Fix curly quotes and use <a id> for the FAQ anchors - #1010
Merged
Conversation
Follow-up to the FML to Markdown conversion. maven-site-plugin 3.21.0 strips the name= attribute from an inline HTML anchor; 3.22.0 keeps it. On 3.21.0 every anchor on the page is therefore lost silently, on a green build. This repo builds with a parent that pulls 3.22.0, so the pages are correct today, but <a id> is the form that survives either way - and Xhtml5BaseParser reads Attribute.ID first, falling back to NAME only if it is absent, so id= is the primary path. name= on <a> is obsolete in HTML5 besides. Each anchor stays on its own line; folding one into the heading text suppresses the section's own generated id. Straight quotes in the prose are written as ". flexmark's smart-punctuation extension turns a straight quote in Markdown into a typographic one, so the converted page was showing "Old Archetype 1.0.x" with curly quotes where the FML original, being XML, rendered them straight. That is a visible-text change the anchor check cannot see. Verified by rendering faq.html before and after: the anchor set is unchanged and still a superset of what the original FML page served, the <head> is byte-identical, the visible text matches the FML original word for word, and the site log reports no "used more than once" warning. Generated-by: Claude Opus 5 (1M context)
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Follow-up to the FML → Markdown FAQ conversion, which is already merged here. Two small
fixes, one of which is a visible-text regression I introduced.
1. A visible-text regression — curly quotes
flexmark's smart-punctuation extension rewrites a straight
"in Markdown prose as atypographic quote. FML, being XML, rendered those quotes straight. So the converted page
shows “Old Archetype 1.0.x” where the original showed "Old Archetype 1.0.x".
This is my mistake in the conversion, and the anchor-preservation check could not see it —
it is a text change, not an anchor change. Writing
"in the source restores theoriginal rendering; this is the convention maven-site-plugin's own converted pages already
use.
2.
<a id>instead of<a name>maven-site-plugin 3.21.0 strips the
name=attribute from an inline HTML anchor; 3.22.0keeps it. On 3.21.0 every anchor on the page is lost silently, on a green build. This
repo builds with a parent that pulls 3.22.0, so the page is correct as it stands — but
<a id>survives either way,Xhtml5BaseParserreadsAttribute.IDfirst and only fallsback to
NAME, andname=on<a>is obsolete in HTML5.Anchors stay on their own line; folding one into the heading text suppresses the section's
own generated id.
Verification
Rendered
faq.htmlbefore and after:page served, so no deep link changes;
<head>is byte-identical;it did not before this change;
used more than onceanchor warning.target/sitewas deleted before each build and the page confirmed regenerated, so this isnot a stale-output pass.
Drafted with Claude — please verify