org/#Document-partial
+    pub fn _clear(_: *parser.DocumentHTML) void {}
+    pub fn _captureEvents(_: *parser.DocumentHTML) void {}
+    pub fn _releaseEvents(_: *parser.DocumentHTML) void {}
+
+    pub fn get_fgColor(_: *parser.DocumentHTML) []const u8 {
+        return "";
+    }
+    pub fn set_fgColor(_: *parser.DocumentHTML, _: []const u8) []const u8 {
+        return "";
+    }
+    pub fn get_linkColor(_: *parser.DocumentHTML) []const u8 {
+        return "";
+    }
+    pub fn set_linkColor(_: *parser.DocumentHTML, _: []const u8) []const u8 {
+        return "";
+    }
+    pub fn get_vlinkColor(_: *parser.DocumentHTML) []const u8 {
+        return "";
+    }
+    pub fn set_vlinkColor(_: *parser.DocumentHTML, _: []const u8) []const u8 {
+        return "";
+    }
+    pub fn get_alinkColor(_: *parser.DocumentHTML) []const u8 {
+        return "";
+    }
+    pub fn set_alinkColor(_: *parser.DocumentHTML, _: []const u8) []const u8 {
+        return "";
+    }
+    pub fn get_bgColor(_: *parser.DocumentHTML) []const u8 {
+        return "";
+    }
+    pub fn set_bgColor(_: *parser.DocumentHTML, _: []const u8) []const u8 {
+        return "";
+    }
+
+    // Returns the topmost Element at the specified coordinates (relative to the viewport).
+    // Since LightPanda requires the client to know what they are clicking on we do not return the underlying element at this moment
+    // This can currenty only happen if the first pixel is clicked without having rendered any element. This will change when css properties are supported.
+    // This returns an ElementUnion instead of a *Parser.Element in case the element somehow hasn't passed through the js runtime yet.
+    // While x and y should be f32, here we take i32 since that's what our
+    // "renderer" uses. By specifying i32 here, rather than f32 and doing the
+    // conversion ourself, we rely on v8's type conversion which is both more
+    // flexible (e.g. handles NaN) and will be more consistent with a browser.
+    pub fn _elementFromPoint(_: *parser.DocumentHTML, x: i32, y: i32, page: *Page) !?ElementUnion {
+        const element = page.renderer.getElementAtPosition(x, y) orelse return null;
+        // TODO if pointer-events set to none the underlying element should be returned (parser.documentGetDocumentElement(self.document);?)
+        return try Element.toInterface(element);
+    }
+
+    // Returns an array of all elements at the specified coordinates (relative to the viewport). The elements are ordered from the topmost to the bottommost box of the viewport.
+    // While x and y should be f32, here we take i32 since that's what our
+    // "renderer" uses. By specifying i32 here, rather than f32 and doing the
+    // conversion ourself, we rely on v8's type conversion which is both more
+    // flexible (e.g. handles NaN) and will be more consistent with a browser.
+    pub fn _elementsFromPoint(_: *parser.DocumentHTML, x: i32, y: i32, page: *Page) ![]ElementUnion {
+        const element = page.renderer.getElementAtPosition(x, y) orelse return &.{};
+        // TODO if pointer-events set to none the underlying element should be returned (parser.documentGetDocumentElement(self.document);?)
+
+        var list: std.ArrayListUnmanaged(ElementUnion) = .empty;
+        try list.ensureTotalCapacity(page.call_arena, 3);
+        list.appendAssumeCapacity(try Element.toInterface(element));
+
+        // Since we are using a flat renderer there is no hierarchy of elements. What we do know is that the element is part of the main document.
+        // Thus we can add the HtmlHtmlElement and it's child HTMLBodyElement to the returned list.
+        // TBD Should we instead return every parent that is an element? Note that a child does not physically need to be overlapping the parent.
+        // Should we do a render pass on demand?
+        const doc_elem = try parser.documentGetDocumentElement(parser.documentHTMLToDocument(page.window.document)) orelse {
+            return list.items;
+        };
+        if (try parser.documentHTMLBody(page.window.document)) |body| {
+            list.appendAssumeCapacity(try Element.toInterface(parser.bodyToElement(body)));
+        }
+        list.appendAssumeCapacity(try Element.toInterface(doc_elem));
+        return list.items;
+    }
+
+    pub fn documentIsLoaded(self: *parser.DocumentHTML, page: *Page) !void {
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        state.ready_state = .interactive;
+
+        log.debug(.script_event, "dispatch event", .{
+            .type = "DOMContentLoaded",
+            .source = "document",
+        });
+
+        const evt = try parser.eventCreate();
+        defer parser.eventDestroy(evt);
+        try parser.eventInit(evt, "DOMContentLoaded", .{ .bubbles = true, .cancelable = true });
+        _ = try parser.eventTargetDispatchEvent(parser.toEventTarget(parser.DocumentHTML, self), evt);
+
+        try page.window.dispatchForDocumentTarget(evt);
+    }
+
+    pub fn documentIsComplete(self: *parser.DocumentHTML, page: *Page) !void {
+        const state = try page.getOrCreateNodeState(@ptrCast(@alignCast(self)));
+        state.ready_state = .complete;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: HTML.Document" {
+    try testing.htmlRunner("html/document.html");
+}

src/browser/html/error_event.zig

@@ -0,0 +1,86 @@
+// Copyright (C) 2023-2024  Lightpanda (Selecy SAS)
+//
+// Francis Bouvier <francis@lightpanda.io>
+// Pierre Tachoire <pierre@lightpanda.io>
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License as
+// published by the Free Software Foundation, either version 3 of the
+// License, or (at your option) any later version.
+//
+// This program is distributed in the hope that it will be useful,
+// but WITHOUT ANY WARRANTY; without even the implied warranty of
+// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+// GNU Affero General Public License for more details.
+//
+// You should have received a copy of the GNU Affero General Public License
+// along with this program.  If not, see <https://www.gnu.org/licenses/>.
+const Env = @import("../env.zig").Env;
+const parser = @import("../netsurf.zig");
+
+// https://developer.mozilla.org/en-US/docs/Web/API/ErrorEvent
+pub const ErrorEvent = struct {
+    pub const prototype = *parser.Event;
+    pub const union_make_copy = true;
+
+    proto: parser.Event,
+    message: []const u8,
+    filename: []const u8,
+    lineno: i32,
+    colno: i32,
+    @"error": ?Env.JsObject,
+
+    const ErrorEventInit = struct {
+        message: []const u8 = "",
+        filename: []const u8 = "",
+        lineno: i32 = 0,
+        colno: i32 = 0,
+        @"error": ?Env.JsObject = null,
+    };
+
+    pub fn constructor(event_type: []const u8, opts: ?ErrorEventInit) !ErrorEvent {
+        const event = try parser.eventCreate();
+        defer parser.eventDestroy(event);
+        try parser.eventInit(event, event_type, .{});
+        parser.eventSetInternalType(event, .event);
+
+        const o = opts orelse ErrorEventInit{};
+
+        return .{
+            .proto = event.*,
+            .message = o.message,
+            .filename = o.filename,
+            .lineno = o.lineno,
+            .colno = o.colno,
+            .@"error" = if (o.@"error") |e| try e.persist() else null,
+        };
+    }
+
+    pub fn get_message(self: *const ErrorEvent) []const u8 {
+        return self.message;
+    }
+
+    pub fn get_filename(self: *const ErrorEvent) []const u8 {
+        return self.filename;
+    }
+
+    pub fn get_lineno(self: *const ErrorEvent) i32 {
+        return self.lineno;
+    }
+
+    pub fn get_colno(self: *const ErrorEvent) i32 {
+        return self.colno;
+    }
+
+    pub fn get_error(self: *const ErrorEvent) Env.UndefinedOr(Env.JsObject) {
+        if (self.@"error") |e| {
+            return .{ .value = e };
+        }
+        return .undefined;
+    }
+};
+
+const testing = @import("../../testing.zig");
+test "Browser: HTML.ErrorEvent" {
+    try testing.htmlRunner("html/error_event.html");
+}