Plans for PyPy3 wheels?
Hi all, [I'm sending this to scipy-dev because I'm already subscribed, but I guess someone could bring this discussion to numpy-discussion as well] Lately PyPy3 (I'm only interested in Python 3) has matured a lot and it's becoming more and more interesting with time. The developers invested a lot of effort in improving the compatibility with CPython C extensions, and I'm happy to see that SciPy in particular passes all tests in PyPy3: https://circleci.com/gh/scipy/scipy/8318 There was an effort by Antonio Cuni, a PyPy developer, to produce wheels of popular scientific packages. However, it's still a non official thing, and I tried to link them to OpenBLAS but I'm not skilled (or patient) enough: https://github.com/antocuni/pypy-wheels/pull/6 Are there any plans to officially support PyPy3 (and perhaps PyPy) and upload official wheels to pypi.org? Lack of binary packages is holding me back for playing more with PyPy, and there are still no clear plans on what to do with conda packages: https://github.com/conda-forge/pypy2.7-feedstock/issues/1 Did other SciPy or NumPy developers play with PyPy3? Do you think this would be an interesting thing to have? What needs to be done to get there? Best, -- Juan Luis Cano
On Fri, Aug 24, 2018 at 1:49 AM Juan Luis Cano <juanlu001@gmail.com> wrote:
Hi all,
[I'm sending this to scipy-dev because I'm already subscribed, but I guess someone could bring this discussion to numpy-discussion as well]
Lately PyPy3 (I'm only interested in Python 3) has matured a lot and it's becoming more and more interesting with time. The developers invested a lot of effort in improving the compatibility with CPython C extensions, and I'm happy to see that SciPy in particular passes all tests in PyPy3:
https://circleci.com/gh/scipy/scipy/8318
There was an effort by Antonio Cuni, a PyPy developer, to produce wheels of popular scientific packages. However, it's still a non official thing, and I tried to link them to OpenBLAS but I'm not skilled (or patient) enough:
https://github.com/antocuni/pypy-wheels/pull/6
Are there any plans to officially support PyPy3 (and perhaps PyPy) and upload official wheels to pypi.org?
As far as I can find so quickly, there's not yet a clear ABI tag or policy on ABI stability for Pypy wheels ( https://github.com/pypa/wheel/issues/187#issuecomment-320482719). That would be needed before putting wheels on PyPI. Once that's in place, I think we'd be happy to consider releasing official Pypy >=3.5 wheels. If you or someone else wants to look into the current status on this, that would be useful. Cheers, Ralf
Lack of binary packages is holding me back for playing more with PyPy, and there are still no clear plans on what to do with conda packages:
https://github.com/conda-forge/pypy2.7-feedstock/issues/1
Did other SciPy or NumPy developers play with PyPy3? Do you think this would be an interesting thing to have? What needs to be done to get there?
Best,
-- Juan Luis Cano _______________________________________________ SciPy-Dev mailing list SciPy-Dev@python.org https://mail.python.org/mailman/listinfo/scipy-dev
pe, 2018-08-24 kello 12:26 -0700, Ralf Gommers kirjoitti: [clip]
As far as I can find so quickly, there's not yet a clear ABI tag or policy on ABI stability for Pypy wheels ( https://github.com/pypa/wheel/issues/187#issuecomment-320482719). That would be needed before putting wheels on PyPI. Once that's in place, I think we'd be happy to consider releasing official Pypy >=3.5 wheels.
If you or someone else wants to look into the current status on this, that would be useful.
The second issue as I understand is that there are unsolved issues in manylinux1 compatibility (https://bitbucket.org/pypy/pypy/issues/2617/) which is another hurdle for Linux wheels. Numpy/Scipy specifically shouldn't have any unusual needs here, so the efforts should probably be first directed at getting the general binary wheel infrastructure up for Pypy. Pauli
Hi, Thanks for your responses! Antonio Cuni clarified on Twitter that the major blocker is manylinux1 compatibility: https://twitter.com/antocuni/status/1034082028290011137 But today I learned that there's an upcoming standard, manylinux2010, that will supersede manylinux1 and will be based on CentOS 6 instead of 5, hence making things easier: https://github.com/pypa/manylinux/issues/179 Somebody pinged all the pending pull requests yesterday and apparently there are no major blockers, so perhaps we can speak soon about building Linux wheels with the new format, and hopefully PyPy will be compatible to it. On Fri, Aug 24, 2018 at 4:38 PM Pauli Virtanen <pav@iki.fi> wrote:
pe, 2018-08-24 kello 12:26 -0700, Ralf Gommers kirjoitti: [clip]
As far as I can find so quickly, there's not yet a clear ABI tag or policy on ABI stability for Pypy wheels ( https://github.com/pypa/wheel/issues/187#issuecomment-320482719). That would be needed before putting wheels on PyPI. Once that's in place, I think we'd be happy to consider releasing official Pypy >=3.5 wheels.
If you or someone else wants to look into the current status on this, that would be useful.
The second issue as I understand is that there are unsolved issues in manylinux1 compatibility (https://bitbucket.org/pypy/pypy/issues/2617/) which is another hurdle for Linux wheels.
Numpy/Scipy specifically shouldn't have any unusual needs here, so the efforts should probably be first directed at getting the general binary wheel infrastructure up for Pypy.
Pauli
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@python.org https://mail.python.org/mailman/listinfo/scipy-dev
-- Juan Luis Cano
participants (3)
-
Juan Luis Cano -
Pauli Virtanen -
Ralf Gommers