You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@arrow.apache.org by Niklas B <ni...@enplore.com> on 2020/10/22 11:16:11 UTC

Arrow on PyPy3 patch

Hi,

I’ve been (together with the PyPy team) working on getting arrow to build on PyPy3. I’m not looking for full feature capability, but specifically getting it to work with pandas read_parquet/to_parquet which it now does. There were a few roadblocks solved by the awesome Matti Picus on the PyPy Team and we now have a patch that successful builds pyarrow on PyPy3. PyPy3 side has already been patched. 

The patch for pyarrow is on https://gist.githubusercontent.com/mattip/c9c8398b58721ae5893dc8134c353f28/raw/0daff3e11ceed6dcde485a56e6b8bd2b7ca48bbc/gistfile1.txt

A Dockerfile which builds everything is available on https://github.com/bivald/pyarrow-on-pypy3/blob/feature/latest-pypy-latest-pyarrow/Dockerfile (https://github.com/bivald/pyarrow-on-pypy3/tree/feature/latest-pypy-latest-pyarrow)

A surprisingly amount of tests passes (such as all the parquet tests when I tested it last) but some other areas segfaults, but neither of them is super important for me right now.

Would the arrow project be open to a PR with the above patch, even though it doesn’t give you full PyPy support?

Regards,
Niklas

Re: Arrow on PyPy3 patch

Posted by Wes McKinney <we...@gmail.com>.
Either way having a Dockerfile in the project to test with PyPy sounds like
a good idea.

On Thu, Oct 22, 2020 at 6:37 AM Antoine Pitrou <an...@python.org> wrote:

>
> We can, but we cannot be expected to act if something breaks.  So this
> would be wasting CPU resources for little use.
>
> Regards
>
> Antoine.
>
>
> Le 22/10/2020 à 13:33, Krisztián Szűcs a écrit :
> > On Thu, Oct 22, 2020 at 1:21 PM Antoine Pitrou <an...@python.org>
> wrote:
> >>
> >>
> >> Hi Niklas,
> >>
> >> Le 22/10/2020 à 13:16, Niklas B a écrit :
> >>>
> >>> Would the arrow project be open to a PR with the above patch, even
> though it doesn’t give you full PyPy support?
> >>
> >> I think that would be ok, but you'll have to provide maintenance when
> >> needed, because I think we're unlikely to make PyPy mandatory in our CI
> >> setup.
> > We can set up a nightly build though, just need to select the tests
> > cases which are expected to pass.
> >>
> >> Regards
> >>
> >> Antoine.
>

Re: Arrow on PyPy3 patch

Posted by Antoine Pitrou <an...@python.org>.
We can, but we cannot be expected to act if something breaks.  So this
would be wasting CPU resources for little use.

Regards

Antoine.


Le 22/10/2020 à 13:33, Krisztián Szűcs a écrit :
> On Thu, Oct 22, 2020 at 1:21 PM Antoine Pitrou <an...@python.org> wrote:
>>
>>
>> Hi Niklas,
>>
>> Le 22/10/2020 à 13:16, Niklas B a écrit :
>>>
>>> Would the arrow project be open to a PR with the above patch, even though it doesn’t give you full PyPy support?
>>
>> I think that would be ok, but you'll have to provide maintenance when
>> needed, because I think we're unlikely to make PyPy mandatory in our CI
>> setup.
> We can set up a nightly build though, just need to select the tests
> cases which are expected to pass.
>>
>> Regards
>>
>> Antoine.

Re: Arrow on PyPy3 patch

Posted by Krisztián Szűcs <sz...@gmail.com>.
On Thu, Oct 22, 2020 at 1:21 PM Antoine Pitrou <an...@python.org> wrote:
>
>
> Hi Niklas,
>
> Le 22/10/2020 à 13:16, Niklas B a écrit :
> >
> > Would the arrow project be open to a PR with the above patch, even though it doesn’t give you full PyPy support?
>
> I think that would be ok, but you'll have to provide maintenance when
> needed, because I think we're unlikely to make PyPy mandatory in our CI
> setup.
We can set up a nightly build though, just need to select the tests
cases which are expected to pass.
>
> Regards
>
> Antoine.

Re: Arrow on PyPy3 patch

Posted by Antoine Pitrou <an...@python.org>.
Hi Niklas,

Le 22/10/2020 à 13:16, Niklas B a écrit :
> 
> Would the arrow project be open to a PR with the above patch, even though it doesn’t give you full PyPy support?

I think that would be ok, but you'll have to provide maintenance when
needed, because I think we're unlikely to make PyPy mandatory in our CI
setup.

Regards

Antoine.