From alex.suhan at gmail.com Wed Jun 1 14:14:46 2011 From: alex.suhan at gmail.com (=?UTF-8?B?QWxleCDFnnVoYW4=?=) Date: Wed, 01 Jun 2011 15:14:46 +0300 Subject: [pypy-dev] JIT for scripts Message-ID: <4DE62D36.4050909@gmail.com> Hello. PyPy works great for our PHP JIT interpreter, with nice speedups for most of the processing-intensive (with loops), shootout-ish scripts. However, I feel that short scripts running often could benefit from tracing as well if we make the interpreter ?persistent? -- that is, keep all the scripts in memory, create a jump (with a variable target) for dispatching the requested script and a backward jump (to the fore-mentioned dispatch jump) at the end of each script. This article: http://morepypy.blogspot.com/2010/11/improving-memory-behaviour-to-make-self.html refers to the loop_longevity JIT parameter, which suggests it's possible to have alternative executions between different scripts and PyPy will simply pick the appropriate trace as long as it's not too old. Other than the obvious duct taping, are there any caveats to this solution? -- asuhan From fijall at gmail.com Wed Jun 1 15:03:57 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Wed, 1 Jun 2011 15:03:57 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: <4DE62D36.4050909@gmail.com> References: <4DE62D36.4050909@gmail.com> Message-ID: On Wed, Jun 1, 2011 at 2:14 PM, Alex ?uhan wrote: > Hello. > > PyPy works great for our PHP JIT interpreter, with nice speedups for most of > the processing-intensive (with loops), shootout-ish scripts. However, I feel > that short scripts running often could benefit from tracing as well if we > make the interpreter ?persistent? -- that is, keep all the scripts in > memory, create a jump (with a variable target) for dispatching the requested > script and a backward jump (to the fore-mentioned dispatch jump) at the end > of each script. This article: > http://morepypy.blogspot.com/2010/11/improving-memory-behaviour-to-make-self.html > refers to the loop_longevity JIT parameter, which suggests it's possible to > have alternative executions between different scripts and PyPy will simply > pick the appropriate trace as long as it's not too old. Hi, that's great to hear! Can you think about keeping stuff somehow in-process (I think mod_python and mod_wsgi do some tricks along these lines for example) instead of starting new process for each script? That requires a need of cleaning up global state, but you'll get jit compiled assembler staying in memory. PS. I might not be understanding what are your proposed solution Cheers, fijal > > Other than the obvious duct taping, are there any caveats to this solution? > > -- > asuhan > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From arigo at tunes.org Wed Jun 1 15:05:56 2011 From: arigo at tunes.org (Armin Rigo) Date: Wed, 1 Jun 2011 15:05:56 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: <4DE62D36.4050909@gmail.com> References: <4DE62D36.4050909@gmail.com> Message-ID: Hi Alex, On Wed, Jun 1, 2011 at 2:14 PM, Alex ?uhan wrote: > PyPy works great for our PHP JIT interpreter Great to hear :-) > Other than the obvious duct taping, are there any caveats to this solution? Not that I can think of. It sounds like a good solution, or let's say a good workaround, for the way PyPy does things. A bient?t, Armin. From fijall at gmail.com Wed Jun 1 15:13:34 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Wed, 1 Jun 2011 15:13:34 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> Message-ID: > PS. I might not be understanding what are your proposed solution > Ok, seems I'm slow today. Does sound like a good solution, provided you clean up the global state in a nice manner. Cheers, fijal From alex.suhan at gmail.com Wed Jun 1 16:36:15 2011 From: alex.suhan at gmail.com (=?UTF-8?B?QWxleCDFnnVoYW4=?=) Date: Wed, 01 Jun 2011 17:36:15 +0300 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> Message-ID: <4DE64E5F.7030306@gmail.com> On 06/01/2011 04:05 PM, Armin Rigo wrote: > Hi Alex, > > On Wed, Jun 1, 2011 at 2:14 PM, Alex ?uhan wrote: >> PyPy works great for our PHP JIT interpreter > > Great to hear :-) > >> Other than the obvious duct taping, are there any caveats to this solution? > > Not that I can think of. It sounds like a good solution, or let's say > a good workaround, for the way PyPy does things. > > > A bient?t, > > Armin. Thank you both for the answers. Short (but often) running scripts is an unsettling real-world scenario for web apps, so it's good to have at least an acceptable workaround. We're also struggling a bit with calls speed. At the moment we're not too far behind Zend (inside 30% on killer microbenchmarks - e.g. naive Fibonacci), but catching up with the whole language will surely make the matter worse. Seems like PyPy isn't a big fan of recursion either: $ time pypy ~/fibo.py 3524578 real 0m3.763s user 0m3.740s sys 0m0.012s $ time python ~/fibo.py 3524578 real 0m1.373s user 0m1.356s sys 0m0.008s (Debian Unstable 64 bit, PyPy 1.5 vs CPython 2.6.6) Do you settle with this or do you plan to use complementary JIT techniques besides tracing? I guess those won't be framework-ish and reusable for other interpreters, anyway. -- asuhan From fijall at gmail.com Wed Jun 1 16:49:33 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Wed, 1 Jun 2011 16:49:33 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: <4DE64E5F.7030306@gmail.com> References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> Message-ID: On Wed, Jun 1, 2011 at 4:36 PM, Alex ?uhan wrote: > On 06/01/2011 04:05 PM, Armin Rigo wrote: >> >> Hi Alex, >> >> On Wed, Jun 1, 2011 at 2:14 PM, Alex ?uhan ?wrote: >>> >>> PyPy works great for our PHP JIT interpreter >> >> Great to hear :-) >> >>> Other than the obvious duct taping, are there any caveats to this >>> solution? >> >> Not that I can think of. ?It sounds like a good solution, or let's say >> a good workaround, for the way PyPy does things. >> >> >> A bient?t, >> >> Armin. > > Thank you both for the answers. Short (but often) running scripts is an > unsettling real-world scenario for web apps, so it's good to have at least > an acceptable workaround. > > We're also struggling a bit with calls speed. At the moment we're not too > far behind Zend (inside 30% on killer microbenchmarks - e.g. naive > Fibonacci), but catching up with the whole language will surely make the > matter worse. Seems like PyPy isn't a big fan of recursion either: > > $ time pypy ~/fibo.py > 3524578 > > real ? ?0m3.763s > user ? ?0m3.740s > sys ? ? 0m0.012s > > $ time python ~/fibo.py > 3524578 > > real ? ?0m1.373s > user ? ?0m1.356s > sys ? ? 0m0.008s > > (Debian Unstable 64 bit, PyPy 1.5 vs CPython 2.6.6) > > Do you settle with this or do you plan to use complementary JIT techniques > besides tracing? I guess those won't be framework-ish and reusable for other > interpreters, anyway. We don't particularly care about fibonacci (which is the worst-case scenario here, recursion works better in cases where functions are bigger). However there is a plan how to go about it (without reimplementing the JIT). It's even relatively simple. It's just that fibonacci wasn't our major goal, but it does pop up every now and then :-) And yes, it'll be reusable in other interpreters without trouble. PS. Why "catching up on language" would make things slower btw? Cheers, fijal > > -- > asuhan > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From tbaldridge at gmail.com Wed Jun 1 17:02:39 2011 From: tbaldridge at gmail.com (Timothy Baldridge) Date: Wed, 1 Jun 2011 10:02:39 -0500 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> Message-ID: We don't particularly care about fibonacci (which is the worst-case > scenario here, recursion works better in cases where functions are > bigger). However there is a plan how to go about it (without > reimplementing the JIT). It's even relatively simple. It's just that > fibonacci wasn't our major goal, but it does pop up every now and then > :-) And yes, it'll be reusable in other interpreters without trouble. > > What's the problem with recursive functions? I'm writing a interpreter that uses trampolining internally to implement tail-calls, and I'm wondering if I'll hit similar performance issues. As I'm working on a functional language my functions will tend to be quite small, and tail-calls quite common. Timothy -------------- next part -------------- An HTML attachment was scrubbed... URL: From alex.suhan at gmail.com Wed Jun 1 17:06:14 2011 From: alex.suhan at gmail.com (=?UTF-8?B?QWxleCDFnnVoYW4=?=) Date: Wed, 01 Jun 2011 18:06:14 +0300 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> Message-ID: <4DE65566.7000604@gmail.com> On 06/01/2011 05:49 PM, Maciej Fijalkowski wrote: > We don't particularly care about fibonacci (which is the worst-case > scenario here, recursion works better in cases where functions are > bigger). However there is a plan how to go about it (without > reimplementing the JIT). It's even relatively simple. It's just that > fibonacci wasn't our major goal, but it does pop up every now and then > :-) And yes, it'll be reusable in other interpreters without trouble. > > PS. Why "catching up on language" would make things slower btw? > > Cheers, > fijal > Yeah, I knew that Fibonacci will suck big time. The binary-trees from shootout works quite well and I don't care too much about Fibonacci and the likes either as long as more real-world things work better. About PS: it's just a bad feeling - we have most of the crazy things in place so maybe it won't be that bad after all. To my knowledge, there's nothing yet to be added which will break the performance, but you can never be 100% sure. -- asuhan From lac at openend.se Wed Jun 1 17:51:15 2011 From: lac at openend.se (Laura Creighton) Date: Wed, 01 Jun 2011 17:51:15 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: Message from =?UTF-8?B?QWxleCDFnnVoYW4=?= of "Wed, 01 Jun 2011 18:06:14 +0300." <4DE65566.7000604@gmail.com> References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> <4DE65566.7000604@gmail.com> Message-ID: <201106011551.p51FpFA8027766@theraft.openend.se> Great to hear about PHP. I hope you don't mind a question. Were you able to reuse the Python object model for your PHP interpreter, or is the PHP object model too different? Laura From alex.suhan at gmail.com Wed Jun 1 18:20:58 2011 From: alex.suhan at gmail.com (=?UTF-8?B?QWxleCDFnnVoYW4=?=) Date: Wed, 01 Jun 2011 19:20:58 +0300 Subject: [pypy-dev] JIT for scripts In-Reply-To: <201106011551.p51FpFA8027766@theraft.openend.se> References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> <4DE65566.7000604@gmail.com> <201106011551.p51FpFA8027766@theraft.openend.se> Message-ID: <4DE666EA.2080502@gmail.com> On 06/01/2011 06:51 PM, Laura Creighton wrote: > Great to hear about PHP. I hope you don't mind a question. Were you > able to reuse the Python object model for your PHP interpreter, or > is the PHP object model too different? > > Laura > > No, we weren't. We're reusing the Zend frontend and running (fast) the resulting bytecode, therefore we need to accept its idiosyncrasies. The references are way too weird to match the Python object model, I think. Also, PHP is not as expressive as Python, therefore reusing all the cleverness of the PyPy object model might be overkill. Long story short, we're doing our best to match Zend's craziness while running faster. -- asuhan From dimaqq at gmail.com Wed Jun 1 18:33:42 2011 From: dimaqq at gmail.com (Dima Tisnek) Date: Wed, 1 Jun 2011 09:33:42 -0700 Subject: [pypy-dev] JIT for scripts In-Reply-To: <4DE62D36.4050909@gmail.com> References: <4DE62D36.4050909@gmail.com> Message-ID: On 1 June 2011 05:14, Alex ?uhan wrote: > Hello. > > PyPy works great for our PHP JIT interpreter, with nice speedups for most of > the processing-intensive (with loops), shootout-ish scripts. However, I feel > that short scripts running often could benefit from tracing as well if we > make the interpreter ?persistent? -- that is, keep all the scripts in > memory, create a jump (with a variable target) for dispatching the requested > script and a backward jump (to the fore-mentioned dispatch jump) at the end > of each script. This article: > http://morepypy.blogspot.com/2010/11/improving-memory-behaviour-to-make-self.html > refers to the loop_longevity JIT parameter, which suggests it's possible to > have alternative executions between different scripts and PyPy will simply > pick the appropriate trace as long as it's not too old. > > Other than the obvious duct taping, are there any caveats to this solution? > > -- > asuhan > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > Would it be practical to add compiled traces to xcache cache? (or .pyo/.pyj in python-speak) d. From alex.suhan at gmail.com Wed Jun 1 18:42:07 2011 From: alex.suhan at gmail.com (=?UTF-8?B?QWxleCDFnnVoYW4=?=) Date: Wed, 01 Jun 2011 19:42:07 +0300 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> Message-ID: <4DE66BDF.7070906@gmail.com> On 06/01/2011 07:33 PM, Dima Tisnek wrote: > On 1 June 2011 05:14, Alex ?uhan wrote: >> Hello. >> >> PyPy works great for our PHP JIT interpreter, with nice speedups for most of >> the processing-intensive (with loops), shootout-ish scripts. However, I feel >> that short scripts running often could benefit from tracing as well if we >> make the interpreter ?persistent? -- that is, keep all the scripts in >> memory, create a jump (with a variable target) for dispatching the requested >> script and a backward jump (to the fore-mentioned dispatch jump) at the end >> of each script. This article: >> http://morepypy.blogspot.com/2010/11/improving-memory-behaviour-to-make-self.html >> refers to the loop_longevity JIT parameter, which suggests it's possible to >> have alternative executions between different scripts and PyPy will simply >> pick the appropriate trace as long as it's not too old. >> >> Other than the obvious duct taping, are there any caveats to this solution? >> >> -- >> asuhan >> _______________________________________________ >> pypy-dev mailing list >> pypy-dev at python.org >> http://mail.python.org/mailman/listinfo/pypy-dev >> > > Would it be practical to add compiled traces to xcache cache? (or > .pyo/.pyj in python-speak) > > d. We're using APC, not XCache, but I see your point. Sounds good, but it'll need a lot of tinkering with PyPy and the opcode cache implementation. Until we get somewhere (the whole language at the very least) we'll keep it as dumb as possible. Yeah, it'd be much better this way than an ugly concatenation of all the scripts with jumps for dispatching them, but we can't afford this luxury for the moment. -- asuhan From alex.gaynor at gmail.com Wed Jun 1 18:46:26 2011 From: alex.gaynor at gmail.com (Alex Gaynor) Date: Wed, 1 Jun 2011 09:46:26 -0700 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> Message-ID: On Wed, Jun 1, 2011 at 9:33 AM, Dima Tisnek wrote: > On 1 June 2011 05:14, Alex ?uhan wrote: > > Hello. > > > > PyPy works great for our PHP JIT interpreter, with nice speedups for most > of > > the processing-intensive (with loops), shootout-ish scripts. However, I > feel > > that short scripts running often could benefit from tracing as well if we > > make the interpreter ?persistent? -- that is, keep all the scripts in > > memory, create a jump (with a variable target) for dispatching the > requested > > script and a backward jump (to the fore-mentioned dispatch jump) at the > end > > of each script. This article: > > > http://morepypy.blogspot.com/2010/11/improving-memory-behaviour-to-make-self.html > > refers to the loop_longevity JIT parameter, which suggests it's possible > to > > have alternative executions between different scripts and PyPy will > simply > > pick the appropriate trace as long as it's not too old. > > > > Other than the obvious duct taping, are there any caveats to this > solution? > > > > -- > > asuhan > > _______________________________________________ > > pypy-dev mailing list > > pypy-dev at python.org > > http://mail.python.org/mailman/listinfo/pypy-dev > > > > Would it be practical to add compiled traces to xcache cache? (or > .pyo/.pyj in python-speak) > > d. > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > No, compiled traces contain lots of pointers into the memory of a program, not symbolic references. Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero -------------- next part -------------- An HTML attachment was scrubbed... URL: From cfbolz at gmx.de Thu Jun 2 00:28:03 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Thu, 02 Jun 2011 00:28:03 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> Message-ID: <4DE6BCF3.1000001@gmx.de> On 06/01/2011 05:02 PM, Timothy Baldridge wrote: > We don't particularly care about fibonacci (which is the worst-case > > scenario here, recursion works better in cases where functions are > bigger). However there is a plan how to go about it (without > reimplementing the JIT). It's even relatively simple. It's just that > fibonacci wasn't our major goal, but it does pop up every now and then > :-) And yes, it'll be reusable in other interpreters without trouble. > > What's the problem with recursive functions? I'm writing a interpreter > that uses trampolining internally to implement tail-calls, and I'm > wondering if I'll hit similar performance issues. No, you won't. The trampoline function is the main loop of the interpreter, and the tail calls look like loops to the JIT then. Both Pyrolog and other experiments are implemented that way and it works fine (after some twiddling). > As I'm working on a > functional language my functions will tend to be quite small, and > tail-calls quite common. Sounds exciting! What language are you implementing? I've been playing with Haskell the past three days actually. Carl Friedrich From cfbolz at gmx.de Thu Jun 2 00:29:04 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Thu, 02 Jun 2011 00:29:04 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: <4DE666EA.2080502@gmail.com> References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> <4DE65566.7000604@gmail.com> <201106011551.p51FpFA8027766@theraft.openend.se> <4DE666EA.2080502@gmail.com> Message-ID: <4DE6BD30.1090101@gmx.de> On 06/01/2011 06:20 PM, Alex ?uhan wrote: > On 06/01/2011 06:51 PM, Laura Creighton wrote: >> Great to hear about PHP. I hope you don't mind a question. Were you >> able to reuse the Python object model for your PHP interpreter, or >> is the PHP object model too different? >> >> Laura >> >> > > No, we weren't. We're reusing the Zend frontend and running (fast) the > resulting bytecode, therefore we need to accept its idiosyncrasies. The > references are way too weird to match the Python object model, I think. > Also, PHP is not as expressive as Python, therefore reusing all the > cleverness of the PyPy object model might be overkill. Long story short, > we're doing our best to match Zend's craziness while running faster. Is your implementation actually open source? Is it online somewhere? Cheers, Carl Friedrich From holger at merlinux.eu Thu Jun 2 06:47:53 2011 From: holger at merlinux.eu (holger krekel) Date: Thu, 2 Jun 2011 04:47:53 +0000 Subject: [pypy-dev] JIT for scripts In-Reply-To: <4DE6BD30.1090101@gmx.de> References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> <4DE65566.7000604@gmail.com> <201106011551.p51FpFA8027766@theraft.openend.se> <4DE666EA.2080502@gmail.com> <4DE6BD30.1090101@gmx.de> Message-ID: <20110602044753.GK20287@merlinux.eu> On Thu, Jun 02, 2011 at 00:29 +0200, Carl Friedrich Bolz wrote: > On 06/01/2011 06:20 PM, Alex ?uhan wrote: > >On 06/01/2011 06:51 PM, Laura Creighton wrote: > >>Great to hear about PHP. I hope you don't mind a question. Were you > >>able to reuse the Python object model for your PHP interpreter, or > >>is the PHP object model too different? > >> > >>Laura > >> > >> > > > >No, we weren't. We're reusing the Zend frontend and running (fast) the > >resulting bytecode, therefore we need to accept its idiosyncrasies. The > >references are way too weird to match the Python object model, I think. > >Also, PHP is not as expressive as Python, therefore reusing all the > >cleverness of the PyPy object model might be overkill. Long story short, > >we're doing our best to match Zend's craziness while running faster. > > Is your implementation actually open source? Is it online somewhere? it's at https://bitbucket.org/asuhan/happy holger From cfbolz at gmx.de Thu Jun 2 08:31:59 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Thu, 02 Jun 2011 08:31:59 +0200 Subject: [pypy-dev] JIT for scripts In-Reply-To: <20110602044753.GK20287@merlinux.eu> References: <4DE62D36.4050909@gmail.com> <4DE64E5F.7030306@gmail.com> <4DE65566.7000604@gmail.com> <201106011551.p51FpFA8027766@theraft.openend.se> <4DE666EA.2080502@gmail.com> <4DE6BD30.1090101@gmx.de> <20110602044753.GK20287@merlinux.eu> Message-ID: <4DE72E5F.3060800@gmx.de> On 06/02/2011 06:47 AM, holger krekel wrote: >> Is your implementation actually open source? Is it online somewhere? > > it's at https://bitbucket.org/asuhan/happy Great, thank you very much. Carl Friedrich From arigo at tunes.org Thu Jun 2 19:05:13 2011 From: arigo at tunes.org (Armin Rigo) Date: Thu, 2 Jun 2011 19:05:13 +0200 Subject: [pypy-dev] unable to translate with current head of default In-Reply-To: References: <4DDCAEF2.5040100@gmail.com> Message-ID: Hi Wim, On Tue, May 31, 2011 at 8:46 PM, wrote: > ==16631== Process terminating with default action of signal 11 (SIGSEGV) > ==16631== ?Access not within mapped region at address 0x4 > ==16631== ? ?at 0x8AADB0C: pypy_g_OptHeap_emitting_operation It definitely crashes during optimization of the trace produced by the JIT. I imagine that we could get somewhere by carefully inspecting the unoptimized operations that lead to the crash. Try enabling the dump with "PYPYLOG=jit-log-unopt:log", and then reproducing the crash. This should ideally leave behind a file called "log" with all the unoptimized loops seen so far, and the faulty one is the last one (i.e. the last section between "{jit-log-unopt" and "jit-log-unopt}"). A bient?t, Armin. From tbaldridge at gmail.com Thu Jun 2 20:11:11 2011 From: tbaldridge at gmail.com (Timothy Baldridge) Date: Thu, 2 Jun 2011 13:11:11 -0500 Subject: [pypy-dev] [translation:ERROR] AttributeError': 'FrozenDesc' object has no attribute 'rowkey' Message-ID: What does error mean, when running the translator against my rpython code? [translation:ERROR] ?AttributeError': 'FrozenDesc' object has no attribute 'rowkey' Can anyone give me some insight? Thanks. Timothy The code: class?Runtime(Obj): ????def?__init__(self): ????????Obj.__init__(self) ????????self.defs?=?dict() ????????self.vars?=?[] The Error: [translation:ERROR] Error: [translation:ERROR] ?Traceback (most recent call last): [translation:ERROR] ? ?File "./pypy/translator/goal/translate.py", line 306, in main [translation:ERROR] ? ? drv.proceed(goals) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\translator\driver.py", line 809, in proceed [translation:ERROR] ? ? return self._execute(goals, task_skip = self._maybe_skip()) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\translator\tool\taskeng ine.py", line 116, in _execute [translation:ERROR] ? ? res = self._do(goal, taskcallable, *args, **kwds) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\translator\driver.py", line 286, in _do [translation:ERROR] ? ? res = func() [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\translator\driver.py", line 323, in task_annotate [translation:ERROR] ? ? s = annotator.build_types(self.entry_point, self.inputtypes) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 103, in build_types [translation:ERROR] ? ? return self.build_graph_types(flowgraph, inputcells, complete_now=complete_n ow) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 194, in build_graph_types [translation:ERROR] ? ? self.complete() [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 250, in complete [translation:ERROR] ? ? self.processblock(graph, block) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 448, in processblock [translation:ERROR] ? ? self.flowin(graph, block) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 508, in flowin [translation:ERROR] ? ? self.consider_op(block.operations[i]) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 710, in consider_op [translation:ERROR] ? ? raise_nicer_exception(op, str(graph)) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\annrpython.p y", line 707, in consider_op [translation:ERROR] ? ? resultcell = consider_meth(*argcells) [translation:ERROR] ? ?File "<121-codegen c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation \annrpython.py:745>", line 3, in consider_op_simple_call [translation:ERROR] ? ? return arg.simple_call(*args) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\unaryop.py", ?line 175, in simple_call [translation:ERROR] ? ? return obj.call(getbookkeeper().build_args("simple_call", args_s)) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\unaryop.py", ?line 695, in call [translation:ERROR] ? ? return bookkeeper.pbc_call(pbc, args) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\bookkeeper.p y", line 640, in pbc_call [translation:ERROR] ? ? first.mergecallfamilies(*descs[1:]) [translation:ERROR] ? ?File "c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\description. py", line 151, in mergecallfamilies [translation:ERROR] ? ? changed, rep, callfamily = call_families.find(self.rowkey()) [translation:ERROR] ?AttributeError': 'FrozenDesc' object has no attribute 'rowkey' [translation:ERROR] ? ? .. v0 = simple_call((type dict)) [translation:ERROR] ? ? .. '(runtime.runtime:7)Runtime.__init__' [translation:ERROR] Processing block: [translation:ERROR] ?block at 9 is a [translation:ERROR] ?in (runtime.runtime:7)Runtime.__init__ [translation:ERROR] ?containing the following operations: [translation:ERROR] ? ? ? ?v1 = simple_call((instancemethod __init__), self_0) [translation:ERROR] ? ? ? ?v0 = simple_call((type dict)) [translation:ERROR] ? ? ? ?v2 = setattr(self_0, ('defs'), v0) [translation:ERROR] ? ? ? ?v3 = newlist() [translation:ERROR] ? ? ? ?v4 = setattr(self_0, ('vars'), v3) [translation:ERROR] ?--end-- [translation] start debugger... > c:\projects\pypy\pypy-1.5-src\pypy-1.5-src\pypy\annotation\description.py(151)mergecallfamilies() -> changed, rep, callfamily = call_families.find(self.rowkey()) (Pdb+) -- ?One of the main causes of the fall of the Roman Empire was that?lacking zero?they had no way to indicate successful termination of their C programs.? (Robert Firth) From benjamin at python.org Thu Jun 2 20:29:29 2011 From: benjamin at python.org (Benjamin Peterson) Date: Thu, 2 Jun 2011 13:29:29 -0500 Subject: [pypy-dev] [translation:ERROR] AttributeError': 'FrozenDesc' object has no attribute 'rowkey' In-Reply-To: References: Message-ID: 2011/6/2 Timothy Baldridge : > What does error mean, when running the translator against my rpython code? > > [translation:ERROR] ?AttributeError': 'FrozenDesc' object has no > attribute 'rowkey' > > Can anyone give me some insight? > > Thanks. > > Timothy > > > The code: > > class?Runtime(Obj): > ????def?__init__(self): > ????????Obj.__init__(self) > ????????self.defs?=?dict() There's no dict() type in RPython. Use {}. (Yes, obscure). -- Regards, Benjamin From tbaldridge at gmail.com Thu Jun 2 21:21:04 2011 From: tbaldridge at gmail.com (Timothy Baldridge) Date: Thu, 2 Jun 2011 14:21:04 -0500 Subject: [pypy-dev] [translation:ERROR] AttributeError': 'FrozenDesc' object has no attribute 'rowkey' In-Reply-To: References: Message-ID: > > There's no dict() type in RPython. Use {}. (Yes, obscure). > > I would have never guessed. On the plus side, your explanation did give me a good laugh. Timothy -------------- next part -------------- An HTML attachment was scrubbed... URL: From wlavrijsen at lbl.gov Fri Jun 3 01:43:30 2011 From: wlavrijsen at lbl.gov (wlavrijsen at lbl.gov) Date: Thu, 2 Jun 2011 16:43:30 -0700 (PDT) Subject: [pypy-dev] unable to translate with current head of default In-Reply-To: References: <4DDCAEF2.5040100@gmail.com> Message-ID: Hi Armin, > It definitely crashes during optimization of the trace produced by the > JIT. I imagine that we could get somewhere by carefully inspecting > the unoptimized operations that lead to the crash. Try enabling the > dump with "PYPYLOG=jit-log-unopt:log", and then reproducing the crash. > This should ideally leave behind a file called "log" with all the > unoptimized loops seen so far, and the faulty one is the last one > (i.e. the last section between "{jit-log-unopt" and "jit-log-unopt}"). thanks; but to get something useful, I had to use jit-log-noopt: (i.e. output to stderr, since buffers were not flushed; and slightly different filter). Since readline in PyPy is a .py, I was able to reduce the loop that exhibits a segfault to this simple script: for i in range(1001): unicode("aap") although the trace is still far from simple, at least it's a lot shorter than what I got before. :) I've attached the trace.log file. And again, the problem optimization was "heap." For comparison purposes, I grabbed the latest "green" nightly (44610) and made a trace with that one, too (the prebuild editions don't crash). That file is trace_nightly.log (also attached). It is quite a bit shorter. Anything in there that rings a bell? Thanks, Wim -- WLavrijsen at lbl.gov -- +1 (510) 486 6411 -- www.lavrijsen.net -------------- next part -------------- A non-text attachment was scrubbed... Name: trace.log Type: text/x-log Size: 16420 bytes Desc: URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: trace_nightly.log Type: text/x-log Size: 15435 bytes Desc: URL: From caleb.hattingh at gmail.com Sun Jun 5 16:18:53 2011 From: caleb.hattingh at gmail.com (Caleb Hattingh) Date: Sun, 5 Jun 2011 16:18:53 +0200 Subject: [pypy-dev] pypy 1.5 on Windows: sqlite connection args Message-ID: In lib-pypy/_sqlite3.py, line 239 (in Win 1.5 stable release) in class Connection(object): def __init__(self, database, isolation_level="", detect_types=0, timeout=None, cached_statements=None, factory=None): ...might need to become: def __init__(self, database, isolation_level="", detect_types=0, timeout=None, cached_statements=None, factory=None, *check_same_thread=False *): I am testing the web framework web2py against pypy, and this issue came up. I don't know too much about sqlite itself. Spent a lot of time modifying code to explicitly close file handles because of the different behaviour of the ref counting cpython vs. the gc pypy. -------------- next part -------------- An HTML attachment was scrubbed... URL: From pypy at pocketnix.org Mon Jun 6 15:19:13 2011 From: pypy at pocketnix.org (Da_Blitz) Date: Mon, 6 Jun 2011 13:19:13 +0000 Subject: [pypy-dev] os.tmpnam warnings Message-ID: <20110606131913.GA25550@pocketnix.org> Hi pypy 1.5 does not display a warning when using the os.tempnam and os.tmpnam functions. use of these functions is not recommended as they can cause security issues and hence python issues a RuntimeWarning below is a patch to app_posix.py to make it act more like cpython ------------------------------------------ diff -r b590cf6de419 pypy/module/posix/app_posix.py --- a/pypy/module/posix/app_posix.py Fri Apr 29 17:42:40 2011 +0200 +++ b/pypy/module/posix/app_posix.py Mon Jun 06 23:15:15 2011 +1000 @@ -107,6 +107,10 @@ def tmpnam(): """Return an absolute pathname of a file that did not exist at the time the call is made.""" + from warnings import warn + from exceptions import RuntimeWarning + warn(RuntimeWarning("tmpnam is a potential security risk to your program")) + import tempfile return tempfile.mktemp() @@ -114,6 +118,10 @@ """Return an absolute pathname of a file that did not exist at the time the call is made. The directory and a prefix may be specified as strings; they may be omitted or None if not needed.""" + from warnings import warn + from exceptions import RuntimeWarning + warn(RuntimeWarning("tempnam is a potential security risk to your program")) + import tempfile return tempfile.mktemp('', prefix or 'tmp', dir) From pypy at pocketnix.org Mon Jun 6 15:32:37 2011 From: pypy at pocketnix.org (Da_Blitz) Date: Mon, 6 Jun 2011 13:32:37 +0000 Subject: [pypy-dev] setpgrp and getpgrp platform:errors Message-ID: <20110606133237.GB25550@pocketnix.org> Hi below is a patch that removes these errors during translation or when using py.py i am poking with things i do not fully understand here and would prefer it if someone told me i was an idiot and that args were passed to the functions for some reason (eg compatibility with the BSD's). according to the man page on linux (debian) there are 2 calling conventions, the System V and the BSD the os.setpgrp and os.getpgrp do not appear to take arguments and so i just chopped them off and "Magic Happened" (tm) feel free to point out somthing i missed/broke/caused meltdown on ------------------------- diff -r b590cf6de419 pypy/rpython/module/ll_os.py --- a/pypy/rpython/module/ll_os.py Fri Apr 29 17:42:40 2011 +0200 +++ b/pypy/rpython/module/ll_os.py Mon Jun 06 23:26:35 2011 +1000 @@ -191,13 +191,13 @@ if hasattr(os, 'getpgrp'): self.GETPGRP_HAVE_ARG = platform.checkcompiles( - "getpgrp(0)", + "getpgrp()", '#include ', []) if hasattr(os, 'setpgrp'): self.SETPGRP_HAVE_ARG = platform.checkcompiles( - "setpgrp(0,0)", + "setpgrp()", '#include ', []) ( From fijall at gmail.com Mon Jun 6 15:35:01 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Mon, 6 Jun 2011 15:35:01 +0200 Subject: [pypy-dev] os.tmpnam warnings In-Reply-To: <20110606131913.GA25550@pocketnix.org> References: <20110606131913.GA25550@pocketnix.org> Message-ID: Can you post this to issue tracker? Patches on the mailing lists tend to get forgotten... On Mon, Jun 6, 2011 at 3:19 PM, Da_Blitz wrote: > Hi > > pypy 1.5 does not display a warning when using the os.tempnam and > os.tmpnam functions. use of these functions is not recommended as they > can cause security issues and hence python issues a RuntimeWarning > > > below is a patch to app_posix.py to make it act more like cpython > > ------------------------------------------ > > diff -r b590cf6de419 pypy/module/posix/app_posix.py > --- a/pypy/module/posix/app_posix.py ? ?Fri Apr 29 17:42:40 2011 +0200 > +++ b/pypy/module/posix/app_posix.py ? ?Mon Jun 06 23:15:15 2011 +1000 > @@ -107,6 +107,10 @@ > ?def tmpnam(): > ? ? """Return an absolute pathname of a file that did not exist at the > ? ? time the call is made.""" > + ? ?from warnings import warn > + ? ?from exceptions import RuntimeWarning > + ? ?warn(RuntimeWarning("tmpnam is a potential security risk to your program")) > + > ? ? import tempfile > ? ? return tempfile.mktemp() > > @@ -114,6 +118,10 @@ > ? ? """Return an absolute pathname of a file that did not exist at the > ? ? time the call is made. ?The directory and a prefix may be specified > ? ? as strings; they may be omitted or None if not needed.""" > + ? ?from warnings import warn > + ? ?from exceptions import RuntimeWarning > + ? ?warn(RuntimeWarning("tempnam is a potential security risk to your program")) > + > ? ? import tempfile > ? ? return tempfile.mktemp('', prefix or 'tmp', dir) > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From benjamin at python.org Mon Jun 6 15:37:53 2011 From: benjamin at python.org (Benjamin Peterson) Date: Mon, 6 Jun 2011 08:37:53 -0500 Subject: [pypy-dev] os.tmpnam warnings In-Reply-To: <20110606131913.GA25550@pocketnix.org> References: <20110606131913.GA25550@pocketnix.org> Message-ID: 2011/6/6 Da_Blitz : > Hi > > pypy 1.5 does not display a warning when using the os.tempnam and > os.tmpnam functions. use of these functions is not recommended as they > can cause security issues and hence python issues a RuntimeWarning > > > below is a patch to app_posix.py to make it act more like cpython > > ------------------------------------------ > > diff -r b590cf6de419 pypy/module/posix/app_posix.py > --- a/pypy/module/posix/app_posix.py ? ?Fri Apr 29 17:42:40 2011 +0200 > +++ b/pypy/module/posix/app_posix.py ? ?Mon Jun 06 23:15:15 2011 +1000 > @@ -107,6 +107,10 @@ > ?def tmpnam(): > ? ? """Return an absolute pathname of a file that did not exist at the > ? ? time the call is made.""" > + ? ?from warnings import warn > + ? ?from exceptions import RuntimeWarning > + ? ?warn(RuntimeWarning("tmpnam is a potential security risk to your program")) > + > ? ? import tempfile > ? ? return tempfile.mktemp() > > @@ -114,6 +118,10 @@ > ? ? """Return an absolute pathname of a file that did not exist at the > ? ? time the call is made. ?The directory and a prefix may be specified > ? ? as strings; they may be omitted or None if not needed.""" > + ? ?from warnings import warn > + ? ?from exceptions import RuntimeWarning When you do post to the tracker, note RuntimeWarning is builtin. -- Regards, Benjamin From amauryfa at gmail.com Mon Jun 6 15:55:08 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Mon, 6 Jun 2011 15:55:08 +0200 Subject: [pypy-dev] os.tmpnam warnings In-Reply-To: References: <20110606131913.GA25550@pocketnix.org> Message-ID: Hi, 2011/6/6 Benjamin Peterson : > 2011/6/6 Da_Blitz : >> ? ? """Return an absolute pathname of a file that did not exist at the >> ? ? time the call is made. ?The directory and a prefix may be specified >> ? ? as strings; they may be omitted or None if not needed.""" >> + ? ?from warnings import warn >> + ? ?from exceptions import RuntimeWarning > > When you do post to the tracker, note RuntimeWarning is builtin. Also ensure that the warning message points to the *caller* of os.tempnam(). (I don't remember whether applevel functions are visible in the Traceback) -- Amaury Forgeot d'Arc From cfbolz at gmx.de Mon Jun 6 19:57:36 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Mon, 06 Jun 2011 19:57:36 +0200 Subject: [pypy-dev] CFP: IWTC 2011: First International Workshop on Trace Compilation Message-ID: <4DED1510.6080001@gmx.de> IWTC 2011: First International Workshop on Trace Compilation August 24, 2011, Kongens Lyngby, Denmark Colocated with PPPJ 2011 === SCOPE === The first International Workshop on Trace Compilation (IWTC) aims to bring together researchers and practitioners who study trace compilation and optimization techniques within statically and dynamically typed programming languages. The workshop will be co-located with PPPJ 2011, August 24-26, at the Technical University of Denmark, Kongens Lyngby, Denmark. We invite papers on all topics related to trace compilation, including but not limited to: Trace recording and compilation Trace compilation for static and dynamic languages VM integration of trace compilers Novel optimizations for trace based compilers Novel intermediate representations for trace based compilers Formal aspects of traces Integration of trace compilers and method compilers Traces in binary translation systems Trace collection and analysis Testing of trace based compilers Industry / experience papers Tool demonstration === IMPORTANT DATES === June 20, 2011 Paper Submission July 15, 2011 Author Notification August 1, 2011 Camera Ready Submissions === SUBMISSION GUIDELINES === All papers must represent original and unpublished work that is not currently under review. Papers will be evaluated according to their significance, originality, technical content, style, clarity, and relevance to the conference. At least one author of each accepted paper is required to attend the conference and present the paper. Papers should range between 4 and 6 pages. Please number the pages. The workshop proceedings will be published as part of the ACM International Proceedings Series and will be disseminated through the ACM Digital Library. The required format for the submission is the ACM SIGPLAN Proceedings Style, using 9-point font size. The author information containing the Latex and Word templates can be found at http://www.sigplan.org/authorInformation.htm. Please apply the ACM Computing Classification categories and terms. The submission website is at http://www.sigplan.org/authorInformation.htm. === PROGRAM COMMITTEE === - Michael Bebenita, University of California, Irvine (USA), chair - Borys J. Bradel, Altera Corporation (Canada) - Michael Franz, University of California, Irvine (USA) - Carl Friedrich Bolz, Heinrich-Heine-University D?sseldorf (Germany) - Andreas Gal, Mozilla Corporation (USA) - Ian Rogers, Azul Systems (USA) - Nikolai Tillman, Microsoft Corporation (USA) - Christian Wimmer, Oracle Corporation (USA) - Peng Wu, IBM T.J.Watson Research Center (USA) - Mathew Zaleski, University of Toronto (Canada) From tbaldridge at gmail.com Mon Jun 6 20:18:21 2011 From: tbaldridge at gmail.com (Timothy Baldridge) Date: Mon, 6 Jun 2011 13:18:21 -0500 Subject: [pypy-dev] Arrays in rpython Message-ID: Is there a way in rpython to specify that a given list should be implemented as an array? I want to create a list of objects exactly 32 entries long. I could use [] but it seems like that would be slower than using an array. In CPython I can create an array of primitive types, but I'm looking to create a array of object types. Timothy -- ?One of the main causes of the fall of the Roman Empire was that?lacking zero?they had no way to indicate successful termination of their C programs.? (Robert Firth) From arigo at tunes.org Mon Jun 6 20:50:22 2011 From: arigo at tunes.org (Armin Rigo) Date: Mon, 6 Jun 2011 20:50:22 +0200 Subject: [pypy-dev] Compiling PyPy 1.5 fails on Slackware 13.37 In-Reply-To: <20110512102059.GA5341@kiras> References: <20110512095915.GA3960@kiras> <20110512102059.GA5341@kiras> Message-ID: Hi, The problem should be fixed now (017e187b2716). Can you try again? Hopefully I didn't break Windows or Mac builds, but I cannot easily be sure. Feel free to blame me if I did :-) A bient?t, Armin. From amauryfa at gmail.com Mon Jun 6 20:52:40 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Mon, 6 Jun 2011 20:52:40 +0200 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: Hi, 2011/6/6 Timothy Baldridge : > Is there a way in rpython to specify that a given list should be > implemented as an array? I want to create a list of objects exactly 32 > entries long. I could use [] but it seems like that would be slower > than using an array. In CPython I can create an array of primitive > types, but I'm looking to create a array of object types. In case you don't already know: in Python (and RPython), lists are implemented as resizeable arrays, i.e. they look more like a C++ std::vector and not a std::list. This said, IIRC RPython has special code for fixed-size lists. I think you just have to avoid all resizing functions. This means that the list must be created with statements like array_of_ints = [0] * 32 array_of_objects = [None] * 32 -- Amaury Forgeot d'Arc From tbaldridge at gmail.com Mon Jun 6 20:54:56 2011 From: tbaldridge at gmail.com (Timothy Baldridge) Date: Mon, 6 Jun 2011 13:54:56 -0500 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: > In case you don't already know: in Python (and RPython), lists are > implemented as > resizeable arrays, i.e. they look more like a C++ std::vector and not > a std::list. > > This said, IIRC RPython has special code for fixed-size lists. > I think you just have to avoid all resizing functions. > This means that the list must be created with statements like > ? ?array_of_ints = [0] * 32 > ? ?array_of_objects = [None] * 32 Excellent! Exactly what I was looking for. Thanks. Timothy -- ?One of the main causes of the fall of the Roman Empire was that?lacking zero?they had no way to indicate successful termination of their C programs.? (Robert Firth) From anto.cuni at gmail.com Mon Jun 6 20:58:59 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Mon, 06 Jun 2011 20:58:59 +0200 Subject: [pypy-dev] [pypy-commit] pypy default: (arigo, bivab) implement longlong2float and float2longlong in a way that is more close to the C standard. On ARM/32bit this code was causing a reodering of instructions that filled one of the two words with garbage In-Reply-To: <20110606160755.50F28820AE@wyvern.cs.uni-duesseldorf.de> References: <20110606160755.50F28820AE@wyvern.cs.uni-duesseldorf.de> Message-ID: <4DED2373.1040504@gmail.com> On 06/06/11 18:07, bivab wrote: > static double pypy__longlong2float(long long x) { > + int i; > + double dd; > char *p = (char*)&x; > - return *((double*)p); > + char *d = (char*)ⅆ > + for(i = 0; i < 8; i++) { > + d[i] = p[i]; > + } > + return dd; speaking of portability, what about using sizeof(double)/sizeof(char) instead of hardcoding 8? ciao, Anto From romain.py at gmail.com Tue Jun 7 04:40:11 2011 From: romain.py at gmail.com (Romain Guillebert) Date: Tue, 7 Jun 2011 04:40:11 +0200 Subject: [pypy-dev] [GSoC] CTypes backend for Cython aiming PyPy - Week 2 Message-ID: <20110607024011.GA4162@ubuntu> Hi I summarized the second week of my Summer of Code project in this blog post: http://rguillebert.blogspot.com/2011/06/cython-backend-aiming-pypy-week-2.html Cheers Romain From hakan at debian.org Tue Jun 7 07:58:49 2011 From: hakan at debian.org (Hakan Ardo) Date: Tue, 7 Jun 2011 07:58:49 +0200 Subject: [pypy-dev] CFP: IWTC 2011: First International Workshop on Trace Compilation In-Reply-To: <4DED1510.6080001@gmx.de> References: <4DED1510.6080001@gmx.de> Message-ID: Hi, would it be a good idea to submit something about our unrolling efforts here? I'd be happy to put in some typing work for that. However I've never published in this field before so I would suggest a joint effort where someone else takes more responsibility for the structure of the paper. On Mon, Jun 6, 2011 at 7:57 PM, Carl Friedrich Bolz wrote: > IWTC 2011: First International Workshop on Trace Compilation > > August 24, 2011, Kongens Lyngby, Denmark > > Colocated with PPPJ 2011 > > === SCOPE === > > The first International Workshop on Trace Compilation (IWTC) aims to > bring together researchers and practitioners who study trace compilation > and optimization techniques within statically and dynamically typed > programming languages. The workshop will be co-located with PPPJ 2011, > August 24-26, at the Technical University of Denmark, Kongens Lyngby, > Denmark. > > We invite papers on all topics related to trace compilation, including > but not limited to: > > Trace recording and compilation > Trace compilation for static and dynamic languages > VM integration of trace compilers > Novel optimizations for trace based compilers > Novel intermediate representations for trace based compilers > Formal aspects of traces > Integration of trace compilers and method compilers > Traces in binary translation systems > Trace collection and analysis > Testing of trace based compilers > Industry / experience papers > Tool demonstration > > === IMPORTANT DATES === > > June 20, 2011 ? Paper Submission > July 15, 2011 ? Author Notification > August 1, 2011 ?Camera Ready Submissions > > === SUBMISSION GUIDELINES === > > All papers must represent original and unpublished work that is not > currently under review. Papers will be evaluated according to their > significance, originality, technical content, style, clarity, and > relevance to the conference. > > At least one author of each accepted paper is required to attend the > conference and present the paper. > > Papers should range between 4 and 6 pages. Please number the pages. > The workshop proceedings will be published as part of the ACM > International Proceedings Series and will be disseminated through the > ACM Digital Library. The required format for the submission is the > ACM SIGPLAN Proceedings Style, using 9-point font size. The author > information containing the Latex and Word templates can be found at > http://www.sigplan.org/authorInformation.htm. Please apply the ACM > Computing Classification categories and terms. The submission website > is at http://www.sigplan.org/authorInformation.htm. > > === PROGRAM COMMITTEE === > > - Michael Bebenita, University of California, Irvine (USA), chair > - Borys J. Bradel, Altera Corporation (Canada) > - Michael Franz, University of California, Irvine (USA) > - Carl Friedrich Bolz, Heinrich-Heine-University D?sseldorf (Germany) > - Andreas Gal, Mozilla Corporation (USA) > - Ian Rogers, Azul Systems (USA) > - Nikolai Tillman, Microsoft Corporation (USA) > - Christian Wimmer, Oracle Corporation (USA) > - Peng Wu, IBM T.J.Watson Research Center (USA) > - Mathew Zaleski, University of Toronto (Canada) > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > -- H?kan Ard? From fijall at gmail.com Tue Jun 7 08:40:49 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Tue, 7 Jun 2011 08:40:49 +0200 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: On Mon, Jun 6, 2011 at 8:54 PM, Timothy Baldridge wrote: >> In case you don't already know: in Python (and RPython), lists are >> implemented as >> resizeable arrays, i.e. they look more like a C++ std::vector and not >> a std::list. >> >> This said, IIRC RPython has special code for fixed-size lists. >> I think you just have to avoid all resizing functions. >> This means that the list must be created with statements like >> ? ?array_of_ints = [0] * 32 >> ? ?array_of_objects = [None] * 32 > > > Excellent! Exactly what I was looking for. Thanks. > > Timothy > There is also pypy.rlib.debug.make_sure_not_resized which would make annotation explode if you by chance resize this list. From fijall at gmail.com Tue Jun 7 08:42:46 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Tue, 7 Jun 2011 08:42:46 +0200 Subject: [pypy-dev] CFP: IWTC 2011: First International Workshop on Trace Compilation In-Reply-To: References: <4DED1510.6080001@gmx.de> Message-ID: On Tue, Jun 7, 2011 at 7:58 AM, Hakan Ardo wrote: > Hi, > would it be a good idea to submit something about our unrolling > efforts here? I'd be happy to put in some typing work for that. > However I've never published in this field before so I would suggest a > joint effort where someone else takes more responsibility for the > structure of the paper. I can maybe help? PS. Carl - can you use @python.org address instead? > > On Mon, Jun 6, 2011 at 7:57 PM, Carl Friedrich Bolz wrote: >> IWTC 2011: First International Workshop on Trace Compilation >> >> August 24, 2011, Kongens Lyngby, Denmark >> >> Colocated with PPPJ 2011 >> >> === SCOPE === >> >> The first International Workshop on Trace Compilation (IWTC) aims to >> bring together researchers and practitioners who study trace compilation >> and optimization techniques within statically and dynamically typed >> programming languages. The workshop will be co-located with PPPJ 2011, >> August 24-26, at the Technical University of Denmark, Kongens Lyngby, >> Denmark. >> >> We invite papers on all topics related to trace compilation, including >> but not limited to: >> >> Trace recording and compilation >> Trace compilation for static and dynamic languages >> VM integration of trace compilers >> Novel optimizations for trace based compilers >> Novel intermediate representations for trace based compilers >> Formal aspects of traces >> Integration of trace compilers and method compilers >> Traces in binary translation systems >> Trace collection and analysis >> Testing of trace based compilers >> Industry / experience papers >> Tool demonstration >> >> === IMPORTANT DATES === >> >> June 20, 2011 ? Paper Submission >> July 15, 2011 ? Author Notification >> August 1, 2011 ?Camera Ready Submissions >> >> === SUBMISSION GUIDELINES === >> >> All papers must represent original and unpublished work that is not >> currently under review. Papers will be evaluated according to their >> significance, originality, technical content, style, clarity, and >> relevance to the conference. >> >> At least one author of each accepted paper is required to attend the >> conference and present the paper. >> >> Papers should range between 4 and 6 pages. Please number the pages. >> The workshop proceedings will be published as part of the ACM >> International Proceedings Series and will be disseminated through the >> ACM Digital Library. The required format for the submission is the >> ACM SIGPLAN Proceedings Style, using 9-point font size. The author >> information containing the Latex and Word templates can be found at >> http://www.sigplan.org/authorInformation.htm. Please apply the ACM >> Computing Classification categories and terms. The submission website >> is at http://www.sigplan.org/authorInformation.htm. >> >> === PROGRAM COMMITTEE === >> >> - Michael Bebenita, University of California, Irvine (USA), chair >> - Borys J. Bradel, Altera Corporation (Canada) >> - Michael Franz, University of California, Irvine (USA) >> - Carl Friedrich Bolz, Heinrich-Heine-University D?sseldorf (Germany) >> - Andreas Gal, Mozilla Corporation (USA) >> - Ian Rogers, Azul Systems (USA) >> - Nikolai Tillman, Microsoft Corporation (USA) >> - Christian Wimmer, Oracle Corporation (USA) >> - Peng Wu, IBM T.J.Watson Research Center (USA) >> - Mathew Zaleski, University of Toronto (Canada) >> _______________________________________________ >> pypy-dev mailing list >> pypy-dev at python.org >> http://mail.python.org/mailman/listinfo/pypy-dev >> > > > > -- > H?kan Ard? > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From arigo at tunes.org Tue Jun 7 15:38:22 2011 From: arigo at tunes.org (Armin Rigo) Date: Tue, 7 Jun 2011 15:38:22 +0200 Subject: [pypy-dev] [pypy-commit] pypy default: (arigo, bivab) implement longlong2float and float2longlong in a way that is more close to the C standard. On ARM/32bit this code was causing a reodering of instructions that filled one of the two words with Message-ID: Hi Anto, On Mon, Jun 6, 2011 at 8:58 PM, Antonio Cuni wrote: > speaking of portability, what about using sizeof(double)/sizeof(char) instead > of hardcoding 8? That change was not about portability, but about C compilers getting too clever at optimizing. These two functions are (supposed) to be only used on machines with 8-byte long longs and doubles. I could add an assert, checking that sizeof(double) == sizeof(long long) == 8. Armin. From anto.cuni at gmail.com Tue Jun 7 16:02:13 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Tue, 07 Jun 2011 16:02:13 +0200 Subject: [pypy-dev] [pypy-commit] pypy default: (arigo, bivab) implement longlong2float and float2longlong in a way that is more close to the C standard. On ARM/32bit this code was causing a reodering of instructions that filled one of the two words with In-Reply-To: References: Message-ID: <4DEE2F65.4060400@gmail.com> On 07/06/11 15:38, Armin Rigo wrote: > That change was not about portability, but about C compilers getting > too clever at optimizing. These two functions are (supposed) to be > only used on machines with 8-byte long longs and doubles. I could add > an assert, checking that sizeof(double) == sizeof(long long) == 8. yes, that sounds like a good idea. Not that there are any machine around where the assert would fail, AFAIK, but better to play on the safe side. ciao, Anto From jan at jandecaluwe.com Tue Jun 7 17:02:44 2011 From: jan at jandecaluwe.com (Jan Decaluwe) Date: Tue, 07 Jun 2011 17:02:44 +0200 Subject: [pypy-dev] Great results for MyHDL simulations Message-ID: <4DEE3D94.1020408@jandecaluwe.com> I am seeing great improvements for MyHDL simulations by using PyPy, and I have written a page about it: http://www.myhdl.org/doku.php/performance I have asked for feedback from the MyHDL community to confirm these results, before creating some more buzz in the hardware design world :-) Thanks! Jan Decaluwe -- Jan Decaluwe - Resources bvba - http://www.jandecaluwe.com Python as a HDL: http://www.myhdl.org VHDL development, the modern way: http://www.sigasi.com Analog design automation: http://www.mephisto-da.com World-class digital design: http://www.easics.com From anto.cuni at gmail.com Tue Jun 7 17:23:19 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Tue, 07 Jun 2011 17:23:19 +0200 Subject: [pypy-dev] Great results for MyHDL simulations In-Reply-To: <4DEE3D94.1020408@jandecaluwe.com> References: <4DEE3D94.1020408@jandecaluwe.com> Message-ID: <4DEE4267.3020304@gmail.com> On 07/06/11 17:02, Jan Decaluwe wrote: > I am seeing great improvements for MyHDL simulations > by using PyPy, and I have written a page about it: > > http://www.myhdl.org/doku.php/performance Hello Jan, this is really nice to hear :-) Did you try to run the benchmarks with a more recent version of PyPy? According to this chart, we are up to 30% faster than 1.5 on some benchmarks, so you might get even better results: http://speed.pypy.org/comparison/?exe=1%2B301%2C1%2BL&ben=1%2C27%2C2%2C25%2C3%2C4%2C5%2C22%2C6%2C7%2C8%2C23%2C24%2C9%2C10%2C11%2C12%2C13%2C14%2C15%2C16%2C17%2C18%2C19%2C20%2C26&env=1&hor=false&bas=1%2B301&chart=normal+bars ciao, Anto From santhosh.edukulla at gmail.com Tue Jun 7 17:53:52 2011 From: santhosh.edukulla at gmail.com (Santhosh Edukulla) Date: Tue, 7 Jun 2011 21:23:52 +0530 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: Actually, there is also a module by name "Array" import array t=array.array(I,size) Here t is a list and constrained more as an integer array. Check docs for more help. Thx. Santhosh On 6/7/11, Maciej Fijalkowski wrote: > On Mon, Jun 6, 2011 at 8:54 PM, Timothy Baldridge > wrote: >>> In case you don't already know: in Python (and RPython), lists are >>> implemented as >>> resizeable arrays, i.e. they look more like a C++ std::vector and not >>> a std::list. >>> >>> This said, IIRC RPython has special code for fixed-size lists. >>> I think you just have to avoid all resizing functions. >>> This means that the list must be created with statements like >>> ? ?array_of_ints = [0] * 32 >>> ? ?array_of_objects = [None] * 32 >> >> >> Excellent! Exactly what I was looking for. Thanks. >> >> Timothy >> > > There is also pypy.rlib.debug.make_sure_not_resized which would make > annotation explode if you by chance resize this list. > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From amauryfa at gmail.com Tue Jun 7 18:03:04 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Tue, 7 Jun 2011 18:03:04 +0200 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: 2011/6/7 Santhosh Edukulla : > Actually, there is also a module by name "Array" > > import array > t=array.array(I,size) But is this allowed in RPython code? -- Amaury Forgeot d'Arc From tbaldridge at gmail.com Tue Jun 7 18:37:15 2011 From: tbaldridge at gmail.com (Timothy Baldridge) Date: Tue, 7 Jun 2011 11:37:15 -0500 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: >> Actually, there is also a module by name "Array" >> >> import array >> t=array.array(I,size) And if you notice, it doesn't support arrays of objects, which was what my original problem was Timothy From jan at jandecaluwe.com Tue Jun 7 23:40:55 2011 From: jan at jandecaluwe.com (Jan Decaluwe) Date: Tue, 07 Jun 2011 23:40:55 +0200 Subject: [pypy-dev] Great results for MyHDL simulations In-Reply-To: <4DEE4267.3020304@gmail.com> References: <4DEE3D94.1020408@jandecaluwe.com> <4DEE4267.3020304@gmail.com> Message-ID: <4DEE9AE7.6020005@jandecaluwe.com> On 06/07/2011 05:23 PM, Antonio Cuni wrote: > On 07/06/11 17:02, Jan Decaluwe wrote: >> I am seeing great improvements for MyHDL simulations >> by using PyPy, and I have written a page about it: >> >> http://www.myhdl.org/doku.php/performance > > Hello Jan, > > this is really nice to hear :-) > > Did you try to run the benchmarks with a more recent version of PyPy? > According to this chart, we are up to 30% faster than 1.5 on some benchmarks, > so you might get even better results: Not yet, I like to leave some further excitement for later :-) From now on, I plan to track the evolution of my benchmarks with official PyPy releases. Jan -- Jan Decaluwe - Resources bvba - http://www.jandecaluwe.com Python as a HDL: http://www.myhdl.org VHDL development, the modern way: http://www.sigasi.com Analog design automation: http://www.mephisto-da.com World-class digital design: http://www.easics.com From notifications-noreply at bitbucket.org Wed Jun 8 00:23:44 2011 From: notifications-noreply at bitbucket.org (Bitbucket) Date: Tue, 07 Jun 2011 22:23:44 -0000 Subject: [pypy-dev] Notification: pypy-test Message-ID: <20110607222344.5999.23098@bitbucket01.managed.contegix.com> You have received a notification from xoraxax. Hi, I forked pypy. My fork is at https://bitbucket.org/xoraxax/pypy-test. -- Change your notification settings at https://bitbucket.org/account/notifications/ From notifications-noreply at bitbucket.org Wed Jun 8 00:26:52 2011 From: notifications-noreply at bitbucket.org (Bitbucket) Date: Tue, 07 Jun 2011 22:26:52 -0000 Subject: [pypy-dev] Notification: pypy-test2 Message-ID: <20110607222652.21688.12023@bitbucket02.managed.contegix.com> You have received a notification from xoraxax. Hi, I forked pypy. My fork is at https://bitbucket.org/xoraxax/pypy-test2. -- Change your notification settings at https://bitbucket.org/account/notifications/ From cfbolz at gmx.de Wed Jun 8 11:01:18 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Wed, 08 Jun 2011 11:01:18 +0200 Subject: [pypy-dev] Arrays in rpython In-Reply-To: References: Message-ID: <4DEF3A5E.3060107@gmx.de> On 06/07/2011 06:03 PM, Amaury Forgeot d'Arc wrote: > 2011/6/7 Santhosh Edukulla: >> Actually, there is also a module by name "Array" >> >> import array >> t=array.array(I,size) > > But is this allowed in RPython code? no, it is not. Cheers, Carl Friedrich From pa.basso at gmail.com Wed Jun 8 11:15:43 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Wed, 8 Jun 2011 11:15:43 +0200 Subject: [pypy-dev] Interactive fails on pypy Message-ID: Hi, I tryed to run the attached code with pypy but it fails (it works correctly with python 2.6). Can anybody tell me why? Thanks. P -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Interactive.py Type: application/octet-stream Size: 874 bytes Desc: not available URL: From pa.basso at gmail.com Wed Jun 8 11:27:48 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Wed, 8 Jun 2011 11:27:48 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Yep... Are there plans to implement the missing functions? Thanks, P 2011/6/8 Caleb Hattingh > On 8 June 2011 11:15, Paolo Basso wrote: > >> I tryed to run the attached code with pypy but it fails (it works >> correctly with python 2.6). >> Can anybody tell me why? Thanks. >> > > Error is this: > > 0 > Traceback (most recent call last): > File "app_main.py", line 53, in run_toplevel > File "test.py", line 25, in > it.ESC_test("TEST",1.0) > File "test.py", line 14, in ESC_test > if msvcrt.kbhit(): > AttributeError: 'module' object has no attribute 'kbhit' > > Maybe this message in lib_pypy/msvcrt.py line 7 has something to do with > it: > > # XXX incomplete: implemented only functions needed by subprocess.py > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From amauryfa at gmail.com Wed Jun 8 11:43:37 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Wed, 8 Jun 2011 11:43:37 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Hi, 2011/6/8 Paolo Basso : > if msvcrt.kbhit(): > AttributeError: 'module' object has no attribute 'kbhit' The msvcrt module is implemented in pure Python (in pypy_lib/msvcrt.py) This function seems easy to add. Patches are welcome! -- Amaury Forgeot d'Arc From eventh at gmail.com Wed Jun 8 17:16:27 2011 From: eventh at gmail.com (Even) Date: Wed, 8 Jun 2011 17:16:27 +0200 Subject: [pypy-dev] M.Sc. student introducing myself Message-ID: Hello, I'm a computer science student who intent to work on PyPy for my M.Sc. thesis. I have been a passionate Python programmer for seven years, and I additionally knows C, C++, Java, Oz, Lisp and Assembler. My academic interests are algorithms, compilers and evolutionary algorithms, and I have completed classes which covers these topics, which means I know what compilers does and why, and the optimizations they can do. I'm pleased that I have been allowed to focus my thesis on PyPy, but I don't yet know what the scope of my thesis will be. At the moment numpy support in PyPy looks interesting, and I have read the blogs/mails regarding this topic. I won't start any actual work on my thesis until the new year, but I hope to spend summer and autumn getting familiar with the project and contributing in any way I can. So far I have only read through the dev docs, and compiled PyPy from source (which took 4.6 hours with cpython and 3.7h with pypy on my laptop). I have been following this list since easter, and I can be found in #pypy on freenode. -- Even Wiik eventh at gmail.com From theller at ctypes.org Wed Jun 8 17:56:25 2011 From: theller at ctypes.org (Thomas Heller) Date: Wed, 08 Jun 2011 17:56:25 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Am 08.06.2011 11:27, schrieb Paolo Basso: > Yep... Are there plans to implement the missing functions? > Thanks, > P > > 2011/6/8 Caleb Hattingh > > > On 8 June 2011 11:15, Paolo Basso > wrote: > > I tryed to run the attached code with pypy but it fails (it > works correctly with python 2.6). > Can anybody tell me why? Thanks. > > > Error is this: > > 0 > Traceback (most recent call last): > File "app_main.py", line 53, in run_toplevel > File "test.py", line 25, in > it.ESC_test("TEST",1.0) > File "test.py", line 14, in ESC_test > if msvcrt.kbhit(): > AttributeError: 'module' object has no attribute 'kbhit' > Here is a patch against pypy 1.5, only slightly tested. Thomas diff -u c:\pypy\lib_pypy\msvcrt.py.orig c:\pypy\lib_pypy\msvcrt.py --- c:\pypy\lib_pypy\msvcrt.py.orig Wed Jun 08 15:53:02 2011 +++ c:\pypy\lib_pypy\msvcrt.py Wed Jun 08 15:53:02 2011 @@ -38,6 +38,43 @@ _locking.argtypes = [ctypes.c_int, ctypes.c_int, ctypes.c_int] _locking.restype = ctypes.c_int +kbhit = _c._kbhit +kbhit.argtypes = [] +kbhit.restype = ctypes.c_int + +getch = _c._getch +getch.argtypes = [] +getch.restype = ctypes.c_char + +getwch = _c._getwch +getwch.argtypes = [] +getwch.restype = ctypes.c_wchar + +getche = _c._getche +getche.argtypes = [] +getche.restype = ctypes.c_char + +getwche = _c._getwche +getwche.argtypes = [] +getwche.restype = ctypes.c_wchar + +putch = _c._putch +putch.restype = None +putch.argtypes = [ctypes.c_char] + +putwch = _c._putwch +putwch.restype = None +putwch.argtypes = [ctypes.c_char] + +ungetch = _c._ungetch +ungetch.restype = None +ungetch.argtypes = [ctypes.c_char] + +ungetwch = _c._ungetwch +ungetwch.restype = None +ungetwch.argtypes = [ctypes.c_wchar] + + @builtinify def locking(fd, mode, nbytes): '''lock or unlock a number of bytes in a file.''' From amauryfa at gmail.com Wed Jun 8 18:28:27 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Wed, 8 Jun 2011 18:28:27 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Hello Thomas! Still working with ctypes? ;-) 2011/6/8 Thomas Heller : > +putwch.argtypes = [ctypes.c_char] It's probably a c_wchar here... Otherwise it looks good! -- Amaury Forgeot d'Arc From theller at ctypes.org Wed Jun 8 19:04:29 2011 From: theller at ctypes.org (Thomas Heller) Date: Wed, 08 Jun 2011 19:04:29 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Am 08.06.2011 18:28, schrieb Amaury Forgeot d'Arc: > Hello Thomas! > > 2011/6/8 Thomas Heller: >> +putwch.argtypes = [ctypes.c_char] > > It's probably a c_wchar here... Sure, sorry for that. > > Still working with ctypes? ;-) > With ctypes - yes, of course. It's the foundation of some frameworks that we use in our company. On ctypes - no, not currently. With the exception of pypy - I was totally speechless yesterday when I tried out pypy for the first time, and found that most of our software works fine with pypy instead of cPython after disabling some too fancy stuff. Really great work of the pypy crowd! Well, comtypes doesn't work with pypy, but it uses some really deep things from ctypes that is implemented differently in pypy. Maybe I can fix that ;-) Thanks, Thomas From amauryfa at gmail.com Wed Jun 8 20:09:31 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Wed, 8 Jun 2011 20:09:31 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: 2011/6/8 Thomas Heller : > Well, comtypes doesn't work with pypy, but it uses some really deep > things from ctypes that is implemented differently in pypy. ?Maybe > I can fix that ;-) A few years ago I did implement many things for comtypes. I went to the point where it could generate Python wrappers for COM classes (I tried on some Excel class IIRC) Unfortunately, none of these features are tested or documented, so it's more than likely that the recent evolution of pypy's ctypes broke it. -- Amaury Forgeot d'Arc From chris0wj at gmail.com Wed Jun 8 21:34:02 2011 From: chris0wj at gmail.com (Chris Wj) Date: Wed, 8 Jun 2011 15:34:02 -0400 Subject: [pypy-dev] M.Sc. student introducing myself In-Reply-To: References: Message-ID: Awesome, keep us posted! On Wed, Jun 8, 2011 at 11:16 AM, Even wrote: > Hello, I'm a computer science student who intent to work on PyPy for > my M.Sc. thesis. > > I have been a passionate Python programmer for seven years, and I > additionally knows C, C++, Java, Oz, Lisp and Assembler. > > My academic interests are algorithms, compilers and evolutionary > algorithms, and I have completed classes which covers these topics, > which means I know what compilers does and why, and the optimizations > they can do. > > I'm pleased that I have been allowed to focus my thesis on PyPy, but I > don't yet know what the scope of my thesis will be. At the moment > numpy support in PyPy looks interesting, and I have read the > blogs/mails regarding this topic. > > I won't start any actual work on my thesis until the new year, but I > hope to spend summer and autumn getting familiar with the project and > contributing in any way I can. > So far I have only read through the dev docs, and compiled PyPy from > source (which took 4.6 hours with cpython and 3.7h with pypy on my > laptop). > > I have been following this list since easter, and I can be found in > #pypy on freenode. > -- > Even Wiik > eventh at gmail.com > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From arigo at tunes.org Thu Jun 9 07:11:15 2011 From: arigo at tunes.org (Armin Rigo) Date: Thu, 9 Jun 2011 07:11:15 +0200 Subject: [pypy-dev] setpgrp and getpgrp platform:errors In-Reply-To: <20110606133237.GB25550@pocketnix.org> References: <20110606133237.GB25550@pocketnix.org> Message-ID: Hi Da_Blitz, On Mon, Jun 6, 2011 at 3:32 PM, Da_Blitz wrote: > according to the man page on linux (debian) there are 2 calling > conventions, the System V and the BSD the os.setpgrp and os.getpgrp do > not appear to take arguments Yes, it's done for the BSD. According to my Linux manpage and to the CPython source code, getpgrp() and setpgrp() take arguments in their BSD version. So the original code that you changed works fine; I think it's just that people are getting scared about the errors printed in red. /me disables logging these errors -- they are not even warnings. A bient?t, Armin. From arigo at tunes.org Thu Jun 9 07:31:05 2011 From: arigo at tunes.org (Armin Rigo) Date: Thu, 9 Jun 2011 07:31:05 +0200 Subject: [pypy-dev] os.tmpnam warnings In-Reply-To: References: <20110606131913.GA25550@pocketnix.org> Message-ID: Hi Amaury, On Mon, Jun 6, 2011 at 3:55 PM, Amaury Forgeot d'Arc wrote: > Also ensure that the warning message points to the *caller* of os.tempnam(). > (I don't remember whether applevel functions are visible in the Traceback) They aren't; sys._getframe() ignores them completely. Armin. From fijall at gmail.com Thu Jun 9 09:00:16 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Thu, 9 Jun 2011 09:00:16 +0200 Subject: [pypy-dev] M.Sc. student introducing myself In-Reply-To: References: Message-ID: On Wed, Jun 8, 2011 at 5:16 PM, Even wrote: > Hello, I'm a computer science student who intent to work on PyPy for > my M.Sc. thesis. > > I have been a passionate Python programmer for seven years, and I > additionally knows C, C++, Java, Oz, Lisp and Assembler. > > My academic interests are algorithms, compilers and evolutionary > algorithms, and I have completed classes which covers these topics, > which means I know what compilers does and why, and the optimizations > they can do. > > I'm pleased that I have been allowed to focus my thesis on PyPy, but I > don't yet know what the scope of my thesis will be. At the moment > numpy support in PyPy looks interesting, and I have read the > blogs/mails regarding this topic. > > I won't start any actual work on my thesis until the new year, but I > hope to spend summer and autumn getting familiar with the project and > contributing in any way I can. > So far I have only read through the dev docs, and compiled PyPy from > source (which took 4.6 hours with cpython and 3.7h with pypy on my > laptop). > > I have been following this list since easter, and I can be found in > #pypy on freenode. Hi Geat to hear, welcome! We recently have put up a list of interesting mid-size project for PyPy, feel free to have a look: http://pypy.readthedocs.org/en/latest/project-ideas.html (except the website seems to be down for me at the moment) > -- > Even Wiik > eventh at gmail.com > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From theller at ctypes.org Thu Jun 9 15:54:24 2011 From: theller at ctypes.org (Thomas Heller) Date: Thu, 09 Jun 2011 15:54:24 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. Message-ID: Hello, here are patches against pypy 1.5 so that comtypes works with it (I hope it is acceptable to post patches here; I have no time to learn mercurial at the moment). Thanks, Thomas -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: comtypes-patches.diff URL: From amauryfa at gmail.com Thu Jun 9 15:58:18 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Thu, 9 Jun 2011 15:58:18 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: References: Message-ID: 2011/6/9 Thomas Heller : > here are patches against pypy 1.5 so that comtypes works with it > (I hope it is acceptable to post patches here; I have no time to > learn mercurial at the moment). Please use our nice bug tracker so that the patch does not get lost https://bugs.pypy.org/ -- Amaury Forgeot d'Arc From lac at openend.se Thu Jun 9 16:04:00 2011 From: lac at openend.se (Laura Creighton) Date: Thu, 09 Jun 2011 16:04:00 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: Message from Thomas Heller of "Thu, 09 Jun 2011 15:54:24 +0200." References: Message-ID: <201106091404.p59E408C016690@theraft.openend.se> Thank you for the patches, which are most appreciated. If you have any more of them, can you send them to the bug tracker https://bugs.pypy.org/ ? Things posted to the mailing list sometimes get forgotten. Thank you, and thank you for the patches, Laura From pa.basso at gmail.com Thu Jun 9 16:11:53 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Thu, 9 Jun 2011 16:11:53 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Thanks for the patch, is there a nightly build of pypy 1.5 somewhere or I have to build it myself from updated source? In this second case is there a "tutorial" on how to build the source? Thanks, P P.S. I look forward to attend the pypy sprint in Genoa at the end of this month, at least I can learn a little bit more there :) 2011/6/8 Amaury Forgeot d'Arc > 2011/6/8 Thomas Heller : > > Well, comtypes doesn't work with pypy, but it uses some really deep > > things from ctypes that is implemented differently in pypy. Maybe > > I can fix that ;-) > > A few years ago I did implement many things for comtypes. > I went to the point where it could generate Python wrappers for COM classes > (I tried on some Excel class IIRC) > > Unfortunately, none of these features are tested or documented, > so it's more than likely that the recent evolution of pypy's ctypes broke > it. > > -- > Amaury Forgeot d'Arc > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > -------------- next part -------------- An HTML attachment was scrubbed... URL: From amauryfa at gmail.com Thu Jun 9 16:21:29 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Thu, 9 Jun 2011 16:21:29 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Hi, 2011/6/9 Paolo Basso : > Thanks for the patch, is there a nightly build of pypy 1.5 somewhere or I > have to build it myself from updated source? In this second case is there a > "tutorial" on how to build the source? Unfortunately we cannot build win32 versions at the moment. Translating pypy yourself is not too difficult: http://doc.pypy.org/en/latest/getting-started-python.html?translating-the-pypy-python-interpreter http://doc.pypy.org/en/latest/windows.html OTOH, the previous patch only touches the pure Python standard library, and does not need another compilation! Just add the lines to your existing version of lib_pypy\msvcrt.py -- Amaury Forgeot d'Arc From anto.cuni at gmail.com Thu Jun 9 16:32:15 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Thu, 09 Jun 2011 16:32:15 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: References: Message-ID: <4DF0D96F.3030604@gmail.com> On 09/06/11 15:54, Thomas Heller wrote: > Hello, > > here are patches against pypy 1.5 so that comtypes works with it > (I hope it is acceptable to post patches here; I have no time to Hello Thomas, thank you for the patch. However, I fear that it might not work properly with the trunk version, because of the heavy refactoring that happened in the jitypes2 branch. Could you please try it with a newwer version of pypy? ciao, Anto From theller at ctypes.org Thu Jun 9 16:44:30 2011 From: theller at ctypes.org (Thomas Heller) Date: Thu, 09 Jun 2011 16:44:30 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: <201106091404.p59E408C016690@theraft.openend.se> References: <201106091404.p59E408C016690@theraft.openend.se> Message-ID: Am 09.06.2011 16:04, schrieb Laura Creighton: > Thank you for the patches, which are most appreciated. If you have > any more of them, can you send them to the bug tracker > https://bugs.pypy.org/ ? Things posted to the mailing list sometimes > get forgotten. Sure: https://bugs.pypy.org/issue743 Thomas From theller at ctypes.org Thu Jun 9 16:45:35 2011 From: theller at ctypes.org (Thomas Heller) Date: Thu, 09 Jun 2011 16:45:35 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: <4DF0D96F.3030604@gmail.com> References: <4DF0D96F.3030604@gmail.com> Message-ID: Am 09.06.2011 16:32, schrieb Antonio Cuni: > On 09/06/11 15:54, Thomas Heller wrote: >> Hello, >> >> here are patches against pypy 1.5 so that comtypes works with it >> (I hope it is acceptable to post patches here; I have no time to > > Hello Thomas, > > thank you for the patch. However, I fear that it might not work properly with > the trunk version, because of the heavy refactoring that happened in the > jitypes2 branch. Could you please try it with a newwer version of pypy? I have never build pypy myself, so I would prefer to wait for some windows binaries before trying this. Thomas From theller at ctypes.org Thu Jun 9 16:52:55 2011 From: theller at ctypes.org (Thomas Heller) Date: Thu, 09 Jun 2011 16:52:55 +0200 Subject: [pypy-dev] Understanding pypy Message-ID: I have read the blog posts about 'writing an interpreter with pypy' with large interest, but not tried the code myself. It seems that if I restrict myself to the RPython subset, then I can write programs with python syntax, and compile them to standalone executables, optionally containing a jit compiler. Is that correct? Are there any runtime dependencies? Thanks, Thomas From cfbolz at gmx.de Thu Jun 9 18:51:42 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Thu, 09 Jun 2011 18:51:42 +0200 Subject: [pypy-dev] Understanding pypy In-Reply-To: References: Message-ID: <4DF0FA1E.8060609@gmx.de> On 06/09/2011 04:52 PM, Thomas Heller wrote: > I have read the blog posts about 'writing an interpreter with pypy' with > large interest, but not tried the code myself. > > It seems that if I restrict myself to the RPython subset, then I can > write programs with python syntax, and compile them to standalone > executables, optionally containing a jit compiler. Is that correct? Yes. > Are there any runtime dependencies? Not necessarily. Of course you need to have the libraries around that you use in your RPython code, e.g. bz2 for the RPython bz2 support. Carl Friedrich From pa.basso at gmail.com Fri Jun 10 01:00:01 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Fri, 10 Jun 2011 01:00:01 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Thanks for your help, I tryed to add the patch but still have the same problem... Attached are the msvcrt module as I modified it and the code I tryed to run. P 2011/6/8 Amaury Forgeot d'Arc > Hi, > > 2011/6/8 Paolo Basso : > > if msvcrt.kbhit(): > > AttributeError: 'module' object has no attribute 'kbhit' > > The msvcrt module is implemented in pure Python (in pypy_lib/msvcrt.py) > This function seems easy to add. > Patches are welcome! > > -- > Amaury Forgeot d'Arc > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: msvcrt.py Type: application/octet-stream Size: 1957 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Interactive(fails in pypy).py Type: application/octet-stream Size: 874 bytes Desc: not available URL: From alexandre.fayolle at logilab.fr Fri Jun 10 09:17:08 2011 From: alexandre.fayolle at logilab.fr (Alexandre Fayolle) Date: Fri, 10 Jun 2011 09:17:08 +0200 Subject: [pypy-dev] [Pyro] Anyone with Pypy experience? In-Reply-To: <4DF136ED.80103@razorvine.net> References: <4DF136ED.80103@razorvine.net> Message-ID: <201106100917.08918.alexandre.fayolle@logilab.fr> On Thursday 09 June 2011 23:11:09 Irmen de Jong wrote: > Hi, > > I tried to run Pyro4's test suite in Pypy. > It didn't work very well (better on Linux than on Windows, but still). > > Is there anyone here that has experience with Pypy? I think you should ask on the pypy-dev mailing list (CC'ed) or on the #pypy irc channel. You will certainly get answers for the kind folks over there. -- Alexandre Fayolle LOGILAB, Paris (France) Formations Python, CubicWeb, Debian : http://www.logilab.fr/formations D?veloppement logiciel sur mesure : http://www.logilab.fr/services Informatique scientifique: http://www.logilab.fr/science From marius at pov.lt Fri Jun 10 09:49:50 2011 From: marius at pov.lt (Marius Gedminas) Date: Fri, 10 Jun 2011 10:49:50 +0300 Subject: [pypy-dev] Performance challenge of mutating class variables In-Reply-To: <4DD26684.5010303@gmail.com> References: <4DCEC6DD.8050601@gmx.de> <4DD26684.5010303@gmail.com> Message-ID: <20110610074950.GA1426@fridge.pov.lt> On Tue, May 17, 2011 at 02:13:56PM +0200, Antonio Cuni wrote: > On 17/05/11 14:11, Maciej Fijalkowski wrote: > > Which reminds me that we should *really* start running python 2.7 as a > > baseline python > > for this, it should be "enough" to upgrade ubuntu on tannit, which comes with > python2.7. Unless we decide that we want to stay with the LTS release. If you want to stay with an LTS, you can get Python 2.7 from this "quasi-official" PPA: https://launchpad.net/~pythoneers/+archive/lts There's also another PPA, called "deadsnakes", with more Python versions available: https://launchpad.net/~fkrull/+archive/deadsnakes Marius Gedminas -- Mosher's Law of Software Engineering: Don't worry if it doesn't work right. If everything did, you'd be out of a job. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 190 bytes Desc: Digital signature URL: From cfbolz at gmx.de Fri Jun 10 14:11:01 2011 From: cfbolz at gmx.de (Carl Friedrich Bolz) Date: Fri, 10 Jun 2011 14:11:01 +0200 Subject: [pypy-dev] M.Sc. student introducing myself In-Reply-To: References: Message-ID: <4DF209D5.1000205@gmx.de> Hi Even, On 06/08/2011 05:16 PM, Even wrote: > Hello, I'm a computer science student who intent to work on PyPy for > my M.Sc. thesis. > > I have been a passionate Python programmer for seven years, and I > additionally knows C, C++, Java, Oz, Lisp and Assembler. > > My academic interests are algorithms, compilers and evolutionary > algorithms, and I have completed classes which covers these topics, > which means I know what compilers does and why, and the optimizations > they can do. > > I'm pleased that I have been allowed to focus my thesis on PyPy, but I > don't yet know what the scope of my thesis will be. At the moment > numpy support in PyPy looks interesting, and I have read the > blogs/mails regarding this topic. Welcome to the project! It would be great if your thesis topic would be numpy support, that's a feature we really want to have. If you have any questions or need any help, please tell us. > I won't start any actual work on my thesis until the new year, but I > hope to spend summer and autumn getting familiar with the project and > contributing in any way I can. > So far I have only read through the dev docs, and compiled PyPy from > source (which took 4.6 hours with cpython and 3.7h with pypy on my > laptop). Note that in practice we rarely do translations on our own machines, but use our build servers. If you need access to one of these machines, please ask on the IRC channel. Cheers, Carl Friedrich From pa.basso at gmail.com Fri Jun 10 18:00:07 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Fri, 10 Jun 2011 18:00:07 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: Ok, I solved the problem, the patch works great (thanks Thomas) but the problem was a conflict with the python 2.6 library from which the readline.py module was loaded. Is it normal that it happens or it a problem of my configuration? I solved by using the sys.path.insert(0, "path of the pypy_lib in my drive") but I wonder if there is a better way so that I don't need to add that line every time. Thanks, P 2011/6/10 Paolo Basso > Thanks for your help, I tryed to add the patch but still have the same > problem... > > Attached are the msvcrt module as I modified it and the code I tryed to > run. > > P > > 2011/6/8 Amaury Forgeot d'Arc > >> Hi, >> >> 2011/6/8 Paolo Basso : >> > if msvcrt.kbhit(): >> > AttributeError: 'module' object has no attribute 'kbhit' >> >> The msvcrt module is implemented in pure Python (in pypy_lib/msvcrt.py) >> This function seems easy to add. >> Patches are welcome! >> >> -- >> Amaury Forgeot d'Arc >> > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From anto.cuni at gmail.com Fri Jun 10 18:15:59 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Fri, 10 Jun 2011 18:15:59 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: Message-ID: <4DF2433F.3020607@gmail.com> On 10/06/11 18:00, Paolo Basso wrote: > Ok, I solved the problem, the patch works great (thanks Thomas) but the > problem was a conflict with the python 2.6 library from which the readline.py > module was loaded. Is it normal that it happens or it a problem of my > configuration? uhm, it seems that something is wrong. PyPy should not try to load the modules of cpython's stdlib (it doesn't even know if/where it is installed!) What is your PYTHONPATH? And your sys.path? ciao, Anto From anto.cuni at gmail.com Fri Jun 10 18:17:53 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Fri, 10 Jun 2011 18:17:53 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: References: <4DF0D96F.3030604@gmail.com> Message-ID: <4DF243B1.2000706@gmail.com> On 09/06/11 16:45, Thomas Heller wrote: > I have never build pypy myself, so I would prefer to wait for some > windows binaries before trying this. I have (hopefully :-)) fixed the windows buildslave, so we should start having nigthly builds again. There is a translation going on right now: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/79 unless the translation is broken, it should be ready in ~1hr. ciao, Anto From pa.basso at gmail.com Fri Jun 10 18:32:00 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Fri, 10 Jun 2011 18:32:00 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: <4DF2433F.3020607@gmail.com> References: <4DF2433F.3020607@gmail.com> Message-ID: sys.path da pypy: ['', 'C:\\Python26\\DLLs', 'C:\\Python26\\Lib', 'C:\\Python26\\Lib\\site-package s', 'C:\\PyPy\\pypy-1.5.0a0-win32', 'C:\\PyPy\\pypy-1.5.0a0-win32\\lib_pypy', 'C :\\PyPy\\pypy-1.5.0a0-win32\\lib-python\\modified-2.7', 'C:\\PyPy\\pypy-1.5.0a0- win32\\lib-python\\2.7', 'C:\\PyPy\\pypy-1.5.0a0-win32\\lib-python\\modified-2.7 \\lib-tk' , 'C:\\PyPy\\pypy-1.5.0a0-win32\\lib-python\\2.7\\lib-tk'] PYTHONPATH C:\Python26\DLLs;C:\Python26\Lib;C:\Python26\Lib\site-packages; 2011/6/10 Antonio Cuni > On 10/06/11 18:00, Paolo Basso wrote: > > Ok, I solved the problem, the patch works great (thanks Thomas) but the > > problem was a conflict with the python 2.6 library from which the > readline.py > > module was loaded. Is it normal that it happens or it a problem of my > > configuration? > > uhm, it seems that something is wrong. PyPy should not try to load the > modules of cpython's stdlib (it doesn't even know if/where it is > installed!) > > What is your PYTHONPATH? And your sys.path? > > ciao, > Anto > -------------- next part -------------- An HTML attachment was scrubbed... URL: From anto.cuni at gmail.com Fri Jun 10 19:02:51 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Fri, 10 Jun 2011 19:02:51 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: References: <4DF2433F.3020607@gmail.com> Message-ID: <4DF24E3B.7060906@gmail.com> On 10/06/11 18:32, Paolo Basso wrote: > PYTHONPATH > > C:\Python26\DLLs;C:\Python26\Lib;C:\Python26\Lib\site-packages; this is the problem. You are explicitly telling pypy to import modules from cpython's stdlib. Why do you have such a pythonpath, btw? It's not needed, CPython can determine its own automatically. ciao, Anto From pa.basso at gmail.com Fri Jun 10 20:52:21 2011 From: pa.basso at gmail.com (Paolo Basso) Date: Fri, 10 Jun 2011 20:52:21 +0200 Subject: [pypy-dev] Interactive fails on pypy In-Reply-To: <4DF24E3B.7060906@gmail.com> References: <4DF2433F.3020607@gmail.com> <4DF24E3B.7060906@gmail.com> Message-ID: > Why do you have such a pythonpath, btw? I have no idea... However I removed the PYTHONPATH and now everything works fine, thanks. P 2011/6/10 Antonio Cuni > On 10/06/11 18:32, Paolo Basso wrote: > > PYTHONPATH > > > > C:\Python26\DLLs;C:\Python26\Lib;C:\Python26\Lib\site-packages; > > this is the problem. You are explicitly telling pypy to import modules from > cpython's stdlib. > > Why do you have such a pythonpath, btw? It's not needed, CPython can > determine > its own automatically. > > ciao, > Anto > -------------- next part -------------- An HTML attachment was scrubbed... URL: From anto.cuni at gmail.com Fri Jun 10 21:30:04 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Fri, 10 Jun 2011 21:30:04 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: <4DF243B1.2000706@gmail.com> References: <4DF0D96F.3030604@gmail.com> <4DF243B1.2000706@gmail.com> Message-ID: <4DF270BC.6060507@gmail.com> On 10/06/11 18:17, Antonio Cuni wrote: > There is a translation going on right now: > http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/79 > > unless the translation is broken, it should be ready in ~1hr. ok, the translation is broken :-( (in case anyone wants to fix it. Hint, hint :-)) From arigo at tunes.org Sun Jun 12 12:50:46 2011 From: arigo at tunes.org (Armin Rigo) Date: Sun, 12 Jun 2011 12:50:46 +0200 Subject: [pypy-dev] Win32: track_and_esp.s Message-ID: Hi Amaury, Sorry, track_and_esp.s is most probably broken on Windows due to 017e187b2716. This revision contains, among several things, a small simplification of the code because it was (probably) not possible any more to see "and esp, xxx" from gcc. So, of course, we need to partially and carefully revert this simplification... A bient?t, Armin. From fijall at gmail.com Mon Jun 13 09:19:12 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Mon, 13 Jun 2011 09:19:12 +0200 Subject: [pypy-dev] [pypy-commit] pypy default: Ah sorry, re 69cadfd7c8e1. Found how to reproduce (translate -Ojit, In-Reply-To: <20110611155323.E4C95820AE@wyvern.cs.uni-duesseldorf.de> References: <20110611155323.E4C95820AE@wyvern.cs.uni-duesseldorf.de> Message-ID: On Sat, Jun 11, 2011 at 5:53 PM, arigo wrote: > Author: Armin Rigo > Branch: > Changeset: r44886:0e02f7346cf0 > Date: 2011-06-11 17:55 +0200 > http://bitbucket.org/pypy/pypy/changeset/0e02f7346cf0/ > > Log: ? ?Ah sorry, re 69cadfd7c8e1. Found how to reproduce (translate -Ojit, > ? ? ? ?of course). Add a "correct" fix. SizeDescr also uses tid. Maybe we should have BaseLLDescr simply? > > diff --git a/pypy/jit/backend/llsupport/descr.py b/pypy/jit/backend/llsupport/descr.py > --- a/pypy/jit/backend/llsupport/descr.py > +++ b/pypy/jit/backend/llsupport/descr.py > @@ -1,5 +1,6 @@ > ?import py > ?from pypy.rpython.lltypesystem import lltype, rffi, llmemory, rclass > +from pypy.rpython.lltypesystem.lloperation import llop > ?from pypy.jit.backend.llsupport import symbolic, support > ?from pypy.jit.metainterp.history import AbstractDescr, getkind, BoxInt, BoxPtr > ?from pypy.jit.metainterp.history import BasicFailDescr, LoopToken, BoxFloat > @@ -149,6 +150,7 @@ > > ?class BaseArrayDescr(AbstractDescr): > ? ? _clsname = '' > + ? ?tid = llop.combine_ushort(lltype.Signed, 0, 0) > > ? ? def get_base_size(self, translate_support_code): > ? ? ? ? basesize, _, _ = symbolic.get_array_token(_A, translate_support_code) > _______________________________________________ > pypy-commit mailing list > pypy-commit at python.org > http://mail.python.org/mailman/listinfo/pypy-commit > From alexandre.fayolle at logilab.fr Mon Jun 13 11:21:32 2011 From: alexandre.fayolle at logilab.fr (Alexandre Fayolle) Date: Mon, 13 Jun 2011 11:21:32 +0200 Subject: [pypy-dev] [Pyro] Anyone with Pypy experience? In-Reply-To: <4DF262A1.9020708@razorvine.net> References: <4DF136ED.80103@razorvine.net> <201106100917.08918.alexandre.fayolle@logilab.fr> <4DF262A1.9020708@razorvine.net> Message-ID: <201106131121.33302.alexandre.fayolle@logilab.fr> On Friday 10 June 2011 20:29:53 Irmen de Jong wrote: > On 10-6-2011 9:17, Alexandre Fayolle wrote: > > On Thursday 09 June 2011 23:11:09 Irmen de Jong wrote: > >> Hi, > >> > >> I tried to run Pyro4's test suite in Pypy. > >> It didn't work very well (better on Linux than on Windows, but still). > >> > >> Is there anyone here that has experience with Pypy? > > > > I think you should ask on the pypy-dev mailing list (CC'ed) or on the > > #pypy irc channel. You will certainly get answers for the kind folks > > over there. > > Thanks Alexandre, I will probably do that once I've narrowed down the > things that didn't work. I've found and fixed 1 problem so far; Pypy > doesn't seem to like unicode keyword arguments; "TypeError: keyword cannot > be encoded to ascii". This used to be an issue with python < 2.6.? (http://bugs.python.org/issue2646). It may not have been changed in recent versions of pypy. -- Alexandre Fayolle LOGILAB, Paris (France) Formations Python, CubicWeb, Debian : http://www.logilab.fr/formations D?veloppement logiciel sur mesure : http://www.logilab.fr/services Informatique scientifique: http://www.logilab.fr/science From amauryfa at gmail.com Mon Jun 13 12:23:00 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Mon, 13 Jun 2011 12:23:00 +0200 Subject: [pypy-dev] [Pyro] Anyone with Pypy experience? In-Reply-To: <201106131121.33302.alexandre.fayolle@logilab.fr> References: <4DF136ED.80103@razorvine.net> <201106100917.08918.alexandre.fayolle@logilab.fr> <4DF262A1.9020708@razorvine.net> <201106131121.33302.alexandre.fayolle@logilab.fr> Message-ID: Hi, 2011/6/13 Alexandre Fayolle : >> Thanks Alexandre, I will probably do that once I've narrowed down the >> things that didn't work. I've found and fixed 1 problem so far; Pypy >> doesn't seem to like unicode keyword arguments; "TypeError: keyword cannot >> be encoded to ascii". > > This used to be an issue with python < 2.6.? > (http://bugs.python.org/issue2646). It may not have been changed in recent > versions of pypy. Please file a ticket on https://bugs.pypy.org It looks that a CPython test is not precise enough... -- Amaury Forgeot d'Arc From andreas at unstable.nl Mon Jun 13 14:05:43 2011 From: andreas at unstable.nl (Andreas) Date: Mon, 13 Jun 2011 14:05:43 +0200 Subject: [pypy-dev] static binaries please ... Message-ID: <20110613120543.GA5579@unstable.nl> I have just spent an hour compiling various libraries to get the pypy binaries to work (64bit, RHEL3). Probable because the distribution is old (I don't administer it) the shared libraries that pypy requires were not available. Compiling them by hand is a hassle (have to find options to enable shared objects etc.). Please distribute only _static_ binaries, the gains of shared libriares are completely insignificant compared to the hassle of when it doesn't work. Thanks ... -- -- Andreas [ http://unstable.nl | gopher://unstable.nl ] From santagada at gmail.com Mon Jun 13 16:44:07 2011 From: santagada at gmail.com (Leonardo Santagada) Date: Mon, 13 Jun 2011 11:44:07 -0300 Subject: [pypy-dev] static binaries please ... In-Reply-To: <20110613120543.GA5579@unstable.nl> References: <20110613120543.GA5579@unstable.nl> Message-ID: On Mon, Jun 13, 2011 at 9:05 AM, Andreas wrote: > I have just spent an hour compiling various libraries to get the pypy > binaries to work (64bit, RHEL3). Probable because the distribution is old (I > don't administer it) the shared libraries that pypy requires were not > available. Compiling them by hand is a hassle (have to find options to enable > shared objects etc.). > > Please distribute only _static_ binaries, the gains of shared libriares are > completely insignificant compared to the hassle of when it doesn't work. +1 -- Leonardo Santagada From dmalcolm at redhat.com Mon Jun 13 21:03:10 2011 From: dmalcolm at redhat.com (David Malcolm) Date: Mon, 13 Jun 2011 15:03:10 -0400 Subject: [pypy-dev] static binaries please ... In-Reply-To: <20110613120543.GA5579@unstable.nl> References: <20110613120543.GA5579@unstable.nl> Message-ID: <1307991790.16924.5533.camel@surprise> On Mon, 2011-06-13 at 14:05 +0200, Andreas wrote: > I have just spent an hour compiling various libraries to get the pypy > binaries to work (64bit, RHEL3). Probable because the distribution is old (I > don't administer it) the shared libraries that pypy requires were not > available. Compiling them by hand is a hassle (have to find options to enable > shared objects etc.). > > Please distribute only _static_ binaries, the gains of shared libriares are > completely insignificant compared to the hassle of when it doesn't work. I know this is a slightly different approach to the one in your mail, but FWIW, I've packaged pypy (1.5) in rpm form for Fedora, and it may be possible to rebuild the pypy src.rpm on RHEL3. Caveat: I've not tried it, though I've had some success with this src.rpm on RHEL5 and RHEL6 (can take two hours or more to build, and you may want to disable the %check section since this kills the build if certain test failures happen). The specfile can be seen here: http://pkgs.fedoraproject.org/gitweb/?p=pypy.git;a=tree and the src.rpm here: http://kojipkgs.fedoraproject.org/packages/pypy/1.5/1.fc15/src/pypy-1.5-1.fc15.src.rpm in case you want to try it. Hope this is helpful Dave From andreas at unstable.nl Mon Jun 13 21:16:15 2011 From: andreas at unstable.nl (Andreas) Date: Mon, 13 Jun 2011 21:16:15 +0200 Subject: [pypy-dev] static binaries please ... In-Reply-To: <1307991790.16924.5533.camel@surprise> References: <20110613120543.GA5579@unstable.nl> <1307991790.16924.5533.camel@surprise> Message-ID: <20110613191614.GC5579@unstable.nl> On Mon, Jun 13, 2011 at 03:03:10PM -0400, David Malcolm wrote: > I know this is a slightly different approach to the one in your mail, > but FWIW, I've packaged pypy (1.5) in rpm form for Fedora, and it may be > possible to rebuild the pypy src.rpm on RHEL3. Caveat: I've not tried > it, though I've had some success with this src.rpm on RHEL5 and RHEL6 > (can take two hours or more to build, and you may want to disable the > %check > section since this kills the build if certain test failures happen). Thanks, but I reckon I can't do stuff with rpm as I don't have root rights on the box. Either way it's working OK now after manually compiling the dependencies (you do need to enable shared libraries for some of them). A minor annoyance is that it gives warnings about the OpenSSL version -- but how am I supposed to know which of the 20 openssl 0.9.8 versions was used to compile pypy?! Something completely different, this trick works fine in CPython: import sys sys.stdout = codecs.getwriter('utf8')(sys.stdout) print "[some unicode]" For pypy needed to replace the prints with sys.stdout.write calls. -- -- Andreas [ http://unstable.nl | gopher://unstable.nl ] From arigo at tunes.org Mon Jun 13 21:24:13 2011 From: arigo at tunes.org (Armin Rigo) Date: Mon, 13 Jun 2011 21:24:13 +0200 Subject: [pypy-dev] static binaries please ... In-Reply-To: <20110613191614.GC5579@unstable.nl> References: <20110613120543.GA5579@unstable.nl> <1307991790.16924.5533.camel@surprise> <20110613191614.GC5579@unstable.nl> Message-ID: Hi Andreas, On Mon, Jun 13, 2011 at 9:16 PM, Andreas wrote: > Something completely different, this trick works fine in CPython: > > ? ?import sys > ? ?sys.stdout = codecs.getwriter('utf8')(sys.stdout) > ? ? ? ?print "[some unicode]" Already fixed between 1.5 and trunk (unless you have another example that crashes than the one I just tried). Thanks anyway :-) A bient?t, Armin. From arigo at tunes.org Tue Jun 14 16:00:44 2011 From: arigo at tunes.org (Armin Rigo) Date: Tue, 14 Jun 2011 16:00:44 +0200 Subject: [pypy-dev] [Pyro] Anyone with Pypy experience? In-Reply-To: References: <4DF136ED.80103@razorvine.net> <201106100917.08918.alexandre.fayolle@logilab.fr> <4DF262A1.9020708@razorvine.net> <201106131121.33302.alexandre.fayolle@logilab.fr> Message-ID: Hi, On Mon, Jun 13, 2011 at 12:23 PM, Amaury Forgeot d'Arc wrote: > Please file a ticket on https://bugs.pypy.org > It looks that a CPython test is not precise enough... Fixed by benjamin (and reviewed by me). Armin From van.lindberg at gmail.com Tue Jun 14 17:46:25 2011 From: van.lindberg at gmail.com (VanL) Date: Tue, 14 Jun 2011 10:46:25 -0500 Subject: [pypy-dev] libpypy.so? Message-ID: <4DF78251.1010300@gmail.com> Is it possible to create an embeddable libpypy.so/dll? Are there any visible entry points into code created by the pypy toolchain? Thanks, Vna From benjamin at python.org Tue Jun 14 17:49:36 2011 From: benjamin at python.org (Benjamin Peterson) Date: Tue, 14 Jun 2011 10:49:36 -0500 Subject: [pypy-dev] libpypy.so? In-Reply-To: <4DF78251.1010300@gmail.com> References: <4DF78251.1010300@gmail.com> Message-ID: 2011/6/14 VanL : > Is it possible to create an embeddable libpypy.so/dll? Are there any visible > entry points into code created by the pypy toolchain? No. -- Regards, Benjamin From van.lindberg at gmail.com Tue Jun 14 20:34:09 2011 From: van.lindberg at gmail.com (VanL) Date: Tue, 14 Jun 2011 13:34:09 -0500 Subject: [pypy-dev] libpypy.so? In-Reply-To: References: <4DF78251.1010300@gmail.com> Message-ID: <4DF7A9A1.1060904@gmail.com> Thanks for the reply. On 6/14/2011 10:49 AM, Benjamin Peterson wrote: > 2011/6/14 VanL: >> Is it possible to create an embeddable libpypy.so/dll? Are there any visible >> entry points into code created by the pypy toolchain? > No In examining this, I noted the translation option -- shared (http://codespeak.net/pypy/dist/pypy/doc/config/translation.shared.html). Shame on me for not seeing this earlier, but what is the difference between this and libpypy.so? Is this not working? From benjamin at python.org Tue Jun 14 20:35:28 2011 From: benjamin at python.org (Benjamin Peterson) Date: Tue, 14 Jun 2011 13:35:28 -0500 Subject: [pypy-dev] libpypy.so? In-Reply-To: <4DF7A9A1.1060904@gmail.com> References: <4DF78251.1010300@gmail.com> <4DF7A9A1.1060904@gmail.com> Message-ID: 2011/6/14 VanL : > Thanks for the reply. > > On 6/14/2011 10:49 AM, Benjamin Peterson wrote: >> >> 2011/6/14 VanL: >>> >>> Is it possible to create an embeddable libpypy.so/dll? Are there any >>> visible >>> entry points into code created by the pypy toolchain? >> >> No > > In examining this, I noted the translation option -- shared > (http://codespeak.net/pypy/dist/pypy/doc/config/translation.shared.html). > Shame on me for not seeing this earlier, but what is the difference between > this and libpypy.so? Is this not working? Sorry, I should have clarified. You should be able to build shared, but there is no public C interface. -- Regards, Benjamin From amauryfa at gmail.com Tue Jun 14 20:48:58 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Tue, 14 Jun 2011 20:48:58 +0200 Subject: [pypy-dev] libpypy.so? In-Reply-To: References: <4DF78251.1010300@gmail.com> <4DF7A9A1.1060904@gmail.com> Message-ID: 2011/6/14 Benjamin Peterson : >> In examining this, I noted the translation option -- shared >> (http://codespeak.net/pypy/dist/pypy/doc/config/translation.shared.html). >> Shame on me for not seeing this earlier, but what is the difference between >> this and libpypy.so? Is this not working? > > Sorry, I should have clarified. You should be able to build shared, > but there is no public C interface. Really? on Windows at least, the shared library exposes all the cpyext API... but Py_Initialize() is missing: pypy is not yet ready for embedding! -- Amaury Forgeot d'Arc From benjamin at python.org Tue Jun 14 21:04:51 2011 From: benjamin at python.org (Benjamin Peterson) Date: Tue, 14 Jun 2011 14:04:51 -0500 Subject: [pypy-dev] libpypy.so? In-Reply-To: References: <4DF78251.1010300@gmail.com> <4DF7A9A1.1060904@gmail.com> Message-ID: 2011/6/14 Amaury Forgeot d'Arc : > 2011/6/14 Benjamin Peterson : >>> In examining this, I noted the translation option -- shared >>> (http://codespeak.net/pypy/dist/pypy/doc/config/translation.shared.html). >>> Shame on me for not seeing this earlier, but what is the difference between >>> this and libpypy.so? Is this not working? >> >> Sorry, I should have clarified. You should be able to build shared, >> but there is no public C interface. > > Really? on Windows at least, the shared library exposes all the cpyext API... > but Py_Initialize() is missing: pypy is not yet ready for embedding! I should have mentioned cpyext. I thought he was referring solely to a pure-PyPy interface, though. -- Regards, Benjamin From arigo at tunes.org Wed Jun 15 13:22:11 2011 From: arigo at tunes.org (Armin Rigo) Date: Wed, 15 Jun 2011 13:22:11 +0200 Subject: [pypy-dev] Win32: track_and_esp.s In-Reply-To: References: Message-ID: Hi, On Sun, Jun 12, 2011 at 12:50 PM, Armin Rigo wrote: > more to see "and esp, xxx" from gcc. ?So, of course, we need to > partially and carefully revert this simplification... Done, I think. Please tell me if the Windows build still cannot use asmgcc. A bient?t, Armin. From amauryfa at gmail.com Wed Jun 15 13:40:10 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Wed, 15 Jun 2011 13:40:10 +0200 Subject: [pypy-dev] Win32: track_and_esp.s In-Reply-To: References: Message-ID: Hi, 2011/6/15 Armin Rigo : > Done, I think. ?Please tell me if the Windows build still cannot use asmgcc. Yes, I've seen changeset 3549050ab199. But I tried the same change last week-end, and test_asmgcroot crashed with a segfault. Anyway the win32 buildbot now uses --asmgcroot=framework; and at the moment it is stuck because some process is still running... -- Amaury Forgeot d'Arc From emailgregn at googlemail.com Wed Jun 15 14:29:46 2011 From: emailgregn at googlemail.com (emailgregn at googlemail.com) Date: Wed, 15 Jun 2011 12:29:46 +0000 Subject: [pypy-dev] Ootypes Message-ID: <1706411655-1308140988-cardhu_decombobulator_blackberry.rim.net-756381689-@b4.c8.bise7.blackberry> Pretty please, can ootypes get some love in the upcoming sprint? I've gotten through the cli-jit thesis and have the brainfeck tutorial compiling to C, but neither of the CLI or JVM backends work on those tutorial examples. Setting --opt=0 gets further but not all the way there. I'd like to try a parrot PIR backend but without ootypes I'm in over my head. Thank you Sent from my BlackBerry? wireless device From arigo at tunes.org Wed Jun 15 17:04:12 2011 From: arigo at tunes.org (Armin Rigo) Date: Wed, 15 Jun 2011 17:04:12 +0200 Subject: [pypy-dev] Ootypes In-Reply-To: <1706411655-1308140988-cardhu_decombobulator_blackberry.rim.net-756381689-@b4.c8.bise7.blackberry> References: <1706411655-1308140988-cardhu_decombobulator_blackberry.rim.net-756381689-@b4.c8.bise7.blackberry> Message-ID: Hi Gregn, On Wed, Jun 15, 2011 at 2:29 PM, wrote: > Pretty please, can ootypes get some love in the upcoming sprint? It depends on who is there. If you are there and willing to work on it, then sure, we are willing to help you :-) A bient?t, Armin. From cremes.devlist at mac.com Thu Jun 16 04:32:48 2011 From: cremes.devlist at mac.com (Chuck Remes) Date: Wed, 15 Jun 2011 21:32:48 -0500 Subject: [pypy-dev] codespeed setup Message-ID: <6DFAAB8F-CF0C-4065-A85F-DA25CDDED40A@mac.com> I'm working on setting up codespeed for several Ruby projects. The codespeed project points at this list as the place to ask questions and get help. >From the README on the project page (https://github.com/tobami/codespeed) it says that codespeed needs to clone any repositories locally that it is tracking. > Note: Git and Mercurial need to locally clone the repository. That means that your codespeed/speedcenter/repos directory will need to be owned by the server. In the case of a typical Apache installation, you'll need to type sudo chown www-data:www-data codespeed/speedcenter/repos This implies that serving up all web requests to get access to the stats happens on the *same* box that is also running the benchmarks. I would think this would skew results over time especially if there is heavy web traffic. I would like to avoid having the resource consumption from web hosting impact the benchmarking output. Once I'm ready to take this into production, is there a django configuration change I can make so that the box offloads all web hosting to another host? Ideally, as commits are detected in the repositories and the benchmarks are rerun and the results posted, codespeed would be able to regen a bunch of static files and ship them off to the web server. Am I on the right track with this or am I missing something obvious? cr From fijall at gmail.com Thu Jun 16 09:12:04 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Thu, 16 Jun 2011 09:12:04 +0200 Subject: [pypy-dev] codespeed setup In-Reply-To: <6DFAAB8F-CF0C-4065-A85F-DA25CDDED40A@mac.com> References: <6DFAAB8F-CF0C-4065-A85F-DA25CDDED40A@mac.com> Message-ID: On Thu, Jun 16, 2011 at 4:32 AM, Chuck Remes wrote: > I'm working on setting up codespeed for several Ruby projects. The codespeed project points at this list as the place to ask questions and get help. There is a codespeed-dev list somewhere as well, it is ok to ask here though. > > >From the README on the project page (https://github.com/tobami/codespeed) it says that codespeed needs to clone any repositories locally that it is tracking. > >> Note: Git and Mercurial need to locally clone the repository. That means that your codespeed/speedcenter/repos directory will need to be owned by the server. In the case of a typical Apache installation, you'll need to type sudo chown www-data:www-data codespeed/speedcenter/repos > > This implies that serving up all web requests to get access to the stats happens on the *same* box that is also running the benchmarks. I would think this would skew results over time especially if there is heavy web traffic. I would like to avoid having the resource consumption from web hosting impact the benchmarking output. No, it does not imply that in any way. Requiring local copy doesn't mean you have to have benchmarks run there. > > Once I'm ready to take this into production, is there a django configuration change I can make so that the box offloads all web hosting to another host? Ideally, as commits are detected in the repositories and the benchmarks are rerun and the results posted, codespeed would be able to regen a bunch of static files and ship them off to the web server. > > Am I on the right track with this or am I missing something obvious? Pages are a bit too dynamic for that. You can request particular ranges of data etc. In theory it's maybe possible, but I don't think with the current codebase. > > cr > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From tobami at googlemail.com Thu Jun 16 09:18:07 2011 From: tobami at googlemail.com (Miquel Torres) Date: Thu, 16 Jun 2011 09:18:07 +0200 Subject: [pypy-dev] codespeed setup In-Reply-To: References: <6DFAAB8F-CF0C-4065-A85F-DA25CDDED40A@mac.com> Message-ID: I have answered on the Codespeed discussion group. Maciej is right that you should actually run bechmarks on a separate, clean box, which then posts results to the Codespeed instance. Miquel 2011/6/16 Maciej Fijalkowski : > On Thu, Jun 16, 2011 at 4:32 AM, Chuck Remes wrote: >> I'm working on setting up codespeed for several Ruby projects. The codespeed project points at this list as the place to ask questions and get help. > > There is a codespeed-dev list somewhere as well, it is ok to ask here though. > >> >> >From the README on the project page (https://github.com/tobami/codespeed) it says that codespeed needs to clone any repositories locally that it is tracking. >> >>> Note: Git and Mercurial need to locally clone the repository. That means that your codespeed/speedcenter/repos directory will need to be owned by the server. In the case of a typical Apache installation, you'll need to type sudo chown www-data:www-data codespeed/speedcenter/repos >> >> This implies that serving up all web requests to get access to the stats happens on the *same* box that is also running the benchmarks. I would think this would skew results over time especially if there is heavy web traffic. I would like to avoid having the resource consumption from web hosting impact the benchmarking output. > > No, it does not imply that in any way. Requiring local copy doesn't > mean you have to have benchmarks run there. > >> >> Once I'm ready to take this into production, is there a django configuration change I can make so that the box offloads all web hosting to another host? Ideally, as commits are detected in the repositories and the benchmarks are rerun and the results posted, codespeed would be able to regen a bunch of static files and ship them off to the web server. >> >> Am I on the right track with this or am I missing something obvious? > > Pages are a bit too dynamic for that. You can request particular > ranges of data etc. In theory it's maybe possible, but I don't think > with the current codebase. > >> >> cr >> >> _______________________________________________ >> pypy-dev mailing list >> pypy-dev at python.org >> http://mail.python.org/mailman/listinfo/pypy-dev >> > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From lac at openend.se Thu Jun 16 16:01:25 2011 From: lac at openend.se (Laura Creighton) Date: Thu, 16 Jun 2011 16:01:25 +0200 Subject: [pypy-dev] PyCon Finland Oct 17-18 in Turku -- CFP Message-ID: <201106161401.p5GE1PZo031308@theraft.openend.se> We should give a talk. It's even _close_ to some of us. :-) Laura ------- Forwarded Message Return-Path: conferences-bounces+lac=openend.se at python.org Delivery-Date: Thu Jun 16 14:39:35 2011 From: Jyry Suvilehto To: tiedotus at python.fi, pigfi at googlegroups.com, python-stockholm at googlegroups.com Subject: [Conferences] PyCon Finland 2011 Call For Proposals PyCon Finland will take place October 17-18 in Turku. The first day will feature presentations and the second is reserved for sprints. We are currently accepting proposals for both talks and sprints. If you would like to give a presentation, organize a sprint or see presentations on a particular topic, please see instructions at http://python.fi/pyconfi. The deadline for proposals is 1.8. The organizers will notify accepted presenters and sprint coordinators by 14.8. The presentation slots will be 40 minutes + 10 minutes of discussion at the end. Shared sessions are also possible. The language for the presentations should be English to encourage international participation. We are also looking for sponsors for the event. If you are interested in sponsoring, please contact Python Finland at hallitus at python.fi for details about sponsorship packages. - -- Jyry Suvilehto Spokesperson, Python Finland _______________________________________________ Conferences mailing list: Conferences at python.org http://mail.python.org/mailman/listinfo/conferences This is an open list with open archives; sensitive or confidential information should not be discussed here. ------- End of Forwarded Message From theller at ctypes.org Fri Jun 17 20:31:59 2011 From: theller at ctypes.org (Thomas Heller) Date: Fri, 17 Jun 2011 20:31:59 +0200 Subject: [pypy-dev] Patch: comtypes works with these changes. In-Reply-To: References: <4DF0D96F.3030604@gmail.com> Message-ID: Am 09.06.2011 16:45, schrieb Thomas Heller: > I have never build pypy myself, so I would prefer to wait for some > windows binaries before trying this. I have now built pypy from the repository - the build did only work with -O2 not with -Ojit. And I have uploaded a new patch plus unittest: https://bugs.pypy.org/issue743 Thomas From tismer at stackless.com Wed Jun 22 19:27:57 2011 From: tismer at stackless.com (Christian Tismer) Date: Wed, 22 Jun 2011 19:27:57 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org Message-ID: <4E02261D.9000602@stackless.com> Hi friends, the subject line says it all... I'm in the progress of updating stackless to use mercurial on python.org and talked to Martin v. Loewis who pointed out the restrictions of Bitbucket. Besides the impression that Bitbucket is pretty slow, it is also not possible to add our own hooks to it. I'm pretty sure python.org would be happy to host PyPy. Is there any good reason why we don't ask and move to python.org? cheers - chris -- Christian Tismer :^) tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ From benjamin at python.org Wed Jun 22 19:30:34 2011 From: benjamin at python.org (Benjamin Peterson) Date: Wed, 22 Jun 2011 12:30:34 -0500 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <4E02261D.9000602@stackless.com> References: <4E02261D.9000602@stackless.com> Message-ID: 2011/6/22 Christian Tismer : > Hi friends, > > the subject line says it all... > > I'm in the progress of updating stackless to use mercurial on > python.org and talked to Martin v. Loewis who pointed out > the restrictions of Bitbucket. > Besides the impression that Bitbucket is pretty slow, it is also not > possible to add our own hooks to it. The impression that Bitbucket is slow? What does that mean? I don't find python.org any faster than Bitbucket. > > I'm pretty sure python.org would be happy to host PyPy. > Is there any good reason why we don't ask and move to python.org? I didn't do the mercurial transition, but I'm pretty happy with bitbucket. -- Regards, Benjamin From fijall at gmail.com Wed Jun 22 19:43:56 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Wed, 22 Jun 2011 19:43:56 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: References: <4E02261D.9000602@stackless.com> Message-ID: On Wed, Jun 22, 2011 at 7:30 PM, Benjamin Peterson wrote: > 2011/6/22 Christian Tismer : >> Hi friends, >> >> the subject line says it all... >> >> I'm in the progress of updating stackless to use mercurial on >> python.org and talked to Martin v. Loewis who pointed out >> the restrictions of Bitbucket. >> Besides the impression that Bitbucket is pretty slow, it is also not >> possible to add our own hooks to it. > > The impression that Bitbucket is slow? What does that mean? I don't > find python.org any faster than Bitbucket. > >> >> I'm pretty sure python.org would be happy to host PyPy. >> Is there any good reason why we don't ask and move to python.org? > > I didn't do the mercurial transition, but I'm pretty happy with bitbucket. > I think there are good reasons for and against, but as of now we don't need to do anything to stay on bitbucket, which is good enough. Besides they sent us free shirts! From tismer at stackless.com Thu Jun 23 17:51:42 2011 From: tismer at stackless.com (Christian Tismer) Date: Thu, 23 Jun 2011 17:51:42 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: References: <4E02261D.9000602@stackless.com> Message-ID: <4E03610E.9060706@stackless.com> On 6/22/11 7:30 PM, Benjamin Peterson wrote: > 2011/6/22 Christian Tismer: >> Hi friends, >> >> the subject line says it all... >> >> I'm in the progress of updating stackless to use mercurial on >> python.org and talked to Martin v. Loewis who pointed out >> the restrictions of Bitbucket. >> Besides the impression that Bitbucket is pretty slow, it is also not >> possible to add our own hooks to it. > The impression that Bitbucket is slow? What does that mean? I don't > find python.org any faster than Bitbucket. > >> I'm pretty sure python.org would be happy to host PyPy. >> Is there any good reason why we don't ask and move to python.org? > I didn't do the mercurial transition, but I'm pretty happy with bitbucket. This is no answer but an opinion ;-) I was asking why we don't use python.org instead of bitbucket. Before, we had codespeak.net which was very convenient because I knew all relevant people in person. Python used sourceforge before, but preferred to have the freedom to host their data themselves. By using python.org, PyPy would have similar convenience as before. Therefore my question: What makes bitbucket the better choice over python.org, despite free t-shirts? (which might be an important reason for some :-) ) still wondering -- chris -- Christian Tismer :^) tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ From fijall at gmail.com Thu Jun 23 18:52:30 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Thu, 23 Jun 2011 18:52:30 +0200 Subject: [pypy-dev] [pypy-commit] pypy default: No cookie for fijal :-) Two heads on default. In-Reply-To: <20110623160349.9F476820AE@wyvern.cs.uni-duesseldorf.de> References: <20110623160349.9F476820AE@wyvern.cs.uni-duesseldorf.de> Message-ID: Pfff I kind of suspect the "fulfill pull request" to do push -f. I plead ignorance and clicking buttons on the web :( On Thu, Jun 23, 2011 at 6:03 PM, arigo wrote: > Author: Armin Rigo > Branch: > Changeset: r45085:cf73cef5f51f > Date: 2011-06-23 18:08 +0200 > http://bitbucket.org/pypy/pypy/changeset/cf73cef5f51f/ > > Log: ? ?No cookie for fijal :-) Two heads on default. > > _______________________________________________ > pypy-commit mailing list > pypy-commit at python.org > http://mail.python.org/mailman/listinfo/pypy-commit > From alex.gaynor at gmail.com Thu Jun 23 18:54:31 2011 From: alex.gaynor at gmail.com (Alex Gaynor) Date: Thu, 23 Jun 2011 09:54:31 -0700 Subject: [pypy-dev] [pypy-commit] pypy default: No cookie for fijal :-) Two heads on default. In-Reply-To: References: <20110623160349.9F476820AE@wyvern.cs.uni-duesseldorf.de> Message-ID: If just using a pull request caused this we should let Jesper/bitbucket know. Alex On Thu, Jun 23, 2011 at 9:52 AM, Maciej Fijalkowski wrote: > Pfff > > I kind of suspect the "fulfill pull request" to do push -f. I plead > ignorance and clicking buttons on the web :( > > On Thu, Jun 23, 2011 at 6:03 PM, arigo wrote: > > Author: Armin Rigo > > Branch: > > Changeset: r45085:cf73cef5f51f > > Date: 2011-06-23 18:08 +0200 > > http://bitbucket.org/pypy/pypy/changeset/cf73cef5f51f/ > > > > Log: No cookie for fijal :-) Two heads on default. > > > > _______________________________________________ > > pypy-commit mailing list > > pypy-commit at python.org > > http://mail.python.org/mailman/listinfo/pypy-commit > > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero -------------- next part -------------- An HTML attachment was scrubbed... URL: From alex.gaynor at gmail.com Sat Jun 25 03:11:34 2011 From: alex.gaynor at gmail.com (Alex Gaynor) Date: Fri, 24 Jun 2011 18:11:34 -0700 Subject: [pypy-dev] Debugging "cannot find gc roots!" Message-ID: Hi all (especially Armin and Amaury :-)), In running my application I hit a "cannot find gc roots!", it reproduces reliably, but it's a pain to isolate because it occurs in a very large application, running under paste, with threads :( What is the appropriate way to try to debug this? Running under gdb the program still prints out the error and aborts, however gdb seems to think execution completed normally, and as such doesn't drop me in anyplace. What is the general strategy for debugging such an issue, or would I be better off to just use shadow stack? Thanks, Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero -------------- next part -------------- An HTML attachment was scrubbed... URL: From arigo at tunes.org Sat Jun 25 10:11:32 2011 From: arigo at tunes.org (Armin Rigo) Date: Sat, 25 Jun 2011 10:11:32 +0200 Subject: [pypy-dev] Debugging "cannot find gc roots!" In-Reply-To: References: Message-ID: Hi Alex, On Sat, Jun 25, 2011 at 3:11 AM, Alex Gaynor wrote: > In running my application I hit a "cannot find gc roots!", it reproduces > reliably, but it's a pain to isolate because it occurs in a very large > application, running under paste, with threads :( You should rebuild pypy with "make lldebug". Then gdb should stop as soon as the crash is found. Armin From davide.setti at gmail.com Sat Jun 25 14:07:12 2011 From: davide.setti at gmail.com (Davide Setti) Date: Sat, 25 Jun 2011 14:07:12 +0200 Subject: [pypy-dev] pypy and PIL Message-ID: Hi all, i'm trying to install PIL 1.1.7 on pypy 1.5.0, but when i run python setup.py build_ext i get an exception, due to distutils. The difference is that: $ python -c "from distutils import sysconfig; print sysconfig.get_config_var('CFLAGS')" -fno-strict-aliasing -O3 -march=core2 -msse4.1 -w -pipe -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes $ pypy -c "from distutils import sysconfig; print sysconfig.get_config_var('CFLAGS')" None I use osX. You can find the traceback on https://gist.github.com/1046409 Thanks -- Davide Setti blog: http://blog.flatlandia.eu code: http://github.com/vad From holger at merlinux.eu Sun Jun 26 12:27:50 2011 From: holger at merlinux.eu (holger krekel) Date: Sun, 26 Jun 2011 10:27:50 +0000 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <4E03610E.9060706@stackless.com> References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> Message-ID: <20110626102750.GB20287@merlinux.eu> Hi Christian, On Thu, Jun 23, 2011 at 17:51 +0200, Christian Tismer wrote: > On 6/22/11 7:30 PM, Benjamin Peterson wrote: > >2011/6/22 Christian Tismer: > >>Hi friends, > >> > >>the subject line says it all... > >> > >>I'm in the progress of updating stackless to use mercurial on > >>python.org and talked to Martin v. Loewis who pointed out > >>the restrictions of Bitbucket. > >>Besides the impression that Bitbucket is pretty slow, it is also not > >>possible to add our own hooks to it. > >The impression that Bitbucket is slow? What does that mean? I don't > >find python.org any faster than Bitbucket. > > > >>I'm pretty sure python.org would be happy to host PyPy. > >>Is there any good reason why we don't ask and move to python.org? > >I didn't do the mercurial transition, but I'm pretty happy with bitbucket. > > This is no answer but an opinion ;-) > > I was asking why we don't use python.org instead of bitbucket. I don't remember any big comparison analysis. Some people pushed for bitbucket, a number was using it already, and the others didn't mind. The main effort and focus was on the conversion of the svn repository, anyway. > Before, we had codespeak.net which was very convenient because > I knew all relevant people in person. We have some personal contacts to bitbucket - they actually sponsor an unlimited plan for PyPy. Moreover, some pypy devs wanted a hosting solution where we do not depend on private connects or work but can rather rely on a company basing their business on such hosting. > Python used sourceforge before, but preferred to have the freedom > to host their data themselves. > By using python.org, PyPy would have similar convenience as > before. > Therefore my question: What makes bitbucket the better choice over > python.org, despite free t-shirts? (which might be an important reason > for some :-) ) PyPy has by now quite some integration code wrt to bitbucket. It seems all are quite happy with bitbucket services as it stands. So seen from now the question probably rather is why we should move anywhere else. best, holger > still wondering -- chris > > -- > Christian Tismer :^) > tismerysoft GmbH : Have a break! Take a ride on Python's > Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ > 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ > work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 > PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 > whom do you want to sponsor today? http://www.stackless.com/ > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From tismer at stackless.com Sun Jun 26 13:37:54 2011 From: tismer at stackless.com (Christian Tismer) Date: Sun, 26 Jun 2011 13:37:54 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <20110626102750.GB20287@merlinux.eu> References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> Message-ID: <4E071A12.9040200@stackless.com> On 6/26/11 12:27 PM, holger krekel wrote: > Hi Christian, > > On Thu, Jun 23, 2011 at 17:51 +0200, Christian Tismer wrote: >> On 6/22/11 7:30 PM, Benjamin Peterson wrote: >>> 2011/6/22 Christian Tismer: >>>> Hi friends, >>>> >>>> the subject line says it all... >>>> >>>> I'm in the progress of updating stackless to use mercurial on >>>> python.org and talked to Martin v. Loewis who pointed out >>>> the restrictions of Bitbucket. >>>> Besides the impression that Bitbucket is pretty slow, it is also not >>>> possible to add our own hooks to it. >>> The impression that Bitbucket is slow? What does that mean? I don't >>> find python.org any faster than Bitbucket. >>> >>>> I'm pretty sure python.org would be happy to host PyPy. >>>> Is there any good reason why we don't ask and move to python.org? >>> I didn't do the mercurial transition, but I'm pretty happy with bitbucket. >> This is no answer but an opinion ;-) >> >> I was asking why we don't use python.org instead of bitbucket. > I don't remember any big comparison analysis. Some people > pushed for bitbucket, a number was using it already, and the others > didn't mind. The main effort and focus was on the conversion of the > svn repository, anyway. > >> Before, we had codespeak.net which was very convenient because >> I knew all relevant people in person. > We have some personal contacts to bitbucket - they actually sponsor an > unlimited plan for PyPy. Moreover, some pypy devs wanted a hosting > solution where we do not depend on private connects or work but can > rather rely on a company basing their business on such hosting. > Well, I understand that all. Maybe I was implicitly assuming that everybody felt like me: It is an honor for Stackless Python to live on python.org, and probably also a positive sign, like some acceptance by core python. That made me wonder. If I had a chance to use python.org instead of anything else, I'd always prefer python.org, unless it has a significant drawback, or they told me "no, go somewhere else" ;-) >> Python used sourceforge before, but preferred to have the freedom >> to host their data themselves. >> By using python.org, PyPy would have similar convenience as >> before. >> Therefore my question: What makes bitbucket the better choice over >> python.org, despite free t-shirts? (which might be an important reason >> for some :-) ) > PyPy has by now quite some integration code wrt to bitbucket. > It seems all are quite happy with bitbucket services as it stands. > So seen from now the question probably rather is why we should > move anywhere else. Well, as said, I see a positive political effect in moving to python.org that I (personally) would not underestimate. But PyPy is maybe popular enough that my point doesn't really exist, or even vice versa - maybe the distinction is even welcome. ;-) cheers - chris -- Christian Tismer :^) tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ From holger at merlinux.eu Sun Jun 26 14:21:19 2011 From: holger at merlinux.eu (holger krekel) Date: Sun, 26 Jun 2011 12:21:19 +0000 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <4E071A12.9040200@stackless.com> References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> <4E071A12.9040200@stackless.com> Message-ID: <20110626122119.GE20287@merlinux.eu> On Sun, Jun 26, 2011 at 13:37 +0200, Christian Tismer wrote: > On 6/26/11 12:27 PM, holger krekel wrote: > >Hi Christian, > > > >On Thu, Jun 23, 2011 at 17:51 +0200, Christian Tismer wrote: > >>On 6/22/11 7:30 PM, Benjamin Peterson wrote: > >>>2011/6/22 Christian Tismer: > >>>>Hi friends, > >>>> > >>>>the subject line says it all... > >>>> > >>>>I'm in the progress of updating stackless to use mercurial on > >>>>python.org and talked to Martin v. Loewis who pointed out > >>>>the restrictions of Bitbucket. > >>>>Besides the impression that Bitbucket is pretty slow, it is also not > >>>>possible to add our own hooks to it. > >>>The impression that Bitbucket is slow? What does that mean? I don't > >>>find python.org any faster than Bitbucket. > >>> > >>>>I'm pretty sure python.org would be happy to host PyPy. > >>>>Is there any good reason why we don't ask and move to python.org? > >>>I didn't do the mercurial transition, but I'm pretty happy with bitbucket. > >>This is no answer but an opinion ;-) > >> > >>I was asking why we don't use python.org instead of bitbucket. > >I don't remember any big comparison analysis. Some people > >pushed for bitbucket, a number was using it already, and the others > >didn't mind. The main effort and focus was on the conversion of the > >svn repository, anyway. > > > >>Before, we had codespeak.net which was very convenient because > >>I knew all relevant people in person. > >We have some personal contacts to bitbucket - they actually sponsor an > >unlimited plan for PyPy. Moreover, some pypy devs wanted a hosting > >solution where we do not depend on private connects or work but can > >rather rely on a company basing their business on such hosting. > > > > Well, I understand that all. > Maybe I was implicitly assuming that everybody felt like me: > It is an honor for Stackless Python to live on python.org, and > probably also a positive sign, like some acceptance by core python. > > That made me wonder. If I had a chance to use python.org instead > of anything else, I'd always prefer python.org, unless it has a significant > drawback, or they told me "no, go somewhere else" ;-) > >>Python used sourceforge before, but preferred to have the freedom > >>to host their data themselves. > >>By using python.org, PyPy would have similar convenience as > >>before. > >>Therefore my question: What makes bitbucket the better choice over > >>python.org, despite free t-shirts? (which might be an important reason > >>for some :-) ) > >PyPy has by now quite some integration code wrt to bitbucket. > >It seems all are quite happy with bitbucket services as it stands. > >So seen from now the question probably rather is why we should > >move anywhere else. > > Well, as said, I see a positive political effect in moving to python.org > that I (personally) would not underestimate. I agree, it would have this positive political effect so that is clearly on the pro side. It did when i moved the mailing lists to mail.python.org. For what i know, we are welcome to move our repositories to python.org. > But PyPy is maybe popular enough that my point doesn't really exist, > or even vice versa - maybe the distinction is even welcome. ;-) Heh, dunno. At this stage it would cause some pain but not much technical gain to again move somewhere else. And i think a number of pypy devs are happy with not having to care or think about infrastructure issues or change of dev habits. cheers, holger holger > cheers - chris > > -- > Christian Tismer :^) > tismerysoft GmbH : Have a break! Take a ride on Python's > Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ > 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ > work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 > PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 > whom do you want to sponsor today? http://www.stackless.com/ > From anto.cuni at gmail.com Sun Jun 26 14:36:39 2011 From: anto.cuni at gmail.com (Antonio Cuni) Date: Sun, 26 Jun 2011 14:36:39 +0200 Subject: [pypy-dev] [pypy-commit] pypy default: Fix test_pypy_c. In-Reply-To: <20110626113135.98F6482178@wyvern.cs.uni-duesseldorf.de> References: <20110626113135.98F6482178@wyvern.cs.uni-duesseldorf.de> Message-ID: <4E0727D7.20803@gmail.com> On 26/06/11 13:31, arigo wrote: > Author: Armin Rigo > Branch: > Changeset: r45136:ff284eff23fd > Date: 2011-06-26 13:37 +0200 > http://bitbucket.org/pypy/pypy/changeset/ff284eff23fd/ > > Log: Fix test_pypy_c. > > diff --git a/pypy/module/pypyjit/test_pypy_c/test_array.py b/pypy/module/pypyjit/test_pypy_c/test_array.py > --- a/pypy/module/pypyjit/test_pypy_c/test_array.py > +++ b/pypy/module/pypyjit/test_pypy_c/test_array.py > @@ -46,7 +46,7 @@ > guard_no_overflow(descr=) > i18 = int_add(i7, 1) > --TICK-- > - jump(p0, p1, p2, p3, p4, p5, p6, i18, i16, i9, i10, descr=) > + jump(p0, p1, p2, p3, p4, p5, i18, i16, p8, i9, i10, descr=) > """) note for the future: you can also write "..." if you don't care about matching the parameter list: jump(..., descr=) ciao, Anto From sdouche at gmail.com Sun Jun 26 15:15:42 2011 From: sdouche at gmail.com (Sebastien Douche) Date: Sun, 26 Jun 2011 15:15:42 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <4E071A12.9040200@stackless.com> References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> <4E071A12.9040200@stackless.com> Message-ID: On Sun, Jun 26, 2011 at 13:37, Christian Tismer wrote: > Well, I understand that all. > Maybe I was implicitly assuming that everybody felt like me: > It is an honor for Stackless Python to live on python.org, and > probably also a positive sign, like some acceptance by core python. >From a user point of view, I feel the same thing. > That made me wonder. If I had a chance to use python.org instead > of anything else, I'd always prefer python.org, unless it has a significant > drawback, or they told me "no, go somewhere else" ;-) A solution could be to use bitbucket for the day to day job, and push on python.org regularly (each release?). The cost is very low. -- Sebastien Douche Twitter : @sdouche From wilk at flibuste.net Sun Jun 26 19:17:59 2011 From: wilk at flibuste.net (William) Date: Sun, 26 Jun 2011 17:17:59 +0000 (UTC) Subject: [pypy-dev] Bitbucket? Why not python.org References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> <4E071A12.9040200@stackless.com> Message-ID: On 26-06-2011, Christian Tismer wrote: > Maybe I was implicitly assuming that everybody felt like me: > It is an honor for Stackless Python to live on python.org, and > probably also a positive sign, like some acceptance by core python. As a simple user, this argument catch me also... -- William Dod? - http://flibuste.net Informaticien Ind?pendant From Ronny.Pfannschmidt at gmx.de Mon Jun 27 07:40:23 2011 From: Ronny.Pfannschmidt at gmx.de (Ronny Pfannschmidt) Date: Mon, 27 Jun 2011 07:40:23 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <20110626122119.GE20287@merlinux.eu> References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> <4E071A12.9040200@stackless.com> <20110626122119.GE20287@merlinux.eu> Message-ID: <1309153223.15293.2.camel@Klappe2> Hi, having a synced repo on python.org might be a help however i wouldn't want push-permission handling there, since we are probably a bit ore lax about giving new talents access to the repo. Simply having a synced repo on pythong.org might have a better effect for visibility. -- Ronny On Sun, 2011-06-26 at 12:21 +0000, holger krekel wrote: > On Sun, Jun 26, 2011 at 13:37 +0200, Christian Tismer wrote: > > On 6/26/11 12:27 PM, holger krekel wrote: > > >Hi Christian, > > > > > >On Thu, Jun 23, 2011 at 17:51 +0200, Christian Tismer wrote: > > >>On 6/22/11 7:30 PM, Benjamin Peterson wrote: > > >>>2011/6/22 Christian Tismer: > > >>>>Hi friends, > > >>>> > > >>>>the subject line says it all... > > >>>> > > >>>>I'm in the progress of updating stackless to use mercurial on > > >>>>python.org and talked to Martin v. Loewis who pointed out > > >>>>the restrictions of Bitbucket. > > >>>>Besides the impression that Bitbucket is pretty slow, it is also not > > >>>>possible to add our own hooks to it. > > >>>The impression that Bitbucket is slow? What does that mean? I don't > > >>>find python.org any faster than Bitbucket. > > >>> > > >>>>I'm pretty sure python.org would be happy to host PyPy. > > >>>>Is there any good reason why we don't ask and move to python.org? > > >>>I didn't do the mercurial transition, but I'm pretty happy with bitbucket. > > >>This is no answer but an opinion ;-) > > >> > > >>I was asking why we don't use python.org instead of bitbucket. > > >I don't remember any big comparison analysis. Some people > > >pushed for bitbucket, a number was using it already, and the others > > >didn't mind. The main effort and focus was on the conversion of the > > >svn repository, anyway. > > > > > >>Before, we had codespeak.net which was very convenient because > > >>I knew all relevant people in person. > > >We have some personal contacts to bitbucket - they actually sponsor an > > >unlimited plan for PyPy. Moreover, some pypy devs wanted a hosting > > >solution where we do not depend on private connects or work but can > > >rather rely on a company basing their business on such hosting. > > > > > > > Well, I understand that all. > > Maybe I was implicitly assuming that everybody felt like me: > > It is an honor for Stackless Python to live on python.org, and > > probably also a positive sign, like some acceptance by core python. > > > > That made me wonder. If I had a chance to use python.org instead > > of anything else, I'd always prefer python.org, unless it has a significant > > drawback, or they told me "no, go somewhere else" ;-) > > >>Python used sourceforge before, but preferred to have the freedom > > >>to host their data themselves. > > >>By using python.org, PyPy would have similar convenience as > > >>before. > > >>Therefore my question: What makes bitbucket the better choice over > > >>python.org, despite free t-shirts? (which might be an important reason > > >>for some :-) ) > > >PyPy has by now quite some integration code wrt to bitbucket. > > >It seems all are quite happy with bitbucket services as it stands. > > >So seen from now the question probably rather is why we should > > >move anywhere else. > > > > Well, as said, I see a positive political effect in moving to python.org > > that I (personally) would not underestimate. > > I agree, it would have this positive political effect so that is clearly > on the pro side. It did when i moved the mailing lists to mail.python.org. > > For what i know, we are welcome to move our repositories to python.org. > > > But PyPy is maybe popular enough that my point doesn't really exist, > > or even vice versa - maybe the distinction is even welcome. ;-) > > Heh, dunno. At this stage it would cause some pain but not much technical > gain to again move somewhere else. And i think a number of pypy devs > are happy with not having to care or think about infrastructure issues > or change of dev habits. > > cheers, > holger > > > > holger > > > cheers - chris > > > > -- > > Christian Tismer :^) > > tismerysoft GmbH : Have a break! Take a ride on Python's > > Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ > > 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ > > work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 > > PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 > > whom do you want to sponsor today? http://www.stackless.com/ > > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 836 bytes Desc: This is a digitally signed message part URL: From alex.gaynor at gmail.com Tue Jun 28 00:45:15 2011 From: alex.gaynor at gmail.com (Alex Gaynor) Date: Mon, 27 Jun 2011 15:45:15 -0700 Subject: [pypy-dev] Debugging "cannot find gc roots!" In-Reply-To: References: Message-ID: On Sat, Jun 25, 2011 at 1:11 AM, Armin Rigo wrote: > Hi Alex, > > On Sat, Jun 25, 2011 at 3:11 AM, Alex Gaynor > wrote: > > In running my application I hit a "cannot find gc roots!", it reproduces > > reliably, but it's a pain to isolate because it occurs in a very large > > application, running under paste, with threads :( > > You should rebuild pypy with "make lldebug". Then gdb should stop as > soon as the crash is found. > > > Armin > Hrm, this is pretty strange, running an lldebug build, not under gdb reliably produces a crash: alex at devalex:~/ans/venv/answeb$ ANS_INSTANCE=dwc pypy ../lib/pypy/bin/paster serve --reload alex-dev.ini Starting subprocess with file monitor overriding ans instance based on environment: dwc Starting server in PID 6809. serving on 0.0.0.0:5050 view at http://127.0.0.1:5050 cannot find gc roots! debug_alloc.h: 9 mallocs left (use PYPY_ALLOC=1 to see the list) However, running under gdb results in broken pip errors from a socket (this is a web app in case that wasn't clear): alex at devalex:~/ans/venv/answeb$ ANS_INSTANCE=dwc gdb --args pypy ../lib/pypy/bin/paster serve --reload alex-dev.ini GNU gdb (GDB) 7.1-ubuntu Copyright (C) 2010 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "x86_64-linux-gnu". For bug reporting instructions, please see: ... Reading symbols from /home/alex/ans/venv/bin/pypy...(no debugging symbols found)...done. (gdb) run Starting program: /home/alex/ans/venv/bin/pypy ../lib/pypy/bin/paster serve --reload alex-dev.ini warning: the debug information found in "/lib/ld-2.11.1.so" does not match "/lib/ld-linux.so.2" (CRC mismatch). warning: the debug information found in "/usr/lib/debug//lib/ld-2.11.1.so" does not match "/lib/ld-linux.so.2" (CRC mismatch). warning: the debug information found in "/usr/lib/debug/lib/ld-2.11.1.so" does not match "/lib/ld-linux.so.2" (CRC mismatch). [Thread debugging using libthread_db enabled] Starting subprocess with file monitor overriding ans instance based on environment: dwc Starting server in PID 6743. serving on 0.0.0.0:5050 view at http://127.0.0.1:5050 ---------------------------------------- Exception happened during processing of request from ('10.160.27.96', 44587) Traceback (most recent call last): File "/home/alex/ans/venv/lib/pypy/site-packages/paste/httpserver.py", line 1068, in process_request_in_thread self.finish_request(request, client_address) File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line 323, in finish_request self.RequestHandlerClass(request, client_address, self) File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line 641, in __init__ self.finish() File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line 694, in finish self.wfile.flush() File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", line 348, in flush self._sock.sendall(view[write_offset:write_offset+buffer_size]) File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", line 266, in sendall self._sock.sendall(data, flags) error: [Errno 32] Broken pipe ---------------------------------------- ---------------------------------------- Exception happened during processing of request from ('10.160.27.96', 44608) Traceback (most recent call last): File "/home/alex/ans/venv/lib/pypy/site-packages/paste/httpserver.py", line 1068, in process_request_in_thread self.finish_request(request, client_address) File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line 323, in finish_request self.RequestHandlerClass(request, client_address, self) File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line 641, in __init__ self.finish() File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line 694, in finish self.wfile.flush() File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", line 348, in flush self._sock.sendall(view[write_offset:write_offset+buffer_size]) File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", line 266, in sendall self._sock.sendall(data, flags) error: [Errno 32] Broken pipe ---------------------------------------- ---------------------------------------- The only thing worth noting is that this a 32-bit PyPy on a 64-bit OS (the gdb is also 64-bits, so if that matters the gdb runs may not be valid). I'm not sure whether I see the socket issue with gdb because its dieing earlier than the gc roots issue, or if it's because the error doesn't appear for some reason, and therefore it gets farther and hits a second issue. Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero -------------- next part -------------- An HTML attachment was scrubbed... URL: From aguirrea at gmail.com Tue Jun 28 01:49:09 2011 From: aguirrea at gmail.com (Andres Aguirre) Date: Mon, 27 Jun 2011 20:49:09 -0300 Subject: [pypy-dev] butialo activity Message-ID: Hi to everyone. There's a new activit, I hope that most of you will enjoy, it's called Butialo (http://activities.sugarlabs.org/en-US/sugar/addon/4457) Butialo is a IDE based on PyPy that allows programming the Buti? robot (http://www.fing.edu.uy/inco/proyectos/butia/) with Lua. Lua is a very fast, powerfull and easy to use language. The IDE provides autodetects the configuration of your Butia robot, and provides snippets of code showing how to read sensors and do stuff. Best regards -- /\ndr?s From alex.gaynor at gmail.com Tue Jun 28 02:09:14 2011 From: alex.gaynor at gmail.com (Alex Gaynor) Date: Mon, 27 Jun 2011 17:09:14 -0700 Subject: [pypy-dev] Debugging "cannot find gc roots!" In-Reply-To: References: Message-ID: On Mon, Jun 27, 2011 at 3:45 PM, Alex Gaynor wrote: > > > On Sat, Jun 25, 2011 at 1:11 AM, Armin Rigo wrote: > >> Hi Alex, >> >> On Sat, Jun 25, 2011 at 3:11 AM, Alex Gaynor >> wrote: >> > In running my application I hit a "cannot find gc roots!", it reproduces >> > reliably, but it's a pain to isolate because it occurs in a very large >> > application, running under paste, with threads :( >> >> You should rebuild pypy with "make lldebug". Then gdb should stop as >> soon as the crash is found. >> >> >> Armin >> > > Hrm, this is pretty strange, running an lldebug build, not under gdb > reliably produces a crash: > > alex at devalex:~/ans/venv/answeb$ ANS_INSTANCE=dwc pypy > ../lib/pypy/bin/paster serve --reload alex-dev.ini > Starting subprocess with file monitor > overriding ans instance based on environment: dwc > Starting server in PID 6809. > serving on 0.0.0.0:5050 view at http://127.0.0.1:5050 > cannot find gc roots! > debug_alloc.h: 9 mallocs left (use PYPY_ALLOC=1 to see the list) > > However, running under gdb results in broken pip errors from a socket (this > is a web app in case that wasn't clear): > > alex at devalex:~/ans/venv/answeb$ ANS_INSTANCE=dwc gdb --args pypy > ../lib/pypy/bin/paster serve --reload alex-dev.ini > GNU gdb (GDB) 7.1-ubuntu > Copyright (C) 2010 Free Software Foundation, Inc. > License GPLv3+: GNU GPL version 3 or later < > http://gnu.org/licenses/gpl.html> > This is free software: you are free to change and redistribute it. > There is NO WARRANTY, to the extent permitted by law. Type "show copying" > and "show warranty" for details. > This GDB was configured as "x86_64-linux-gnu". > For bug reporting instructions, please see: > ... > Reading symbols from /home/alex/ans/venv/bin/pypy...(no debugging symbols > found)...done. > (gdb) run > Starting program: /home/alex/ans/venv/bin/pypy ../lib/pypy/bin/paster serve > --reload alex-dev.ini > warning: the debug information found in "/lib/ld-2.11.1.so" does not match > "/lib/ld-linux.so.2" (CRC mismatch). > > warning: the debug information found in "/usr/lib/debug//lib/ld-2.11.1.so" > does not match "/lib/ld-linux.so.2" (CRC mismatch). > > warning: the debug information found in "/usr/lib/debug/lib/ld-2.11.1.so" > does not match "/lib/ld-linux.so.2" (CRC mismatch). > > [Thread debugging using libthread_db enabled] > Starting subprocess with file monitor > overriding ans instance based on environment: dwc > Starting server in PID 6743. > serving on 0.0.0.0:5050 view at http://127.0.0.1:5050 > ---------------------------------------- > Exception happened during processing of request from ('10.160.27.96', > 44587) > Traceback (most recent call last): > File "/home/alex/ans/venv/lib/pypy/site-packages/paste/httpserver.py", > line 1068, in process_request_in_thread > self.finish_request(request, client_address) > File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line > 323, in finish_request > self.RequestHandlerClass(request, client_address, self) > File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line > 641, in __init__ > self.finish() > File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line > 694, in finish > self.wfile.flush() > File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", > line 348, in flush > self._sock.sendall(view[write_offset:write_offset+buffer_size]) > File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", > line 266, in sendall > self._sock.sendall(data, flags) > error: [Errno 32] Broken pipe > ---------------------------------------- > ---------------------------------------- > Exception happened during processing of request from ('10.160.27.96', > 44608) > Traceback (most recent call last): > File "/home/alex/ans/venv/lib/pypy/site-packages/paste/httpserver.py", > line 1068, in process_request_in_thread > self.finish_request(request, client_address) > File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line > 323, in finish_request > self.RequestHandlerClass(request, client_address, self) > File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line > 641, in __init__ > self.finish() > File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line > 694, in finish > self.wfile.flush() > File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", > line 348, in flush > self._sock.sendall(view[write_offset:write_offset+buffer_size]) > File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", > line 266, in sendall > self._sock.sendall(data, flags) > error: [Errno 32] Broken pipe > ---------------------------------------- > ---------------------------------------- > > > The only thing worth noting is that this a 32-bit PyPy on a 64-bit OS (the > gdb is also 64-bits, so if that matters the gdb runs may not be valid). I'm > not sure whether I see the socket issue with gdb because its dieing earlier > than the gc roots issue, or if it's because the error doesn't appear for > some reason, and therefore it gets farther and hits a second issue. > > > Alex > > > -- > "I disapprove of what you say, but I will defend to the death your right to > say it." -- Evelyn Beatrice Hall (summarizing Voltaire) > "The people's good is the highest law." -- Cicero > > FWIW under shadowstack I'm seeing the broken pipe error as well, so I guess gdb is just making the error not happen. Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) "The people's good is the highest law." -- Cicero -------------- next part -------------- An HTML attachment was scrubbed... URL: From fijall at gmail.com Tue Jun 28 09:33:51 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Tue, 28 Jun 2011 09:33:51 +0200 Subject: [pypy-dev] butialo activity In-Reply-To: References: Message-ID: On Tue, Jun 28, 2011 at 1:49 AM, Andres Aguirre wrote: > Hi to everyone. > There's a new activit, I hope that most of you will enjoy, it's called > Butialo (http://activities.sugarlabs.org/en-US/sugar/addon/4457) > Butialo is a IDE based on PyPy that allows programming the Buti? robot > (http://www.fing.edu.uy/inco/proyectos/butia/) > with Lua. Lua is a very fast, powerfull and easy to use language. The > IDE provides autodetects the configuration of your Butia robot, and > provides snippets of code showing how to read sensors and do stuff. > Best regards Does it have anything at all to do with PyPy? > > -- > /\ndr?s > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > From fijall at gmail.com Tue Jun 28 09:36:08 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Tue, 28 Jun 2011 09:36:08 +0200 Subject: [pypy-dev] Debugging "cannot find gc roots!" In-Reply-To: References: Message-ID: On Tue, Jun 28, 2011 at 2:09 AM, Alex Gaynor wrote: > > > On Mon, Jun 27, 2011 at 3:45 PM, Alex Gaynor wrote: >> >> >> On Sat, Jun 25, 2011 at 1:11 AM, Armin Rigo wrote: >>> >>> Hi Alex, >>> >>> On Sat, Jun 25, 2011 at 3:11 AM, Alex Gaynor >>> wrote: >>> > In running my application I hit a "cannot find gc roots!", it >>> > reproduces >>> > reliably, but it's a pain to isolate because it occurs in a very large >>> > application, running under paste, with threads :( >>> >>> You should rebuild pypy with "make lldebug". ?Then gdb should stop as >>> soon as the crash is found. >>> >>> >>> Armin >> >> Hrm, this is pretty strange, running an lldebug build, not under gdb >> reliably produces a crash: >> >> alex at devalex:~/ans/venv/answeb$ ANS_INSTANCE=dwc pypy >> ../lib/pypy/bin/paster serve --reload alex-dev.ini >> Starting subprocess with file monitor >> overriding ans instance based on environment: dwc >> Starting server in PID 6809. >> serving on 0.0.0.0:5050 view at http://127.0.0.1:5050 >> cannot find gc roots! >> debug_alloc.h: 9 mallocs left (use PYPY_ALLOC=1 to see the list) >> >> However, running under gdb results in broken pip errors from a socket >> (this is a web app in case that wasn't clear): >> >> alex at devalex:~/ans/venv/answeb$ ANS_INSTANCE=dwc gdb --args pypy >> ../lib/pypy/bin/paster serve --reload alex-dev.ini >> GNU gdb (GDB) 7.1-ubuntu >> Copyright (C) 2010 Free Software Foundation, Inc. >> License GPLv3+: GNU GPL version 3 or later >> >> This is free software: you are free to change and redistribute it. >> There is NO WARRANTY, to the extent permitted by law.? Type "show copying" >> and "show warranty" for details. >> This GDB was configured as "x86_64-linux-gnu". >> For bug reporting instructions, please see: >> ... >> Reading symbols from /home/alex/ans/venv/bin/pypy...(no debugging symbols >> found)...done. >> (gdb) run >> Starting program: /home/alex/ans/venv/bin/pypy ../lib/pypy/bin/paster >> serve --reload alex-dev.ini >> warning: the debug information found in "/lib/ld-2.11.1.so" does not match >> "/lib/ld-linux.so.2" (CRC mismatch). >> >> warning: the debug information found in "/usr/lib/debug//lib/ld-2.11.1.so" >> does not match "/lib/ld-linux.so.2" (CRC mismatch). >> >> warning: the debug information found in "/usr/lib/debug/lib/ld-2.11.1.so" >> does not match "/lib/ld-linux.so.2" (CRC mismatch). >> >> [Thread debugging using libthread_db enabled] >> Starting subprocess with file monitor >> overriding ans instance based on environment: dwc >> Starting server in PID 6743. >> serving on 0.0.0.0:5050 view at http://127.0.0.1:5050 >> ---------------------------------------- >> Exception happened during processing of request from ('10.160.27.96', >> 44587) >> Traceback (most recent call last): >> ? File "/home/alex/ans/venv/lib/pypy/site-packages/paste/httpserver.py", >> line 1068, in process_request_in_thread >> ??? self.finish_request(request, client_address) >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line >> 323, in finish_request >> ??? self.RequestHandlerClass(request, client_address, self) >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line >> 641, in __init__ >> ??? self.finish() >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line >> 694, in finish >> ??? self.wfile.flush() >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", >> line 348, in flush >> ??? self._sock.sendall(view[write_offset:write_offset+buffer_size]) >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", >> line 266, in sendall >> ??? self._sock.sendall(data, flags) >> error: [Errno 32] Broken pipe >> ---------------------------------------- >> ---------------------------------------- >> Exception happened during processing of request from ('10.160.27.96', >> 44608) >> Traceback (most recent call last): >> ? File "/home/alex/ans/venv/lib/pypy/site-packages/paste/httpserver.py", >> line 1068, in process_request_in_thread >> ??? self.finish_request(request, client_address) >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line >> 323, in finish_request >> ??? self.RequestHandlerClass(request, client_address, self) >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line >> 641, in __init__ >> ??? self.finish() >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/2.7/SocketServer.py", line >> 694, in finish >> ??? self.wfile.flush() >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", >> line 348, in flush >> ??? self._sock.sendall(view[write_offset:write_offset+buffer_size]) >> ? File "/home/alex/ans/venv/lib/pypy/lib-python/modified-2.7/socket.py", >> line 266, in sendall >> ??? self._sock.sendall(data, flags) >> error: [Errno 32] Broken pipe >> ---------------------------------------- >> ---------------------------------------- >> >> >> The only thing worth noting is that this a 32-bit PyPy on a 64-bit OS (the >> gdb is also 64-bits, so if that matters the gdb runs may not be valid).? I'm >> not sure whether I see the socket issue with gdb because its dieing earlier >> than the gc roots issue, or if it's because the error doesn't appear for >> some reason, and therefore it gets farther and hits a second issue. >> >> Alex >> >> >> -- >> "I disapprove of what you say, but I will defend to the death your right >> to say it." -- Evelyn Beatrice Hall (summarizing Voltaire) >> "The people's good is the highest law." -- Cicero >> > > FWIW under shadowstack I'm seeing the broken pipe error as well, so I guess > gdb is just making the error not happen. > > Alex > > -- > "I disapprove of what you say, but I will defend to the death your right to > say it." -- Evelyn Beatrice Hall (summarizing Voltaire) > "The people's good is the highest law." -- Cicero > > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > > You have to make gdb ignore SIGPIPE signals (pass it to application), then stuff can work normally. You can also just do ulimit -c and then do gdb pypy-c core From lac at openend.se Tue Jun 28 09:34:39 2011 From: lac at openend.se (Laura Creighton) Date: Tue, 28 Jun 2011 09:34:39 +0200 Subject: [pypy-dev] butialo activity In-Reply-To: Message from Andres Aguirre of "Mon, 27 Jun 2011 20:49:09 -0300." References: Message-ID: <201106280734.p5S7Ydgx016901@theraft.openend.se> In a message of Mon, 27 Jun 2011 20:49:09 -0300, Andres Aguirre writes: >Hi to everyone. >There's a new activit, I hope that most of you will enjoy, it's called >Butialo (http://activities.sugarlabs.org/en-US/sugar/addon/4457) >Butialo is a IDE based on PyPy that allows programming the Buti=E1 robot >(http://www.fing.edu.uy/inco/proyectos/butia/) >with Lua. Lua is a very fast, powerfull and easy to use language. The >IDE provides autodetects the configuration of your Butia robot, and >provides snippets of code showing how to read sensors and do stuff. >Best regards > >-- = > >/\ndr=E9s I looked at those links, and the robot looks like a lot of fun. But I am still confused, a bit. What exactly did you use pypy for? Laura From chef at ghum.de Tue Jun 28 09:47:44 2011 From: chef at ghum.de (Massa, Harald Armin) Date: Tue, 28 Jun 2011 09:47:44 +0200 Subject: [pypy-dev] pypy seems not to read site.py Message-ID: Hello! I just installed pypy-1.5.0a0-win32, and modified site.py within lib-python\2.7\site.py with no effect. Even entering a big "print haaaaaaaalllllloooo" at the beginning does not do anything. Does pypy not use / read / need site.py? I know there is a -S to NOT import site. But... how to run site on startup? Harald -- GHUM GmbH Harald Armin Massa Spielberger Stra?e 49 70435 Stuttgart 0173/9409607 Amtsgericht Stuttgart, HRB 734971 - persuadere. et programmare -------------- next part -------------- An HTML attachment was scrubbed... URL: From fijall at gmail.com Tue Jun 28 09:57:52 2011 From: fijall at gmail.com (Maciej Fijalkowski) Date: Tue, 28 Jun 2011 09:57:52 +0200 Subject: [pypy-dev] pypy seems not to read site.py In-Reply-To: References: Message-ID: On Tue, Jun 28, 2011 at 9:47 AM, Massa, Harald Armin wrote: > Hello! > I just installed pypy-1.5.0a0-win32, and modified site.py within > lib-python\2.7\site.py > with no effect. > Even entering a big "print haaaaaaaalllllloooo" at the beginning does not do > anything. > Does pypy not use / read / need site.py? I know there is a -S to NOT import > site. > But... how to run site on startup? > Harald It's a little messiness with our python library setup. Modules from modified-2.7 shadow modules from 2.7, this is mostly to avoid having to hack them inline and be able to submit patches back upstream. So site.py is one example - modify one in modified and it'll work. Cheers and sorry for confusion, fijal > -- > GHUM GmbH > Harald Armin Massa > Spielberger Stra?e 49 > 70435 Stuttgart > 0173/9409607 > > Amtsgericht Stuttgart, HRB 734971 > - > persuadere. > et programmare > > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > > From amauryfa at gmail.com Tue Jun 28 10:35:05 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Tue, 28 Jun 2011 10:35:05 +0200 Subject: [pypy-dev] butialo activity In-Reply-To: <201106280734.p5S7Ydgx016901@theraft.openend.se> References: <201106280734.p5S7Ydgx016901@theraft.openend.se> Message-ID: 2011/6/28 Laura Creighton : >>Hi to everyone. >>There's a new activit, I hope that most of you will enjoy, it's called >>Butialo (http://activities.sugarlabs.org/en-US/sugar/addon/4457) >>Butialo is a IDE based on PyPy that allows programming the Buti=E1 robot >>(http://www.fing.edu.uy/inco/proyectos/butia/) >>with Lua. Lua is a very fast, powerfull and easy to use language. The >>IDE provides autodetects the configuration of your Butia robot, and >>provides snippets of code showing how to read sensors and do stuff. >>Best regards > > I looked at those links, and the robot looks ?like a lot of fun. ?But > I am still confused, a bit. ?What exactly did you use pypy for? This seems to be the wrong mailing list. Some links refer to the "Pippy" program, which is part or the OLPC (One Laptop Per Child) project. -- Amaury Forgeot d'Arc From tismer at stackless.com Tue Jun 28 13:17:45 2011 From: tismer at stackless.com (Christian Tismer) Date: Tue, 28 Jun 2011 13:17:45 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> <4E071A12.9040200@stackless.com> Message-ID: <4E09B859.9040105@stackless.com> On 6/26/11 3:15 PM, Sebastien Douche wrote: > On Sun, Jun 26, 2011 at 13:37, Christian Tismer wrote: >> Well, I understand that all. >> Maybe I was implicitly assuming that everybody felt like me: >> It is an honor for Stackless Python to live on python.org, and >> probably also a positive sign, like some acceptance by core python. > From a user point of view, I feel the same thing. > >> That made me wonder. If I had a chance to use python.org instead >> of anything else, I'd always prefer python.org, unless it has a significant >> drawback, or they told me "no, go somewhere else" ;-) > A solution could be to use bitbucket for the day to day job, and push > on python.org regularly (each release?). The cost is very low. > Yes, this sounds like a sensible solution. Thanks for all comments so far, I appreciate this! If it turns out to be the best solution anyway, well, then we could do this for the stackless repos too... I will discuss this with the pypy guys, too - we are sprinting right now in Genua, and also ask the core stackless people. cheers -- chris -- Christian Tismer :^) tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ From tismer at stackless.com Tue Jun 28 13:40:15 2011 From: tismer at stackless.com (Christian Tismer) Date: Tue, 28 Jun 2011 13:40:15 +0200 Subject: [pypy-dev] Bitbucket? Why not python.org In-Reply-To: <1309153223.15293.2.camel@Klappe2> References: <4E02261D.9000602@stackless.com> <4E03610E.9060706@stackless.com> <20110626102750.GB20287@merlinux.eu> <4E071A12.9040200@stackless.com> <20110626122119.GE20287@merlinux.eu> <1309153223.15293.2.camel@Klappe2> Message-ID: <4E09BD9F.2040008@stackless.com> Hi Ronny, I heard you are very knowledgeable on hg and helped pypy quite a lot. May I ask you for some help for stackless as well? Are you on the stackless list? Well, I'll contact you directly. cheers - chris On 6/27/11 7:40 AM, Ronny Pfannschmidt wrote: > Hi, > > having a synced repo on python.org might be a help > however i wouldn't want push-permission handling there, > since we are probably a bit ore lax about giving new talents access to > the repo. > > Simply having a synced repo on pythong.org might have a better effect > for visibility. > > -- Ronny > > On Sun, 2011-06-26 at 12:21 +0000, holger krekel wrote: >> On Sun, Jun 26, 2011 at 13:37 +0200, Christian Tismer wrote: >>> On 6/26/11 12:27 PM, holger krekel wrote: >>>> Hi Christian, >>>> >>>> On Thu, Jun 23, 2011 at 17:51 +0200, Christian Tismer wrote: >>>>> On 6/22/11 7:30 PM, Benjamin Peterson wrote: >>>>>> 2011/6/22 Christian Tismer: >>>>>>> Hi friends, >>>>>>> >>>>>>> the subject line says it all... >>>>>>> >>>>>>> I'm in the progress of updating stackless to use mercurial on >>>>>>> python.org and talked to Martin v. Loewis who pointed out >>>>>>> the restrictions of Bitbucket. >>>>>>> Besides the impression that Bitbucket is pretty slow, it is also not >>>>>>> possible to add our own hooks to it. >>>>>> The impression that Bitbucket is slow? What does that mean? I don't >>>>>> find python.org any faster than Bitbucket. >>>>>> >>>>>>> I'm pretty sure python.org would be happy to host PyPy. >>>>>>> Is there any good reason why we don't ask and move to python.org? >>>>>> I didn't do the mercurial transition, but I'm pretty happy with bitbucket. >>>>> This is no answer but an opinion ;-) >>>>> >>>>> I was asking why we don't use python.org instead of bitbucket. >>>> I don't remember any big comparison analysis. Some people >>>> pushed for bitbucket, a number was using it already, and the others >>>> didn't mind. The main effort and focus was on the conversion of the >>>> svn repository, anyway. >>>> >>>>> Before, we had codespeak.net which was very convenient because >>>>> I knew all relevant people in person. >>>> We have some personal contacts to bitbucket - they actually sponsor an >>>> unlimited plan for PyPy. Moreover, some pypy devs wanted a hosting >>>> solution where we do not depend on private connects or work but can >>>> rather rely on a company basing their business on such hosting. >>>> >>> Well, I understand that all. >>> Maybe I was implicitly assuming that everybody felt like me: >>> It is an honor for Stackless Python to live on python.org, and >>> probably also a positive sign, like some acceptance by core python. >>> >>> That made me wonder. If I had a chance to use python.org instead >>> of anything else, I'd always prefer python.org, unless it has a significant >>> drawback, or they told me "no, go somewhere else" ;-) >>>>> Python used sourceforge before, but preferred to have the freedom >>>>> to host their data themselves. >>>>> By using python.org, PyPy would have similar convenience as >>>>> before. >>>>> Therefore my question: What makes bitbucket the better choice over >>>>> python.org, despite free t-shirts? (which might be an important reason >>>>> for some :-) ) >>>> PyPy has by now quite some integration code wrt to bitbucket. >>>> It seems all are quite happy with bitbucket services as it stands. >>>> So seen from now the question probably rather is why we should >>>> move anywhere else. >>> Well, as said, I see a positive political effect in moving to python.org >>> that I (personally) would not underestimate. >> I agree, it would have this positive political effect so that is clearly >> on the pro side. It did when i moved the mailing lists to mail.python.org. >> >> For what i know, we are welcome to move our repositories to python.org. >> >>> But PyPy is maybe popular enough that my point doesn't really exist, >>> or even vice versa - maybe the distinction is even welcome. ;-) >> Heh, dunno. At this stage it would cause some pain but not much technical >> gain to again move somewhere else. And i think a number of pypy devs >> are happy with not having to care or think about infrastructure issues >> or change of dev habits. >> >> cheers, >> holger >> >> >> >> holger >> >>> cheers - chris >>> >>> -- >>> Christian Tismer :^) >>> tismerysoft GmbH : Have a break! Take a ride on Python's >>> Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ >>> 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ >>> work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 >>> PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 >>> whom do you want to sponsor today? http://www.stackless.com/ >>> >> _______________________________________________ >> pypy-dev mailing list >> pypy-dev at python.org >> http://mail.python.org/mailman/listinfo/pypy-dev -- Christian Tismer :^) tismerysoft GmbH : Have a break! Take a ride on Python's Johannes-Niemeyer-Weg 9A : *Starship* http://starship.python.net/ 14109 Berlin : PGP key -> http://wwwkeys.pgp.net/ work +49 30 802 86 56 mobile +49 173 24 18 776 fax +49 30 80 90 57 05 PGP 0x57F3BF04 9064 F4E1 D754 C2FF 1619 305B C09C 5A3B 57F3 BF04 whom do you want to sponsor today? http://www.stackless.com/ From aguirrea at gmail.com Tue Jun 28 13:03:31 2011 From: aguirrea at gmail.com (=?ISO-8859-1?Q?Andr=E9s?= Aguirre) Date: Tue, 28 Jun 2011 08:03:31 -0300 Subject: [pypy-dev] butialo activity In-Reply-To: References: <201106280734.p5S7Ydgx016901@theraft.openend.se> Message-ID: <1309259011.4238.463.camel@andres-laptop2> Yes, is an IDE based on pippy, sorry for the OT, it must be the pippy mailing list, not the pypy. regards andres El mar, 28-06-2011 a las 10:35 +0200, Amaury Forgeot d'Arc escribi?: > Pippy From chef at ghum.de Tue Jun 28 17:40:29 2011 From: chef at ghum.de (Massa, Harald Armin) Date: Tue, 28 Jun 2011 17:40:29 +0200 Subject: [pypy-dev] bounties for pypy Message-ID: has anyone allready setup a bounties page for pypy-enabling modules? Or had that idea and trashed it, because bounties won't be motivating any of the pypy-capable developers? I am dreaming of something like: module: py2exe person_a pledging x money units for an adaption person_b pledging y money units for an adaption person_c pledging z money units for an adaption collectivly there are (x+y+z) money units pledged; (x+y+z) is over the boredome-treshhold, and it gets realised. a pypy-able guy claims the work, the pledges are collected, and after finishing the money is given to the work-doer. Quite sure there are existing commercial entitities who do exactly this, but is there a subarea for pypy? Or am I just missing the link from the main page? Harald -- GHUM GmbH Harald Armin Massa Spielberger Stra?e 49 70435 Stuttgart 0173/9409607 Amtsgericht Stuttgart, HRB 734971 - persuadere. et programmare -------------- next part -------------- An HTML attachment was scrubbed... URL: From lac at openend.se Tue Jun 28 19:51:44 2011 From: lac at openend.se (Laura Creighton) Date: Tue, 28 Jun 2011 19:51:44 +0200 Subject: [pypy-dev] bounties for pypy In-Reply-To: Message from "Massa, Harald Armin" of "Tue, 28 Jun 2011 17:40:29 +0200." References: Message-ID: <201106281751.p5SHpixC014361@theraft.openend.se> In a message of Tue, 28 Jun 2011 17:40:29 +0200, "Massa, Harald Armin" writes: Hi Harald! Note that we are pypy-dev at python.org these days. >has anyone already setup a bounties page for pypy-enabling modules? No. We just had the idea of doing it very recently. >Or had that idea and trashed it, because bounties won't be motivating any >of the pypy-capable developers? No, but it took the reception of Europython, where I met many people who would pay for GIL removal, for instance, before I got the idea that crowdsourcing would work for us. So we have been discussing this at this sprint. >I am dreaming of something like: > > >module: py2exe > >person_a pledging x money units for an adaption >person_b pledging y money units for an adaption >person_c pledging z money units for an adaption > >collectivly there are (x+y+z) money units pledged; (x+y+z) is over the >boredome-treshhold, and it gets realised. > >a pypy-able guy claims the work, the pledges are collected, and after >finishing the money is given to the work-doer. I think we want to say who wants to do the work, and how soon they could start before we collect the pledges. For things that are complicated, like 'kill the GIL' and 'numpy integration' we would need to do the spelling out of exactly what it is that we would be willing to do. I'm also open to the idea of crowdsourcing the idea of adding feature requests. I'd rather get paid up front, as well, for things that take months and or years to do. >Quite sure there are existing commercial entitities who do exactly this, >but >is there a subarea for pypy? Or am I just missing the link from the main >page? After 4 days of looking I have not found a commercial entity that does this. But I have been talking to fundedbyme, a Swedish based international competitor for kickstarter. The nice thing is that fundedbyme is a Django app. I've been talking to one of the founders, and they would like to share code with us, or set up a way to benefit open source programs in general. When the PyPy Sprint is over, and I am done a few days of vacation, I will return to Sweden, and go meet with these people. They are pypy fans already. In the meantime, if you find a commercial entity which is doing what we want already, do let me know. For me it is a matter of balancing the benefit of being paid by a hundred people in 25 Euro chunks, vs the hassle of having to return money already pledged via paypal or some service if the feature doesn't receive enough funding, (if we collect before the work starts) or going after deadbeat promisers who never pay (if we collect after the work is done). What was clear to me was that there was enough public support for PyPy that we could really live on the cash donations of people wanting to pay what they could to get feature X. The community based support is there. So now I just have to find a way to implement it. > >Harald Send ideas this way. Laura From fuzzyman at gmail.com Tue Jun 28 19:58:11 2011 From: fuzzyman at gmail.com (Michael Foord) Date: Tue, 28 Jun 2011 18:58:11 +0100 Subject: [pypy-dev] bounties for pypy In-Reply-To: <201106281751.p5SHpixC014361@theraft.openend.se> References: <201106281751.p5SHpixC014361@theraft.openend.se> Message-ID: On 28 June 2011 18:51, Laura Creighton wrote: > In a message of Tue, 28 Jun 2011 17:40:29 +0200, "Massa, Harald Armin" > writes: > > Hi Harald! Note that we are pypy-dev at python.org these days. > > >has anyone already setup a bounties page for pypy-enabling modules? > > No. We just had the idea of doing it very recently. > > >Or had that idea and trashed it, because bounties won't be motivating any > >of the pypy-capable developers? > > No, but it took the reception of Europython, where I met many people who > would pay for GIL removal, for instance, before I got the idea that > crowdsourcing would work for us. So we have been discussing this at > this sprint. > > >I am dreaming of something like: > > > > > >module: py2exe > > > >person_a pledging x money units for an adaption > >person_b pledging y money units for an adaption > >person_c pledging z money units for an adaption > > > >collectivly there are (x+y+z) money units pledged; (x+y+z) is over the > >boredome-treshhold, and it gets realised. > > > >a pypy-able guy claims the work, the pledges are collected, and after > >finishing the money is given to the work-doer. > > I think we want to say who wants to do the work, and how soon they > could start before we collect the pledges. For things that are > complicated, like 'kill the GIL' and 'numpy integration' we would > need to do the spelling out of exactly what it is that we would > be willing to do. I'm also open to the idea of crowdsourcing the > idea of adding feature requests. > > I'd rather get paid up front, as well, for things that take months and > or years to do. > > >Quite sure there are existing commercial entitities who do exactly this, > >but > >is there a subarea for pypy? Or am I just missing the link from the main > >page? > > After 4 days of looking I have not found a commercial entity that does > this. But I have been talking to fundedbyme, a Swedish based international > competitor for kickstarter. The nice thing is that fundedbyme is a Django > app. I've been talking to one of the founders, and they would like to > share > code with us, or set up a way to benefit open source programs in general. > When the PyPy Sprint is over, and I am done a few days of vacation, I > will return to Sweden, and go meet with these people. They are pypy > fans already. > Using either kickstarter of fundedbyme to raise money for pypy sounds like a great idea. I'm sure that a lot of the python community would put in *some money* for *specific* goals if it was easy. Michael > > In the meantime, if you find a commercial entity which is doing what > we want already, do let me know. For me it is a matter of balancing > the benefit of being paid by a hundred people in 25 Euro chunks, vs > the hassle of having to return money already pledged via paypal or > some service if the feature doesn't receive enough funding, (if we > collect before the work starts) or going after deadbeat promisers who > never pay (if we collect after the work is done). What was clear to me > was that there was enough public support for PyPy that we could really > live on the cash donations of people wanting to pay what they could > to get feature X. The community based support is there. > > So now I just have to find a way to implement it. > > > > >Harald > > Send ideas this way. > > Laura > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > -- http://www.voidspace.org.uk/ May you do good and not evil May you find forgiveness for yourself and forgive others May you share freely, never taking more than you give. -- the sqlite blessing http://www.sqlite.org/different.html -------------- next part -------------- An HTML attachment was scrubbed... URL: From holger at merlinux.eu Tue Jun 28 20:08:12 2011 From: holger at merlinux.eu (holger krekel) Date: Tue, 28 Jun 2011 18:08:12 +0000 Subject: [pypy-dev] bounties for pypy In-Reply-To: References: Message-ID: <20110628180812.GL20287@merlinux.eu> Hi Harald, On Tue, Jun 28, 2011 at 17:40 +0200, Massa, Harald Armin wrote: > has anyone allready setup a bounties page for pypy-enabling modules? > > Or had that idea and trashed it, because bounties won't be motivating any of > the pypy-capable developers? The idea has been around in some discussions but no concrete actions was taken. I wonder, aren't there websites who organise the work style you describe below already? best, holger > I am dreaming of something like: > > > module: py2exe > > person_a pledging x money units for an adaption > person_b pledging y money units for an adaption > person_c pledging z money units for an adaption > > collectivly there are (x+y+z) money units pledged; (x+y+z) is over the > boredome-treshhold, and it gets realised. > > a pypy-able guy claims the work, the pledges are collected, and after > finishing the money is given to the work-doer. > > Quite sure there are existing commercial entitities who do exactly this, but > is there a subarea for pypy? Or am I just missing the link from the main > page? > > Harald > -- > GHUM GmbH > Harald Armin Massa > Spielberger Stra?e 49 > 70435 Stuttgart > 0173/9409607 > > Amtsgericht Stuttgart, HRB 734971 > - > persuadere. > et programmare > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev From chef at ghum.de Tue Jun 28 21:43:58 2011 From: chef at ghum.de (Massa, Harald Armin) Date: Tue, 28 Jun 2011 21:43:58 +0200 Subject: [pypy-dev] bounties for pypy In-Reply-To: <201106281751.p5SHpixC014361@theraft.openend.se> References: <201106281751.p5SHpixC014361@theraft.openend.se> Message-ID: Hello Laura, Hi Harald! Note that we are pypy-dev at python.org these days. > > yepp, allready in the process of updating my autocomplete :) > No. We just had the idea of doing it very recently. > aaaa... so there was a gnostic fields of information .... > > No, but it took the reception of Europython, where I met many people who > would pay for GIL removal, for instance, before I got the idea that > crowdsourcing would work for us. So we have been discussing this at > this sprint. > > that is great. > >I think we want to say who wants to do the work, and how soon they > >could start before we collect the pledges. Hmmmm .. I hope my english does not work against me: "pledge" I used for "person x commits to pay amount z when the feature is being realised"; so that developers can tune in as soon as enough money is promised for a feature. The idea was also to possibly attract new developers ... for example, if there would be "10 days in money" for adapting py2exe, I am sure many would jump to solve this puzzle. complicated, like 'kill the GIL' and 'numpy integration' we would > need to do the spelling out of exactly what it is that we would > be willing to do. definitely. > I'd rather get paid up front, as well, for things that take months and > or years to do. > > my dream was of a trustee service: after somebody commits to do the work, the pledgers have to pay to a trustee. then the work is done. then the trustee pays the worker. Higher levels would be "pay per completion" or similiar. >In the meantime, if you find a commercial entity which is doing what > >we want already, do let me know. For me it is a matter of balancing > >the benefit of being paid by a hundred people in 25 Euro chunks, vs > >the hassle of having to return money already pledged via paypal or > >some service if the feature doesn't receive enough funding, (if we > >collect before the work starts) or going after deadbeat promisers who > >never pay (if we collect after the work is done). and I added the idea of a trustee. Which would put the situation "not enough people pay" to the trustee, he would have to pay back. Hmmm.... a structure could be: - service provider does the technical stuff, as in: # website # collect pledges # handly project description # collect money # distribute money after featuer completion - PSF / pypy-foundation / whateverfoundation provides the trust Thanks for confirming the need for such a thing! Harald -- GHUM GmbH Harald Armin Massa Spielberger Stra?e 49 70435 Stuttgart 0173/9409607 Amtsgericht Stuttgart, HRB 734971 - persuadere. et programmare -------------- next part -------------- An HTML attachment was scrubbed... URL: From lac at openend.se Wed Jun 29 09:53:22 2011 From: lac at openend.se (Laura Creighton) Date: Wed, 29 Jun 2011 09:53:22 +0200 Subject: [pypy-dev] bounties for pypy In-Reply-To: Message from "Massa, Harald Armin" of "Tue, 28 Jun 2011 21:43:58 +0200." References: <201106281751.p5SHpixC014361@theraft.openend.se> Message-ID: <201106290753.p5T7rMKh002901@theraft.openend.se> >The idea was also to possibly attract new developers ... for example, if >there would be "10 days in money" for adapting py2exe, I am sure many wou >ld >jump to solve this puzzle. This is sort of a bad example. Because py2exe embeds CPython, and we wouldn't want to do that. So what we would probably want to do is to make some general tool that willmake a windows binary, or a mac one, and get rid of the need for bzfreeze and friends. So now you are looking at a general embedding solution, and that is more than 10 days worth of work. But I get the idea. >> my dream was of a trustee service: after somebody commits to do the wor >k, >the pledgers have to pay to a trustee. then the work is done. then the >trustee pays the worker. This is one of the things I want to talk with fundedbyme about. But having an explicit trustee is a new idea. I think the pypy core developers are already rather well trusted in this community, but this may be important to new developers who aren't as well known. And it handles the problem' of 'I got sick and cannot do this any more' more gracefully than other solutions. >Hmmm.... a structure could be: > >- service provider does the technical stuff, as in: > # website > # collect pledges > # handle project description > # collect money > # distribute money after feature completion fundedbyme has sort of indicated an interst in doing this (except they were talking about distribution before, and I was leaving project description to the project, not outsiders). I will follow up on this when I get back home to Sweden. >- PSF / pypy-foundation / whateverfoundation provides the trust > >Thanks for confirming the need for such a thing! > >Harald Thanks once again for seeing a marketing solution that nerds like us often miss. Laura From ian at ianozsvald.com Wed Jun 29 15:16:37 2011 From: ian at ianozsvald.com (Ian Ozsvald) Date: Wed, 29 Jun 2011 14:16:37 +0100 Subject: [pypy-dev] bounties for pypy In-Reply-To: <201106290753.p5T7rMKh002901@theraft.openend.se> References: <201106281751.p5SHpixC014361@theraft.openend.se> <201106290753.p5T7rMKh002901@theraft.openend.se> Message-ID: I'm glad this thread is up. Laura - I'm the chap from Armin's talk who offered a monthly retainer for a year towards numpy integration (in my mind I'm offering ?50/month for 12 months). I spoke to you later and you mentioned the flattr site but having to do it each month is a bit of a pain (I know it is simple but I don't want to think about it...). So, for the record, I have ?600 sitting here with someone's name on it, I'll account for it as a marketing expense or something out of my company. I'm a one man consultancy, PyPy doesn't directly help me as I'm an A.I./science researcher (so I need numpy, IPython, matplotlib etc) but I believe strongly that it will help all of Python (and me in part) over time, so it is worth pledging some of my earnings towards the goal of eventual numpy integration. If I can pledge it to someone or a project then that's cool, if I should just move the money to someone's account then that's cool too. I'm quite happy to have my name down as ContributorNumber1ForNumpy if it helps you spread the word. Ian. ps. I posted the v0.1 PDF of my High Performance Python tutorial this morning (it is based on my EuroPython training session). It has a section on PyPy and I'd happily accept input if that section should be expanded: http://ianozsvald.com/2011/06/29/high-performance-python-tutorial-v0-1-from-my-4-hour-tutorial-at-europython-2011/ On 29 June 2011 08:53, Laura Creighton wrote: > >>The idea was also to possibly attract new developers ... for example, if >>there would be "10 days in money" for adapting py2exe, I am sure many wou >>ld >>jump to solve this puzzle. > > This is sort of a bad example. ?Because py2exe embeds CPython, and > we wouldn't want to do that. ?So what we would probably want to do is > to make some general tool that willmake a windows binary, or a > mac one, and get rid of the need for bzfreeze and friends. ?So now > you are looking at a general embedding solution, and that is more > than 10 days worth of work. > > But I get ?the idea. > > >>> my dream was of a trustee service: after somebody commits to do the wor >>k, >>the pledgers have to pay to a trustee. then the work is done. then the >>trustee pays the worker. > > This is one of the things I want to talk with fundedbyme about. ?But > having an explicit trustee is a new idea. ?I think the pypy core > developers are already rather well trusted in this community, but > this may be important to new developers who aren't as well known. > And it handles the problem' of 'I got sick and cannot do this ?any > more' more gracefully than other solutions. > >>Hmmm.... a structure could be: >> >>- service provider does the technical stuff, as in: >> ? # website >> ? # collect pledges >> ? # handle project description >> ? # collect money >> ? # distribute money after feature completion > > fundedbyme has sort of indicated an interst in doing this (except > they were talking about distribution before, and I was leaving > project description to the project, not outsiders). ?I will follow > up on this when I get back home to Sweden. > >>- PSF / pypy-foundation / whateverfoundation provides the trust >> >>Thanks for confirming the need for such a thing! >> >>Harald > > Thanks once again for seeing a marketing solution that nerds like > us often miss. > > Laura > _______________________________________________ > pypy-dev mailing list > pypy-dev at python.org > http://mail.python.org/mailman/listinfo/pypy-dev > -- Ian Ozsvald (A.I. researcher, screencaster) ian at IanOzsvald.com http://IanOzsvald.com http://SocialTiesApp.com/ http://MorConsulting.com/ http://blog.AICookbook.com/ http://TheScreencastingHandbook.com http://FivePoundApp.com/ http://twitter.com/IanOzsvald From lac at openend.se Thu Jun 30 09:09:27 2011 From: lac at openend.se (Laura Creighton) Date: Thu, 30 Jun 2011 09:09:27 +0200 Subject: [pypy-dev] New Fast machine for benchmarks!! Message-ID: <201106300709.p5U79REJ016166@theraft.openend.se> Yippee!! Everybody please thank Jesse Noller when you see him. :-) Laura ------- Forwarded Message Date: Thu, Jun 30, 2011 at 12:05 AM Subject: speed.python.org machine online To: Python-Dev , pypy-dev at codespeak.net I've posted a more expansive entry on my blog: http://jessenoller.com/2011/06/29/announcing-the-new-speed-python-org-machine/ But the short version, that as discussed at the VM and language summit, we now have a hosted machine dedicated to the running of cross-interpreter speed tests, etc. The hardware was generously donate by HP and the hosting provided, again, free, by OSU/OSL. DL380 HP DL380G7 X5670 LFF (2U) Dual HP NC382i Dual Port Mul?ti?func?tion Giga?bit Server Adapters HP Smart Array P410i/1GB FBWC Controller 4x 4GB (1x4GB) Dual Rank x4 PC3-10600 (DDR3-1333) Reg?is?tered CAS-9 Mem?ory Kit 2x HP 750W Com?mon Slot Gold Hot Plug Power Sup?ply Kit HP iLO Advanced includ?ing 1yr 24x7 Tech?ni?cal Sup?port and Updates Elec?tronic ?License 4x HP 300GB 6G SAS 15K rpm LFF (3.5-inch) Dual Port Enter?prise 3yr War?ranty Hard Drive 2 ? HP DL380 G7 Intel? Xeon? X5680 (3.33GHz/6-core/130W/12MB) FIO Proces?sor Kit With hyperthreading on, the machine has 24 cores, and handily translates pypy using cpython 2.7 in about half the time it typically takes. I am looking forward to handing this over to the team who will be running with the project from here on out - special thanks to Van, Bob Gobeille at HP and the entire OSU/OSL team. jesse ------- End of Forwarded Message From alexandre.fayolle at logilab.fr Thu Jun 30 09:26:50 2011 From: alexandre.fayolle at logilab.fr (Alexandre Fayolle) Date: Thu, 30 Jun 2011 09:26:50 +0200 Subject: [pypy-dev] algorithm used for float -> str conversion Message-ID: <201106300926.50520.alexandre.fayolle@logilab.fr> Hello, I just came across http://www.serpentine.com/blog/2011/06/29/here-be-dragons- advances-in-problems-you-didnt-even-know-you-had/ which talk about a paper submitted last year about a fast float -> str conversion routine. I don't know what is currently used in pypy (or in Python actually), probably the underlying libc's implementation, but this could be something interesting for people writing lots of floats in csv files. Cheers -- Alexandre Fayolle LOGILAB, Paris (France) Formations Python, CubicWeb, Debian : http://www.logilab.fr/formations D?veloppement logiciel sur mesure : http://www.logilab.fr/services Informatique scientifique: http://www.logilab.fr/science From amauryfa at gmail.com Thu Jun 30 10:24:00 2011 From: amauryfa at gmail.com (Amaury Forgeot d'Arc) Date: Thu, 30 Jun 2011 10:24:00 +0200 Subject: [pypy-dev] algorithm used for float -> str conversion In-Reply-To: <201106300926.50520.alexandre.fayolle@logilab.fr> References: <201106300926.50520.alexandre.fayolle@logilab.fr> Message-ID: Hi, 2011/6/30 Alexandre Fayolle : > I just came across http://www.serpentine.com/blog/2011/06/29/here-be-dragons- > advances-in-problems-you-didnt-even-know-you-had/ which talk about a paper > submitted last year about a fast float -> str conversion routine. I don't know > what is currently used in pypy (or in Python actually), probably the > underlying libc's implementation, but this could be something interesting for > people writing lots of floats in csv files. Someone reported the same post on the cpython tracker, and I've just filed a separate issue about this idea. http://bugs.python.org/issue12450 Note that cpython2.7 (and pypy 1.5) already uses a specific algorithm to convert float to strings: a slightly customized version of David Gay's dtoa.c: http://www.netlib.org/fp/dtoa.c it is already faster and more accurate than many libc implementations. -- Amaury Forgeot d'Arc From alexandre.fayolle at logilab.fr Thu Jun 30 11:27:09 2011 From: alexandre.fayolle at logilab.fr (Alexandre Fayolle) Date: Thu, 30 Jun 2011 11:27:09 +0200 Subject: [pypy-dev] algorithm used for float -> str conversion In-Reply-To: References: <201106300926.50520.alexandre.fayolle@logilab.fr> Message-ID: <201106301127.09927.alexandre.fayolle@logilab.fr> On Thursday 30 June 2011 10:24:00 Amaury Forgeot d'Arc wrote: > Hi, > > 2011/6/30 Alexandre Fayolle : > > I just came across > > http://www.serpentine.com/blog/2011/06/29/here-be-dragons- > > advances-in-problems-you-didnt-even-know-you-had/ which talk about a > > paper submitted last year about a fast float -> str conversion routine. > > I don't know what is currently used in pypy (or in Python actually), > > probably the underlying libc's implementation, but this could be > > something interesting for people writing lots of floats in csv files. > > Someone reported the same post on the cpython tracker, > and I've just filed a separate issue about this idea. > http://bugs.python.org/issue12450 > > Note that cpython2.7 (and pypy 1.5) already uses a specific algorithm > to convert float to strings: a slightly customized version of David Gay's > dtoa.c: http://www.netlib.org/fp/dtoa.c > it is already faster and more accurate than many libc implementations. Nice! Thanks a lot for the time machinesque quick answer :-) -- Alexandre Fayolle LOGILAB, Paris (France) Formations Python, CubicWeb, Debian : http://www.logilab.fr/formations D?veloppement logiciel sur mesure : http://www.logilab.fr/services Informatique scientifique: http://www.logilab.fr/science From bokr at oz.net Thu Jun 30 15:21:07 2011 From: bokr at oz.net (Bengt Richter) Date: Thu, 30 Jun 2011 15:21:07 +0200 Subject: [pypy-dev] bounties for pypy In-Reply-To: <201106281751.p5SHpixC014361@theraft.openend.se> References: <201106281751.p5SHpixC014361@theraft.openend.se> Message-ID: On 06/28/2011 07:51 PM Laura Creighton wrote: > In a message of Tue, 28 Jun 2011 17:40:29 +0200, "Massa, Harald Armin" writes: > > Hi Harald! Note that we are pypy-dev at python.org these days. > >> has anyone already setup a bounties page for pypy-enabling modules? > > No. We just had the idea of doing it very recently. > >> Or had that idea and trashed it, because bounties won't be motivating any >> of the pypy-capable developers? > > No, but it took the reception of Europython, where I met many people who > would pay for GIL removal, for instance, before I got the idea that > crowdsourcing would work for us. So we have been discussing this at > this sprint. > >> I am dreaming of something like: >> >> >> module: py2exe >> >> person_a pledging x money units for an adaption >> person_b pledging y money units for an adaption >> person_c pledging z money units for an adaption >> >> collectivly there are (x+y+z) money units pledged; (x+y+z) is over the >> boredome-treshhold, and it gets realised. >> >> a pypy-able guy claims the work, the pledges are collected, and after >> finishing the money is given to the work-doer. > > I think we want to say who wants to do the work, and how soon they > could start before we collect the pledges. For things that are > complicated, like 'kill the GIL' and 'numpy integration' we would > need to do the spelling out of exactly what it is that we would > be willing to do. I'm also open to the idea of crowdsourcing the > idea of adding feature requests. > > I'd rather get paid up front, as well, for things that take months and > or years to do. > >> Quite sure there are existing commercial entitities who do exactly this, >> but >> is there a subarea for pypy? Or am I just missing the link from the main >> page? > > After 4 days of looking I have not found a commercial entity that does > this. But I have been talking to fundedbyme, a Swedish based international > competitor for kickstarter. The nice thing is that fundedbyme is a Django > app. I've been talking to one of the founders, and they would like to share > code with us, or set up a way to benefit open source programs in general. > When the PyPy Sprint is over, and I am done a few days of vacation, I > will return to Sweden, and go meet with these people. They are pypy > fans already. > > In the meantime, if you find a commercial entity which is doing what > we want already, do let me know. For me it is a matter of balancing > the benefit of being paid by a hundred people in 25 Euro chunks, vs > the hassle of having to return money already pledged via paypal or > some service if the feature doesn't receive enough funding, (if we > collect before the work starts) or going after deadbeat promisers who > never pay (if we collect after the work is done). What was clear to me > was that there was enough public support for PyPy that we could really > live on the cash donations of people wanting to pay what they could > to get feature X. The community based support is there. > > So now I just have to find a way to implement it. > >> >> Harald > > Send ideas this way. > > Laura Hm, my (very) few past posts here have found their way around into my newsgroup in-basket, but this one did not. I am guessing it might be because I used "reply-all" and deleted (Newsgroup:) gmane.comp.python.pypy (newsgroup) in favor of leaving just (CC:) pypy-dev in the addressee list. I subscribed via gmane, IIRC, so perhaps I am filtered out if I send to pypy-dev at python.org? Perhaps the above observation is useful, if not the body of my post, repeated below, with this now posted solely to gmane.comp.python.pypy: (We'll see what happens ;-) _______________________________________________________________ A thought (or two): Since it is hard to define end goals exactly, perhaps it would be useful to conceive of a development goal as a later stage in a process, where partial goals are defined as achieved when a new version of an incrementally expanded test suite is passed. Thus passing tests of a certain level can be used as an objective acceptance criterion for distributing rewards defined for that stage. Of course, conceiving and developing tests is also reward-worthy work, and I am not sure how to define objective measurable results for their completion prior to the s/w they test ;-) The main idea is just to break down the work into units whose satisfactory completion can objectively be demonstrated as progress occurs and contributors perhaps only work on certain phases or tests they are interested in (or feel they can make easy money knocking out ;-) This should also make it easier for an interested party to offer a bounty for a specialized feature, by introducing a specialized test, which could be e.g., achieving a given speed on a special benchmark on a particular platform, and/or anything of particular interest to him or her. A public website tying pledges and actual escrowed funds held by the "trustee" to test suite versions (identified with hashing version control), would make level(s) of interest(s) apparent and definitions clear. I guess some bounty offerers may wish to set licensing criteria as well. Is there room for that? Hm, wonder how to write a test program to verify special bounty criteria such as licensing? (Extract and verify by hash delimited licensing boilerplate in the sources or the contributions? Verify existence of a digitally signed legal document in a designated registry? Etc.?) I hope the FOSS ethos can overcome the difficulties of sharing rewards equitably when unequal contributions are all necessary for passing a testing milestone, but none is alone sufficient ;-) Where determining proportion of credit and contributions could be contested, perhaps some thought could preempt unhappy squabbles? Peer voting? How to avoid the temptations of negative competition (e.g. withholding information or its use so as to hobble competitors, to the detriment[1] of the common good) if personal monetary gain depends on "winning" a contest for credit? These are just some thoughts in case the reward process scales up and gets complicated (not intended as a proposal for how to complicate things ;-) Regards, Bengt Richter [1] BTW, this is IMO the essential flaw in the patent/copyright exclusive-rights method of rewarding inventors and authors: it rewards by way of an artificial "right" to create artificial scarcity -- an artificiality particularly obvious in the digital domain. Better IMO to reward authors and inventors by a guaranteed minimum share of eventual profits in the market, rather than creating impediments to the use of ideas in the market through privileged powers of denial that trolls can exploit.