Add FixedTuple member enforcing a given number of items

[MatthieuDartiailh]

No description provided.

[codecov]

Codecov Report

Merging #211 (fb808da) into main (330b8fc) will increase coverage by 0.09%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #211      +/-   ##
==========================================
+ Coverage   97.57%   97.67%   +0.09%     
==========================================
  Files          24       24              
  Lines        1114     1075      -39     
  Branches      189      174      -15     
==========================================
- Hits         1087     1050      -37     
+ Misses         13       12       -1     
+ Partials       14       13       -1     

[MatthieuDartiailh]

@frmdstryr do you have an interest in this feature and do you want to review ?

[frmdstryr]

This could be moved to after the size check

[frmdstryr]

This is a really nice addition! A quick test and it seems to work fine.

Instead of adding a new member, what do you think about changing the existing Tuple to have a variable_length parameter that defaults to True if there is only one or no items? I think the original Tuple member will get practically no use if FixedTuple it is added as is.

It would be really nice if the type error would show the whole format eg tuple[int, bool] instead of %s-tuple, but I realize that's more complicated to do in cpp.

The unrelated formatting changes could probably be a separate commit.

[MatthieuDartiailh]

That's a good point. I will look into it. The error message may indeed be more challenging but I will look into it.

[MatthieuDartiailh]

So the issue with re-using the same member is that it is not backward compatible. Changes to how we interpret annotations makes the system stricter but it bothers me less since it should not break anything that was previously valid.

[frmdstryr]

I am normally against breaking backwards compatibility but in this case I think it is worth it. I'm so used to using the tuple[x, y, z] with typing I actually found an invalid use of the current Tuple that passes what should be the second type as the default value. Maybe others can chime in but do whatever.

Edit: After thinking about it more, i'd say just leave it as a separate member.

[MatthieuDartiailh]

Thanks @frmdstryr I will keep two separate members.

Please note that if you use annotations instead of the actual members tuple[x, y, z] will use FixedTuple which should fix your issue.

[MatthieuDartiailh]

I forgot to address your comments (forgot to re-read the whole discussion). I will come back to it once I am done with #210