Branch: refs/heads/main
Home: https://github.com/WebKit/WebKit
Commit: 294b5b56d930d94057b56f016baf33826b9bd250
https://github.com/WebKit/WebKit/commit/294b5b56d930d94057b56f016baf33826b9bd250
Author: Anne van Kesteren <[email protected]>
Date: 2026-09-11 (Fri, 11 Sep 2026)
Changed paths:
M
LayoutTests/imported/w3c/web-platform-tests/html/syntax/parsing/cdata-in-integration-point-fragment-expected.txt
M
LayoutTests/imported/w3c/web-platform-tests/html/syntax/parsing/cdata-in-integration-point-fragment.html
M
LayoutTests/imported/w3c/web-platform-tests/html/syntax/parsing/html5lib_url_file=plain-text-unsafe-expected.txt
M Source/WebCore/html/parser/HTMLTreeBuilder.cpp
M Source/WebCore/html/parser/HTMLTreeBuilder.h
Log Message:
-----------
U+0000 NULL should be replaced with U+FFFD when fragment parsing in foreign
content
https://bugs.webkit.org/show_bug.cgi?id=323867
Reviewed by Chris Dumez.
The tokenizer's forceNullCharacterReplacement flag is computed from the
adjusted current
node after each token is processed. When fragment parsing the adjusted current
node is the
context element from the start, but the first characters are tokenized before
any token
reaches the tree builder, so NULL in the first character run was dropped
instead of being
replaced with U+FFFD.
Compute the flags when the fragment tree builder is constructed as well. This
also fixes
shouldAllowCDATA, so a CDATA section at the start of the fragment is no longer
parsed as a
bogus comment.
The extracted helper is ALWAYS_INLINE because constructTree() calls it for
every token.
Tests upstream: https://github.com/web-platform-tests/wpt/pull/62593
Canonical link: https://commits.webkit.org/320909@main
To unsubscribe from these emails, change your notification settings at
https://github.com/WebKit/WebKit/settings/notifications
