Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.56%. Comparing base (68c62e5) to head (8c5db54).

❗ Current head 8c5db54 differs from pull request most recent head d447151

Please upload reports for the commit d447151 to get more accurate results.

@@            Coverage Diff             @@
##             main     #265      +/-   ##
==========================================
+ Coverage   80.66%   84.56%   +3.89%     
==========================================
  Files          10       10              
  Lines         631      570      -61     
  Branches      131      121      -10     
==========================================
- Hits          509      482      -27     
+ Misses         84       56      -28     
+ Partials       38       32       -6     

Good evening @brabster ,

Thank you for your interest in this project, for investigating and even proposing a solution, it's great!

Indeed, it's a... pragmatic approach! It's a problem I've encountered before, and I opted for a... phlegmatic approach! Pragmatic > phlegmatic, so why not?

Before doing a more in-depth review, have you looked at the Mkdocs utilities for handling this kind of case? After all, it's part of the Markdown generator's job to manage HTML tags when it's about to convert. Because ideally you don't want to add homemade code that's complex to maintain, as it can't handle every possible and unimaginable case. Especially with regex.

Hi there! I looked in a couple of places for a solution but came up empty - mkdocs/catalog and pymdown-extensions. MagicLink doesn't look helpful. I'm not seeing anything in markdown extensions either but perhaps a fairly trivial markdown extension to write and avoid the nasty post-processing approach. I recall I started down this road initially and got stuck because my mkdocs-material blog has some complexity in the way it produces URLs - blog posts get yyyy/mm/dd and images don't, for example. Can't just glue the site URL on, it wont work. I'll have a go if I have time and see what I find out

Hmm, first attempt not going well. Found PathConverter but not able to get it working so far:

elif (
            abstract_delimiter
            and (
                excerpt_separator_position := in_page.markdown.find(abstract_delimiter)
            )
            > -1
        ):
            return markdown.markdown(
                    in_page.markdown[:excerpt_separator_position],
                    output_format="html5",
                    extensions=['pymdownx.pathconverter'],
                    extension_configs={
                        'pymdownx.pathconverter': {
                            'base_url': 'https://foo.bar'
                        }
                    })

I've also remembers that there's other little things that I have in mkdocs-material like snippets that don't get converted in the feed but do if you process the html like this

<description><p><img alt="A photo of a crab on a night dive in the red sea. Credit: me" src="./assets/phone.webp"></p><p>Why I've been trialling GitHub Codespaces as a more secure alternative to local development. I never expected to be pushing changes from my phone!</p><p>--8&lt;-- "ee.md"</p></description>

I remember now I was seeing little things not working correctly like this and it seemed you'd need to mirror the mkdocs-material configuration somehow - at that point, just processing the output HTML from the standard processing seemed like a pretty good idea!

Seems like this is just a fairly hard problem - pretty much every one of the feeds we are dropping into Slack have some kind of mess in the descriptions, even Medium.

That failing test is one that checks a description length below a fixed size - naturally my change makes that test fail because there's an h1 wrapper with an id attr now - don't think it's a real problem

To solve my remaining problems, I've added an additional step to clean up some of the wrapper tags that can appear in mkdocs-material - glightbox and grid-cards.

I wouldn't take it as-is if I were you. If you're interested in a version of it, perhaps protected by a specific opt-in config to avoid breaking anyone and naturally more tests, I can probably do that without much trouble. I'd rather clean it up and have it supported than effectively fork it, and I'm not seeing a better option right now 🤷

Hi @brabster

As you may have noticed, I've made a set of improvements lately and published them in a 1.13.0 version. Just to say that I did not forget your PR but I'm still a bit confused about how to implement it without slowing mkdocs build and I don't understand why you don't use the excerpt function, which lets you cut content where you want it?

EDIT: hum, sorry, I've spent too much time away from this project and my answer is a little off the mark. I'll discuss on the issue on the issue and I'll get back here when it's clear enougth.

Quality Gate passed

Issues
1 New issue
0 Accepted issues

Measures
0 Security Hotspots
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarCloud

Quality Gate passed

Issues
1 New issue
0 Accepted issues

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

See analysis details on SonarCloud