RLS: 1.1.3

[simonjayhawkins]

Tracking issue for the 1.1.3 release.

https://github.com/pandas-dev/pandas/milestone/77

Please do not remove/change milestones from these issues without a note explaining the reasoning (changing milestones doesn't trigger notification)

During issue triage, regressions from 1.0.5 onwards should be milestoned 1.1.3 in the first instance. cc @pandas-dev/pandas-triage

[simonjayhawkins]

arbitrary scheduled date of September 29, 2020 for release.

https://pandas.pydata.org/pandas-docs/dev/whatsnew/v1.1.3.html

There are a few more PRs that imo are very close. I think a release on Monday/Tuesday is viable. And then start 1.1.4. I think there are enough issues outstanding to plan another patch release rather than delaying 1.1.3 to catch more fixes.

wheels are currently building MacPython/pandas-wheels#100 and pip test and conda test passing https://github.com/simonjayhawkins/pandas-release/actions/runs/268749912 but we don't yet have 3.9 wheels.

@pandas-dev/pandas-core

[jreback]

yeah 3.9 wheels are good

[simonjayhawkins]

i've had a brief look at what is involved in building 3.9 wheels.

Numpy is building wheels for linux see MacPython/numpy-wheels#99

@jreback Is not having wheels a blocker?

[TomAugspurger]

We can always add 3.9 wheels later.

[jreback]

why is building wheels for 3.9 hard?

w/o wheels we don't have 3.9 support - so yes it's a blocker

[jreback]

but yes could be added after

[simonjayhawkins]

why is building wheels for 3.9 hard?

for someone who knows what to do. probably not. I am not that person.

[simonjayhawkins]

release delayed by #36552 and I've got issues with google today (maybe my ISP) which is a PITA since I use google mail for github notifications so can put off till tomorrow.

[simonjayhawkins]

w/o wheels we don't have 3.9 support

maybe better to remove the mention of 3.9 support from the whatsnew until we actually have support.

[bashtage]

If it builds from source and passes tests than it seems reasonable to leave in the mention of 3.9, even if wheels aren't available yet.

[simonjayhawkins]

MacOS wheel build currently failing MacPython/pandas-wheels#100

error: RPC failed; curl 18 transfer closed with outstanding read data remaining
fatal: the remote end hung up unexpectedly
Error: Fetching /usr/local/Homebrew/Library/Taps/homebrew/homebrew-core failed!

restarted but same failure. will try again later.

[simonjayhawkins]

MacOS wheel build currently failing MacPython/pandas-wheels#100

all release readiness checks now passing.

#36552 has not yet been merged. if there is a green light for release, will start release without it.

way forward on py3.9 support?

[jreback]

@simonjayhawkins what's the status here? I would move all current issues on 1.1.3 to 1.1.4

[simonjayhawkins]

The blocker is py39 wheels

The most important release on 1.1.x is the last one. 1.1.3 will be short-lived.

I propose reverting the backport of py39 support (and bumping cython) and including it in the next 1.1.x release (3-4 weeks)

otherwise ready to release anytime.

[simontorres]

Looking forward to this release :D

[TomAugspurger]

The blocker is py39 wheels

What's the status on py39 wheels? MacPython/pandas-wheels#100 is the blocker? I can do a bit of work here if needed.

[simonjayhawkins]

MacPython/pandas-wheels#100 is just for testing wheel build and am now closing.

There are no numpy wheels on PyPI for 3.9. Do they need to be available first?

also numpy is only build wheels for linux on nightly. so presumably need Windows and MacOS numpy wheels?

1.1.3 tagged. will release (in Github) after building documentation.

I can do a bit of work here if needed.

lmk if you want me to stop the release process.

[simonjayhawkins]

@TomAugspurger looks like need to tweak release scripts if want to include aarch64 wheels.

am about to release on PyPI without them, for now. so will leave this issue open.

[TomAugspurger]

OK thanks. I'm happy to wait till NumPy 3.9 wheels are out. I don't think we should build against an unreleased version.

[jreback]

i think we should release
and then when can build wheels for 3.9 can add then

[jreback]

but source is fine for 3.9 so shouldn't hold up

[TomAugspurger]

NumPy will issue a new release (1.19.3) with wheels for 3.9. Once that's out we can build pandas 1.1.3 for it: http://numpy-discussion.10968.n7.nabble.com/Python-3-9-is-out-tp48505p48507.html

[frenzymadness]

Python 3.9.0 final has been released yesterday so it would be nice to have wheels for it. Thank you for working on it.

[simonjayhawkins]

@TomAugspurger looks like need to tweak release scripts if want to include aarch64 wheels.

am about to release on PyPI without them, for now. so will leave this issue open.

aarch64 wheels for pandas 1.1.3 on PyPi, xref MacPython/pandas-wheels#102 (comment) thanks @TomAugspurger

[jreback]

are the dev docs just not updated anymore? https://dev.pandas.io/docs/

[simonjayhawkins]

moved to https://pandas.pydata.org/pandas-docs/dev/

[jreback]

ok, let me create an issue to update the dev docs then :->

[TomAugspurger]

3.9 wheels are up on PyPI.

[simonjayhawkins]

nice. installed on Windows

(py39pandas) C:\Users\simon>pip install pandas
Collecting pandas
  Downloading pandas-1.1.3-cp39-cp39-win_amd64.whl (8.9 MB)
     |████████████████████████████████| 8.9 MB 1.1 MB/s
Collecting pytz>=2017.2
  Downloading pytz-2020.1-py2.py3-none-any.whl (510 kB)
     |████████████████████████████████| 510 kB 1.1 MB/s
Collecting python-dateutil>=2.7.3
  Downloading python_dateutil-2.8.1-py2.py3-none-any.whl (227 kB)
     |████████████████████████████████| 227 kB 1.1 MB/s
Collecting numpy>=1.15.4
  Downloading numpy-1.19.3-cp39-cp39-win_amd64.whl (13.3 MB)
     |████████████████████████████████| 13.3 MB 1.1 MB/s
Collecting six>=1.5
  Downloading six-1.15.0-py2.py3-none-any.whl (10 kB)
Installing collected packages: pytz, six, python-dateutil, numpy, pandas
Successfully installed numpy-1.19.3 pandas-1.1.3 python-dateutil-2.8.1 pytz-2020.1 six-1.15.0

(py39pandas) C:\Users\simon>python
Python 3.9.0 | packaged by conda-forge | (default, Oct 14 2020, 22:54:35) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import pandas as pd
>>> pd.__version__
'1.1.3'
>>>