Markdown syntax check

[nijel]

Is your feature request related to a problem? Please describe.
Many projects use Weblate to edit text files using markdown (and extracted through po4a).

Describe the solution you'd like
Weblate should provide quality check for markdown markup.

Additional context
Example project can be seen at https://hosted.weblate.org/projects/f-droid/website-posts/

Example po file: https://gitlab.com/fdroid/fdroid-website/blob/master/po/_posts.cs.po

[emmapeel2]

someone mentioned https://github.com/MihaZupan/MarkdownValidator

[eighthave]

I have a very basic script that check links in Markdown. I could probably port those to Weblate checks pretty easily. But a more complete solution would be nicer. Here is my script:

• https://gitlab.com/fdroid/fdroid-website/blob/1.72/tools/check-page-links.py

@matrixcoffee FYI

[nijel]

Thank you for your report, the issue you have reported has just been fixed. I've implemented checks which I think make sense (links, link references and inline markup).

• In case you see problem with the fix, please comment on this issue.
• In case you see similar problem, please open separate issue.
• If you are happy with the outcome, consider supporting Weblate by donating.

[sajolida]

Hi @nijel, I'm really happy to see this implemented!

I'm not really good at reading Python, so can you clarify if these checks cover, for example, missing or extra * for bold and italic?

For example:

• Good Markdown: This is **bold** and this is *italic*.
• Bad Markdown: This **bold* is broken.

[nijel]

So far it does check if markup in translation is same as in source.

[sajolida]

Of course! This sounds like an excellent approach :)

[eighthave]

One thing I see a lot in translations is a space between ] and ( that breaks the Markdown link syntax, e.g. '[Použij Weblate] (https://weblate.org/). That would be a good test case to have for this.

[nijel]

It does catch this (just to be sure, I've just added this to the testsuite).

[eighthave]

The other one that I see quite commonly is that translators translate the strings in the URL portion of the Markdown link. That of course breaks the link. That's also in https://gitlab.com/fdroid/fdroid-website/blob/1.72/tools/check-page-links.py Basically, I store all links that are in the source string, then for each link found in the translation string, I check to make sure it is in the list of source string links.

[eighthave]

The hard part with that check is that some sites might want to change all links to force the page to use the localized version, but I haven't seen that case in the stuff I've worked out.

[nijel]

Currently it doesn't trigger on changed URL, exactly for reason that it might be valid to change them to locale specific in some cases.

[eighthave]

On our site, we almost always use relative links for internal links, then the locale is inherited from the page, and doesn't need to be in the actual link itself. I think this is a relatively common setup for localized websites. And so far, no one has found a localized link for any of the external links that are included.

[emmapeel2]

With ikiwiki most of the links are locale-agnostic, but the inlines need to be changed (add the locale at the end as .es with Spanish)

[nijel]

Okay, maybe it makes sense to check relative links for being same.

[marcbria]

Sorry @nijel is this included in last weblate relase?
In other words, is weblate markdown ready?

[nijel]

What do you mean by "markdown ready"? It can validate markdown syntax in the strings, it can not parse markdown files, see #3106

[marcbria]

I was asking if is able to work with ".md" files, so you answered my question.
Thanks a lot.