From dcallahan at fb.com  Tue Aug  1 16:14:13 2017
From: dcallahan at fb.com (David Callahan)
Date: Tue, 1 Aug 2017 20:14:13 +0000
Subject: [pypy-dev] pypy3.5 / boost
Message-ID: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>

I was exploring the boost C++ library support for Python 3 extensions with Python3.5 using a built-from-source copy of pypy using the py3.5 branch.  Below are the errors I encountered.

Are these expected problems that someone is working on?
Should someone be working on them?
Or is this intended behavior and I should adapt boost?

Thanks
--david


libs/python/src/exec.cpp: In function ?boost::python::api::object boost::python::exec_file(boost::python::str, boost::python::api::object, boost::python::api::object)?:
libs/python/src/exec.cpp:89:30: error: ?_Py_fopen? was not declared in this scope
   FILE *fs = _Py_fopen(f, "r");
                              ^

libs/python/src/wrapper.cpp: In member function ?boost::python::override boost::python::detail::wrapper_base::get\_override(const char*, PyTypeObject*) const?:
libs/python/src/wrapper.cpp:28:50: error: ?struct PyMethodObject? has no member named ?im_self?
>                   && ((PyMethodObject*)m.get())->im_self == this->m_self
                                                  ^
libs/python/src/wrapper.cpp:37:61: error: ?struct PyMethodObject? has no member named ?im_func?
               if (borrowed_f != ((PyMethodObject*)m.get())->im_func)
                                                             ^



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170801/aac77b98/attachment.html>

From michal.moiseev at firmitas.solutions  Wed Aug  2 06:02:23 2017
From: michal.moiseev at firmitas.solutions (Michal Moiseev)
Date: Wed, 2 Aug 2017 13:02:23 +0300
Subject: [pypy-dev] pyzmq fails on sending bytearray on pypy3
Message-ID: <CAH97920C0HbVoAnJDRgoNYtDP36wayaRrjPMP56Ak0ngGSMOgw@mail.gmail.com>

Hello,
our projects runs on python 3.4, and when trying to run them with pypy3
we run into this problem:
when trying to send bytearray with send()
like:
clientSocket.send(bytearray(some_bytes))
it fails on:
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/opt/pypy3-v5.8.0-src/site-packages/zmq/backend/cffi/socket.py",
line 209, in send
    c_message = ffi.new('char[]', message)
  File "/opt/pypy3-v5.8.0-src/lib_pypy/cffi/api.py", line 250, in new
    return self._backend.newp(cdecl, init)
TypeError: 'bytearray' object cannot be interpreted as an integer
it behaves differently when running with python3.4 (Cpython) - which works
i'm not sure if the problem is with the the pyzmq library or the cffi
thank you in advance

--
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170802/9e6ec9f0/attachment.html>

From michal.moiseev at firmitas.solutions  Thu Aug  3 07:49:32 2017
From: michal.moiseev at firmitas.solutions (Michal Moiseev)
Date: Thu, 3 Aug 2017 14:49:32 +0300
Subject: [pypy-dev] pyzmq fails on sending bytearray on pypy3
In-Reply-To: <CAH97920C0HbVoAnJDRgoNYtDP36wayaRrjPMP56Ak0ngGSMOgw@mail.gmail.com>
References: <CAH97920C0HbVoAnJDRgoNYtDP36wayaRrjPMP56Ak0ngGSMOgw@mail.gmail.com>
Message-ID: <CAH97922bq2+sFDcuzOFzc2XFRrS87SQEJzwVSbwSan9iB0cNPA@mail.gmail.com>

do you know what could be the problem?

On Wed, Aug 2, 2017 at 1:02 PM, Michal Moiseev <
michal.moiseev at firmitas.solutions> wrote:

> Hello,
> our projects runs on python 3.4, and when trying to run them with pypy3
> we run into this problem:
> when trying to send bytearray with send()
> like:
> clientSocket.send(bytearray(some_bytes))
> it fails on:
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/opt/pypy3-v5.8.0-src/site-packages/zmq/backend/cffi/socket.py",
> line 209, in send
>     c_message = ffi.new('char[]', message)
>   File "/opt/pypy3-v5.8.0-src/lib_pypy/cffi/api.py", line 250, in new
>     return self._backend.newp(cdecl, init)
> TypeError: 'bytearray' object cannot be interpreted as an integer
> it behaves differently when running with python3.4 (Cpython) - which works
> i'm not sure if the problem is with the the pyzmq library or the cffi
> thank you in advance
>
> --
>
>


--
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170803/a5625c29/attachment.html>

From cfbolz at gmx.de  Thu Aug  3 10:06:51 2017
From: cfbolz at gmx.de (Carl Friedrich Bolz)
Date: Thu, 3 Aug 2017 16:06:51 +0200
Subject: [pypy-dev] pyzmq fails on sending bytearray on pypy3
In-Reply-To: <CAH97920C0HbVoAnJDRgoNYtDP36wayaRrjPMP56Ak0ngGSMOgw@mail.gmail.com>
References: <CAH97920C0HbVoAnJDRgoNYtDP36wayaRrjPMP56Ak0ngGSMOgw@mail.gmail.com>
Message-ID: <4c08041b-6743-ab27-e85a-840804509a0f@gmx.de>

Hello Michal,

sorry, this is not enough information for us to debug the problem. We
would need detailed instructions how to reproduce it, not just a traceback:

http://doc.pypy.org/en/latest/faq.html#how-should-i-report-a-bug


As a guess: It sounds a bit like the socket.send method in zmq with the
cython backend (which is used by CPython) accepts a bytearray as the
message, while the cffi backend (used by PyPy) does not. In this case,
it would be a zmq bug and you should report it with them.

Cheers,

Carl Friedrich

On 02/08/17 12:02, Michal Moiseev wrote:
> Hello,
> our projects runs on python 3.4, and when trying to run them with pypy3
> we run into this problem:
> when trying to send bytearray with send()
> like:
> clientSocket.send(bytearray(some_bytes))
> it fails on:
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
>   File "/opt/pypy3-v5.8.0-src/site-packages/zmq/backend/cffi/socket.py",
> line 209, in send
>     c_message = ffi.new('char[]', message)
>   File "/opt/pypy3-v5.8.0-src/lib_pypy/cffi/api.py", line 250, in new
>     return self._backend.newp(cdecl, init)
> TypeError: 'bytearray' object cannot be interpreted as an integer
> it behaves differently when running with python3.4 (Cpython) - which works
> i'm not sure if the problem is with the the pyzmq library or the cffi 
> thank you in advance
> 
> -- 
> 
> 
> 
> _______________________________________________
> pypy-dev mailing list
> pypy-dev at python.org
> https://mail.python.org/mailman/listinfo/pypy-dev
> 


From tbaldridge at gmail.com  Thu Aug  3 19:08:41 2017
From: tbaldridge at gmail.com (Timothy Baldridge)
Date: Thu, 3 Aug 2017 17:08:41 -0600
Subject: [pypy-dev] Delimited Continuations in PyPy?
Message-ID: <CAL36E+u3X9oCbZ_scZjT8yFLz2ksMoP02f8meqrJTLd5A9J3YA@mail.gmail.com>

I recently came across this paper:
https://www.microsoft.com/en-us/research/publication/implementing-algebraic-effects-c/

The gist is that they use setjmp/longjmp to create delimited continuations
in compliant C99.

It seems like this method is fairly close to what PyPy already does with
stacklets. Any reason why this couldn't be ported to PyPy? Of course
there'd have to be some sort of stack integration as with stacklets, but it
seems that that mostly boils down to making sure the GC has a way to trace
the bits of the stack contained in the continuations.

I'm considering trying my hand at using their library via rffi, and copying
what I need from _stacklet_shadowstack.py. But perhaps there's something
I'm missing that makes this a no-go from the start.


Thanks,

Timothy Baldridge
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170803/64b8faaa/attachment.html>

From armin.rigo at gmail.com  Fri Aug  4 03:51:11 2017
From: armin.rigo at gmail.com (Armin Rigo)
Date: Fri, 4 Aug 2017 09:51:11 +0200
Subject: [pypy-dev] Delimited Continuations in PyPy?
In-Reply-To: <CAL36E+u3X9oCbZ_scZjT8yFLz2ksMoP02f8meqrJTLd5A9J3YA@mail.gmail.com>
References: <CAL36E+u3X9oCbZ_scZjT8yFLz2ksMoP02f8meqrJTLd5A9J3YA@mail.gmail.com>
Message-ID: <CAMSv6X2jEFTKs9hcJq7OXJ417_vh=caynbDCCPLt9Ko-qgL+_A@mail.gmail.com>

Hi Timothy,

On 4 August 2017 at 01:08, Timothy Baldridge <tbaldridge at gmail.com> wrote:
> It seems like this method is fairly close to what PyPy already does with
> stacklets. Any reason why this couldn't be ported to PyPy?

My first reaction is that this *is* stacklets.  Can you describe in
two words what there is more?


A bient?t,

Armin.

From wlavrijsen at lbl.gov  Fri Aug  4 14:30:10 2017
From: wlavrijsen at lbl.gov (wlavrijsen at lbl.gov)
Date: Fri, 4 Aug 2017 11:30:10 -0700 (PDT)
Subject: [pypy-dev] pypy3.5 / boost
In-Reply-To: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>
References: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>
Message-ID: <alpine.LSU.2.11.1708041124090.1775@cp3-jyni>

David,

since I didn't see a reply, I'm having a stab at it ...

There is patch file for wrapper.cpp:

   pypy/module/cpyext/patches/boost.patch

which looks to be what you want, as it modifies the 'offending' code. I've
not tested it.

And out of curiosity, why boost? cpyext needs to do a mapping of internal
representations, which can be costly, and boost.python isn't known for
speed either?

Best regards,
            Wim
-- 
WLavrijsen at lbl.gov    --    +1 (510) 486 6411    --    www.lavrijsen.net

From dcallahan at fb.com  Fri Aug  4 15:55:35 2017
From: dcallahan at fb.com (David Callahan)
Date: Fri, 4 Aug 2017 19:55:35 +0000
Subject: [pypy-dev] pypy3.5 / boost
In-Reply-To: <alpine.LSU.2.11.1708041124090.1775@cp3-jyni>
References: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>
 <alpine.LSU.2.11.1708041124090.1775@cp3-jyni>
Message-ID: <E89B96C7-BF59-4C82-8081-8523D222CFCD@fb.com>

Thanks. Did you look into the missing Py_fopen as well?

To answer your question, Facebook has a large python application which leverages an even larger C++ code base which is exposed to python with iva sadly large number of different ways including Boost.  We are interested in using PyPy but need to work through a bunch of these issues to even evaluate.
--david


On 8/4/17, 11:30 AM, "wlavrijsen at lbl.gov" <wlavrijsen at lbl.gov> wrote:

    David,
    
    since I didn't see a reply, I'm having a stab at it ...
    
    There is patch file for wrapper.cpp:
    
       pypy/module/cpyext/patches/boost.patch
    
    which looks to be what you want, as it modifies the 'offending' code. I've
    not tested it.
    
    And out of curiosity, why boost? cpyext needs to do a mapping of internal
    representations, which can be costly, and boost.python isn't known for
    speed either?
    
    Best regards,
                Wim
    -- 
    WLavrijsen at lbl.gov    --    +1 (510) 486 6411    --    www.lavrijsen.net
    


From dcallahan at fb.com  Fri Aug  4 16:18:37 2017
From: dcallahan at fb.com (David Callahan)
Date: Fri, 4 Aug 2017 20:18:37 +0000
Subject: [pypy-dev] pypy3.5 / boost
In-Reply-To: <E89B96C7-BF59-4C82-8081-8523D222CFCD@fb.com>
References: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>
 <alpine.LSU.2.11.1708041124090.1775@cp3-jyni>
 <E89B96C7-BF59-4C82-8081-8523D222CFCD@fb.com>
Message-ID: <B58DE3CE-EDE7-4E70-9B6D-48088307B578@fb.com>

Wow, this is a 7-year old issue with boost; https://svn.boost.org/trac10/ticket/4125   ( 

On 8/4/17, 12:55 PM, "David Callahan" <dcallahan at fb.com> wrote:

    Thanks. Did you look into the missing Py_fopen as well?
    
    To answer your question, Facebook has a large python application which leverages an even larger C++ code base which is exposed to python with iva sadly large number of different ways including Boost.  We are interested in using PyPy but need to work through a bunch of these issues to even evaluate.
    --david
    
    
    On 8/4/17, 11:30 AM, "wlavrijsen at lbl.gov" <wlavrijsen at lbl.gov> wrote:
    
        David,
        
        since I didn't see a reply, I'm having a stab at it ...
        
        There is patch file for wrapper.cpp:
        
           pypy/module/cpyext/patches/boost.patch
        
        which looks to be what you want, as it modifies the 'offending' code. I've
        not tested it.
        
        And out of curiosity, why boost? cpyext needs to do a mapping of internal
        representations, which can be costly, and boost.python isn't known for
        speed either?
        
        Best regards,
                    Wim
        -- 
        WLavrijsen at lbl.gov    --    +1 (510) 486 6411    --    www.lavrijsen.net
        
    
    


From dcallahan at fb.com  Fri Aug  4 16:28:37 2017
From: dcallahan at fb.com (David Callahan)
Date: Fri, 4 Aug 2017 20:28:37 +0000
Subject: [pypy-dev] pypy3.5 / boost
In-Reply-To: <B58DE3CE-EDE7-4E70-9B6D-48088307B578@fb.com>
References: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>
 <alpine.LSU.2.11.1708041124090.1775@cp3-jyni>
 <E89B96C7-BF59-4C82-8081-8523D222CFCD@fb.com>
 <B58DE3CE-EDE7-4E70-9B6D-48088307B578@fb.com>
Message-ID: <C468AAD3-03DB-48C0-95F0-07074FC49A2E@fb.com>

Never mind about Py_fopen, I see you gave me a workaround.

On 8/4/17, 1:18 PM, "David Callahan" <dcallahan at fb.com> wrote:

    Wow, this is a 7-year old issue with boost; https://svn.boost.org/trac10/ticket/4125   ( 
    
    On 8/4/17, 12:55 PM, "David Callahan" <dcallahan at fb.com> wrote:
    
        Thanks. Did you look into the missing Py_fopen as well?
        
        To answer your question, Facebook has a large python application which leverages an even larger C++ code base which is exposed to python with iva sadly large number of different ways including Boost.  We are interested in using PyPy but need to work through a bunch of these issues to even evaluate.
        --david
        
        
        On 8/4/17, 11:30 AM, "wlavrijsen at lbl.gov" <wlavrijsen at lbl.gov> wrote:
        
            David,
            
            since I didn't see a reply, I'm having a stab at it ...
            
            There is patch file for wrapper.cpp:
            
               pypy/module/cpyext/patches/boost.patch
            
            which looks to be what you want, as it modifies the 'offending' code. I've
            not tested it.
            
            And out of curiosity, why boost? cpyext needs to do a mapping of internal
            representations, which can be costly, and boost.python isn't known for
            speed either?
            
            Best regards,
                        Wim
            -- 
            WLavrijsen at lbl.gov    --    +1 (510) 486 6411    --    www.lavrijsen.net
            
        
        
    
    


From wlavrijsen at lbl.gov  Fri Aug  4 17:08:31 2017
From: wlavrijsen at lbl.gov (wlavrijsen at lbl.gov)
Date: Fri, 4 Aug 2017 14:08:31 -0700 (PDT)
Subject: [pypy-dev] pypy3.5 / boost
In-Reply-To: <B58DE3CE-EDE7-4E70-9B6D-48088307B578@fb.com>
References: <75071BA8-15B8-4265-A696-161BDD5BB433@fb.com>
 <alpine.LSU.2.11.1708041124090.1775@cp3-jyni>
 <E89B96C7-BF59-4C82-8081-8523D222CFCD@fb.com>
 <B58DE3CE-EDE7-4E70-9B6D-48088307B578@fb.com>
Message-ID: <alpine.LSU.2.11.1708041357080.1775@cp3-jyni>

David,

> Wow, this is a 7-year old issue with boost;
> https://svn.boost.org/trac10/ticket/4125   (

well, I don't think boost.python has had any serious development effort
behind it since 2004. (Although, the reference manual says (c) 2002-2005, so
maybe there's been a bit.)

>    To answer your question, Facebook has a large python application which
> leverages an even larger C++ code base which is exposed to python with iva
> sadly large number of different ways including Boost.

Yes, no shortage of bindings options. In our field, we took inventory in 2003,
and still ended up doing our own thing (is now: http://cppyy.readthedocs.io/).
As it stood, it never gave anyone a reason to try anything else. That is now
changing, with more external software being used. Lots of fun.

Best regards,
            Wim
-- 
WLavrijsen at lbl.gov    --    +1 (510) 486 6411    --    www.lavrijsen.net

From zhao.meide at gmail.com  Fri Aug  4 18:00:29 2017
From: zhao.meide at gmail.com (Meide Zhao)
Date: Fri, 4 Aug 2017 15:00:29 -0700
Subject: [pypy-dev] Pypy seems not so active recently since I can't find any
 talk related to pypy in PYCON 2017
Message-ID: <CAMCbCqK221Xqwbpp-hm34BS1XELLW1WoQmqCvSHW_G5D18AnOA@mail.gmail.com>

Hi all,

I'm looking for a recent talk on pypy but can't find one in PYCON 2017. It
seems pypy is not so hot recently. Can someone give us status and the
future direction on pypy? It seems cpython will stay active for a long time
but pypy which seems better than llvm is going to the right direction for
compiler technology.

Thanks,
Meide
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170804/79cad91b/attachment.html>

From rymg19 at gmail.com  Fri Aug  4 20:40:38 2017
From: rymg19 at gmail.com (rymg19 at gmail.com)
Date: Sat, 5 Aug 2017 02:40:38 +0200
Subject: [pypy-dev] Pypy seems not so active recently since I can't find
 any talk related to pypy in PYCON 2017
In-Reply-To: <<CAMCbCqK221Xqwbpp-hm34BS1XELLW1WoQmqCvSHW_G5D18AnOA@mail.gmail.com>>
Message-ID: <CAO41-mPsvawxbHUZd2+HdbS_tJSvTspTWR4=gA6GAqrKYajEqQ@mail.gmail.com>

Not so active?

https://morepypy.blogspot.com/2017/06/pypy-v58-released.html

https://morepypy.blogspot.com/2017/07/binary-wheels-for-pypy.html?


--
Ryan (????)
Yoko Shimomura, ryo (supercell/EGOIST), Hiroyuki Sawano >> everyone
elsehttp://refi64.com

On Aug 4, 2017 at 5:02 PM, <Meide Zhao <zhao.meide at gmail.com>> wrote:

Hi all,

I'm looking for a recent talk on pypy but can't find one in PYCON 2017. It
seems pypy is not so hot recently. Can someone give us status and the
future direction on pypy? It seems cpython will stay active for a long time
but pypy which seems better than llvm is going to the right direction for
compiler technology.

Thanks,
Meide
_______________________________________________ pypy-dev mailing list
pypy-dev at python.org https://mail.python.org/mailman/listinfo/pypy-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170805/74988660/attachment.html>

From armin.rigo at gmail.com  Sat Aug  5 03:07:04 2017
From: armin.rigo at gmail.com (Armin Rigo)
Date: Sat, 5 Aug 2017 09:07:04 +0200
Subject: [pypy-dev] Pypy seems not so active recently since I can't find
 any talk related to pypy in PYCON 2017
In-Reply-To: <CAMCbCqK221Xqwbpp-hm34BS1XELLW1WoQmqCvSHW_G5D18AnOA@mail.gmail.com>
References: <CAMCbCqK221Xqwbpp-hm34BS1XELLW1WoQmqCvSHW_G5D18AnOA@mail.gmail.com>
Message-ID: <CAMSv6X2J-hceWA-Q7O3W5Uf_wD0TDYSMmCyFKy4WsT5zwc8RUA@mail.gmail.com>

Hi,

On 5 August 2017 at 00:00, Meide Zhao <zhao.meide at gmail.com> wrote:
> I'm looking for a recent talk on pypy but can't find one in PYCON 2017. It
> seems pypy is not so hot recently.

That's because we're mostly not travelling to the US any more.  You
have to look at other conferences like EuroPython.


A bient?t,

Armin.

From dcallahan at fb.com  Thu Aug 10 13:21:32 2017
From: dcallahan at fb.com (David Callahan)
Date: Thu, 10 Aug 2017 17:21:32 +0000
Subject: [pypy-dev] mixed embedding and cpyext access`
Message-ID: <AA9796EB-AE43-46EA-8C28-CFE96E47E6E3@fb.com>

The Boost libraries in their test jigs use a mixture of an embedded API and direct access to the  methods in Python.h.
Below is an example which compiles with  the py3.5 branch, correctly prints ?hello world? but then faults with an message:
?Invalid RPython operation (NULL ptr or bad array index)?.

Is this kind of mixing of the two interfaces supported?
Thanks
--david

#include <Python.h>
#include <PyPy.h>

#include <assert.h>

int main() {
  rpython_startup_code();
  int err = pypy_setup_home((char*)"/home/dcallahan/local/pypy",
                            /*verbbose*/1);
  pypy_execute_source("print('hello world')");
  const char *s = "'abcdefg'.upper()";
  PyObject *obj = ::PyUnicode_FromString(s);
  return 0;
}


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170810/6b40b6fe/attachment.html>

From stuaxo2 at yahoo.com  Thu Aug 10 19:31:07 2017
From: stuaxo2 at yahoo.com (Stuart Axon)
Date: Thu, 10 Aug 2017 23:31:07 +0000 (UTC)
Subject: [pypy-dev] Cocos2D - works for me.
References: <832738331.1802682.1502407867985.ref@mail.yahoo.com>
Message-ID: <832738331.1802682.1502407867985@mail.yahoo.com>

Hi,?? Today I was trying out Cocos2D in PyPy, it seems work :)
I don't have any of my own projects to try, but all the examples that come with it seem to work, I also tried some tutorials which seemed to work as well.

Not sure if I use use it enough to be able to say "it works", but I wasn't able to break it,
los-cocos/cocos

| 
| 
| 
|  |  |

 |

 |
| 
|  | 
los-cocos/cocos

cocos - graphic library for games and multimedia, for python language on PC-like hardware
 |

 |

 |




S++
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170810/b25cc3de/attachment.html>

From armin.rigo at gmail.com  Fri Aug 11 10:34:53 2017
From: armin.rigo at gmail.com (Armin Rigo)
Date: Fri, 11 Aug 2017 16:34:53 +0200
Subject: [pypy-dev] mixed embedding and cpyext access`
In-Reply-To: <AA9796EB-AE43-46EA-8C28-CFE96E47E6E3@fb.com>
References: <AA9796EB-AE43-46EA-8C28-CFE96E47E6E3@fb.com>
Message-ID: <CAMSv6X3gqb2nUtzJA1dc5KuwXBkkdeOE6bo9_K+JWkK1Bb-hng@mail.gmail.com>

Hi David,

On 10 August 2017 at 19:21, David Callahan <dcallahan at fb.com> wrote:
> The Boost libraries in their test jigs use a mixture of an embedded API and
> direct access to the  methods in Python.h.
>
> Below is an example which compiles with  the py3.5 branch, correctly prints
> ?hello world? but then faults with an message:
>
> ?Invalid RPython operation (NULL ptr or bad array index)?.
>
>
>
> Is this kind of mixing of the two interfaces supported?

No.  Our cpyext emulates the C API but doesn't support the embedding
mode of CPython.


A bient?t,

Armin.

From stuaxo2 at yahoo.com  Mon Aug 14 18:57:04 2017
From: stuaxo2 at yahoo.com (Stuart Axon)
Date: Mon, 14 Aug 2017 22:57:04 +0000 (UTC)
Subject: [pypy-dev] =?utf-8?q?PyBaseExceptionObject_has_no_member_named_?=
 =?utf-8?b?4oCYYXJnc+KAmQ==?=
References: <461190651.2013712.1502751424823.ref@mail.yahoo.com>
Message-ID: <461190651.2013712.1502751424823@mail.yahoo.com>

Found this trying to compile pycairo, is it worth opening a bug about? (I realise there is CairoCFFI, but they don't have feature parity - OTOH, CFFI probably is the way to go eventually).


$ ?python setup.py install
running install
running build
running build_py
creating build
creating build/lib.linux-x86_64-2.7
creating build/lib.linux-x86_64-2.7/cairo
copying cairo/__init__.py -> build/lib.linux-x86_64-2.7/cairo
running build_ext
building 'cairo._cairo' extension
creating build/temp.linux-x86_64-2.7
creating build/temp.linux-x86_64-2.7/cairo
cc -pthread -DNDEBUG -O2 -fPIC -I/usr/include/cairo -I/usr/include/glib-2.0 -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng16 -I/home/stu/.virtualenvs/pypy-nightly/include -c cairo/device.c -o build/temp.linux-x86_64-2.7/cairo/device.o -fno-strict-aliasing
cc -pthread -DNDEBUG -O2 -fPIC -I/usr/include/cairo -I/usr/include/glib-2.0 -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng16 -I/home/stu/.virtualenvs/pypy-nightly/include -c cairo/bufferproxy.c -o build/temp.linux-x86_64-2.7/cairo/bufferproxy.o -fno-strict-aliasing
cc -pthread -DNDEBUG -O2 -fPIC -I/usr/include/cairo -I/usr/include/glib-2.0 -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/libpng16 -I/home/stu/.virtualenvs/pypy-nightly/include -c cairo/error.c -o build/temp.linux-x86_64-2.7/cairo/error.o -fno-strict-aliasing
In file included from /home/stu/.virtualenvs/pypy-nightly/include/Python.h:81:0,
???????????????? from cairo/error.c:32:
cairo/error.c: In function ?error_init?:
cairo/error.c:111:35: error: ?PyBaseExceptionObject {aka struct <anonymous>}? has no member named ?args?
???? if(PyTuple_GET_SIZE(self->base.args) >= 2) {
?????????????????????????????????? ^
/home/stu/.virtualenvs/pypy-nightly/include/object.h:61:39: note: in definition of macro ?Py_SIZE?
 #define Py_SIZE(ob)??(((PyVarObject*)(ob))->ob_size)
?????????????????????????????????????? ^~
cairo/error.c:111:8: note: in expansion of macro ?PyTuple_GET_SIZE?
???? if(PyTuple_GET_SIZE(self->base.args) >= 2) {
????????^~~~~~~~~~~~~~~~
In file included from /home/stu/.virtualenvs/pypy-nightly/include/Python.h:108:0,
???????????????? from cairo/error.c:32:
cairo/error.c:112:49: error: ?PyBaseExceptionObject {aka struct <anonymous>}? has no member named ?args?
???????? status_obj = PyTuple_GET_ITEM(self->base.args, 1);
???????????????????????????????????????????????? ^
/home/stu/.virtualenvs/pypy-nightly/include/tupleobject.h:23:53: note: in definition of macro ?PyTuple_GET_ITEM?
 #define PyTuple_GET_ITEM(op, i) (((PyTupleObject *)(op))->ob_item[i])
???????????????????????????????????????????????????? ^~
In file included from /home/stu/.virtualenvs/pypy-nightly/include/Python.h:81:0,
???????????????? from cairo/error.c:32:
cairo/error.c: In function ?error_str?:
cairo/error.c:152:36: error: ?PyBaseExceptionObject {aka struct <anonymous>}? has no member named ?args?
???? if (PyTuple_GET_SIZE(self->base.args) >= 1) {
????????????????????????????????????^
/home/stu/.virtualenvs/pypy-nightly/include/object.h:61:39: note: in definition of macro ?Py_SIZE?
 #define Py_SIZE(ob)??(((PyVarObject*)(ob))->ob_size)
?????????????????????????????????????? ^~
cairo/error.c:152:9: note: in expansion of macro ?PyTuple_GET_SIZE?
???? if (PyTuple_GET_SIZE(self->base.args) >= 1) {
???????? ^~~~~~~~~~~~~~~~
In file included from /home/stu/.virtualenvs/pypy-nightly/include/Python.h:108:0,
???????????????? from cairo/error.c:32:
cairo/error.c:153:56: error: ?PyBaseExceptionObject {aka struct <anonymous>}? has no member named ?args?
???????? return PyObject_Str(PyTuple_GET_ITEM(self->base.args, 0));
????????????????????????????????????????????????????????^
/home/stu/.virtualenvs/pypy-nightly/include/tupleobject.h:23:53: note: in definition of macro ?PyTuple_GET_ITEM?
 #define PyTuple_GET_ITEM(op, i) (((PyTupleObject *)(op))->ob_item[i])
???????????????????????????????????????????????????? ^~
error: command 'cc' failed with exit status 1
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170814/242a6ecb/attachment.html>

From brad.kish at arcticwolf.com  Mon Aug 14 12:10:04 2017
From: brad.kish at arcticwolf.com (Brad Kish)
Date: Mon, 14 Aug 2017 16:10:04 +0000
Subject: [pypy-dev] GIL removal
Message-ID: <FFB5C88D-0FE3-4B05-8E31-46F6D1D5A19D@arcticwolf.com>

The blog post on the GIL removal work seems to indicate that the work is gated on ?interest of the community and the commercial partners?.

How is this interest going to be measured?

We use pypy in production at Arctic Wolf Networks, and have several cases where this change could benefit our use cases.

We would be interested in helping support this project.

Brad.

From fijall at gmail.com  Tue Aug 15 12:25:17 2017
From: fijall at gmail.com (Maciej Fijalkowski)
Date: Tue, 15 Aug 2017 18:25:17 +0200
Subject: [pypy-dev] GIL removal
In-Reply-To: <FFB5C88D-0FE3-4B05-8E31-46F6D1D5A19D@arcticwolf.com>
References: <FFB5C88D-0FE3-4B05-8E31-46F6D1D5A19D@arcticwolf.com>
Message-ID: <CAK5idxQOONgzmY_L7veAQF2YBY3+0UCag97qJLzmxQvs_N7YsQ@mail.gmail.com>

Hi Brad

Thanks for your mail! We're pleased to hear that pypy is getting used
and is giving benefits :-)

The interest will be measured in "can we find someone to fund
$50/$100k proposal". We already have $15k backing plus $20k that we
have left over for STM. Can we find someone willing to fund the rest?
I'm more than happy to draft a contract (we have a commercial entity)
to work on that topic.

Best regards,
Maciej Fijalkowski

On Mon, Aug 14, 2017 at 6:10 PM, Brad Kish <brad.kish at arcticwolf.com> wrote:
> The blog post on the GIL removal work seems to indicate that the work is gated on ?interest of the community and the commercial partners?.
>
> How is this interest going to be measured?
>
> We use pypy in production at Arctic Wolf Networks, and have several cases where this change could benefit our use cases.
>
> We would be interested in helping support this project.
>
> Brad.
> _______________________________________________
> pypy-dev mailing list
> pypy-dev at python.org
> https://mail.python.org/mailman/listinfo/pypy-dev

From dynamicgl at gmail.com  Sun Aug 20 00:29:46 2017
From: dynamicgl at gmail.com (Gelin Yan)
Date: Sun, 20 Aug 2017 12:29:46 +0800
Subject: [pypy-dev] Does pypy 58 arm version support numpy?
Message-ID: <CABkOF6QSGu0rrOPJSpC_EOYvhYN+O9qkv+rMzwBGvjUo6_XvXA@mail.gmail.com>

Hi All

    When I tried to use pip to install numpy, I got some compiler
failtures. I want to know whether pypy arm version fully support numpy.

OS: raspbian
Gcc version: 4.9.2
Pypy version: 5.8
Hardware: raspberry pi 3


Regards

gelin yan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/pypy-dev/attachments/20170820/1885bb09/attachment.html>

From fijall at gmail.com  Tue Aug 29 09:15:34 2017
From: fijall at gmail.com (Maciej Fijalkowski)
Date: Tue, 29 Aug 2017 15:15:34 +0200
Subject: [pypy-dev] 
	=?utf-8?q?PyBaseExceptionObject_has_no_member_named_?=
	=?utf-8?b?4oCYYXJnc+KAmQ==?=
In-Reply-To: <461190651.2013712.1502751424823@mail.yahoo.com>
References: <461190651.2013712.1502751424823.ref@mail.yahoo.com>
 <461190651.2013712.1502751424823@mail.yahoo.com>
Message-ID: <CAK5idxSRt=ok2EdFF5Pio47fZCLHPygMz4yUWGnCddHZBia4Bw@mail.gmail.com>

Hi Stuart.

PyCairo should not abuse the API by directly accessing members of
structures, this is not supported on pypy

Cheers,
fijal

On Tue, Aug 15, 2017 at 12:57 AM, Stuart Axon via pypy-dev
<pypy-dev at python.org> wrote:
> Found this trying to compile pycairo, is it worth opening a bug about? (I
> realise there is CairoCFFI, but they don't have feature parity - OTOH, CFFI
> probably is the way to go eventually).
>
>
> $ ?python setup.py install
> running install
> running build
> running build_py
> creating build
> creating build/lib.linux-x86_64-2.7
> creating build/lib.linux-x86_64-2.7/cairo
> copying cairo/__init__.py -> build/lib.linux-x86_64-2.7/cairo
> running build_ext
> building 'cairo._cairo' extension
> creating build/temp.linux-x86_64-2.7
> creating build/temp.linux-x86_64-2.7/cairo
> cc -pthread -DNDEBUG -O2 -fPIC -I/usr/include/cairo -I/usr/include/glib-2.0
> -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -I/usr/include/pixman-1
> -I/usr/include/freetype2 -I/usr/include/libpng16
> -I/home/stu/.virtualenvs/pypy-nightly/include -c cairo/device.c -o
> build/temp.linux-x86_64-2.7/cairo/device.o -fno-strict-aliasing
> cc -pthread -DNDEBUG -O2 -fPIC -I/usr/include/cairo -I/usr/include/glib-2.0
> -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -I/usr/include/pixman-1
> -I/usr/include/freetype2 -I/usr/include/libpng16
> -I/home/stu/.virtualenvs/pypy-nightly/include -c cairo/bufferproxy.c -o
> build/temp.linux-x86_64-2.7/cairo/bufferproxy.o -fno-strict-aliasing
> cc -pthread -DNDEBUG -O2 -fPIC -I/usr/include/cairo -I/usr/include/glib-2.0
> -I/usr/lib/x86_64-linux-gnu/glib-2.0/include -I/usr/include/pixman-1
> -I/usr/include/freetype2 -I/usr/include/libpng16
> -I/home/stu/.virtualenvs/pypy-nightly/include -c cairo/error.c -o
> build/temp.linux-x86_64-2.7/cairo/error.o -fno-strict-aliasing
> In file included from
> /home/stu/.virtualenvs/pypy-nightly/include/Python.h:81:0,
>                  from cairo/error.c:32:
> cairo/error.c: In function ?error_init?:
> cairo/error.c:111:35: error: ?PyBaseExceptionObject {aka struct
> <anonymous>}? has no member named ?args?
>      if(PyTuple_GET_SIZE(self->base.args) >= 2) {
>                                    ^
> /home/stu/.virtualenvs/pypy-nightly/include/object.h:61:39: note: in
> definition of macro ?Py_SIZE?
> #define Py_SIZE(ob)  (((PyVarObject*)(ob))->ob_size)
>                                        ^~
> cairo/error.c:111:8: note: in expansion of macro ?PyTuple_GET_SIZE?
>      if(PyTuple_GET_SIZE(self->base.args) >= 2) {
>         ^~~~~~~~~~~~~~~~
> In file included from
> /home/stu/.virtualenvs/pypy-nightly/include/Python.h:108:0,
>                  from cairo/error.c:32:
> cairo/error.c:112:49: error: ?PyBaseExceptionObject {aka struct
> <anonymous>}? has no member named ?args?
>          status_obj = PyTuple_GET_ITEM(self->base.args, 1);
>                                                  ^
> /home/stu/.virtualenvs/pypy-nightly/include/tupleobject.h:23:53: note: in
> definition of macro ?PyTuple_GET_ITEM?
> #define PyTuple_GET_ITEM(op, i) (((PyTupleObject *)(op))->ob_item[i])
>                                                      ^~
> In file included from
> /home/stu/.virtualenvs/pypy-nightly/include/Python.h:81:0,
>                  from cairo/error.c:32:
> cairo/error.c: In function ?error_str?:
> cairo/error.c:152:36: error: ?PyBaseExceptionObject {aka struct
> <anonymous>}? has no member named ?args?
>      if (PyTuple_GET_SIZE(self->base.args) >= 1) {
>                                     ^
> /home/stu/.virtualenvs/pypy-nightly/include/object.h:61:39: note: in
> definition of macro ?Py_SIZE?
> #define Py_SIZE(ob)  (((PyVarObject*)(ob))->ob_size)
>                                        ^~
> cairo/error.c:152:9: note: in expansion of macro ?PyTuple_GET_SIZE?
>      if (PyTuple_GET_SIZE(self->base.args) >= 1) {
>          ^~~~~~~~~~~~~~~~
> In file included from
> /home/stu/.virtualenvs/pypy-nightly/include/Python.h:108:0,
>                  from cairo/error.c:32:
> cairo/error.c:153:56: error: ?PyBaseExceptionObject {aka struct
> <anonymous>}? has no member named ?args?
>          return PyObject_Str(PyTuple_GET_ITEM(self->base.args, 0));
>                                                         ^
> /home/stu/.virtualenvs/pypy-nightly/include/tupleobject.h:23:53: note: in
> definition of macro ?PyTuple_GET_ITEM?
> #define PyTuple_GET_ITEM(op, i) (((PyTupleObject *)(op))->ob_item[i])
>                                                      ^~
> error: command 'cc' failed with exit status 1
>
> _______________________________________________
> pypy-dev mailing list
> pypy-dev at python.org
> https://mail.python.org/mailman/listinfo/pypy-dev
>