slachiewicz opened a new pull request, #147: URL: https://github.com/apache/maven-doxia-converter/pull/147
Follow-up to #146, which I should have included there — it merged while I was still pushing. `tools/normalize-site-page.py` as merged reduces a page to its visible text and link targets. That means a page which has lost its **title, authors and date** compares equal to its own pre-migration self, on every page, with no diff at all. This is not a hypothetical failure mode. It is how a large APT→Markdown conversion shipped with the metadata of several hundred pages silently dropped: the body matched everywhere, so nothing flagged it. The fix is both halves of the cause: - **The script** now emits the `<title>` and the `author`/`date` `<meta>` content alongside the body, so a `diff` catches the loss. - **The guide** now says to keep the YAML front matter the converter writes — it carries that metadata out of the APT header — and that it must be the first bytes of the file, with the licence header below it, because the Markdown parser only looks for front matter when the source starts with `---`. Deleting it makes the parser fall back to the first heading for the title, so a page headed *Release Notes* under project *Modello* retitles itself `Modello – Modello`. The "Checking the result" section now leads with comparing the `<head>`, not just the body. **Verified:** `mvn site` builds clean, all eight section headings still render, and the normaliser emits the `[TITLE:...]` token on the generated page. <!-- Generated with assistance from Claude Opus 5; see the Generated-by trailer on the commit. --> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
