[Zerthick]

Description

Resolve double-encoded HTML entities (e.g. &, <, >) in RSS feed title and description fields.

MkDocs pre-escapes content, and then Jinja's |e filter escapes again, producing double-encoded entities. Using html.unescape() in Python before template rendering ensures |e performs a single correct escape.

Changes

• Add html_module.unescape() for title and description in plugin.py
• Add test fixture with special characters (& < >)
• Add test verifying no double-encoded entities in output

---

Split from: #451

---

💡 AI Contribution Level: 8 — Based on the VisiData AI contribution scale, this was an AI-assisted contribution where the AI performed the code splitting and PR creation with minimal human guidance.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.18%. Comparing base (1b5e630) to head (ada0a8d).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #456      +/-   ##
==========================================
+ Coverage   83.16%   83.18%   +0.02%     
==========================================
  Files          11       11              
  Lines         784      785       +1     
  Branches      133      133              
==========================================
+ Hits          652      653       +1     
  Misses         86       86              
  Partials       46       46              

Flag	Coverage Δ
unittests	83.18% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
mkdocs_rss_plugin/plugin.py	89.50% <100.00%> (+0.06%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Guts]

Thanks for this one.

The description part, however, introduces a regression, and the test doesn't exercise the bug:

1. description is HTML, not text : get_description_or_abstract() returns page.content or markdown.markdown(...) output in almost every branch, and the JSON Feed publishes it as content_html. In RSS, <description> carries that HTML escaped once for XML, so < in the raw XML is the correct encoding of a < in the HTML. Unescaping it first turns code samples into real markup. With Use `<script>` and `a < b` in code.:

   • main: readers get <p>Use <code>&lt;script&gt;</code> and <code>a &lt; b</code> in code.</p>
   • this PR: readers get <p>Use <code><script></code> and <code>a < b</code> in code.</p>

   For a docs plugin, where inline code is everywhere, that's a blocker. Please keep the unescape on title only (see inline suggestions).

2. fixture doesn't reproduce the bug because its title is set in the front-matter, which MkDocs returns verbatim; only H1-derived titles carry entities. Same for the description meta, which |e already escapes once. The new test passes on main without the fix. I've suggested a fixture with an H1 title instead.

Please also rebase on main.

[Guts]

This module can be removed: it duplicates a full build already done in test_build.py, and assertNotIn("<") rejects the valid encoding of HTML descriptions. A single assertion in test_xml_escaping_in_author covers the title fix.

[Guts]

Short comment to reùember why only the title gets this treatment:

Suggested change

	title=html_module.unescape(page.title) if page.title else None,
	# MkDocs keeps HTML entities in titles extracted from the H1
	title=html_module.unescape(page.title) if page.title else None,

[Guts]

description is HTML (page.content / markdown.markdown() output, published as content_html in the JSON Feed): it must stay escaped once, so no unescape here.