Mailman 3 September 2015 - pypy-dev

About gurobipy
by Luis José Novoa 31 May '18

31 May '18

Hi All, My name is Luis Novoa. Im working on a large scale optimization problem using python/gurobipy for my implementation. I was wondering if any of you in this list has used gurobipy with pypy, or if it is even possible at this time. Thanks in advance, Luis.

9 9

Copy of list
by Tuom Larsen 07 Oct '15

07 Oct '15

Dear PyPy developers! In CPython, a common idiom to copy a list is to use slice, as in: copy = original[:] I noticed that under PyPy 2.6.0 it seems quite a bit slower than using `list` constructor: from timeit import timeit print timeit('b = a[:]', 'a = list(range(100))') # 0.0732719898224 print timeit('b = list(a)', 'a = list(range(100))') # 0.00285792350769 Please, could some comment on what is the preferred way to copy a list under PyPy? Should I try to avoid the "slice way"? Thanks!

4 7

FIY: Basic Block Versioning JIT
by Niki Spahiev 29 Sep '15

29 Sep '15

http://pointersgonewild.com/2015/09/24/basic-block-versioning-my-best-resul… HTH Niki

4 4

Fwd: cffi dlopen
by Roshan Cherian 22 Sep '15

22 Sep '15

Hi Armin/Maciej, Thanks for the reply, following the code sample I wrote up the following: from cffi import FFI ffi = FFI() ffi.set_source("_ical_cffi_ext", """ // passed to the real C compiler #include <libical/ical.h> """, include_dirs=['./include'], # where ical.h is libraries=['ical'], library_dirs=['./lib'],) # or a list of libraries to link with ffi.cdef(""" // some declarations from the man page typedef struct icalcomponent_impl icalcomponent; typedef struct icalproperty_impl icalproperty; typedef ... icalparameter; typedef struct icalvalue_impl icalvalue; typedef ... icaltimezonechange; //really typedef struct _icaltimezonechange icaltimezonechange; struct _icaltimezonechange { int utc_offset; int prev_utc_offset; int year; /**< Actual year, e.g. 2001. */ int month; /**< 1 (Jan) to 12 (Dec). */ int day; int hour; int minute; int second; int is_daylight; }; """) if __name__ == "__main__": ffi.compile() _icaltimezonechange is opaque as its in src/libical/icaltimezone.c a c file and not in in ical.h. However this struct is required for my purpose. However when it compiles I do get errors like: _ical_cffi_ext.c:455: error: dereferencing pointer to incomplete type _ical_cffi_ext.c:457: error: invalid use of undefined type 'struct _icaltimezonechange' _ical_cffi_ext.c:458: error: dereferencing pointer to incomplete type _ical_cffi_ext.c:468: error: invalid application of 'sizeof' to incomplete type 'struct _icaltimezonechange' _ical_cffi_ext.c:468: error: initializer element is not constant _ical_cffi_ext.c:468: error: (near initialization for '_cffi_struct_unions[2].alignment') Can I know what should be done for these cases? Thanks, -Roshan On Tue, Sep 15, 2015 at 8:17 AM, Armin Rigo <arigo(a)tunes.org> wrote: > Hi Roshan, > > On Tue, Sep 15, 2015 at 5:01 PM, Roshan Cherian <cherian.rosh(a)gmail.com> > wrote: > >>>>> libical = ffi.dlopen(ctypes.util.find_library('ical')) > >>>>> > libical.ical_set_unknown_token_handling_setting(libical.ICAL_ASSUME_IANA_TOKEN) > > > > I get an error: AttributeError: ical_set_unknown_token_handling_setting > > This means the library file was found, but it doesn't contain any > symbol of the name "ical_set_unknown_token_handling_setting". It may > be because such a name is actually a macro. In that case you can't > use ffi.dlopen(). > > You should look at ffi.set_source() instead (and *not* ffi.verify(); > Maciej probably mentioned this by mistake, it is now called > set_source()). > > > I haven't tried out ffi.verify I am not sure how to load a shared lib > using > > ffi.verify, I am sorry if my understanding is wrong. > > See > http://cffi.readthedocs.org/en/latest/overview.html#real-example-api-level-… > . In your case you need: > > ffi.set_source("_ical_cffi", > """ #include <ical.h> > """, > include_dirs=['./include'], # or where ical.h is > libraries=['ical'], > library_dirs=['./lib'], # or where libical.so is > ) > > You also need the same ffi.cdef() as you use now, and ffi.compile() as > shown in the example. The ffi.cdef() can use the syntax ``...`` > (dot-dot-dot) in many places when it is combined with > ffi.set_source(). See the rest of the documentation section for more. > > > A bientôt, > > Armin. >

2 4

Trying to support PyPy on PyAV
by Phyo Arkar 21 Sep '15

21 Sep '15

Mike Boers is trying to support PyAV (python Audio Video editing/merging/converting library based on FFMPEG) . The current master fails 6 of the unit tests: - 4 of them are in __repr__ methods, and seem to be due to basic difference in module name handling. - 1 is with Numpy; I expect there was an API change there. - 2 is with memoryviews; I suspect there are marginal differences in that implementation. https://github.com/mikeboers/PyAV/issues/130 I am forwarding it here. Thanks Phyo.

1 0

State of GUI toolkits with PyPy?
by Eric Driggers 20 Sep '15

20 Sep '15

Looking at rewriting a number of my older python2 scripts to python3 and was hoping in the mean time to also make all of them play nice with PyPy where possible. However a number of them use pyqt when I needed a GUI. From what I can find QT still does not play nice with PyPy yet, however are there any others? Has there been movement on GUI toolkits/libraries with CFFI/CTypes/cpyext support via pypy? The compatibility wiki has not really had those edited since sometime around early 2013... https://bitbucket.org/pypy/compatibility/wiki/Home Wondering is all before I try and see myself and spend a few days trying to build them with pypy.

8 14

pypy 2.6 memory leak bug!! very importent
by 于大伟 16 Sep '15

16 Sep '15

hi, i use pypy 2.6 with BeautifulSoup module and pymongo. i make a spider with the pypy. when i run pypy ,i found a problem ,the pypy use memory higher then 6GB when i run pypy at once , the pypy use only 240mb memory! the screenshot in the attachment!

3 2

speed.pypy.org
by Armin Rigo 16 Sep '15

16 Sep '15

Hi all, Summarizing a few comments on irc: the main page of speed.pypy.org is broken right now. (The rest seems to work.) A bientôt, Armin.

3 2

cffi dlopen
by Roshan Cherian 15 Sep '15

15 Sep '15

Hi Team, I am not sure if I should be asking the question here, however hoping someone could help. we use ffi.cdef for definitions and ffi.dlopen to open shared library. To my understanding ffi.dlopen only works in the presence of a c compiler to dynamically link the shared lib. We have a linux box without gcc in which we are supposed to run this setup. We do have dlopen on the box as man dlopen returns. Could I know if it is at all possible to use ffi.dlopen without a C compiler? Apologies if I am totally off. Thanks, -Roshan

3 4

Re: [pypy-dev] [pypy-commit] pypy default: I *think* the inputargs should always contain InputArgXxx objects that
by Maciej Fijalkowski 15 Sep '15

15 Sep '15

It might contain something else than InputArgXxx (because of e.g. not tracing from the start), but indeed the forwarded should be set to None by the time we reach the backend (maybe we forget it too late?) On Mon, Sep 14, 2015 at 7:31 PM, arigo <noreply(a)buildbot.pypy.org> wrote: > Author: Armin Rigo <arigo(a)tunes.org> > Branch: > Changeset: r79638:cfe5d8e88d3d > Date: 2015-09-14 19:31 +0200 > http://bitbucket.org/pypy/pypy/changeset/cfe5d8e88d3d/ > > Log: I *think* the inputargs should always contain InputArgXxx objects > that are not forwarded. (Hitting the case from test_ll_random) > > diff --git a/rpython/jit/backend/x86/regalloc.py b/rpython/jit/backend/x86/regalloc.py > --- a/rpython/jit/backend/x86/regalloc.py > +++ b/rpython/jit/backend/x86/regalloc.py > @@ -134,6 +134,8 @@ > self.final_jump_op = None > > def _prepare(self, inputargs, operations, allgcrefs): > + for box in inputargs: > + assert box.get_forwarded() is None > cpu = self.assembler.cpu > self.fm = X86FrameManager(cpu.get_baseofs_of_frame_field()) > operations = cpu.gc_ll_descr.rewrite_assembler(cpu, operations, > _______________________________________________ > pypy-commit mailing list > pypy-commit(a)python.org > https://mail.python.org/mailman/listinfo/pypy-commit

2 3