pkgs/sage-conf: Move metadata from setup.cfg to pyproject.toml

[mkoeppe]

This is a trivial "chore" PR. It updates Python metadata to the latest format. No controversies about the current format are known about in the Python community. In a typical open source project, someone in a Maintainer role would open a PR and then immediately merge it, or when receiving such a PR from the outside, quickly review and merge it (examples: my PRs pytest-dev/pytest-mock#410 (merged in within 1 day), pyodide/pyodide#4472, pytest-dev/pytest-xdist#1020,
sagemath/cypari2#158, fplll/fpylll#258, polymake/JuPyMake#2, cvxpy/cvxpy#2276, sagemath/cysignals#177).

• In part split out from Use meson instead of configure for conda install #36489
• Part of Migrate from setup.cfg to pyproject.toml #33577
  • see also pkgs/sagemath-standard: Move metadata from setup.cfg to pyproject.toml #37902

📝 Checklist

• The title is concise, informative, and self-explanatory.
• The description explains in detail what this PR is about.
• I have linked a relevant issue or discussion.
• I have created tests covering the changes.
• I have updated the documentation accordingly.

⌛ Dependencies

• Depends on pkgs/sage-conf_pypi: Repair after #36400, #36435 #36533 (which changes setup.cfg, merged here)
• Depends on Switch more Python toolchain packages to use (platform-independent) wheels #36885

[kiwifb]

There should be links for sage_conf_conda and sage_conf_pypi or do they not show up in github?

[mkoeppe]

These links are already in the repo and there's no change to them on this PR.

[kiwifb]

Scrap that, I thought the file was brand new. It is already linked. Too early on Monday for me to review properly.

[mkoeppe]

Note that there was already pyproject.toml for the build-system information; here we're just adding more sections to the file.

[mkoeppe]

Happy Monday!

[tobiasdiez]

Please make sure that this change is compatible with the usage of pyproject.toml in #36489 (I think that PR uses the default config that you overwrite here).

[mkoeppe]

Sure, there's no reason for that implementation of sage-conf to be incompatible with this pyproject.toml.

[mkoeppe]

Let's get this in please

[tobiasdiez]

Sure, there's no reason for that implementation of sage-conf to be incompatible with this pyproject.toml.

I've tried it with #36489 and its not working as there is no _sage_conf folder anymore. I think its better to stick with the defaults (i.e. either a src or sage_conf folder containing the sources).

[mkoeppe]

That #36489 is not ready cannot be blamed on this PR.

[kiwifb]

LGTM, I do not have any unexpected issue with it in sage-on-gentoo.

[mkoeppe]

Thank you!

[tobiasdiez]

As I've said before, these lines

packages = ["_sage_conf"]
py-modules = ["sage_conf"]

are not okay for me.

(I think its also not polite to cherry pick commits from a PR and to then introduce changes that lead to conflicts with that PR)

[mkoeppe]

As I've said before, these lines

packages = ["_sage_conf"]
py-modules = ["sage_conf"]

are not okay for me.

Tobias, what Python packages and modules are part of the sage-conf distribution is not a matter of opinion or preference, or anything that makes sense to "negotiate". It is just a fact, which was previously declared in setup.cfg; and now it is declared in pyproject.toml.

I also know that you know that.

[tobiasdiez]

Matthias, you are blocking #36489 and on top of that you try to enforce your conventions via this PR. This is not a nice approach.

[mkoeppe]

No, Tobias, as I have just explained, what is declared in pyproject.toml "is not a matter of opinion or preference", and certainly it is not "my convention". It is the standard. Do I need to point out that you opened #33577, which asks for exactly what is done in this PR?

[mkoeppe]

---> sage-abuse (again)

(Edit: I've referred this incident -- abusing of your privileges as a member of the Triage team, and making false and harmful accusations -- to the sage-abuse committee.)

[tobiasdiez]

Matthias, please don't add a positive review label when there is a negative review. Thanks!

I respect that you don't agree with me nor that you want to discuss things with me, but that doesn't imply that my review just disappears.

[mkoeppe]

Lint failure is unrelated, fixed in #36822.

[mkoeppe]

I think we are free to vote against this PR because it further entrenches a bad design, or because we are drunk, or for any other reason.

Yes, that's how voting works. But that does not stop me from reminding community members about their duty to act responsibly.

[github-actions]

Documentation preview for this PR (built with commit 9977681; changes) is ready! 🎉
This preview will update shortly after each push to this PR.

[mkoeppe]

The current tally:

• 4 in favor: @mkoeppe @kwankyu @kiwifb @NathanDunfield
• 3 against: @tobiasdiez @tornaria @orlitzky

Please vote!

[saraedum]

I vote -1 on this PR.

[mkoeppe]

I'll note that I am the developer who has been maintaining these distribution packages in pkgs/. I have built and maintain the templating facility (via m4) because synchronizing the version information and other metadata simplifies this maintenance work; it's a carefully thought out and executed part of my plan for the modularized distributions. I would ask reviewers here on the PR and similar PRs to take this into account.

[mkoeppe]

9 months later, this PR is still open.

The current tally:

• 4 in favor: @mkoeppe @kwankyu @kiwifb @NathanDunfield
• 4 against: @tobiasdiez @tornaria @orlitzky @saraedum

Please vote!

[culler]

This PR seems to replace a pyproject.toml file by an m4 template that can be used to
automatically update version information when generating the pyproject.toml file for a new version. I don't see anything particularly controversial about that. There are surely alternatives to using m4, but Sage is already using m4, so this choice does not make a significant difference.