From robertwb at gmail.com  Mon Jul  3 15:17:25 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Mon, 3 Jul 2017 12:17:25 -0700
Subject: [Cython] Cython 0.26 release
In-Reply-To: <CAChs6_kg4dfj_RTCYZGxv-gw209t8-eN2LzV1XOStFMoGSZkiQ@mail.gmail.com>
References: <CADiQ+QDNx5tVwXWr8R+u7gsY5zwXEpVjE3ovvZ-T+-jxo33HqQ@mail.gmail.com>
 <594A669A.7050505@cage.ugent.be>
 <CADiQ+QAscepzvgCOEH88NYJSNV0BD0su6SkB-4NZ6VgfgFtBtw@mail.gmail.com>
 <594AA604.9040805@cage.ugent.be>
 <CADiQ+QBgg9u0BdQBXuDTrEVXSaFg9g01roohdjzbJ0+ynMSasQ@mail.gmail.com>
 <594ACBDC.4000107@cage.ugent.be>
 <CAChs6_kg4dfj_RTCYZGxv-gw209t8-eN2LzV1XOStFMoGSZkiQ@mail.gmail.com>
Message-ID: <CADiQ+QDPgbqAw_BM4eOOotiQKuXUoT8EpV3ox0Wqy3_PELvFmw@mail.gmail.com>

I have made the bad vtable generation a warning rather than an error
for now to ease the transition. I was hoping to be able to reproduce
(and fix) https://github.com/cython/cython/issues/1744 but still
having no luck so I'm going to proceed with a beta to not stall the
process.

On Fri, Jun 23, 2017 at 6:32 PM, David Roe <roed.math at gmail.com> wrote:
>
>
> On Wed, Jun 21, 2017 at 12:41 PM, Jeroen Demeyer <jdemeyer at cage.ugent.be>
> wrote:
>>
>> On 2017-06-21 20:25, Robert Bradshaw wrote:
>>>
>>> I suppose in that case we just got lucky that no one cimported these
>>> classes and tried to use methods after this point in the vtable...
>>
>>
>> I think that's the case indeed.
>
>
> In fact, Julian Rueth and I eventually did cimport one of these classes
> (Polynomial_generic_dense), eventually leading to
> https://github.com/cython/cython/issues/1732.
>
> A warning does seem like it would make fixing existing code easier.
> David
>>
>>
>>
>> _______________________________________________
>> cython-devel mailing list
>> cython-devel at python.org
>> https://mail.python.org/mailman/listinfo/cython-devel
>
>
>
> _______________________________________________
> cython-devel mailing list
> cython-devel at python.org
> https://mail.python.org/mailman/listinfo/cython-devel
>

From robertwb at gmail.com  Mon Jul  3 16:29:26 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Mon, 3 Jul 2017 13:29:26 -0700
Subject: [Cython] Cython 0.26 beta
Message-ID: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>

There have been a lot of changes since Cython 0.25.2, please test and
report. https://github.com/cython/cython/archive/0.26b0.zip

At this point I anticipate only bugfixes will be merged before the
final release.

From lists at onerussian.com  Thu Jul  6 12:30:54 2017
From: lists at onerussian.com (Yaroslav Halchenko)
Date: Thu, 6 Jul 2017 12:30:54 -0400
Subject: [Cython] 0.25.2b0 with 32bit fixes: (sometimes) Interrupted
 system call in builtin_subtype_methods_T653
In-Reply-To: <CADiQ+QC7gu=xwkvn8bd-tjt0R-9Gwz2zyNP4eLZfna9q30T5cg@mail.gmail.com>
References: <20161130144619.GD23180@onerussian.com>
 <CADiQ+QC7gu=xwkvn8bd-tjt0R-9Gwz2zyNP4eLZfna9q30T5cg@mail.gmail.com>
Message-ID: <20170706163053.GA12287@onerussian.com>

Dear Robert

Sorry that I haven't followed up :-(   FWIW -- I haven't observed it
recently while building 0.26~b0, detected issues for which I have
just filed at https://github.com/cython/cython/issues

Cheers,

On Wed, 30 Nov 2016, Robert Bradshaw wrote:

> When it does reproduce, is it always this particular test?

> On Wed, Nov 30, 2016 at 6:46 AM, Yaroslav Halchenko
> <lists at onerussian.com> wrote:
> > A fresh one I ran into while rebuilding with fixes for 32bit (thanks
> > BTW! sorry for not prompt followup/gratitude)

> > ======================================================================
> > ERROR: runTest (__main__.CythonRunTestCase)
> > compiling (cpp) and running builtin_subtype_methods_T653
> > ----------------------------------------------------------------------
> > Traceback (most recent call last):
> >   File "runtests.py", line 998, in run
> >     self.run_tests(result, ext_so_path)
> >   File "runtests.py", line 1013, in run_tests
> >     self.run_doctests(self.module, result, ext_so_path)
> >   File "runtests.py", line 1023, in run_doctests
> >     run_forked_test(result, run_test, self.shortDescription(), self.fork)
> >   File "runtests.py", line 1076, in run_forked_test
> >     cid, result_code = os.waitpid(child_id, 0)
> > OSError: [Errno 4] Interrupted system call

> > ----------------------------------------------------------------------
> > Ran 581 tests in 188.225s

> > FAILED (errors=1)

> > it is not 100% reproducible.  Rebuilding the package went smooth without this
> > hiccup.  Just wanted to let you know

> > might be related that the tests were ran as


> > set -e; for P in 2.7 3.5; do \
> >  PYTHONPATH=`/bin/ls -d /build/cython-0.25.2~b0/build/lib.*-$P` \
> >   /usr/bin/python$P runtests.py --shard_count=16 \
> >         --no-refnanny -v -v --exclude="(parallel|Debugger|annotate_html)" --work-dir=build/work-dir 2>&1; \
> > done

> > so testing in parallel to speed up the build  of the package.  This one was the
> > only failure I have ran into but it built fine before (but there could have
> > been some changes in the core packages since then)

> > --
> > Yaroslav O. Halchenko
> > Center for Open Neuroscience     http://centerforopenneuroscience.org
> > Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
> > Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
> > WWW:   http://www.linkedin.com/in/yarik
> > _______________________________________________
> > cython-devel mailing list
> > cython-devel at python.org
> > https://mail.python.org/mailman/listinfo/cython-devel
> _______________________________________________
> cython-devel mailing list
> cython-devel at python.org
> https://mail.python.org/mailman/listinfo/cython-devel


-- 
Yaroslav O. Halchenko
Center for Open Neuroscience     http://centerforopenneuroscience.org
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik        

From t3kcit at gmail.com  Fri Jul  7 17:10:07 2017
From: t3kcit at gmail.com (Andreas Mueller)
Date: Fri, 7 Jul 2017 17:10:07 -0400
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
Message-ID: <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>

Hey folks.
So scikit-learn master doesn't build with 0.26b0:
https://github.com/scikit-learn/scikit-learn/issues/9272

Robert already looked into this, but it hasn't been resolved yet:
https://github.com/scikit-learn/scikit-learn/pull/9133

We want to create a release candidate today and release soon(ish).
It would be great if you could give us hand so that our release
builds with your release :)

Cheers,
Andy

On 07/03/2017 04:29 PM, Robert Bradshaw wrote:
> There have been a lot of changes since Cython 0.25.2, please test and
> report. https://github.com/cython/cython/archive/0.26b0.zip
>
> At this point I anticipate only bugfixes will be merged before the
> final release.
> _______________________________________________
> cython-devel mailing list
> cython-devel at python.org
> https://mail.python.org/mailman/listinfo/cython-devel


From robertwb at gmail.com  Mon Jul 10 13:59:20 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Mon, 10 Jul 2017 10:59:20 -0700
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
Message-ID: <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>

https://github.com/scikit-learn/scikit-learn/pull/9311 looks like a
good resolution. The infinite recursion issue has been fixed as well.

On Fri, Jul 7, 2017 at 2:10 PM, Andreas Mueller <t3kcit at gmail.com> wrote:
> Hey folks.
> So scikit-learn master doesn't build with 0.26b0:
> https://github.com/scikit-learn/scikit-learn/issues/9272
>
> Robert already looked into this, but it hasn't been resolved yet:
> https://github.com/scikit-learn/scikit-learn/pull/9133
>
> We want to create a release candidate today and release soon(ish).
> It would be great if you could give us hand so that our release
> builds with your release :)
>
> Cheers,
> Andy
>
>
> On 07/03/2017 04:29 PM, Robert Bradshaw wrote:
>>
>> There have been a lot of changes since Cython 0.25.2, please test and
>> report. https://github.com/cython/cython/archive/0.26b0.zip
>>
>> At this point I anticipate only bugfixes will be merged before the
>> final release.
>> _______________________________________________
>> cython-devel mailing list
>> cython-devel at python.org
>> https://mail.python.org/mailman/listinfo/cython-devel
>
>
> _______________________________________________
> cython-devel mailing list
> cython-devel at python.org
> https://mail.python.org/mailman/listinfo/cython-devel

From robertwb at gmail.com  Mon Jul 10 15:49:30 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Mon, 10 Jul 2017 12:49:30 -0700
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
Message-ID: <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>

Another beta: https://github.com/cython/cython/archive/0.26b1.zip

On Mon, Jul 10, 2017 at 10:59 AM, Robert Bradshaw <robertwb at gmail.com> wrote:
> https://github.com/scikit-learn/scikit-learn/pull/9311 looks like a
> good resolution. The infinite recursion issue has been fixed as well.
>
> On Fri, Jul 7, 2017 at 2:10 PM, Andreas Mueller <t3kcit at gmail.com> wrote:
>> Hey folks.
>> So scikit-learn master doesn't build with 0.26b0:
>> https://github.com/scikit-learn/scikit-learn/issues/9272
>>
>> Robert already looked into this, but it hasn't been resolved yet:
>> https://github.com/scikit-learn/scikit-learn/pull/9133
>>
>> We want to create a release candidate today and release soon(ish).
>> It would be great if you could give us hand so that our release
>> builds with your release :)
>>
>> Cheers,
>> Andy
>>
>>
>> On 07/03/2017 04:29 PM, Robert Bradshaw wrote:
>>>
>>> There have been a lot of changes since Cython 0.25.2, please test and
>>> report. https://github.com/cython/cython/archive/0.26b0.zip
>>>
>>> At this point I anticipate only bugfixes will be merged before the
>>> final release.
>>> _______________________________________________
>>> cython-devel mailing list
>>> cython-devel at python.org
>>> https://mail.python.org/mailman/listinfo/cython-devel
>>
>>
>> _______________________________________________
>> cython-devel mailing list
>> cython-devel at python.org
>> https://mail.python.org/mailman/listinfo/cython-devel

From robertwb at gmail.com  Mon Jul 10 17:01:35 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Mon, 10 Jul 2017 14:01:35 -0700
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
Message-ID: <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>

Scratch that, please try https://github.com/cython/cython/archive/0.26b1.zip

On Mon, Jul 10, 2017 at 12:49 PM, Robert Bradshaw <robertwb at gmail.com> wrote:
> Another beta: https://github.com/cython/cython/archive/0.26b1.zip
>
> On Mon, Jul 10, 2017 at 10:59 AM, Robert Bradshaw <robertwb at gmail.com> wrote:
>> https://github.com/scikit-learn/scikit-learn/pull/9311 looks like a
>> good resolution. The infinite recursion issue has been fixed as well.
>>
>> On Fri, Jul 7, 2017 at 2:10 PM, Andreas Mueller <t3kcit at gmail.com> wrote:
>>> Hey folks.
>>> So scikit-learn master doesn't build with 0.26b0:
>>> https://github.com/scikit-learn/scikit-learn/issues/9272
>>>
>>> Robert already looked into this, but it hasn't been resolved yet:
>>> https://github.com/scikit-learn/scikit-learn/pull/9133
>>>
>>> We want to create a release candidate today and release soon(ish).
>>> It would be great if you could give us hand so that our release
>>> builds with your release :)
>>>
>>> Cheers,
>>> Andy
>>>
>>>
>>> On 07/03/2017 04:29 PM, Robert Bradshaw wrote:
>>>>
>>>> There have been a lot of changes since Cython 0.25.2, please test and
>>>> report. https://github.com/cython/cython/archive/0.26b0.zip
>>>>
>>>> At this point I anticipate only bugfixes will be merged before the
>>>> final release.
>>>> _______________________________________________
>>>> cython-devel mailing list
>>>> cython-devel at python.org
>>>> https://mail.python.org/mailman/listinfo/cython-devel
>>>
>>>
>>> _______________________________________________
>>> cython-devel mailing list
>>> cython-devel at python.org
>>> https://mail.python.org/mailman/listinfo/cython-devel

From robertwb at gmail.com  Mon Jul 10 17:51:06 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Mon, 10 Jul 2017 14:51:06 -0700
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
 <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
Message-ID: <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>

https://github.com/cython/cython/archive/0.26b2.zip

On Mon, Jul 10, 2017 at 2:01 PM, Robert Bradshaw <robertwb at gmail.com> wrote:
> Scratch that, please try https://github.com/cython/cython/archive/0.26b1.zip
>
> On Mon, Jul 10, 2017 at 12:49 PM, Robert Bradshaw <robertwb at gmail.com> wrote:
>> Another beta: https://github.com/cython/cython/archive/0.26b1.zip
>>
>> On Mon, Jul 10, 2017 at 10:59 AM, Robert Bradshaw <robertwb at gmail.com> wrote:
>>> https://github.com/scikit-learn/scikit-learn/pull/9311 looks like a
>>> good resolution. The infinite recursion issue has been fixed as well.
>>>
>>> On Fri, Jul 7, 2017 at 2:10 PM, Andreas Mueller <t3kcit at gmail.com> wrote:
>>>> Hey folks.
>>>> So scikit-learn master doesn't build with 0.26b0:
>>>> https://github.com/scikit-learn/scikit-learn/issues/9272
>>>>
>>>> Robert already looked into this, but it hasn't been resolved yet:
>>>> https://github.com/scikit-learn/scikit-learn/pull/9133
>>>>
>>>> We want to create a release candidate today and release soon(ish).
>>>> It would be great if you could give us hand so that our release
>>>> builds with your release :)
>>>>
>>>> Cheers,
>>>> Andy
>>>>
>>>>
>>>> On 07/03/2017 04:29 PM, Robert Bradshaw wrote:
>>>>>
>>>>> There have been a lot of changes since Cython 0.25.2, please test and
>>>>> report. https://github.com/cython/cython/archive/0.26b0.zip
>>>>>
>>>>> At this point I anticipate only bugfixes will be merged before the
>>>>> final release.
>>>>> _______________________________________________
>>>>> cython-devel mailing list
>>>>> cython-devel at python.org
>>>>> https://mail.python.org/mailman/listinfo/cython-devel
>>>>
>>>>
>>>> _______________________________________________
>>>> cython-devel mailing list
>>>> cython-devel at python.org
>>>> https://mail.python.org/mailman/listinfo/cython-devel

From jdemeyer at cage.ugent.be  Wed Jul 12 07:58:46 2017
From: jdemeyer at cage.ugent.be (Jeroen Demeyer)
Date: Wed, 12 Jul 2017 13:58:46 +0200
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
 <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
 <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>
Message-ID: <59660EF6.7000900@cage.ugent.be>

On 2017-07-10 23:51, Robert Bradshaw wrote:
> https://github.com/cython/cython/archive/0.26b2.zip

Everything fine in SageMath!

I did notice a speed regression and I suspect that
https://github.com/cython/cython/pull/1719 might be the reason.

From robertwb at math.washington.edu  Wed Jul 12 11:13:38 2017
From: robertwb at math.washington.edu (Robert Bradshaw)
Date: Wed, 12 Jul 2017 08:13:38 -0700
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <59660EF6.7000900@cage.ugent.be>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
 <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
 <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>
 <59660EF6.7000900@cage.ugent.be>
Message-ID: <CADiQ+QByHoy6kfKW6Z=75iQTVTOiHVDfbz4b1doBZ9ragiaD6Q@mail.gmail.com>

On Wed, Jul 12, 2017 at 4:58 AM, Jeroen Demeyer <jdemeyer at cage.ugent.be> wrote:
> On 2017-07-10 23:51, Robert Bradshaw wrote:
>>
>> https://github.com/cython/cython/archive/0.26b2.zip
>
>
> Everything fine in SageMath!

Thanks!

> I did notice a speed regression and I suspect that
> https://github.com/cython/cython/pull/1719 might be the reason.

Could also be https://github.com/cython/cython/pull/1728 , though of
course correctness trumps speed.

As for timing, I hope to get an RC out this week, releasing shortly after.

From jdemeyer at cage.ugent.be  Wed Jul 12 16:05:27 2017
From: jdemeyer at cage.ugent.be (Jeroen Demeyer)
Date: Wed, 12 Jul 2017 22:05:27 +0200
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QByHoy6kfKW6Z=75iQTVTOiHVDfbz4b1doBZ9ragiaD6Q@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
 <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
 <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>
 <59660EF6.7000900@cage.ugent.be>
 <CADiQ+QByHoy6kfKW6Z=75iQTVTOiHVDfbz4b1doBZ9ragiaD6Q@mail.gmail.com>
Message-ID: <59668107.2090409@cage.ugent.be>

On 2017-07-12 17:13, Robert Bradshaw wrote:
> Could also be https://github.com/cython/cython/pull/1728 , though of
> course correctness trumps speed.

That would be unlikely since we currently do *NOT* use binding=True in 
SageMath, so that patch should not affect anything.

From robertwb at gmail.com  Fri Jul 14 20:28:38 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Fri, 14 Jul 2017 17:28:38 -0700
Subject: [Cython] Cython 0.26 release candidate.
Message-ID: <CADiQ+QBNPsiSSCbtwfi3aVc-qPx=60CDPG-E_H7zGsygCwFdjg@mail.gmail.com>

Looks like everything's gone smoothly with the beta. I have prepared a
release candidate, likely to become the final release.

The git sources are available at
https://github.com/cython/cython/archive/0.26rc0.zip
I have also created an sdist, up at
http://bradshawfamily.net/cython/Cython-0.26rc0.tar.gz

- Robert

From robertwb at gmail.com  Sat Jul 15 01:58:59 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Fri, 14 Jul 2017 22:58:59 -0700
Subject: [Cython] Cython 0.26 release candidate.
In-Reply-To: <CADiQ+QBNPsiSSCbtwfi3aVc-qPx=60CDPG-E_H7zGsygCwFdjg@mail.gmail.com>
References: <CADiQ+QBNPsiSSCbtwfi3aVc-qPx=60CDPG-E_H7zGsygCwFdjg@mail.gmail.com>
Message-ID: <CADiQ+QDFBXYL3vM9daVMWLP83gAwBS98Z2MKuT=kJ8nCrpHj_g@mail.gmail.com>

Some last-minute commits to unbreak Py3.7. New release candidate at

http://bradshawfamily.net/cython/Cython-0.26rc1.tar.gz (sdist)

or

https://github.com/cython/cython/archive/0.26rc1.zip (git)

On Fri, Jul 14, 2017 at 5:28 PM, Robert Bradshaw <robertwb at gmail.com> wrote:
> Looks like everything's gone smoothly with the beta. I have prepared a
> release candidate, likely to become the final release.
>
> The git sources are available at
> https://github.com/cython/cython/archive/0.26rc0.zip
> I have also created an sdist, up at
> http://bradshawfamily.net/cython/Cython-0.26rc0.tar.gz
>
> - Robert

From stefan_ml at behnel.de  Sun Jul 16 03:22:03 2017
From: stefan_ml at behnel.de (Stefan Behnel)
Date: Sun, 16 Jul 2017 09:22:03 +0200
Subject: [Cython] [cython-users] Cython 0.26 release candidate.
In-Reply-To: <CADiQ+QCgrrpH+omgE=e1S7nA1JF9xut_WYsAbM2fcV=L-Emp6w@mail.gmail.com>
References: <CADiQ+QBNPsiSSCbtwfi3aVc-qPx=60CDPG-E_H7zGsygCwFdjg@mail.gmail.com>
 <5969E413.6020007@cage.ugent.be>
 <CADiQ+QCgrrpH+omgE=e1S7nA1JF9xut_WYsAbM2fcV=L-Emp6w@mail.gmail.com>
Message-ID: <1f6bd745-1d99-6a94-cd5e-17fcaecd2530@behnel.de>

Robert Bradshaw schrieb am 16.07.2017 um 01:19:
> On Sat, Jul 15, 2017 at 2:44 AM, Jeroen Demeyer wrote:
>> Do you intend to do something about the speed regression due to
>> https://github.com/cython/cython/pull/1719
>>
>> I'm happy with a yes/no answer just to know what to expect.
>> 
> Had a chance to take a look at this:
>
https://github.com/cython/cython/commit/5e2dc4f78e2c9b560b07f73e664f3e70f7a8cfc8
> I think this is safe enough to get into the next release.

I'd suggest this change:

https://github.com/cython/cython/pull/1773

Stefan

From robertwb at gmail.com  Tue Jul 18 12:28:53 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Tue, 18 Jul 2017 09:28:53 -0700
Subject: [Cython] [cython-users] Cython 0.26 release candidate.
In-Reply-To: <1f6bd745-1d99-6a94-cd5e-17fcaecd2530@behnel.de>
References: <CADiQ+QBNPsiSSCbtwfi3aVc-qPx=60CDPG-E_H7zGsygCwFdjg@mail.gmail.com>
 <5969E413.6020007@cage.ugent.be>
 <CADiQ+QCgrrpH+omgE=e1S7nA1JF9xut_WYsAbM2fcV=L-Emp6w@mail.gmail.com>
 <1f6bd745-1d99-6a94-cd5e-17fcaecd2530@behnel.de>
Message-ID: <CADiQ+QCk6_upfqFar9XfrQ4w5NKo6aTzCoiTXR=6NNKmKBOLCg@mail.gmail.com>

That change has been merged, as well as a couple of (should be
completely safe) additional declarations and warning fixes. New
release candidate up at
https://github.com/cython/cython/archive/0.26rc2.zip . I plan to make
this the actual release tomorrow.

On Sun, Jul 16, 2017 at 12:22 AM, Stefan Behnel <stefan_ml at behnel.de> wrote:
> Robert Bradshaw schrieb am 16.07.2017 um 01:19:
>> On Sat, Jul 15, 2017 at 2:44 AM, Jeroen Demeyer wrote:
>>> Do you intend to do something about the speed regression due to
>>> https://github.com/cython/cython/pull/1719
>>>
>>> I'm happy with a yes/no answer just to know what to expect.
>>>
>> Had a chance to take a look at this:
>>
> https://github.com/cython/cython/commit/5e2dc4f78e2c9b560b07f73e664f3e70f7a8cfc8
>> I think this is safe enough to get into the next release.
>
> I'd suggest this change:
>
> https://github.com/cython/cython/pull/1773
>
> Stefan
>
> --
>
> ---
> You received this message because you are subscribed to the Google Groups "cython-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to cython-users+unsubscribe at googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

From robertwb at gmail.com  Thu Jul 20 02:07:41 2017
From: robertwb at gmail.com (Robert Bradshaw)
Date: Wed, 19 Jul 2017 23:07:41 -0700
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <59668107.2090409@cage.ugent.be>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
 <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
 <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>
 <59660EF6.7000900@cage.ugent.be>
 <CADiQ+QByHoy6kfKW6Z=75iQTVTOiHVDfbz4b1doBZ9ragiaD6Q@mail.gmail.com>
 <59668107.2090409@cage.ugent.be>
Message-ID: <CADiQ+QDcHoY4r=Rn=OYb=b317wYtYiX48sub3pzowG5hsWTdQA@mail.gmail.com>

The release is out. Thanks everyone for the feedback on and off list.

On Jul 12, 2017 1:05 PM, "Jeroen Demeyer" <jdemeyer at cage.ugent.be> wrote:

> On 2017-07-12 17:13, Robert Bradshaw wrote:
>
>> Could also be https://github.com/cython/cython/pull/1728 , though of
>> course correctness trumps speed.
>>
>
> That would be unlikely since we currently do *NOT* use binding=True in
> SageMath, so that patch should not affect anything.
> _______________________________________________
> cython-devel mailing list
> cython-devel at python.org
> https://mail.python.org/mailman/listinfo/cython-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/cython-devel/attachments/20170719/4061a955/attachment.html>

From stefan_ml at behnel.de  Thu Jul 20 09:36:19 2017
From: stefan_ml at behnel.de (Stefan Behnel)
Date: Thu, 20 Jul 2017 15:36:19 +0200
Subject: [Cython] Cython 0.26 beta
In-Reply-To: <CADiQ+QDcHoY4r=Rn=OYb=b317wYtYiX48sub3pzowG5hsWTdQA@mail.gmail.com>
References: <CADiQ+QBO=ppHEQCmijzC_13pR78dCxaXWRTV8ftefZyjw3a4qw@mail.gmail.com>
 <9892b32c-f704-8cf6-7b21-108aafb3a858@gmail.com>
 <CADiQ+QDebdTyFt9_gm5Tx-yRE4bqnrHXYptsxCUhvCoJm2k5uw@mail.gmail.com>
 <CADiQ+QCP0665XHF9WbmAGKad4coMK8LshGO+0ks0rgkXSTKKDg@mail.gmail.com>
 <CADiQ+QAOJzfLi3BNwytRRs9xK4b2hYEtvw=ZrrNN6RxDHvcVCA@mail.gmail.com>
 <CADiQ+QBLap5p88VE6cQHGMdpzdBdp-aCcNuLxhS3gp_Q3GCV0g@mail.gmail.com>
 <59660EF6.7000900@cage.ugent.be>
 <CADiQ+QByHoy6kfKW6Z=75iQTVTOiHVDfbz4b1doBZ9ragiaD6Q@mail.gmail.com>
 <59668107.2090409@cage.ugent.be>
 <CADiQ+QDcHoY4r=Rn=OYb=b317wYtYiX48sub3pzowG5hsWTdQA@mail.gmail.com>
Message-ID: <02b1e5f6-9290-1b4e-0a27-812d7dbb4457@behnel.de>

Robert Bradshaw schrieb am 20.07.2017 um 08:07:
> The release is out. Thanks everyone for the feedback on and off list.

Thanks, Robert, for preparing it and fixing and reviewing heaps of bugs and
code along the way.

Stefan

From stefan_ml at behnel.de  Sun Jul 23 03:57:21 2017
From: stefan_ml at behnel.de (Stefan Behnel)
Date: Sun, 23 Jul 2017 09:57:21 +0200
Subject: [Cython] PEP 489 impacts "public" declarations (generated module
 header file)
Message-ID: <71ebfd9a-77b9-5a05-b46e-9ded02801ee1@behnel.de>

Hi all!

Here's a question to cython-users. Is anyone of you using the "public"
modifier to export symbols to external code? And are you calling the module
init function from them directly, e.g. when embedding a Cython module in a
C/C++ application? I'd like to get some feedback on how people use this
feature.

Here's why: I'm pretty much done implementing PEP 489 for CPython 3.5+,
which splits the extension module initialisation into multiple steps to
enable more Python-like module features. It changes the behaviour of the
module init function, which no longer does the complete module creation and
initialisation, but only returns a metadata object (PyModuleDef) that
steers the init process.

https://www.python.org/dev/peps/pep-0489/

My implementation is here:

https://github.com/cython/cython/pull/1794

One problem that I found: The docs on `public` declarations show how to
call the (Py2) module init function in order to initialise the module when
embedding:

http://docs.cython.org/en/latest/src/userguide/external_C_code.html#public-declarations

That is no longer enough in Py3.5+. The PEP contains a minimal example of
what needs to be done now, in addition to some new C-API functions that
help doing it:

https://www.python.org/dev/peps/pep-0489/#legacy-init

https://www.python.org/dev/peps/pep-0489/#new-functions

I can see two ways to fix the new situation: let users adapt their code
(assuming that it's probably a rare use case), or export a new public
function that implements the module initialisation process similar to what
Cython does in pre-Py3.5, but probably losing most of the new features,
such as setting a proper value for "__file__" at initialisation time,
although we could accept an (optional?) ModuleSpec as argument.

Exporting such a function adds convenience, at least for some users, at the
price of exporting another public name from Cython modules (at least those
that use public symbols). But it's not clear that such a function would
cover the user's needs, because some might want to influence the process
(that's what the PEP was written for, after all). And since I expect the
number of users to be low, adding a convenience feature for even less of
them might be unhelpful.

I'd also rather avoid putting this function into the header file itself,
because that makes code that uses it highly dependent on the internals,
which might change with newer Cython/Python versions.

So, does anyone use this feature and could explain a bit how you use it to
help us understand what we should consider "convenient" for users?

Stefan