Fix: catch 404s but ignore time out errors

[lwasser]

I think we should be catching 404s but ignoring timeout errors in CI. let's see if that helps

[lwasser]

Okay, this caught some legitimate 404 errors 🚀 and the timeout errors (0) are gone. Let's not merge this until I fix the broken links. Then, maybe we can remove footnotes altogether to avoid the translation issues. I'm not sure footnotes are worth it in tutorial content.

[sneakers-the-rat]

Heck ya.

I think it would be worth diagnosing the footnotes problem - it indicates that something is wrong with how markdown is being parsed in the translated sites, and we wouldn't want that to e.g. break some other kind of markdown element without us knowing about it.

I need to take a look at how/when myst parser is evoked relative to the gettext injection, bc it's super odd that the target footnote rendered and even had the correct backlink, but the forward link wasn't rendered.

[lwasser]

@sneakers-the-rat, thank you! If you have any ideas as to what is happening, that would be great (regarding the footnotes!). I keep looking at it, and I'm not sure how to debug this issue as it's specific to the translations.

And yes! It's really odd that the target is ok, and it's struggling with the footnotes. I hadn't considered that this could be a myst issue! I could post a bug report tomorrow in the myst repo to see if they have any ideas.

[sneakers-the-rat]

Searching issues in myst-parser, found one that looks like the same problems: executablebooks/MyST-Parser#690

and someone had already backlinked to a previous issue of ours where @flpm had resolved it with a venv version update: #390

So maybe there were two bugs in #390? This footnote rendering bug (that has a patch in that issue) as well as the venv version problem?

[flpm]

It makes sense, I wonder though how the issue I was having magically fixed itself.

Not sure if related but I just noticed that our build book CI action is running on a Python 3.9, maybe that forces older versions of the sphinx modules?

[flpm]

Maybe the bug happens for a certain combination of Sphinx module version. That happened in my outdated venv and it also happens in CI because of Python 3.9? Just a crazy theory, I am away from my computer now so I can check 😞