remove usage of the _deprecate_positional_args decorator, and tests for its behavior

[d-v-b]

TODO:

• Add unit tests and/or doctests in docstrings
• Add docstrings and API docs for any new/modified user-facing classes and functions
• New/modified features documented in docs/user-guide/*.rst
• Changes documented as a new file in changes/
• GitHub Actions have all passed
• Test coverage is 100% (Codecov passes)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.61%. Comparing base (7154f98) to head (77a9b16).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3225      +/-   ##
==========================================
- Coverage   94.76%   94.61%   -0.15%     
==========================================
  Files          78       78              
  Lines        8672     8643      -29     
==========================================
- Hits         8218     8178      -40     
- Misses        454      465      +11     

Files with missing lines	Coverage Δ
src/zarr/api/synchronous.py	92.85% <ø> (-1.81%)	⬇️
src/zarr/core/array.py	98.37% <ø> (-0.15%)	⬇️
src/zarr/core/group.py	94.81% <ø> (-0.07%)	⬇️

... and 3 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[d-v-b]

there's a reduction in coverage, but the previously covered lines of code were not actually being tested for correctness -- they were just being run to check the behavior of the positional args decorator. So I think we can disregard the coverage reduction here.

[d-v-b]

this failure is due to recent changes to our hypothesis tests, presumably revealing a real bug, but likely one unrelated to this PR (cc @dcherian)

[dstansby]

This breaks the hypothesis tests and pre-commit...

[d-v-b]

pre-commit is failing, and hypothesis is also reporting a failed test, but not because of this PR

[dstansby]

There were pre-existing pre-commit errors before this PR, but this introduced more. And I'm pretty sure this broke the hypothesis tests? They were working before this PR was merged: https://github.com/zarr-developers/zarr-python/actions/runs/16271528357/job/45940137517

[d-v-b]

compare the changes in this PR (removing a decorator that only raised a warning, and some tests for that decorator) with the pre-commit errors (type-checking failures in obstore and hypothesis tests). I don't see how they are related other than by timing. We should just fix the pre-commit errors (by correcting type hints as needed).

As for the hypothesis test failure, I don't see how that could be related to the changes in this PR since it pertains to deleting groups from a local file system-backed store. we should of course also fix this. but it seems very unlikely to be caused by the changes in this PR.

[dstansby]

I don't see how they are related other than by timing.

If I run pre-commit on the main commit before this was merged (2f64d67), I get 7 errors. If I run it after this was merged (b267ff6) I get 14 errors. I think that's a pretty clear smoking gun that the new errors were caused by this PR, and this PR changed the acceptable call signatures of lots of functions/methods, I don't think it's that unlikely. For the same reason I don't think it's unlikely that the hypothesis failures were caused by this PR.

[d-v-b]

Good detective work! Do you want to work on fixing the mypy problems? And @dcherian do you have any idea what's causing the hypothesis test failures?