Conflict with abbreviations, toc extension, and typeset plugin

[jonaharagon]

Context

Updating to the latest mkdocs-material-insiders breaks some tables of contents on our staging site: https://deploy-preview-2177.preview.privacyguides.dev/en/vpn/#ipv6-support

Bug description

In the Table of Contents with the typeset plugin enabled, if there's an abbreviation in a heading, the abbreviation and any subsequent text doesn't get rendered at all.

If the abbreviation is at the very beginning of the header it does show the entire heading text, but the permalink icon gets erroneously included at the end:

Related links

• Reporting a bug
• https://squidfunk.github.io/mkdocs-material/setup/extensions/python-markdown/?h=toc#table-of-contents
• https://squidfunk.github.io/mkdocs-material/setup/extensions/python-markdown/?h=abbr#abbreviations
• https://squidfunk.github.io/mkdocs-material/reference/?h=typeset#built-in-typeset-plugin
• Navigation: typeset plugin not rendering tooltips for abbreviations #5165
• Material for MkDocs `v9.1.12` appears to break the abbreviation tooltips on hover #5529

Reproduction

example.zip

Steps to reproduce

1. Enable typeset plugin
2. Enable TOC
3. Enable abbreviations
4. Use abbreviation in heading

Browser

No response

Before submitting

• I have read and followed the bug reporting guidelines.
• I have attached links to the documentation, and possibly related issues and discussions.
• I assure that I have removed all customizations before submitting this bug report.
• I have attached a .zip file with a minimal reproduction.

[squidfunk]

Thanks for reporting. So just to be clear – it works if you downgrade to 9.1.14+insiders-4.35.1, right? If so, it is very likely related to the change added in #5536. /cc @pawamoy.

[pawamoy]

Thanks for the ping @squidfunk 🙂
I tried locally, reverting the change from #5536, and the issue persists.

[pawamoy]

I think typeset's regular expressions are responsible here.

Here is an example heading containing the abbreviation, from OP's repro:

<h2 id="othertext-abbreviation-moretext">
  OtherText <abbr title="Test abbreviation!">Abbreviation</abbr> MoreText
  <a class="headerlink" href="#othertext-abbreviation-moretext" title="Permanent link">¶</a>
</h2>

The issue is simply that <abbr starts with <a 😅
So title = re.sub(r"<a.*?>.*?<\/a>$", "", title) removes abbreviation tags.
Quick, unfuzzed fix:

-            title = re.sub(r"^<a.*?>(.*?)<\/a>", r"\1", title)
-            title = re.sub(r"<a.*?>.*?<\/a>$", "", title)
+            title = re.sub(r"^<a\s+.*?>(.*?)<\/a>", r"\1", title)
+            title = re.sub(r"<a\s+.*?>.*?<\/a>$", "", title)

[squidfunk]

Ah, interesting, thanks for debugging! I'll check that asap!

[squidfunk]

Fixed in 72c3d0d5f – I adapted the fix proposed by @pawamoy and made it a little more explicit. Thanks again for the excellent reproduction, that made fixing and checking dead simple. Perfect issue!

[squidfunk]

Released as part of 9.1.15+insiders-4.35.3!