Mailman 3 November 2009 - pypy-dev

Re: [pypy-dev] pypy-c-jit compilation error
by Amaury Forgeot d'Arc Nov. 20, 2009

Nov. 20, 2009

2009/11/19 Leo Trottier <trottier(a)gmail.com>: > attached -- > also (more lodgeit abuse): http://paste.pocoo.org/show/151708/ > Leo I narrowed the problem to one function. copy the attached file somewhere, and run trackgcroot like this: python translator/c/gcc/trackgcroot.py -fdarwin -t function.s I'm still trying to understand the control flow of this function (there are 68 jumps...) and why trackgcroot fails to find the value at the origin of some register. -- Amaury Forgeot d'Arc

3 3

Re: [pypy-dev] pypy-c-jit compilation error
by Amaury Forgeot d'Arc Nov. 19, 2009

Nov. 19, 2009

2009/11/19 Leo Trottier <trottier(a)gmail.com>: > The .s files are typically removed when the build fails. I've managed to > grab one, I think, before it got deleted. > You'll see it fail in the log, presumably in the same way, here: > http://paste.pocoo.org/show/151704/ > I've abused paste.pocoo and posted it > here: http://paste.pocoo.org/show/151706/ > (I've also attached it) Unfortunately, the file you sent has been truncated (before the point where it fails) Can you please run rm implement_9.s make implement_9.s to regenerate it. Cheers, -- Amaury Forgeot d'Arc

1 0

Re: [pypy-dev] pypy-c-jit compilation error
by Maciej Fijalkowski Nov. 19, 2009

Nov. 19, 2009

Thanks. They stay for a while (4 more runs by default), so they're deleted but not immediately. Cheers, fijal On Thu, Nov 19, 2009 at 6:03 PM, Leo Trottier <trottier(a)gmail.com> wrote: > The .s files are typically removed when the build fails. I've managed to > grab one, I think, before it got deleted. > You'll see it fail in the log, presumably in the same way, here: > http://paste.pocoo.org/show/151704/ > I've abused paste.pocoo and posted it > here: http://paste.pocoo.org/show/151706/ > (I've also attached it) > Leo > > On Thu, Nov 19, 2009 at 4:19 AM, Maciej Fijalkowski <fijall(a)gmail.com> > wrote: >> >> Hey. >> >> This seems to be a problem with trackgcroot, can you put up offending >> .s file (implement_5.s) somewhere? If not, send it to me by mail, I'll >> do it. >> >> On Thu, Nov 19, 2009 at 1:13 PM, Leo Trottier <trottier(a)gmail.com> wrote: >> > As requested: >> > http://paste.pocoo.org/show/151641/ >> > >> > On Thu, Nov 19, 2009 at 4:06 AM, Amaury Forgeot d'Arc >> > <amauryfa(a)gmail.com> >> > wrote: >> >> >> >> 2009/11/19 Leo Trottier <trottier(a)gmail.com>: >> >> > [apologies for the spam, but people seem interested] >> >> > On Snow Leopard, after adding the -m32 flags recommended by Leonardo, >> >> > it >> >> > got >> >> > all the way to compilation, and then failed. I had tried to do a >> >> > c-jit >> >> > compile over the latest trunk using 32bit python. >> >> > Here is the full run log: http://paste.pocoo.org/show/151639/ >> >> >> >> Unfortunately, the log does not show all compilation output. >> >> Again you should do it manually: >> >> >> >> cd >> >> >> >> /var/folders/S+/S+mQx368HdO8O9qxkcyoQk+++TI/-Tmp-/usession-trunk-0/testing_1 >> >> make >> >> >> >> -- >> >> Amaury Forgeot d'Arc >> > >> > >> > _______________________________________________ >> > pypy-dev(a)codespeak.net >> > http://codespeak.net/mailman/listinfo/pypy-dev >> > > >

1 0

Re: [pypy-dev] pypy-c-jit compilation error
by Leo Trottier Nov. 19, 2009

Nov. 19, 2009

[apologies for the spam, but people seem interested] On Snow Leopard, after adding the -m32 flags recommended by Leonardo, it got all the way to compilation, and then failed. I had tried to do a c-jit compile over the latest trunk using 32bit python. Here is the full run log: http://paste.pocoo.org/show/151639/ Leo On Wed, Nov 18, 2009 at 7:16 PM, Leonardo wrote: > There is still no backend for x86_64 but you can compile it with the jit in > 32 bit mode. For this you have to make your cpython go to 32bit mode (search > for it on the web, it is different if it is the standard python, the > python.org dmg or the macports one) and apply this patch: > > http://paste.pocoo.org/show/151604/ > > but it is way faster than compiling on a virtual machine. > > I want to work more on this patch and make it force the -m flag (based on > the arch of the python interpreter) to the compiler so it properly works on > snow leopard (that can run both 32 and 64 bit programs independent of kernel > arch). > > If you do please tell me how it went please :) > > On Nov 19, 2009, at 1:01 AM, Leo wrote: > > > I meant to say, is the JIT working in Snow Leopard in 64 bit? > > > > Leo > > > > On Wed, Nov 18, 2009 at 7:00 PM, Leo Trottier <trottier(a)gmail.com> > wrote: > >> Is the working in Snow Leopard in 64 bit? It seemed there were > >> intractable errors in this regard earlier? > >> > >> Leo > >> > >> On Wed, Nov 18, 2009 at 6:40 PM, Leonardo wrote: > >>> Why don't you compile it on osx? most things should be working on osx > now. > >>> > >>> On Nov 18, 2009, at 11:14 PM, Leo wrote: > >>> > >>>> ... perhaps this has been resolved (??) but two compilation attempts > >>>> on the pypy/trunk from earlier today failed with the following error > >>>> message. I'm compiling on a VirtualBox install of Xubuntu on OS-X (2 > >>>> gig allocated, dual-core 1 GHz cpu). > >>>> > >>>> the pypy-dist version compiles fine ... > >>>> > >>>> and/all guidance would be appreciated > >>>> ========================== > >>>> > >>>> [platform:ERROR] make: Entering directory > `/tmp/usession-trunk-1/testing_1' > >>>> [platform:ERROR] gcc -O3 -pthread -fomit-frame-pointer > >>>> -frandom-seed=testing_1.c -o testing_1.s -S testing_1.c > >>>> -I/usr/include/python2.6 -I/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/c -I/usr/include/libffi > >>>> [platform:ERROR] make: Leaving directory > `/tmp/usession-trunk-1/testing_1' > >>>> [Timer] Timings: > >>>> [Timer] annotate --- 623.8 s > >>>> [Timer] rtype_lltype --- 768.0 s > >>>> [Timer] pyjitpl_lltype --- 570.6 s > >>>> [Timer] backendopt_lltype --- 359.7 s > >>>> [Timer] stackcheckinsertion_lltype --- 118.4 s > >>>> [Timer] database_c --- 789.6 s > >>>> [Timer] source_c --- 927.0 s > >>>> [Timer] compile_c --- 2.5 s > >>>> [Timer] =========================================== > >>>> [Timer] Total: --- 4159.4 s > >>>> [translation:ERROR] Error: > >>>> [translation:ERROR] Traceback (most recent call last): > >>>> [translation:ERROR] File "translate.py", line 277, in main > >>>> [translation:ERROR] drv.proceed(goals) > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/driver.py", line 741, in proceed > >>>> [translation:ERROR] return self._execute(goals, task_skip = > >>>> self._maybe_skip()) > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/tool/taskengine.py", line 116, in > >>>> _execute > >>>> [translation:ERROR] res = self._do(goal, taskcallable, *args, > **kwds) > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/driver.py", line 279, in _do > >>>> [translation:ERROR] res = func() > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/driver.py", line 504, in > >>>> task_compile_c > >>>> [translation:ERROR] cbuilder.compile() > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/c/genc.py", line 447, in compile > >>>> [translation:ERROR] > >>>> self.translator.platform.execute_makefile(self.targetdir) > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/platform/posix.py", line 114, in > >>>> execute_makefile > >>>> [translation:ERROR] self._handle_error(returncode, stdout, stderr, > >>>> path.join('make')) > >>>> [translation:ERROR] File "/home/leo/Source > >>>> Code/pypy-trunk/pypy/translator/platform/__init__.py", line 117, in > >>>> _handle_error > >>>> [translation:ERROR] raise CompilationError(stdout, stderr) > >>>> [translation:ERROR] CompilationError: <CompilationError err="make: > >>>> Entering directory `/tmp/usession-trunk-1/testing_1'\ngcc -O3 -pthread > >>>> -fomit-frame-pointer -frandom-seed=testi...e > >>>> Code/pypy-trunk/pypy/translator/c -I/usr/include/libffi\nmake: Leaving > >>>> directory `/tmp/usession-trunk-1/testing_1'\n"> > >>>> _______________________________________________ > >>>> pypy-dev(a)codespeak.net > >>>> http://codespeak.net/mailman/listinfo/pypy-dev > >>> > >>> -- > >>> Leonardo > >>> > >>> > >>> > >>> > >> > > -- > Leonardo > > > >

3 3

pypy-c-jit compilation error
by Leo Trottier Nov. 19, 2009

Nov. 19, 2009

... perhaps this has been resolved (??) but two compilation attempts on the pypy/trunk from earlier today failed with the following error message. I'm compiling on a VirtualBox install of Xubuntu on OS-X (2 gig allocated, dual-core 1 GHz cpu). the pypy-dist version compiles fine ... and/all guidance would be appreciated ========================== [platform:ERROR] make: Entering directory `/tmp/usession-trunk-1/testing_1' [platform:ERROR] gcc -O3 -pthread -fomit-frame-pointer -frandom-seed=testing_1.c -o testing_1.s -S testing_1.c -I/usr/include/python2.6 -I/home/leo/Source Code/pypy-trunk/pypy/translator/c -I/usr/include/libffi [platform:ERROR] make: Leaving directory `/tmp/usession-trunk-1/testing_1' [Timer] Timings: [Timer] annotate --- 623.8 s [Timer] rtype_lltype --- 768.0 s [Timer] pyjitpl_lltype --- 570.6 s [Timer] backendopt_lltype --- 359.7 s [Timer] stackcheckinsertion_lltype --- 118.4 s [Timer] database_c --- 789.6 s [Timer] source_c --- 927.0 s [Timer] compile_c --- 2.5 s [Timer] =========================================== [Timer] Total: --- 4159.4 s [translation:ERROR] Error: [translation:ERROR] Traceback (most recent call last): [translation:ERROR] File "translate.py", line 277, in main [translation:ERROR] drv.proceed(goals) [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/driver.py", line 741, in proceed [translation:ERROR] return self._execute(goals, task_skip = self._maybe_skip()) [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/tool/taskengine.py", line 116, in _execute [translation:ERROR] res = self._do(goal, taskcallable, *args, **kwds) [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/driver.py", line 279, in _do [translation:ERROR] res = func() [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/driver.py", line 504, in task_compile_c [translation:ERROR] cbuilder.compile() [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/c/genc.py", line 447, in compile [translation:ERROR] self.translator.platform.execute_makefile(self.targetdir) [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/platform/posix.py", line 114, in execute_makefile [translation:ERROR] self._handle_error(returncode, stdout, stderr, path.join('make')) [translation:ERROR] File "/home/leo/Source Code/pypy-trunk/pypy/translator/platform/__init__.py", line 117, in _handle_error [translation:ERROR] raise CompilationError(stdout, stderr) [translation:ERROR] CompilationError: <CompilationError err="make: Entering directory `/tmp/usession-trunk-1/testing_1'\ngcc -O3 -pthread -fomit-frame-pointer -frandom-seed=testi...e Code/pypy-trunk/pypy/translator/c -I/usr/include/libffi\nmake: Leaving directory `/tmp/usession-trunk-1/testing_1'\n">

2 1

pypy.objspace.std.model.UnwrapError in socket.epoll.register
by Fiedzia Nov. 11, 2009

Nov. 11, 2009

Out of curiosity i tried to use pypy to run diesel (http://dieselweb.org/lib/). For some reason pypy in trunk doesn't add OutputType property in cStringIO (perhaps it needs some installation or postprocessing because this property is mentioned in cStringIO pypy module). After setting it manually at the beginning of diesel helloworld example, i can see long traceback that ends with: File "/home/fiedzia/soft/pypy/trunk/pypy/objspace/std/fake.py", line 141, in funcrun frame.setfastscope(scope_w) File "/home/fiedzia/soft/pypy/trunk/pypy/objspace/std/fake.py", line 162, in setfastscope raise UnwrapError('calling %s: %s' % (code.cpy_callable, e)) pypy.objspace.std.model.UnwrapError: calling <method 'register' of 'select.epoll' objects>: cannot unwrap <W_ObjectObjectUserSlotsWeakrefDel() instance of <W_TypeObject(_socketobject)>> using select.epoll from pypy console seems to work, but obviously diesel pushes it to the limits. Is there something that i can try to make it working or do i hit something unsupported by pypy? -- Thanks Maciej Dziardziel

2 1

New - untrusted code
by Victor Williamson Nov. 11, 2009

Nov. 11, 2009

Hello Pypy dev, I am researching ways to allow applications to safely import untrusted code in Python without having to run the malicious code in its own process; Pypy may be a good prototyping environment. I want to verify if any work either as an extension or as interpreter changes has been done to handle untrusted imports in Pypy. Thanks and best regards, Victor

2 1

Problem with pypy-c and stacklesssocket.py
by Andrew Francis Nov. 10, 2009

Nov. 10, 2009

Hi Folks: I was testing a simple echoserver with stackless python 2.5.2 and pypy-c. I am running Ubuntu 9.04 and I compiled pypy-c with --stackless and optimization level 1. Under stackless python, the echoserver executes correctly. Under pypy-c, the server runs a while before I get: _scheduler_switch(curr, task) File "/home/andrew/lab/pypy-dist/pypy/lib/stackless.py", line 156, in _scheduler_switch next.switch() File "/home/andrew/lab/pypy-dist/pypy/lib/stackless.py", line 452, in _func func(*argl, **argd) File "/home/andrew/lab/pypy-dist/lib-python/2.5.2/stacklesssocket.py", line 77, in ManageSockets asyncore.poll(0.05) File "/home/andrew/lab/pypy-dist/lib-python/2.5.2/asyncore.py", line 109, in poll is_r = obj.readable() ReferenceError: weakly referenced object no longer exists I am using the stacklesssocket.py from http://stacklessexamples.googlecode.com/svn/trunk/examples/networking/stack… I have also enclosed the echoserver.py and a small test driver.py written with Twisted. I believe one can also use the drivers that come with the Beazley talk: http://www.dabeaz.com/coroutines/blaster.py Meanwhile I am looking over stacklesssocket.py to understand the logic. That said, any suggestions would be appreciated. Cheers, Andrew

2 2

Interesting reading for people generating assembler code
by Jacob Hallén Nov. 8, 2009

Nov. 8, 2009

I think this presentation contains some interesting tidbits and insights that may be of use to people doing machine code generation - perhaps both in the JIT and in the translation end of things. http://www.linux- kongress.org/2009/slides/compiler_survey_felix_von_leitner.pdf Jacob

1 0

pypy jit error
by Ludovico Cavedon Nov. 8, 2009

Nov. 8, 2009

Hi all, I am trying to run pypy with the jit compiler. I want to report some code I have that is failing, while with CPython it is not. I was able to isolate the the problem in a small testcase I am attaching. It is failing after exactly 1000 next() calls to the generator created by b(). File "pypyfail.py", line 6, in b b= (vv[i] for i in [0]) TypeError: object <code object <genexp>, file 'pypyfail.py', line 6> is not callable I built the latest SVN trunk snapshot of pypy with pypy/translator/goal/translate.py -Ojit Btw, the executable is called "pypy-c" rather than "pypy-c-jit", is that ok? Thank you in advance, Ludovico

3 4