advance() asserts that each byte it steps over is either an ASCII byte
or a UTF-8 sequence leader, never a continuation byte (0x80–0xBF).
consumeName() was calling advance(1) for all non-ASCII bytes
('\x80'...'\xFF'), processing multi-byte sequences one byte at a time.
For a two-byte sequence like é (0xC3 0xA9), the second iteration landed
on the continuation byte 0xA9 and triggered the assertion, crashing the
browser in Debug mode.
Fix: replace advance(1) with consumeChar() for all non-ASCII bytes.
consumeChar() reads the lead byte, derives the sequence length via
utf8ByteSequenceLength, and advances the full code point in one step,
so the position never rests on a continuation byte.
Observed on saintcyrlecole.caliceo.com, whose root element carries an
inline style with custom property names containing French accented
characters (--color-store-bulles-été-fg, etc.). The crash aborted JS
execution before the Angular app could render any dynamic content.
Most significantly, if removing from the multi fails, the connection
is added to a "dirty" list for the removal to be retried later. Looking at
the curl source code, remove fails on a recursive call, and we've struggled with
recursive calls before, so I _think_ this might be happening (it fails in other
cases, but I suspect if it _is_ happening, it's for this reason). The retry
happens _after_ `perform`, so it cannot fail for due to recursiveness. If it
fails at this point, we @panic. This is harsh, but it isn't easily recoverable
and before putting effort into it, I'd like to know that it's actually happening.
Fix potential use of undefined when a 401-407 request is received, but no
'WWW-Authenticate' or 'Proxy-Authenticate' header is received.
Don't call `curl_multi_remove_handle` on an easy that hasn't been added yet do
to error. Specifically, if `makeRequest` fails during setup, transfer_conn is
nulled so that `transfer.deinit()` doesn't try to remove the connection. And the
conn is removed from the `in_use` queue and made `available` again.
On Abort, if getting the private fails (extremely unlikely), we now still try
to remove the connection from the multi.
Added a few more fields to the famous "ScriptManager.Header recall" assertion.
This PR introduces a custom CDP domain 'LP' (Lightpanda) to expose browser-specific tools. The first method, 'LP.getMarkdown', allows retrieving a Markdown representation of the DOM or a specific node by its 'nodeId'. This is optimized for AI agents and LLM-based scraping tasks.
This allows us to leverage the Caller.Function.call method, which does type
mapping, caching, etc... and allows the Zig function callback to be written like
any other Zig WebAPI function.
ReadableStream.zig was the only webapi file importing v8 directly.
Extract the repeated newFunctionWithData / callback boilerplate into
js/Local (newFunctionWithData) and js/Caller (initFromHandle,
FunctionCallbackInfo.getData), and update ReadableStream and Context
to use them.
Returns 1 when writable (default high water mark), 0 when closed,
and null when errored, matching the spec behavior for streams
without a custom queuing strategy.
After BOM stripping or when receiving an empty Uint8Array, the
decoded input can be zero-length. Per spec, empty chunks should
produce no output rather than enqueuing an empty string.
When JS called controller.enqueue(42), the value was coerced to the
string "42" because Chunk only had uint8array and string variants.
Add a js_value variant that persists the raw JS value handle, and
expose enqueueValue(js.Value) as the JS-facing enqueue method so
numbers, booleans, and objects round-trip with their original types.
- Use StaticStringMap and enums for method, tool, and resource lookups.
- Implement comptime JSON minification for tool schemas.
- Refactor router and harness to use more efficient buffered polling.
- Consolidate integration tests and add synchronous unit tests.
Replace synchronous queue-draining approach with async promise-based
piping using V8's thenAndCatch callbacks. PipeState struct manages the
async read loop: reader.read() returns a Promise, onReadFulfilled
extracts {done, value}, writes chunks to the writable side, and
recurses via pumpRead() until the stream closes.
Implement synchronous piping that drains queued chunks from a
ReadableStream into a WritableStream. pipeThrough accepts any
{readable, writable} pair (TransformStream, TextDecoderStream, etc.)
and returns the readable side. pipeTo writes all chunks to a
WritableStream and resolves when complete.
Mirrors TextEncoderStream: wraps a TransformStream with a Zig-level
transform that converts Uint8Array chunks to strings. Supports the
same constructor options as TextDecoder (label, fatal, ignoreBOM).
Add the missing Streams API types needed for TextEncoderStream support:
- WritableStream with locked/getWriter, supporting both JS sink callbacks
and internal TransformStream routing
- WritableStreamDefaultWriter with write/close/releaseLock/closed/ready
- WritableStreamDefaultController with error()
- TransformStream with readable/writable accessors, JS transformer
callbacks (start/transform/flush), and Zig-level transform support
- TransformStreamDefaultController with enqueue/error/terminate
- TextEncoderStream that encodes string chunks to UTF-8 Uint8Array
via a Zig-level transform function
