Merge pull request #1642 from lightpanda-io/html-noscript-no-escape

don't escape noscript content on html dump
2026-03-21 20:24:42 +00:00 · 2026-02-25 07:31:26 +08:00
parent f7254ee169 1d8e0629af
commit d2a05bb622
2 changed files with 30 additions and 0 deletions
--- a/src/browser/dump.zig
+++ b/src/browser/dump.zig
@@ -294,6 +294,12 @@ fn shouldEscapeText(node_: ?*Node) bool {
    if (node.is(Node.Element.Html.Script) != null) {
        return false;
    }
+    // When scripting is enabled, <noscript> is a raw text element per the HTML spec
+    // (https://html.spec.whatwg.org/multipage/parsing.html#serialising-html-fragments).
+    // Its text content must not be HTML-escaped during serialization.
+    if (node.is(Node.Element.Html.Generic)) |generic| {
+        if (generic._tag == .noscript) return false;
+    }
    return true;
 }
 fn writeEscapedText(text: []const u8, writer: *std.Io.Writer) !void {
--- a/src/browser/tests/node/noscript_serialization.html
+++ b/src/browser/tests/node/noscript_serialization.html
@@ -0,0 +1,24 @@
+<!DOCTYPE html>
+<script src="../testing.js"></script>
+
+<!-- When scripting is enabled the HTML parser treats <noscript> as a raw text
+     element: its entire content is stored as a single text node containing the
+     raw markup.  Serializing it back (outerHTML / innerHTML) must output that
+     markup as-is without HTML-escaping the angle brackets. -->
+<noscript id="ns1"><h1>Hello</h1><p>World</p></noscript>
+<noscript id="ns2"><div id="bsky_post_summary"><h3>Post</h3><p id="bsky_display_name">Henri Helvetica</p></div></noscript>
+
+<script id="noscript-outerHTML">
+  const ns1 = document.getElementById('ns1');
+  testing.expectEqual('<noscript id="ns1"><h1>Hello</h1><p>World</p></noscript>', ns1.outerHTML);
+</script>
+
+<script id="noscript-innerHTML">
+  const ns2 = document.getElementById('ns2');
+  testing.expectEqual('<div id="bsky_post_summary"><h3>Post</h3><p id="bsky_display_name">Henri Helvetica</p></div>', ns2.innerHTML);
+</script>
+
+<script id="noscript-textContent">
+  // The raw text node content must be the literal HTML markup (not parsed DOM)
+  testing.expectEqual('<h1>Hello</h1><p>World</p>', ns1.firstChild.nodeValue);
+</script>