feat!: bump rust-polars to 0.30.0

[eitsupi]

https://github.com/pola-rs/polars/releases/tag/rs-0.29.0
https://github.com/pola-rs/polars/releases/tag/rs-0.30.0

[eitsupi]

The errors in Rust have been fixed but have not yet passed the R tests.
Feel free to edit this branch as I may not have enough time to work on this.

[etiennebacher]

I think this should close #230

[eitsupi]

I fixed as many of the test failures as I could.

The remaining failures are a problem where date_range_lazy now returns a list, so when converted to R, it becomes a list instead of a vector. (pola-rs/polars#8513)

[sorhawell]

Thanks. I will take a look at it.

[eitsupi]

I was just looking at this myself and Python's date_range behaves quite strangely.
I mean, when eager=True it returns an un-nested Series (r-polars' current behavior, lazy=TRUE/FALSE), but when eager=False (default now) it returns a list type, which I think is more appropriate here given the meaning of range, but anyway it returns something completely different for True and Flase.
Some tests must be removed if this is to be mimicked.

[sorhawell]

@eitsupi

I found polars did a behavior change on date_range such that it produces Eager = False a list of DateRanges.
It is being discussed here

[eitsupi]

I found polars did a behavior change on date_range such that it produces Eager = False a list of DateRanges.
It is being discussed here (pola-rs/polars#9019)

Oh, thank you for finding that.

The change definitely makes sense, but we cannot incorporate that change here because it has not yet been made in the main branch.
Perhaps we can change the lazy=FALSE behavior in this repository using implode? (This is different from the behavior of py-polars, but since it will certainly be changed in py-polars in the future, it does not seem to be a problem to change it ahead of time.)

[sorhawell]

One timezone error remaining and the we have restored previous behavior.

We can update the date_range(s) behavior in a new PR after this has merged

[sorhawell]

tiny detail for reference "fix subtle typo bug" is an incorrect commit msg. There was a unrelated subtle bug (high/low) which did not show, and then I disabled a test on timezone until a future refactor.

[eitsupi]

All tests now pass; I would like to defer the renaming from list to arr on the R side to a follow-up fix as it would be an additional major task.

[etiennebacher]

Looks great, thanks for updating this. It's fine for me to do all the renaming from arr to list in a separate PR, are you also ok with this @sorhawell ?

[etiennebacher]

Would be better to have a proper title here (not just for this function)

[eitsupi]

@etiennebacher Thanks for your review.

Since this PR contains updates to the toolchain and everything needs to be recompiled every time we switch branches, I think it's better to merge it in sooner rather than later, so I'm merging it.
Additional modifications can be made later.......