cfbolz changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://botbot.me/freenode/pypy/ ) | use cffi for calling C | mac OS and Fedora are not Windows
speeder39_ has joined #pypy
marky1991 has quit [Read error: Connection reset by peer]
ilbelkyr has joined #pypy
jcea has quit [Quit: jcea]
jacob22__ has quit [Read error: Connection reset by peer]
jacob22__ has joined #pypy
zmt00 has quit [Quit: Leaving]
jacob22__ has quit [Remote host closed the connection]
jacob22__ has joined #pypy
DIRT has joined #pypy
dddddd has quit [Remote host closed the connection]
<kenaan> mattip unicode-utf8-py3 26082fc25722 /pypy/: fix for MAXUNICODE < 65536
<kenaan> mattip unicode-utf8-py3 a6872f6289d6 /rpython/rlib/runicode.py: for py3, MAXUNICODE always 0x10ffff
<bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1870 [mattip: force build, unicode-utf8-py3]
Guest64224 has joined #pypy
tayfun26 has joined #pypy
speeder39_ has quit [Quit: Connection closed for inactivity]
tayfun26 has quit [Quit: tayfun26]
tayfun26 has joined #pypy
riddle has quit [Ping timeout: 272 seconds]
cjwelborn has quit [Ping timeout: 276 seconds]
cjwelborn has joined #pypy
riddle has joined #pypy
arigato has joined #pypy
arigato has quit [Remote host closed the connection]
antocuni has joined #pypy
<cfbolz> piem: that's a question for arigato (who is on an unrelated sprint atm), I think
<fijal> piem: are you sure you run enough iterations?
inad923 has joined #pypy
lesshaste has joined #pypy
<lesshaste> hi all
antocuni has quit [Ping timeout: 244 seconds]
lritter has joined #pypy
mcepl[m] has left #pypy ["User left"]
xorAxAx has joined #pypy
<piem> cfbolz: thanks, i will be patient
<piem> fijal: each row represent 100 runs on the script. the stdev is in gray.
<fijal> piem: try 100.
<fijal> 1000
<piem> fijal: sure, trying now. what change do you expecte to see?
<fijal> I don't know
<fijal> Jit needs time to warm up
marky1991 has joined #pypy
kanaka has quit [Read error: No route to host]
marky1991 has quit [Ping timeout: 272 seconds]
lritter has quit [Quit: Leaving]
tayfun26 has quit [Quit: tayfun26]
Guest64224 has quit [Ping timeout: 264 seconds]
Rhy0lite has joined #pypy
fryguybob has quit [Quit: leaving]
antocuni has joined #pypy
fryguybob has joined #pypy
adamholmberg has joined #pypy
<mdash> fijal: hi. here's the rffi bits giving me a headache: https://github.com/monte-language/typhon/blob/master/typhon/ruv.py#L628
<mdash> fijal: my primary question here is whether scoped_alloc is compatible with the kind of thing we're trying to do on lines 570 and 589
<mdash> fijal: it seems like ll2ctypes is allocating a new array when making the set_stdio_stream call on line 639
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
adamholmberg has quit [Read error: Connection reset by peer]
adamholmberg has joined #pypy
<piem> fijal: increasing to 1000 shows more variations across runs: https://aubio.org/images/cffi_timings_rPmm50.png
<piem> fijal: for now i'm mostly interested in reducing the reading time for cpython+cffi, which will be the use case for most users
<fijal> Right so cpython cffi would not speed up with more runs
<fijal> But it's also not something that we can do something about
<bbot2> Failure: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1870 [mattip: force build, unicode-utf8-py3]
jcea has joined #pypy
<kenaan> antocuni gc-disable 2db479ad4b70 /: improve the return value of gc.collect_step: now it returns a GcCollectStepStats, the same which is passed to...
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5751 [antocuni: force build, gc-disable]
antocuni has quit [Ping timeout: 250 seconds]
DIRT has quit [Ping timeout: 240 seconds]
Zaab1t has joined #pypy
<bbot2> Success: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5751 [antocuni: force build, gc-disable]
<mattip> ronan: thanks, setting sys.MAXUNICODE made win32 pass many more tests
atomizer_ has joined #pypy
thnee1 has joined #pypy
tazle___ has joined #pypy
ilbelkyr_ has joined #pypy
commandoline_ has joined #pypy
larstiq_ has joined #pypy
dan-- has joined #pypy
dan-- has joined #pypy
dan-- has quit [Changing host]
lesshaste has quit [*.net *.split]
hexa- has quit [*.net *.split]
Hasimir has quit [*.net *.split]
__peke__ has quit [*.net *.split]
marvin_ has quit [*.net *.split]
commandoline has quit [*.net *.split]
LarstiQ has quit [*.net *.split]
dan- has quit [*.net *.split]
atomizer has quit [*.net *.split]
Remi_M has quit [*.net *.split]
thnee has quit [*.net *.split]
tazle has quit [*.net *.split]
ilbelkyr has quit [*.net *.split]
marvin has joined #pypy
dan-- is now known as dan-
Hasimir has joined #pypy
lesshaste has joined #pypy
lesshaste is now known as Guest58261
hexa- has joined #pypy
__peke__ has joined #pypy
Remi_M has joined #pypy
Zaab1t has quit [Quit: bye bye friends]
Guest58261 has quit [Quit: Leaving]
arigato has joined #pypy
ilbelkyr_ is now known as ilbelkyr
Rhy0lite has quit [Quit: Leaving]
jamesaxl has quit [Quit: WeeChat 2.2]
speeder39_ has joined #pypy
<arigato> cfbolz: the issue #2904 does indeed look like some unrolling fun
<cfbolz> arigato: yes :-(
<cfbolz> arigato: :I started reading the trace but gave up for tonight
<arigato> I've posted another tweaked example that might give more clues
<cfbolz> arigato: cool
<bbot2> Started: http://buildbot.pypy.org/builders/own-win-x86-32/builds/1871 [mattip: force build, unicode-utf8-py3]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4181 [mattip: force build, unicode-utf8-py3]
<cfbolz> arigato: note that the trace that gets produced is a bit funny. Do I see it correctly that it starts with the while, and then goes via the break through the for loop back to the while again?
<bbot2> Started: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7034 [mattip: force build, unicode-utf8-py3]
<bbot2> Started: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5752 [mattip: force build, unicode-utf8-py3]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/4181 [mattip: force build, unicode-utf8-py3]
<arigato> cfbolz: I didn't get to know that, I was expecting a backend issue and debugging assembly
<cfbolz> arigato: ah
<cfbolz> arigato: your debug print shows that it's a frontend problem though, right?
<arigato> yes, I gave up when I realized that and posted my current debug_print solution to the issue
<cfbolz> Right
<cfbolz> Fun times ahead :-(
<arigato> and indeed, copying the debug_print at the end of the loop shows that 'height' is equally broken there, which makes sense if the end of the loop is always just before the start of the next iteration
<cfbolz> That height changes at all is just weird
<arigato> yes
<cfbolz> Maybe it's some kind of weird state matching problem
<arigato> as a clue, I think there is somewhere close a short preamble taken
<arigato> unrelated, but I'm sending a mail to everbody on pypy-dev---but with a different subject for everybody. Maybe that will let us track who forwards his mails to the westhost issue tracker
<cfbolz> arigato: ah, you are sending these mails spread out across time to see who it is?
<arigato> no, just with a different subject. They are also spread out across time but that's only because I can't easily send 833 mails at once
<cfbolz> Ah, subject, makes sense
<arigato> the issue tracker's reply contains the same subject as the original mail
<cfbolz> Cool
[Arfrever] has joined #pypy
<arigato> found it
tsutsumi has quit [Quit: Ping timeout (120 seconds)]
tsutsumi has joined #pypy
marky1991 has joined #pypy
marky1991 has quit [Read error: Connection reset by peer]
<arigato> piem: you were asking about cffi, some not completely clear question
<arigato> if by "reading the code generated by cffi" you mean the C code generated by cffi, then yes, it only contains CPython-specific argument parsing if compiled with CPython
<arigato> we PyPy we choose the simpler solution of not providing it, and instead relying on libffi to do the calls when not JITted, and on our JIT when JITted
<arigato> we could have done the same with CPython, but the cost would never be removed, so it's more noticeable
<arigato> regarding your timings, it's hard to conclude anything because they are all close to each other
<arigato> and well, we can't guess what exactly is the code doing
adamholmberg has quit [Remote host closed the connection]
<bbot2> Failure: http://buildbot.pypy.org/builders/own-linux-x86-64/builds/7034 [mattip: force build, unicode-utf8-py3]
<bbot2> Failure: http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/5752 [mattip: force build, unicode-utf8-py3]
arigato has quit [Remote host closed the connection]
TheAdversary has quit [Disconnected by services]
plm has joined #pypy
Khayman has joined #pypy
TheAdversary has joined #pypy
Hasimir has quit [Ping timeout: 272 seconds]
speeder39_ has quit [Quit: Connection closed for inactivity]
adamholmberg has joined #pypy
DIRT has joined #pypy
adamholmberg has quit [Remote host closed the connection]
adamholmberg has joined #pypy
adamholmberg has quit [Ping timeout: 245 seconds]