Report the renderer's tag and attribute drops to HtmlChangeListener (#469) - #480
Merged
Conversation
HtmlStreamRenderer refuses a start tag whose name is not one HTML allows, which an ElementPolicy can produce by renaming, refuses a tag arriving inside literal content it is writing, and leaves off an attribute whose name is not one HTML allows. Each reached only the bad-HTML handler as a message, which PolicyFactory.sanitize wires to Handler.DO_NOTHING, and since the policy had emitted the tag or attribute, HtmlChangeReporter counted it as kept and told the listener nothing. The renderer's drop listener, which since #463 carried literal content it could not emit, now carries these too, as a DropListener with a method per kind. The reporter maps a refused tag onto discardedTag under the input name, treating the attributes the policy rejected on it as the policy's own decisions, as for an attributeless skip; and maps an attribute left off onto discardedAttributes and discardedAttribute, with the value the author wrote where an input copy accounted for it and as the renderer received it where a policy added it. An end tag the renderer refuses is the same loss as its start tag and is not reported again. Fixes #469 Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014Ydng7gBm6Ax5zfwt4vZip
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #469.
HtmlStreamRendererhas three drop sites that reached only the bad-HTML handler, whichPolicyFactory.sanitizewires toHandler.DO_NOTHING: a start tag whose name is not a valid HTML name (anElementPolicyrenaming toa@bproduces one), a tag arriving inside literal content the renderer is writing (a policy renamingdivintostyleproduces one), and an attribute whose name is not a valid HTML name. Since the policy had emitted each,HtmlChangeReportercounted it as kept and the listener heard nothing.The renderer's package-private drop listener, added in #463 for literal content it could not emit, becomes a
DropListenerwith a method per kind, and the reporter maps them onto the existing listener methods:discardedTagunder the input name. Attributes the policy rejected on it are still reported, as they are for a tag skipped for having none: rejecting them was the policy's decision and the tag went for another reason. The end tag the renderer refuses later is the same loss and is not reported again.discardedAttributesanddiscardedAttribute, with the value the author wrote where an input copy accounted for it, and as the renderer received it where a policy added it.The bad-HTML handler keeps receiving its messages unchanged; the existing renderer tests for those messages pass as they are. Five new tests: one on the renderer pins the channel for all three kinds, four on the reporter pin each mapping through
PolicyFactory-built policies.HtmlChangeListener's Javadoc says what is now reported and under which conditions. Verified with./mvnw clean verifyon JDK 11, 17, 21 and 25.One side effect of the literal-content site remains and is out of scope here: the policy's open-element stack still counts a tag the renderer refused as open, so text approved for a nested element inside a renamed
stylelands as raw style content. I will file that separately.🤖 Generated with Claude Code
https://claude.ai/code/session_014Ydng7gBm6Ax5zfwt4vZip