temporal: Fix FLY002: String-join inefficiencies in temporal/t.register tests

[arohanajit]

Fixed FLY002 and used f-strings instead of join in t.register/

[echoix]

@arohanajit Can your titles when making fixes using ruff be more specific? If you're fixing one rule, for one module, it will be way easier to find and understand when looking at the history what was changed.

There's hundreds of rules from all linters in ruff, and hundreds of files here, so I feel like saying "ruff errors" could be applied to dozens and dozens of PRs doing totally different things.

But, at least, you add the error code in the PR, but most often, since your commit messages only mention "update", once merged, that error code isn't searchable, when searching for a commit.

[echoix]

To know if it's specific enough: put yourself as the one that needs to find a commit to revert since something broke. You know what rule change caused the problem, and try to find the commits up to where do your bisect. If it's hard to find these commits, it's maybe not precise enough

[wenzeslaus]

To add to what @echoix is saying, it is clear that you are trying to fix things (including the word "Fix" is great, but not enough). It is also clear which directory or file it is when just a prefix or one name is included (t.register or temporal, but not both). Similarly, it does not add much to know that you are fixing errors, that's sort of expected, but which errors, that's the question.

If it is hard to express all changes in a title and be specific enough at the same time, maybe it is time to split the PR into two.

To @echoix's note on trying to see where the commit message (PR title) appears see (latest commits, history of one file, and blame):

https://github.com/OSGeo/grass/commits/main/
https://github.com/OSGeo/grass/commits/main/scripts/d.shade/d.shade.py
https://github.com/OSGeo/grass/blame/main/scripts/d.shade/d.shade.py

[echoix]

To be clear, the work is not bad ;)
It's to help get it even better and more useful for long-term reference (decades time span)

[arohanajit]

Thanks @wenzeslaus and @echoix I'll keep this in mind in future PRs

[petrasovaa]

Thanks @wenzeslaus and @echoix I'll keep this in mind in future PRs

So are you able to edit the PR title or you don't have permissions?

[petrasovaa]

This part is still failing in the CI

[petrasovaa]

@arohanajit unrelated to this PR, if you check the CI, there is lot's of noisy numpy warnings and ResourceWarnings, could you look into that (not in this PR)? I wonder if there was any issue/discussion about this already, do you know @echoix?

[echoix]

The numpy warnings are from doctests, and that is what is blocking testing with newer OS/Python versions like 3.13

First issue:

• [Bug] Python doctests must be updated for numpy 2+ or some later versions of numpy 1 #4100

Solving this would unblock

• CI(Ubuntu): Install latest version of Python (3.13) for full config #4372.
  See the end for a possible solution (works around by using the old string representation, not a real fix)

Also see:

• style: fix numpy-deprecation-rules (NPY) #4062
• [Bug] pygrass – test_pygrass_doctests fails with Numpy 2 #4116
• CI: update macOS runner dependencies #4120

After all that time, I feel like we might need to decide to not run doctests for every version. It would be impossible to always satisfy all. (Unless the doctests have conditionnals, that would be ugly)

[echoix]

For the resource warnings, at build time I see that there's some related to the files that make documentation (maybe mkhtml), so probably a fix like #5063 needs to be applied, or the devnull one from #4917 (comment)

[echoix]

Otherwise, it might be in doctests that we'll need to look at

[arohanajit]

@arohanajit unrelated to this PR, if you check the CI, there is lot's of noisy numpy warnings and ResourceWarnings, could you look into that (not in this PR)? I wonder if there was any issue/discussion about this already, do you know @echoix?

Got it, I'll start working on this after finishing with all the ruff errros