Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Installation of Pandas in Pip v10.0.0b2 on PyPy fails #20666

Closed
davidjlloyd opened this issue Apr 12, 2018 · 14 comments
Closed

Installation of Pandas in Pip v10.0.0b2 on PyPy fails #20666

davidjlloyd opened this issue Apr 12, 2018 · 14 comments
Labels
Build Library building on various platforms

Comments

@davidjlloyd
Copy link

Code Sample, a copy-pastable example if possible

root@9b34418ce20d:/# /opt/pypyenv/bin/pip install pandas
Looking in indexes: https://devpi.myindex/root/myindex/+simple/
Collecting pandas
  Downloading https://devpi.myindex/root/pypi/+f/c7a/2757b60774825/pandas-0.22.0.tar.gz (11.3MB)
    100% |################################| 11.3MB 1.5MB/s 
  Could not find a version that satisfies the requirement Cython (from versions: )
No matching distribution found for Cython

Problem description

See pypa/pip#5171 for the other side of this issue. Pandas defines a pyproject.toml for PEP 518 compliance. Unfortunately, Pip 10 only partially supports PEP 518; it requires that all dependencies be available as binary distributions which is generally not the case for implementations like PyPy. As such installation fails as above.

@TomAugspurger
Copy link
Contributor

TomAugspurger commented Apr 12, 2018

Is there anything pandas should do here? From my brief skim of pypa/pip#5171 it's not clear what we should change (if anything).

Note, we just followed SciPy here: https://github.com/scipy/scipy/blob/master/pyproject.toml, so they'll have similar issues.

@TomAugspurger TomAugspurger added the Build Library building on various platforms label Apr 12, 2018
@davidjlloyd
Copy link
Author

Unclear. PyPA are suggesting that the implementation of PEP 518 was premature, but I'm not sure I agree with that. At least having this issue reported here will help people find out what's wrong and how to fix it (for reference, you can somewhat avoid the issue by passing the flag --no-build-isolation to Pip 10). I think it is up to Pandas contributors to decide what, if any, action should be taken in response to this upcoming issue.

@TomAugspurger
Copy link
Contributor

TomAugspurger commented Apr 12, 2018 via email

@dinapappor
Copy link

dinapappor commented Apr 17, 2018

This also happens in a virtualenv with python 3.6 (no pypy) on FreeBSD 11.1 (fully patched).

@jorisvandenbossche
Copy link
Member

I think this issue is another reason to remove pyproject.toml for now: #20718 (that will of course only solve it once 0.23.0 is out)

@diraol
Copy link

diraol commented Apr 17, 2018

This also happens on Debian 9, python 3.6 pip 10.0.0, setuptools 39.0.1 and wheel 0.31.0 (linux) while trying to install without binaries.

@jorisvandenbossche
Copy link
Member

You can pass --no-build-isolation flag to pip to avoid this problem (pypa/pip#5229).

@dinapappor
Copy link

Passing --no-build-isolation yields a cc error for me.

@TomAugspurger
Copy link
Contributor

TomAugspurger commented Apr 18, 2018 via email

@dinapappor
Copy link

dinapappor commented Apr 18, 2018 via email

@dinapappor
Copy link

This is the error I am getting:

distutils.errors.LinkError: Command "cc /tmp/tmp86b2sak0/tmp/tmp86b2sak0/source.o -L/usr/local/lib -lblas -o /tmp/tmp86b2sak0/a.out" failed with exit status 1

@dinapappor
Copy link

The previous comment was actually because I already had numpy installed as an os level package in FreeBSD. Seems to work with --no-build-isolation.

@jorisvandenbossche
Copy link
Member

I made an overview issue with the problems: #20775, so let's close this one

@BrunoNun3s
Copy link

I'm trying with "pip download pandas"

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Build Library building on various platforms
Projects
None yet
Development

No branches or pull requests

6 participants