RLS: 1.0.2

[jbrockmendel]

@pandas-dev/pandas-core think end of this week?

[TomAugspurger]

Friday sounds good to me, but I don't have a good sense for where things are at.

Any regressions that we definitely want to see fixed?

Any open PRs we'd like to see merged?

[jbrockmendel]

Any regressions that we definitely want to see fixed?

• assignment with datetime64[ns, UTC] raises TypeError #32395 ill make a PR tomorrow morning
• pandas.DataFrame.reindex_like throws if called from a derived class instance #31925 has a viable-looking PR in _ensure_type should use issubclass #32416.
• Should pd.core.algorithms.match have been removed? #31827 I think we just need to say "yes"
• REGR: Series repr of object Index with bools and NaN is wrong #32146 has a viable PR in BUG: Fixed bug, where pandas._libs.lib.maybe_convert_objects function improperly handled arrays with bools and NaNs #32242
• REGR: Timestamp subtraction raises TypeError, Non informative #31774 mostly just needs a better exception message

Probably no action on

• Pandas breaks warnings module when printing DataFrame #31978 looks like we havent been able to reproduce
• ValueError: buffer source array is read-only #31710 supposedly will be fixed by cython 0.30; would be a pain to fix on our end

Most of the rest of the regression issues from the last few weeks are groupby/agg-related, not obvious what it will take to fix.

[jreback]

#31827 I think we just need to say "yes"

this was long deprecated and it is not clear from the OP what is going on - certainly not 1.02 material

[TomAugspurger]

I'm roughly targeting Wednesday for a release.

[charlesdong1991]

@TomAugspurger sorry to interrupt, maybe #31507 #32480 could also be nice to be included in 1.0.2? I've seen several people raise this issue.

If someone could take a look at PR, I would try my best to give updates on the PR ASAP.

[TomAugspurger]

I plan to tag 1.0.2 after #32658 is merged and backported.

[TomAugspurger]

OK, starting the release.

[TomAugspurger]

FYI, the last thing before we can upload to PyPI is the Windows wheels. @cgohlke do you know if you're able to build them for 1.0.2?

When MacPython/pandas-wheels#74 is merged, windows wheels will be built as well. I wouldn't be surprised if we get issue reports from switching to those, but it'd be nice to have control of this stage too, rather than relying on Christoph's (extremely valuable) generosity.

[TomAugspurger]

OK, all the wheels are built. I'm planning to upload to the wheels today. I think having a release with the fixed regressions is valuable, despite the regression in #32685. And conda-forge is already out there.

[jorisvandenbossche]

@TomAugspurger Thanks a lot for doing the release!

Agreed we can do a 1.0.3 afterwards to include the new regression.

[jreback]

yep thanks @TomAugspurger awesome as always

[TomAugspurger]

Released on PyPI now.