Skip to content

Recognize short comments during encoding prescan - #610

Open
muneikh wants to merge 1 commit into
html5lib:masterfrom
muneikh:fix-prescan-short-comments
Open

muneikh wants to merge 1 commit into
html5lib:masterfrom
muneikh:fix-prescan-short-comments

Conversation

@muneikh

@muneikh muneikh commented Oct 11, 2026 •

Copy link
Copy Markdown

Encoding prescan skips the declaration following <!--> or <!---> because it starts searching for --> after the opener’s two hyphens. The HTML prescan algorithm permits those hyphens to form part of the terminator.

Start the terminator search two bytes earlier. Add direct prescan and binary-input regressions for both short comments, normal-comment controls, and a declaration inside a comment that must remain ignored. Include CHANGES and AUTHORS entries.

Fixes #427.

Validation: the full suite passes with 17,504 passes, 15,885 skips, and 683 expected failures on Python 3.11; all 197 encoding tests and changed-file Flake8 checks pass. Both short-comment cases fail with unchanged source. The test environment pins setuptools below 81 for the project’s existing pkg_resources dependency.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

_inputstream.EncodingParser fails to parse short bogus comments (prescan)

1 participant