[openstack-dev] [requirements] Do we care about pypy for clients (broken by cryptography)
cboylan at sapwetik.org
Wed May 31 15:05:11 UTC 2017
On Wed, May 31, 2017, at 07:39 AM, Sean McGinnis wrote:
> On Wed, May 31, 2017 at 09:47:37AM -0400, Doug Hellmann wrote:
> > Excerpts from Monty Taylor's message of 2017-05-31 07:34:03 -0500:
> > > On 05/31/2017 06:39 AM, Sean McGinnis wrote:
> > > > On Wed, May 31, 2017 at 06:37:02AM -0500, Sean McGinnis wrote:
> > > >>
> > > >> I am not aware of anyone using pypy, and there are other valid working
> > > >> alternatives. I would much rather just drop support for it than redo our
> > > >> crypto functions again.
> > > >>
> > >
> > This question came up recently for the Oslo libraries, and I think we
> > also agreed that pypy support was not being actively maintained.
> > Doug
> Thanks Doug. If oslo does not support pypy, then I think that makes the
> decision for me. I will put up a patch to get rid of that job and stop
> wasting infra resources on it.
Just a couple things (I don't think it changes the decision made).
Cryptography does at least claim to support PyPy (see
https://pypi.python.org/pypi/cryptography/ trove identifiers), so
possibly a bug on their end that should get filed?
Also, our user clients should probably run under as many interpreters as
possible to make life easier for end users; however, they currently
depend on oslo and if oslo doesn't support pypy then likely not
reasonable to do in user clients.
More information about the OpenStack-dev