fix: resolve RSS feed validation errors

[Zerthick]

Summary

Fixes three RSS 2.0 validation errors in the generated feed.

Changes

1. <author> → <dc:creator> for item authors (template)

The RSS 2.0 spec requires <author> to be an email address (e.g., user@example.com). The plugin was using it for display names, which causes validation failures on every feed validator. Switched to <dc:creator> which is the correct element for human-readable names—the dc namespace is already declared in the template.

Before: <author>John Doe</author>
After: <dc:creator>John Doe</dc:creator>

2. Fix double-encoded HTML entities in title/description (Python)

MkDocs pre-escapes page titles and descriptions, then Jinja's |e filter escapes them again, producing &, <, > instead of &, <, >.

Fixed by unescaping in Python (html.unescape()) before passing to the template, so Jinja's |e performs a single correct escape. This is cleaner than template-level workarounds and benefits both RSS and JSON feeds.

Before: & (double-encoded)
After: & (correctly encoded)

3. Guard <enclosure> against None/non-positive length (template)

When the plugin fetches a remote image that returns 404, the length field becomes None, producing invalid XML like <enclosure length="None" />. Added a guard to skip the enclosure element when length is None or non-positive.

Before: {% if item.image is not none %}
After: {% if item.image is not none and item.image[2] is not none and item.image[2] > 0 %}

Tests

Added tests/test_rss_validation.py with three tests covering all fixes:

• test_dc_creator_instead_of_author
• test_no_double_encoded_entities
• test_enclosure_guard_none_length

Plus tests/fixtures/docs/page_with_special_chars.md fixture with & < > characters.

Impact

• Non-breaking: <dc:creator> is a standard RSS extension already supported by all feed readers
• Backward compatible: Existing feeds will just use a more correct element
• Defensive: Prevents invalid XML when remote images fail to load

[Guts]

Thanks you and your LLM for this contribution but can you ask it to comply with contributing guidelines please?

Also, are you sure that changes are relevant? Can you share your prompt and give an idea about the level of AI involvement here please?

[Zerthick]

I'm using Material for MkDocs for my own personal blog, I noticed 3 distinct errors in my RSS feed when put in a validator such as: https://www.rssboard.org/rss-validator/

The author issue seems to be a longstanding one: #192 Material for MkDocs typically uses a human readable name for the author, not an email causing validation to fail.

The other two seemed to be edge cases I encountered. Such as having a & character in my article title or an image preview failing to load.

Based on that scale I'd say this is a Level 8 contribution. I did not have a specific prompt, I gave my LLM the errors I was receiving in the validator and told it to resolve them, applied it's fixes, re-validated the feed, until all the errors were resolved.

I can confirm that the forked version is working correctly on my blog now and has resolved the feed and formatting issues I was observing before. https://www.datadelver.com/feed_rss_created.xml

I prompted my LLM to add unit tests, are there any other specific modifications to the PR you would like to see?

[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

[Zerthick]

@Guts You can view some of these validation errors on your own feed as well: