Documentation preview for this PR (built with commit d315182; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

I applied your suggestions.
And also merge drinfeld_tau in this ticket for compatibility.

It's really tedious and frustrating going through each build or test fail individually, just to inevitably note that the problem comes either from internal machinery or a contribution that was not caught. I really hope the maintainers do something about that.

PR approved.

Why are you deleting conftest.py?

Why are you deleting conftest.py?

Sorry. It always disappear by itself, I don't know why.

It's really tedious and frustrating going through each build or test fail individually, just to inevitably note that the problem comes either from internal machinery or a contribution that was not caught. I really hope the maintainers do something about that.

PR approved.

I'm frustrated too. I tried to implement #39539 as a workaround, but it isn't working and I don't know why.

For now, I mostly just look at the meson one and if say at most 2 of them are failing then it's probably just flaky tests.

This doesn't work with build-doc or build-doc-pdf however, there's only one run of it. Doesn't help that you need to search through a huge log file for the error.

Why are you deleting conftest.py?

Sorry. It always disappear by itself, I don't know why.

Strange. After #39363 there should be no issue with it.

Unless… the pull request itself is responsible for the disappearing file.

As a workaround (I of course encourage proper fix, but we don't have enough people to look at all these bugs, unfortunately) you can

git update-index --assume-unchanged conftest.py

Yes, documentation build failure is legitimate.

I suggest checking out the documentation preview (see message by github-actions above) to see if it renders correctly. At least I see two places where it's wrong

The easiest way to check is to click the "changes is ready" button; however this currently give some false positive, but not so much that it becomes useless.

#40530

actually the established macro is \newcommand{\GF}[1]{\Bold{F}_{#1}}. (personally I prefer mathbb, but just for consistency with other parts of the code base)

(this macro is available for you to use)

Actually, we already use mathbb everywhere in the drinfeld_modules directory.
However, I can switch to GF if you think it's better.

I mean, if people decide mathbb is better than mathbf, it's trivial to change the definition of the GF command at one place.

… but whatever, let's get this in. (if build-doc-pdf is successful and no missing character error)

Thanks very much @user202729 for the help and many tips. It's nice.

Looks like the failures are again not related to the PR.
I give back a positive review. Hope this PR will be merged soon :-)

Hope this PR will be merged soon :-)

Me too!!!