Skip to content

Py39#395

Merged
isuruf merged 24 commits into
conda-forge:masterfrom
ocefpaf:py39
Oct 7, 2020
Merged

Py39#395
isuruf merged 24 commits into
conda-forge:masterfrom
ocefpaf:py39

Conversation

@ocefpaf

@ocefpaf ocefpaf commented Oct 6, 2020

Copy link
Copy Markdown
Member

@conda-forge-linter

Copy link
Copy Markdown

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

Comment thread recipe/meta.yaml Outdated
Comment thread recipe/patches/0001-Darwin-Add-arch-x86_64-to-UnicCCompiler-ld_args.patch Outdated
Co-authored-by: Isuru Fernando <isuruf@gmail.com>
Comment thread recipe/patches/gh21228.patch
Comment thread recipe/sysconfigdata/default/_sysconfigdata_linux.py Outdated
@mingwandroid

mingwandroid commented Oct 7, 2020

Copy link
Copy Markdown
Contributor

Yes, I developed a replacements thing recently just to get away from having to bash script or manually edit so much brittle stuff. Here's an example of it: https://github.com/AnacondaRecipes/qt-feedstock/blob/master/recipe/conda_build_config.yaml#L9-L28

Comment thread recipe/meta.yaml Outdated
run:
- ld_impl_{{ target_platform }} # [linux]
- tzdata
- python-tzdata

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I thought that the idea was to avoid the python package and force Python to find the tzdata noarch generic, no?

@jjhelmus jjhelmus mentioned this pull request Oct 7, 2020
if os.name == 'nt':
_init_non_posix(_CONFIG_VARS)
- _CONFIG_VARS['TZPATH'] = ''
+ _CONFIG_VARS['TZPATH'] = os.path.join(_PREFIX, "share", "zoneinfo")

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

🎉

@ocefpaf

ocefpaf commented Oct 7, 2020

Copy link
Copy Markdown
Member Author

@mingwandroid any final thoughts/consideration? Are we good to merge?

@isuruf

isuruf commented Oct 7, 2020

Copy link
Copy Markdown
Member

@mingwandroid, I removed the hand edited sysconfigdata files as they were outdated and added code to generate the sysconfigdata for the system gcc and the conda compiler. Let me know if you are okay with that.

@mingwandroid

Copy link
Copy Markdown
Contributor

I say go for it (then take cover)

@mingwandroid

Copy link
Copy Markdown
Contributor

Oh one quick thing:

case "$target_Platform" in

does the case of the P matter here?

@isuruf

isuruf commented Oct 7, 2020

Copy link
Copy Markdown
Member

yes, it does. fixed in f988bc8

@ocefpaf

ocefpaf commented Oct 7, 2020

Copy link
Copy Markdown
Member Author

I say go for it (then take cover)

@isuruf

isuruf commented Oct 7, 2020

Copy link
Copy Markdown
Member

Thanks @mingwandroid and @ocefpaf

@isuruf isuruf merged commit e7ea437 into conda-forge:master Oct 7, 2020
@ocefpaf ocefpaf deleted the py39 branch October 7, 2020 20:09
@ocefpaf

ocefpaf commented Oct 7, 2020

Copy link
Copy Markdown
Member Author

And you @isuruf! This one really took a village to get done!

@ocefpaf

ocefpaf commented Oct 7, 2020

Copy link
Copy Markdown
Member Author

Also thanks to @pganssle for guiding us with the tzdata package and @mbargull for creating it!

@mingwandroid

mingwandroid commented Oct 8, 2020

Copy link
Copy Markdown
Contributor

Great to see this land. From my perspective, it's the first time I've worked on a big update to Python within conda-forge in the first instance. That's of course the best way, esp. for the important packages.

It was a huge effort for many of us and of course its completion brings a whole bunch of other stuff to build, but with conda-forge's excellent tooling (and GitHub) that's not too scary a prospect.

For now, I must get 3.9.0 on defaults ASAP (I added tzdata and python-tzdata already - do we need both still? I got a bit confused around that bit!). I'll base it on the recipe here, and rebase the anaconda advertising patch to happen last (this makes patch rebasing easier). On that note, we should also talk about adopting the anaconda advertising method for conda-forge too instead of the conda-forge branding stuff, it is more reliable and involves less file IO at interepreter launch time, pinging @jjhelmus on the reliability matter). CMake detects our tools now (thanks KitWare!) and will increasingly do so so having some more uniformity around this stuff would be good.

Keep on keeping on conda-forge.

@ocefpaf

ocefpaf commented Oct 8, 2020

Copy link
Copy Markdown
Member Author

For now, I must get 3.9.0 on defaults ASAP (I added tzdata and python-tzdata already - do we need both still?

To build python we only need tzdata. We also offer python-tzdata b/c some packages may want to use it. They do share the same zoneinfo data now.

I got a bit confused around that bit!). I'll base it on the recipe here, and rebase the anaconda advertising patch to happen last (this makes patch rebasing easier). On that note, we should also talk about adopting the anaconda advertising method for conda-forge too instead of the conda-forge branding stuff, it is more reliable and involves less file IO at interepreter launch time, pinging @jjhelmus on the reliability matter). CMake detects our tools now (thanks KitWare!) and will increasingly do so so having some more uniformity around this stuff would be good.

Sure. I don't really know what that entails but it sounds like something we could do.

@mbargull

mbargull commented Oct 8, 2020

Copy link
Copy Markdown
Member

Great to see this land. From my perspective, it's the first time I've worked on a big update to Python within conda-forge in the first instance. That's of course the best way, esp. for the important packages.

🎉

(I added tzdata and python-tzdata already - do we need both still? I got a bit confused around that bit!).

Yep, like Felipe said, it makes sense to have both. python-tzdata isn't really important for our Python distribution, because with tzdata we should ship everything zoneinfo needs. But we'll probably see (few?) other packages that may add python-tzdata as a dependency in the future.

Keep on keeping on conda-forge.

🤘


Sure. I don't really know what that entails but it sounds like something we could do.

Same here, agreed! Best thing would probably be to just open a PR whenever you're ready and discuss details there.

@Superharz

Copy link
Copy Markdown

Hello,

I'm running on a 64-Bit Windows 10 v. 1909

I just tried to install python 3.9.0 in a new empty Conda Env (named new2).
However, I always get the following error:

grafik

@mingwandroid

Copy link
Copy Markdown
Contributor

You should use: conda config --set add_pip_as_python_dependency false

.. at least until pip is rebuilt.

(everyone should use this, always IMHO)

@Superharz

Copy link
Copy Markdown

You should use: conda config --set add_pip_as_python_dependency false

.. at least until pip is rebuilt.

(everyone should use this, always IMHO)

@mingwandroid Thank you. That solved my problem.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

How should we handle tzdata/zoneinfo

6 participants