MNT: Drop Python 3.7 and update test matrix again

[pllim]

MNT: Drop Python 3.7

TST: Test against Sphinx 7.2 instead of 7.1

Fix #175

[codecov]

Codecov Report

Merging #177 (8497ebc) into main (42a1900) will not change coverage.
Report is 2 commits behind head on main.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main     #177   +/-   ##
=======================================
  Coverage   89.87%   89.87%           
=======================================
  Files          27       27           
  Lines        1146     1146           
=======================================
  Hits         1030     1030           
  Misses        116      116           

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[bsipocz]

Isn't this considered major release? I would consider keeping this for now.

[pllim]

Given all the breakage from yesterday, I guess, LoL. Can I drop "7.0" then? I don't want to keep a super long list of old Sphinx releases.

[bsipocz]

I would say wait? It's released back in April, and maybe there are still some <7.1 pins around. I don't expect much failures, but then we also don't run the test suite that often that it should be that big of a deal to cover all cases.

Dropping down to a single version for old major numbers sounds good though.

[pllim]

Do you have a preference between 5.3 and 6.2?

[bsipocz]

One of 5.x.y and 6.z.w should be enough I think. Actually I would stay with the oldest on each major versions, but it's not at all a strong preference. for "oldest" though I would very much use the same that is declared as dependency, so would hard wire '4.0.0' rather than allowing it to pick-up the lastest bugfix.

[pllim]

hard wire '4.0.0' rather than allowing it to pick-up the lastest bugfix

Wouldn't it make sense to pick up the bug fixes too? I don't promise it would work with a buggy release for other versions.

[bsipocz]

Wouldn't it make sense to pick up the bug fixes too?

No, I think that would be the same logic of supporting only the latest and greatest version, as that one has the most features and fixes.

My logic here is that the config files state a minimum required version, our task is to make sure that minimum is in fact the minimum and not just aspiration, therefore that minimum should be the one tested.

[pllim]

I think I addressed your comments but I am not sure. Please re-review. Thanks!

[bsipocz]

Oh, I mean keep one of each of the major versions (e.g. don't drop 5.x.y), but one should be enough, no need for multiple between 5.x and 6.y

[pllim]

Okay, what about now?

[bsipocz]

One more inline comment, but it's 🚲 🏠 , so up to you whether to have it or not, the PR is ready to go.

[bsipocz]

Thanks!