Use strict=True with zip to ensure length equality

[DanielYang59]

Summary

• Use strict=True when zip to ensure length equality @janosh, to finish Drop Python 3.9 support #4009 (comment)

TODOs

• Let's see how many tests are failing
• Make sure strict keyword is for zip not other function

[janosh]

thanks @DanielYang59! that was way faster than i expected. 👍

did you notice any cases where strict=True stood out as potentially problematic? knowing pymatgen's test coverage, i assume there are cases where strict=True wouldn't raise a CI warning despite strict=False being required for backwards compatibility

[DanielYang59]

did you notice any cases where strict=True stood out as potentially problematic?

Sorry I didn't really go through each of these replacements (just a batch replace and fixed those CI failures in 89dc247 and b149ebb). Perhaps I should mark it as draft now and fully go through all changes later.

i assume there are cases where strict=True wouldn't raise a CI warning

With strict=True a ValueError would be raised directly instead of warning. So it would be a good thing (reveal untested code) and a bad thing (might lead to many breakages from untested code).

[janosh]

With strict=True a ValueError would be raised directly instead of warning

sorry, i meant error, not warning

So it would be a good thing (reveal untested code) and a bad thing (might lead to many breakages from untested code).

agreed

Perhaps I should mark it as draft now and fully go through all changes later.

not strictly necessary imo. these issues are very hard to spot by just looking at the code. overall less time spent reporting and fixing errors as they occur

[DanielYang59]

not strictly necessary imo. these issues are very hard to spot by just looking at the code. overall less time spent reporting and fixing errors as they occur

Yes that's what I thought... Though sounds a bit irresponsible, but it's quite hard just by eyeballing

[janosh]

thanks as always @DanielYang59!

[DanielYang59]

No problem. Thanks for reviewing. Fingers crossed that this PR wouldn't be pinged a hundred times :)