#pypy IRC log for Friday, 2011-12-09

tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer00:00
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.00:01
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer00:03
tilgovi (~randall@99.13.242.166) joined #pypy.00:03
tilgovi (~randall@99.13.242.166) left irc: Changing host00:03
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.00:03
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 248 seconds00:08
amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 244 seconds00:16
etrepum_ (~bob@accessnat4.mochimedia.net) joined #pypy.00:25
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) left irc: Remote host closed the connection00:27
etrepum (~bob@accessnat4.mochimedia.net) left irc: Ping timeout: 240 seconds00:28
etrepum_ (~bob@accessnat4.mochimedia.net) left irc: Ping timeout: 244 seconds00:30
kennethreitz (~kennethre@204.14.152.118) left irc: Ping timeout: 244 seconds00:32
alistra (alistra@jasminek.net) joined #pypy.00:38
alistraoh, who's going to tell me, why during the compilation, there are ascii fractals written to stdout00:39
nedbat (~nedbat@python/psf/nedbat) joined #pypy.00:39
Alex_Gaynoralistra: http://doc.pypy.org/en/latest/faq.html#why-does-pypy-draw-a-mandelbrot-fractal-while-translating00:40
alistrait's not fun :|00:41
alistraalso haven't tried too hard, but i see clang is not supported :P 00:42
alistra(at least my build crashed with it)00:42
Alex_Gaynorin theory it's supported if you compile with --gcrootfinder=shadowstack00:43
alistrawell i would have to do it by hand (now trying with gentoo's emerge)00:44
gutworth (~benjamin@dhcp-128-36-84-91.central.yale.edu) left irc: Quit: leaving00:44
Alex_Gaynorok, I know nothing about gentoo or the pypy package for it00:44
alistrawell basically gentoo has a package menager where you can compile stuff00:44
alistraand you can CC=clang emerge pypy00:45
alistrathen it uses clang instead of gcc :}00:48
Alex_GaynorI figured as much :)00:48
alistraalso emerge is written i python, would love to try it out with pypy00:49
jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy.00:51
alistrai sure would like a progress bar more than the mandelbrot :300:54
Alex_Gaynoralistra: it'd be like a windows progress bar if we had one :)00:54
alistrai don't understand the joke (if it's one) because i'm not familliar with windows' progress bar00:55
Alex_Gaynorthey tend to jump around.  30 seconds, 10 minutes, done,for no aparent reason00:57
Alex_Gaynorhttp://xkcd.com/612/00:58
alistraah those ones00:58
alistrathe old ones00:58
Alex_Gaynoryeah :)00:58
alistranot sure if they're the same in the new windows00:58
alistrai didn't say nothing about an eta00:58
Action: Alex_Gaynor tires to do as little as possible on windows00:58
alistraas we all do00:59
ericholscher (~eric@173.203.223.63) left irc: Ping timeout: 258 seconds00:59
pjenvey (~pjenvey@underboss.org) left irc: Ping timeout: 255 seconds00:59
pjenvey (~pjenvey@underboss.org) joined #pypy.01:03
ericholscher (~eric@173.203.223.63) joined #pypy.01:03
Husio (husiatyn@oceanic.wsisiz.edu.pl) left irc: Ping timeout: 252 seconds01:05
zain__ (~textual@c-67-160-201-63.hsd1.ca.comcast.net) joined #pypy.01:05
Husio (husiatyn@oceanic.wsisiz.edu.pl) joined #pypy.01:06
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/16401:07
whitelynx|work (~whitelynx@63.241.75.144) left irc: Quit: Ex-Chat01:09
alistrai surely understand why the buildbot, compiling now for half an hour01:14
overminder (overminder@wf126-114.ust.hk) joined #pypy.01:14
derdon (~derdon@p5DE8A137.dip.t-dialin.net) left irc: Remote host closed the connection01:22
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.01:22
Nick change: Gulaway -> Gulopine01:33
mtigas (~mtigas@users.spokesman.com) left irc: Quit: mtigas01:35
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/188201:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/62001:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/63601:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/114901:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/82101:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/146301:36
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/74301:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/28701:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/29901:36
neunon (~snoonan@unaffiliated/neunon) joined #pypy.01:40
xiaochen (xiaochen@nat/redhat/x-bmajhmdgopxhxoym) joined #pypy.01:46
gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy.01:47
mfoord (~anonymous@python/psf/voidspace) left irc: Quit: mfoord01:56
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/28701:58
xiaochen (xiaochen@nat/redhat/x-bmajhmdgopxhxoym) left irc: Quit: leaving02:00
xiaochen (xiaochen@nat/redhat/x-ndgjrxdrmgjjbvxa) joined #pypy.02:03
voidspace (~voidspace@python/psf/voidspace) left irc: Quit: Uhm... gotta go02:05
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer02:06
Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Ping timeout: 240 seconds02:09
linq (~ident@24-246-25-39.cable.teksavvy.com) joined #pypy.02:10
rokujyouhitoma (~rokujyouh@173.227.61.4) left irc: Remote host closed the connection02:28
rokujyouhitoma (~rokujyouh@173.227.61.3) joined #pypy.02:31
rokujyouhitoma (~rokujyouh@173.227.61.3) left irc: Remote host closed the connection02:36
Moku (~John@osbk-4db17084.pool.mediaWays.net) joined #pypy.02:40
Nick change: Moku -> Guest9549702:40
Shanita (~John@osbk-4db1538a.pool.mediaWays.net) left irc: Ping timeout: 240 seconds02:40
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/96902:46
Niedar (gfghfghf@host-173-230-2-133.vanodod.clients.pavlovmedia.com) left irc: Read error: Connection reset by peer02:54
Count_Niedar (gfghfghf@host-173-230-2-133.vanodod.clients.pavlovmedia.com) joined #pypy.02:54
dmalcolm (~david@c-24-61-12-82.hsd1.ma.comcast.net) left irc: Quit: Leaving02:58
gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Quit: Konversation terminated!02:59
linq (~ident@24-246-25-39.cable.teksavvy.com) left irc: Ping timeout: 240 seconds03:00
overminder (overminder@wf126-114.ust.hk) left irc: Ping timeout: 240 seconds03:01
overminder (overminder@wf213-202.ust.hk) joined #pypy.03:06
fprimex_ (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.03:08
Nick change: Gulopine -> Gulaway03:10
Nick change: Gulaway -> Gulopine03:23
thinker (~thinker@203.69.99.17) joined #pypy.03:28
fprimex_ (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: quit03:29
Nick change: zain__ -> zain03:30
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: Leaving03:30
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.03:32
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 260 seconds03:36
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/188203:47
harrison (~quassel@adsl-69-209-212-18.dsl.chcgil.ameritech.net) left irc: Ping timeout: 244 seconds03:48
brutal_chaosFinally. A comparison of the prebuilt 64bit asmgcc pypy-1.7 and a shadowstack (built with clang) benchmark has finished and been submitted to the ml.04:02
ericflo (~ericflo@75.103.8.110) left irc: Quit: ericflo04:18
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: quit04:36
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.04:39
azanella (~azanella@189.6.80.207) left irc: Ping timeout: 252 seconds04:44
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/146304:46
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/82104:49
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: quit05:05
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/63605:07
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.05:10
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/62005:15
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/114905:17
rokujyouhitoma (~rokujyouh@12.51.221.130) joined #pypy.06:14
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/74306:16
Nick change: Gulopine -> Gulaway06:20
anish_ (anish@nat/redhat/x-bzrtwcbiqqfztqwp) joined #pypy.06:20
whitelynx (~whitelynx@75.111.197.204) left irc: Quit: Leaving06:21
mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 244 seconds06:26
nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 244 seconds06:29
mwhudson (~mwh@linaro/mwhudson) left irc: Ping timeout: 252 seconds06:29
arigato (~arigo@89.204.154.78) joined #pypy.06:43
nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.06:43
asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer06:46
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: quit07:04
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/29907:09
overminder (overminder@wf213-202.ust.hk) left irc: Ping timeout: 240 seconds07:11
mwhudson (~mwh@121.73.77.183) joined #pypy.07:13
mwhudson (~mwh@121.73.77.183) left irc: Changing host07:13
mwhudson (~mwh@linaro/mwhudson) joined #pypy.07:13
kenaan12arigo default 111edf7af77df6 15/pypy/jit/metainterp/test/test_del.py: A failing test: keepalive_until_there is not really strong enough.07:13
squiddy (~squiddy@f053084199.adsl.alicedsl.de) joined #pypy.07:17
rokujyouhitoma (~rokujyouh@12.51.221.130) left irc: Remote host closed the connection07:21
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy.07:29
overminder (overminder@wf120-054.ust.hk) joined #pypy.07:36
nedbat (~nedbat@python/psf/nedbat) joined #pypy.07:37
asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy.07:40
fijal (~fijal@197.170.59.231) joined #pypy.07:42
arigatohi07:44
arigatofijal: there is nothing about light finalizers in the jit, so far, I think, unlike what you implied yesterday?07:44
fijaloh ok07:44
fijalmaybe I just wanted to implement them?07:44
arigato:-)07:45
arigatoI'm trying to get keepalives to work in all cases, but it looks like a mess07:45
fijalall cases being?07:45
arigatothrough the jit07:45
arigatoright now the support is minimal, but I haven't even managed to make sure if it's enough or not07:46
Insectoid (~Q@c-98-230-8-37.hsd1.fl.comcast.net) joined #pypy.07:46
InsectoidI've been trying to translate PyPy on my own for a while now. Windows translations are broke and have been for a while, but building the rev of the latest binary release would always blow up with an RPython MemoryError or similar at around 2G of memory usage. Today I sat down and took a look, and it turns out there's a very simple fix that you can incorporate into your build process for releases. With Visual Studio comes a tool, binedit.exe. I07:46
arigatoInsectoid: your line was too long, it was truncated after "binedit.exe"07:47
arigatobut we know about it, yes07:47
InsectoidIf you run binedit /LARGEADDRESSAWARE pypy.exe, PYPY can then use more than 2G of memory without crashing. This resulted in a successful translation on my machine :) I would have reported this on Bitbucket, but I'm blind, and they have an inaccessible CAPTCHA.07:48
arigato:-/07:49
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Ping timeout: 252 seconds07:49
arigatothanks!  I think that Christian Tismer, our Windows developer, already knows about it07:49
MostAwesomeDudeInsectoid: Awesome find. Thank you.07:50
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy.07:51
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 258 seconds07:51
arigatoInsectoid: fwiw, actually the bugs go to http://bugs.pypy.org/, which is not bitbucket and has (I think) no captcha07:52
InsectoidMy mistake. I should have looked instead of just assuming that the codebase and bugtracker were hosted in the same place.07:52
arigatowell it's not really a bug, but it can be filed as an issue anyway.  I'll write it down there07:52
InsectoidShould it throw an https cert error?07:53
arigatoyes :-(07:54
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Remote host closed the connection07:54
InsectoidI've also been trying to get wxPython to build. I needed my own translation to accomplish this because I needed pypy-c.lib. I got a successful build but...07:54
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy.07:54
InsectoidImporting it I get an error dialog that a wx dll cannot be found, and the dll is in the wx directory correctly. Would I report this as a PyPy bug... a WX bug (that won't be fixed) ... ?07:55
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Read error: Connection reset by peer07:55
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy.07:55
kenaan12arigo default 115a510c5ffc55 15/pypy/jit/metainterp/test/test_del.py: bah.07:56
arigatoInsectoid: please report it as a PyPy bug07:57
Insectoidarigato: Okay, thank you.07:58
arigatoit looks like the issue is really on the side of PyPy (on Windows at least)07:58
myf (~myf@li245-175.members.linode.com) joined #pypy.08:00
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Ping timeout: 252 seconds08:00
antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) joined #pypy.08:00
nephics (~nephics@78-72-71-194-no33.tbcn.telia.com) joined #pypy.08:01
nopper (~nopper@ns.patched.biz) left irc: Remote host closed the connection08:02
myfhowdy, anyone's awake at this moment08:03
jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Remote host closed the connection08:06
nopper (~nopper@ns.patched.biz) joined #pypy.08:07
MostAwesomeDudemyf: Yes, but that's not really what you wanted to ask. :308:07
myfyou are awesome!08:08
myfmy question is that if any of you has experience with beautifulsoup in pypy08:08
myfand why is it super slowwww08:09
fijalmyf: no idea08:11
MostAwesomeDudeDo you have an example?08:11
MostAwesomeDudeIs it still slow under lxml?08:11
fijaldoes it use a C extension?08:11
fijalMostAwesomeDude: lxml does not work under pypy08:11
MostAwesomeDudefijal: Aw, really? :c08:11
fijalrly08:12
myfi'm just running it with some script i wrote and time it then compare it with python result....08:12
fijalmyf: maybe you're using different xml parsers or so?08:12
myfso the problem is lxml?08:12
fijalI have a bit no idea08:12
fijalwould be cool to have a reproducible example, so I can have a look08:13
myfwas using some script i wrote for this small startup i used to work, so probably not gonna show it. but it's extremely simply08:13
myfuse beautifulsoup to parse08:14
pjenveyisn't it heavy on regular expressions?08:14
myfno re at all08:14
myfyou mean beautifulsoup?08:14
pjenveyyea08:14
myfit's written in python i think, but I have never looked at the source08:15
myfohhhhh, also in order to download the page there's urllib08:15
fijalmyf: we can't help you if we can't have a look08:17
overminder (overminder@wf120-054.ust.hk) left irc: Read error: Connection reset by peer08:20
overminder (overminder@wf120-054.ust.hk) joined #pypy.08:20
CIA-1303mongoose_q 07roundup * 10#889/Windows binary distribution is missing cpyext header files: 08:21
CIA-13This is more general. wxPython's build process does the same thing. I had to08:21
CIA-13copy pypy-c.lib from %tmp%\usession-unknown-17\testing_1 into the bu ... * 14https://bugs.pypy.org/issue88908:21
pjenveyyea it's written in python, I seem to recall it being re heavy but my memory is from a long time ago08:22
pjenveyif you can narrow down the slowness to a simple benchmark someone could take a look08:24
myfok i've got an example08:26
myfhttp://pastebin.com/zK3jSxFf08:26
myfvery simple script08:26
myfand you can feed just some random html file for comparison08:27
myfmyf@ctesiphon ~> python beautiful.py wea.html 08:27
myf0.00086808204650908:27
myfmyf@ctesiphon ~> pypy beautiful.py wea.html 08:27
myf0.0021789073944108:27
myfthat was my result08:27
xiaochen (xiaochen@nat/redhat/x-ndgjrxdrmgjjbvxa) left irc: Ping timeout: 252 seconds08:29
kirmadoes running time increase linearly with the input size for python, but polynomially for pypy? it could be stupid string concatenation relying on a hackish cpython optimization if so08:31
pjenveythat tells me pypy is 3 times slower on a millisecond benchmark, basically measuring startup time. no time for the jit to warmup08:31
myfyea this is pretty simple08:32
kirmapypy is probably not the best idea for such short-lived processes, anyway08:33
myfbut on my other scrpit to process some page, python does in 6 second and pypy does it in 1208:33
myfi can probably come up with a better one if i spend more time08:33
myfbut oher than hacing jit to warm up, could there be any other speculations?08:34
kirmayou might want to figure out if implementations have different time complexity, or if they more or less differ by a constant multiplier08:35
arigatoyou're expecting a process that takes 0.8 ms to become faster?08:39
arigatodon't look at pypy08:39
arigatoplease come up with an example that takes at least a few seconds08:40
apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) joined #pypy.08:41
myfprhs08:42
myfi'll come up with something tomorrow08:42
myftoo tired to function08:42
myfanyone got coffee08:42
snoonan (~snoonan@2002:43a1:5550:1:f986:8f2c:ba61:c123) joined #pypy.08:43
snoonan (~snoonan@2002:43a1:5550:1:f986:8f2c:ba61:c123) left irc: Client Quit08:43
thinker (~thinker@203.69.99.17) left irc: Remote host closed the connection08:43
kirmadoes spilling coffee on my irc window help? :)08:45
neunon (~snoonan@unaffiliated/neunon) left irc: Ping timeout: 240 seconds08:45
fijal (~fijal@197.170.59.231) left irc: Ping timeout: 248 seconds08:45
myfwhale whale whale08:46
myfthere's a nasty huffingtonpost one08:46
dimazesthi, what would be the best web server to run a flask wsgi app together with pymongo?08:47
dimazeston pypy, ofcourse :)08:47
myfmyf@ctesiphon ~> python beautiful.py http://www.huffingtonpost.com/2011/12/08/richard-cordray-nomination-obama-cfpb_n_1136243.html08:47
myf0.65220093727108:47
myfmyf@ctesiphon ~> pypy beautiful.py http://www.huffingtonpost.com/2011/12/08/richard-cordray-nomination-obama-cfpb_n_1136243.html08:47
myf1.4529001712808:47
myfbut other than that i'd come up with better ideas tomorrow08:47
myfthanks yall08:47
myftill then08:48
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) joined #pypy.08:48
myfhttp://pastebin.com/B7BG1UBh08:49
hruske (~Gasper@188.230.156.183) joined #pypy.08:55
antocuninice, bitbucket's blog uses stringobject.py to showcase the new side-by-side diff08:56
antocunihttp://blog.bitbucket.org/2011/12/08/pull-requests-with-side-by-side-diffs/08:56
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) left irc: Ping timeout: 240 seconds08:57
fijal (~fijal@41.48.188.39) joined #pypy.08:59
myfaight, before i pass out, does half a second benchmark difference significant in this case?09:01
arigatomyf: it depends a bit.  usually yes, but it might also be the case that there are really tons and tons of regexps and code to compile09:02
arigato(in which case it's still an interesting benchmark for speeding up warm-up time)09:02
myfarigato: so regex might be the bottleneck in this case09:02
arigatono, the warm-up time only09:03
myfans someone mentioned that beautifulsoup is heavy on re09:03
kenaan12bivab benchmarks 11dece54731283 15/runner.py: sort benchmark names09:03
arigatoit's hard to guess, you need to run several benchmarks that take longer and longer, and see how it converges09:03
bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy.09:03
myfright09:03
arigatobut yes, your timings look like "it's taking too long on pypy" holds true anyway09:04
xiaochen (xiaochen@nat/redhat/x-bxrrxyanirdlhsmo) joined #pypy.09:05
xiaochen (xiaochen@nat/redhat/x-bxrrxyanirdlhsmo) left irc: Client Quit09:05
myfthanks for the answer, a lot to learn!09:05
xiaochen (xiaochen@nat/redhat/x-kvhpwlnsxntnxafj) joined #pypy.09:07
cfbolzInsectoid: hi09:11
cfbolzInsectoid: I think a bug should be filed about the captcha of bitbucket09:11
cfbolzthat's not really acceptable09:11
G2P (~G2P@fw-asn1.ornis.com) joined #pypy.09:14
arigato (~arigo@89.204.154.78) left irc: Ping timeout: 252 seconds09:14
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) joined #pypy.09:15
fijal (~fijal@41.48.188.39) left irc: Ping timeout: 240 seconds09:15
mwhudson (~mwh@linaro/mwhudson) left irc: Ping timeout: 244 seconds09:15
hruske (~Gasper@188.230.156.183) left irc: Remote host closed the connection09:19
cpets (c358b60a@gateway/web/freenode/ip.195.88.182.10) left irc: Quit: Page closed09:20
Insectoidcfbolz: I agree. I usually have a way to solve them but the service recently shut down.09:22
cfbolzInsectoid: ah, but I guess you need to solve a captcha to file a bug09:22
cfbolzgrr09:22
cfbolzInsectoid: I'll file it09:23
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) left irc: Ping timeout: 248 seconds09:25
lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy.09:26
ronan (~ronan@host86-142-249-87.range86-142.btcentralplus.com) left irc: Remote host closed the connection09:36
fijal (~fijal@197.169.91.5) joined #pypy.09:39
ojii (~ojii@40-34.60-188.cust.bluewin.ch) joined #pypy.09:39
overminder (overminder@wf120-054.ust.hk) left irc: Ping timeout: 240 seconds09:40
asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer__09:41
fijalhey09:42
cpets (c358b60a@gateway/web/freenode/ip.195.88.182.10) joined #pypy.09:45
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy09:51
fijalcfbolz: catch 22 :)09:58
fijalcfbolz: http://www.dukaj.pl/bibliografia/publikacje/TheDedalusBookOfPolishFantasy10:00
fijalif you can reach it *somehow*10:00
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) joined #pypy.10:03
brutal_chaosfijal: your bencmarks await10:03
cfbolzfijal: catch-22 I read recently. probably one of the funniest books ever10:05
brutal_chaoscfbolz: i don't think it was meant to be funnyt10:06
brutal_chaosfunny*]\10:06
brutal_chaosdamn, I cannot type.10:06
brutal_chaosfunny*10:06
cfbolzbrutal_chaos: yes, so?10:06
cfbolzbrutal_chaos: neither was absurd theater meant to be funny, but to express a deep problem. still funny10:07
anish_ (anish@nat/redhat/x-bzrtwcbiqqfztqwp) left irc: Quit: Leaving10:07
cfbolzfijal: the other book seems hard to get indeed10:07
ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) joined #pypy.10:07
fijalbrutal_chaos: my benchmarks?10:08
brutal_chaosllvm/clang vs asmgcc10:08
fijalcfbolz: :(10:08
fijalbrutal_chaos: well, it's slower on clang & shadowstack10:08
fijalcfbolz: I love Dukaj, just got on of his books in polish10:09
fijalmaybe german?10:09
fijalhttp://www.dukaj.pl/bibliografia/publikacje/AufDerStrasseNachOodnadatta10:09
brutal_chaosfijal: well, i didn't know how to understand your stdout....10:09
brutal_chaoslol10:09
cfbolzfijal: that seems cheaper, at least used 95 cents10:10
fijalbrutal_chaos: ah ok, the asmgcc is ~5% faster10:10
fijalsometimes 15% faster10:10
brutal_chaosah alright, good to know.10:10
fijalcfbolz: cool, although there is only one I can recommend there, I don'10:11
fijalt know about others10:11
fijalone story that is10:11
cfbolzfijal: for 95 cents...10:11
fijalcfbolz: sounds like a good deal to me :)10:11
cfbolzfijal: of course there is 3 euro shipping :-)10:12
fijalif you don't like it, I can give you back your 95 cents ;-)10:12
Nick change: wulczer-afk -> wulczer10:12
wulczer (wulczer@wulczer.org) left #pypy ("ERC Version 5.3 (IRC client for Emacs)").10:12
Alex_Gaynorfijal: morning10:14
fijalAlex_Gaynor: hey10:14
fijalcfbolz: I meant catch-22 with bitbucket and bugs btw10:14
fijalAlex_Gaynor: can I make those html5 slides offline btw?10:15
Alex_Gaynorfijal: sure10:15
cfbolzfijal: yes, got that10:15
fijalcfbolz: ok :)10:15
fijalAlex_Gaynor: link to the site?10:15
Alex_Gaynorfijal: http://code.google.com/p/html5slides/ you mean?10:16
fijalyes10:16
fijalis it cool to use?10:16
Action: fijal is making some slides for pug10:16
Alex_Gaynoryeah, the end results look pretty nice10:16
cfbolzfijal: btw, did you see that nbody is now faster by 20% than before, with the introduction of float list?10:16
Alex_Gaynorcfbolz: so the unpacking thing fixed it?10:16
fijalcfbolz: cool!10:17
fijalAlex_Gaynor: I miss being able to come back10:17
Alex_Gaynorfijal: ?10:17
fijalbecause once you had an animation, you can't10:17
Alex_Gaynorah yeah10:17
fijallike bullets or code changes10:17
cfbolzAlex_Gaynor: seems so10:17
Alex_Gaynorcfbolz: I guess benchmarks work then :)10:18
fijalhttp://speed.pypy.org/timeline/?exe=1&base=none&ben=django&env=1&revs=50&equid=off10:18
fijalthis one is still gloomy10:19
fijalcfbolz: btw, some benchmarks like mako are incredibly fast on tannit10:19
fijalrespective to cpython10:19
fijalit's 2x compared to my laptop10:20
fijalpypy/cpython is 2x on tannit, 1.1x on my machine10:20
fijalit has something to do with nursery size (that's mako)10:20
fijalanyway, /me surf break10:20
voidspace (~voidspace@python/psf/voidspace) joined #pypy.10:23
arigato (~arigo@fwstups.cs.uni-duesseldorf.de) joined #pypy.10:23
arigatocan I use tannit to run more benchmarks?10:26
arigatoi.e. is nobody planning to use tannit in the next 2 hours?10:26
overminder (overminder@wf120-054.ust.hk) joined #pypy.10:27
voidspace (~voidspace@python/psf/voidspace) left irc: Read error: No route to host10:29
voidspace (~voidspace@python/psf/voidspace) joined #pypy.10:29
kenaan12arigo temp2 1103796662a8a0 15/pypy/jit/backend/: hg backout 062e9d06c90810:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1150 [12arigo, temp2]10:37
kenaan12arigo SpecialisedTuples 114efbd07c3e55 15/pypy/: hg merge default10:38
xiaochen (xiaochen@nat/redhat/x-kvhpwlnsxntnxafj) left irc: Quit: leaving10:38
khs (~khs@2001:700:300:2120:725a:b6ff:fee5:a44) joined #pypy.10:38
cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) left irc: Quit: Leaving10:44
fijal (~fijal@197.169.91.5) left irc: Read error: Connection reset by peer10:47
bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Remote host closed the connection10:48
bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy.10:48
bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Remote host closed the connection10:48
bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy.10:48
bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Remote host closed the connection10:49
bivab_ (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy.10:49
mfoord (~anonymous@python/psf/voidspace) joined #pypy.10:49
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) left irc: Ping timeout: 244 seconds10:53
Husio (husiatyn@oceanic.wsisiz.edu.pl) left irc: Ping timeout: 244 seconds10:54
Husio (husiatyn@oceanic.wsisiz.edu.pl) joined #pypy.10:55
stakkars_ (~tismer@89.204.139.170) joined #pypy.10:55
stakkars_howdy!10:56
stakkars_what was the translation option for compiling with multi-CPU? I can't find the option in the --help10:56
stakkars_arigato: now I know why my translation isn't _that_ fast: the machine is big, but only 1.6 GHz :-)11:01
stakkars_sigh, good to know11:01
arigatostakkars_: multiple threads: it's not a built-in capability, but "make" accepts a "-j" option on Unix11:02
arigatothis can be set with the env var MAKEFLAGS=-j411:02
stakkars_arigato: ah!?!11:03
arigatoor the translate option  --make-jobs=411:03
arigatobut I guess that the "4" goes nowhere on Windows right now11:03
stakkars_fijal told me -- yes! That is not controlled by us, but a make option? So windows it out ;-)11:04
stakkars_but I could rework the make process to run multiple compiles by splitting it up11:05
stakkars_and then do another make, which then finds all compilations to be ready.11:05
stakkars_but I also could simple leave it alone.....11:06
stakkars_but I also could simple leave it alone&..11:06
arigatolook carefully if "make" already has an equivalent option11:06
stakkars_how much TeraHz does your Laptop have?11:07
stakkars_I calculate you have 2.4 GHz11:08
stakkars_or 2.211:08
arigato2.53GHz11:09
stakkars_ha! all good, then virtualbox is performing fine.11:12
stakkars_and servers tend to have many cores at lower speed. Even my brand-new MacMini has 4 cores, but 2 GHz11:13
Action: stakkars_ craving for STM :->11:13
mikefcWhy might the OSX build on buildbot fail with: "translate translation aborted ( 5 hrs, 32 mins, 32 secs )" 11:14
mikefc?11:14
mikefcobviously it was killed for taking too long. but does that indicate a bug? or just a tempermental build machine?11:15
mikefcoops11:17
mikefcit's that fucking defaultdict test I filed last week11:17
mikefcAlex_Gaynor:  that defaultdict test is causing the osx build to fail i think.11:18
arigato[platform:execute] make -j 4 in /tmp/usession-default-2516/testing_111:18
arigatocommand timed out: 5400 seconds without output, killing pid 5881611:18
arigatoI don't see how this relates to defaultdict?11:19
witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy.11:19
witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) left irc: Client Quit11:19
mikefchttp://buildbot.pypy.org/summary/longrepr?testname=unmodified&builder=pypy-c-jit-macosx-x86-64&build=296&mod=lib-python.2.7.test.test_defaultdict11:19
witulski1 (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy.11:19
Alex_GaynorI can't possiblyimagine the defaultddict thin effected it11:19
kennethreitz (~kennethre@adsl-99-191-194-225.dsl.pltn13.sbcglobal.net) joined #pypy.11:19
mikefc======================================================================11:19
mikefcERROR: test_copy (test.test_defaultdict.TestDefaultDict)11:19
mikefc----------------------------------------------------------------------11:19
mikefcTraceback (most recent call last):11:19
mikefc  File "/Users/pypy/buildslave/pypy-c-jit-macosx-x86-64/build/lib-python/2.7/test/test_defaultdict.py", line 98, in test_copy11:19
mikefc    d2 = d1.copy()11:19
mikefcTypeError: first argument must be callable11:19
arigatomikefc: yes, I fixed this one11:19
mikefcOh. OK. 11:19
arigatoyou're looking at an older failure11:19
mikefcugh. I'm sorry :(11:20
Action: mikefc needs to find a chunk of time to get started on actual pypy hacking instead of just dancing around the edges.11:20
kennethreitz (~kennethre@adsl-99-191-194-225.dsl.pltn13.sbcglobal.net) left irc: Client Quit11:21
kennethreitz (~kennethre@ip65-46-252-62.z252-46-65.customer.algx.net) joined #pypy.11:22
kennethreitz (~kennethre@ip65-46-252-62.z252-46-65.customer.algx.net) left irc: Client Quit11:23
stakkars_arigato: going to upgrade my server, again. Actually I can get a cheaper one11:24
stakkars_with quad core I7 2.6 GHz, and 16 Gb.11:24
stakkars_question:11:24
stakkars_does it make sense to go for a Xeon E3-1275?11:25
stakkars_do we have/need different support for such a CPU, and would it help PyPy to have such a machine to play with?11:26
apollo13quadcore i7 and 16 gb sounds like hetzner11:27
stakkars_sounds like hetzner, walks and quacks like hetzner, _IS_ hetzner11:28
apollo13lol11:28
stakkars_but the apollo13 also needs an upgrade11:28
apollo13stakkars_: hell no, new revisions explode :/11:29
stakkars_:-D11:29
stakkars_so what do people thing: does Xeon make sense, or do we have that already (or don't plan to care?)11:30
stakkars_s/thing/think/11:30
witulski1 (stupsi@fwstups.cs.uni-duesseldorf.de) left #pypy.11:32
overminder (overminder@wf120-054.ust.hk) left irc: Ping timeout: 268 seconds11:33
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1150 [12arigo, temp2]11:36
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1151 [12arigo, SpecialisedTuples]11:36
derdon (~derdon@pD9E1D622.dip.t-dialin.net) joined #pypy.11:37
nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Remote host closed the connection11:51
azanella (~azanella@189.6.80.207) joined #pypy.11:51
K-os (~gabriel@dhcp-205.zbh.uni-hamburg.de) joined #pypy.11:51
Action: arigato managed to run the wrong buildbot12:08
felipejcruz (~felipecru@201.19.235.197) joined #pypy.12:09
bbot213Exception: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1151 [12arigo, SpecialisedTuples]12:09
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/970 [12arigo, temp2]12:10
stakkars_arigato: I'm ordering a server. You can influence me _now_!! ;-)12:10
arigatoI suppose I don't have any more to say than: "why?"12:11
arigatobut I evolved into being a guy that is happy for years with one good machine12:12
stakkars_why new server? because it is faster and cheaper to upgrade.12:12
derdon (~derdon@pD9E1D622.dip.t-dialin.net) left irc: Remote host closed the connection12:12
arigatofeel free.  I don't think I should have any influence on it12:13
stakkars_question is: do we have an advantage from Xeon processor, or is Core I7 enough12:13
arigatoI cannot answer this question12:13
fijal (~fijal@197.104.25.193) joined #pypy.12:13
stakkars_the question was directed to those who work on CPU related things, like insn set. If I upgrade, then to that what's useful for PyPy12:14
stakkars_maybe fijal in interested12:15
felipejcruz (~felipecru@201.19.235.197) left irc: Ping timeout: 244 seconds12:19
cpets (c358b60a@gateway/web/freenode/ip.195.88.182.10) left irc: Ping timeout: 265 seconds12:24
nedbat (~nedbat@python/psf/nedbat) joined #pypy.12:24
felipejcruz (~felipecru@201.19.235.197) joined #pypy.12:32
Nick change: DasIch__ -> DasIch12:36
stakkars_fijal: I'm replacing my root server. Do you have a preference which CPU?12:37
arigatocan someone enlighten me about how sys.path_hooks is supposed to be used?12:37
arigatoI'd like to add an entry to sys.path that doesn't correspond to any filesystem path12:37
stakkars_fijal: is a Xeon processor better for developing PyPy, or is Core I7 fine?12:38
fijalstakkars_: I think the bigger cache the better12:39
fijalotherwise dunno12:39
fijalbut maybe it's just that our heuristics suck12:39
stakkars_better on Xeon?12:39
arigatofijal: how about we try to fix the nursery to 4MB by default?12:40
stakkars_I'm only asking if we have an advantage in the different instruction set. If that's irrelevant, then I'll take an I7, because it costs me less12:40
arigatostakkars_: make sure the machine has at least around 2MB of cache, but I suppose that it's always true nowadays12:42
arigatostakkars_: if you're wondering about instruction set, I don't think we care in the slightest12:42
fijalstakkars_: I think the main difference is that you can put few xeons together in one computer12:42
fijalunlike i712:42
fijalso there is no *actual* difference or it's minor12:42
arigatowell check the cache size12:42
arigatothat also tends to change a little bit12:43
stakkars_yes, I'll do12:43
arigatobut as I said it may not matter12:43
fijalarigato: ok12:43
fijalarigato: it would be better than now12:43
arigatoyes12:43
dracman (~draco@212.255.16.24) joined #pypy.12:44
fijaleh12:45
fijalI'm fed up with martijn writing walls of text on pypy-dev12:45
fijalI'm inclined to answer "it's all nice and possible, but it's very unlikely I'll do anything related"12:45
stakkars_it's always the same with hetzner. You order a great machine, and after one or two yests, it is cheaper to use a new setup. My 2010 server costs 89 Eur/month, 1.5 TB raid1, 24G mem, 1.6 GHz12:45
stakkars_now I can get 3 TB, 16G mem, 2.6 GHz for 49 Eur/month12:46
stakkars_plus 150 setup, amortized after a few months12:47
arigatofijal: yes, that's mostly what I wrote too12:47
fijalarigato: except not explicit enough to stop him from writing another wall of text12:48
fijalbut maybe you can't12:48
arigatowell, never mind that, just write one answer and if the next e-mail doesn't ask new question then you're done12:49
fijal (~fijal@197.104.25.193) left irc: Read error: Connection reset by peer12:49
santagada (~leonardo@189.27.166.216.dynamic.adsl.gvt.net.br) joined #pypy.12:50
Rhy0lite (dje@nat/ibm/x-pnagwetamayeztft) joined #pypy.12:52
CIA-1303arigo 07roundup * 10#958/AMD Opterons don't specify their L3 cache size: 12:52
CIA-13[testing] Changed the default to always use 4MB. It seems better than now in most use12:52
CIA-13cases. More precise testing might give more hints, maybe. * 14https://bugs.pypy.org/issue95812:52
kenaan12arigo default 114c10a84a7da0 15/pypy/rpython/memory/gc/minimark.py: For minimark, use by default 4MB for the nursery size, instead of half the number found by poking in /proc/cpuinfo,...12:53
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 255 seconds12:53
exarkunI want to switch the Twisted website over to run on PyPy, but the buffer case for s* needs to be implemented.  Anyone want to help with that?12:57
felipejcruz (~felipecru@201.19.235.197) left irc: Ping timeout: 244 seconds12:58
felipejcruz (~felipecru@189.24.210.77) joined #pypy.12:59
stakkars_fijal, arigo: 8 Mb cache is ok I think (cheaper option)12:59
stakkars_done, decided12:59
mattip (4fb46e04@gateway/web/freenode/ip.79.180.110.4) joined #pypy.13:00
nedbat (~nedbat@python/psf/nedbat) joined #pypy.13:01
mattiphello. Someone change dlopen which broke the windows build. I created a one line patch here, https://gist.github.com/1451503 but don't really want to push to default. Could one of the core developers merge this please?13:02
stakkars_ok, will do that. Thanks!13:08
felipejcruz (~felipecru@189.24.210.77) left irc: Ping timeout: 252 seconds13:09
felipejcruz (~felipecru@187.102.160.14) joined #pypy.13:09
felipejcruz (~felipecru@187.102.160.14) left irc: Client Quit13:09
stakkars_but I think it is better do keep the interace and write13:10
aboudreault (~alanb@osgeo/member/aboudreault) joined #pypy.13:10
stakkars_def dlopen(name, unused=None)13:10
mattipsee? That's why I didn't want to commit it :)13:20
kenaan12ctismer default 118abfe57358d8 15/pypy/rlib/rdynload.py: windows fix (mattip, stakkars)13:20
Insectoid (~Q@c-98-230-8-37.hsd1.fl.comcast.net) left irc: Ping timeout: 268 seconds13:22
stakkars_actually, interface changes without checking platform dependent code is not very nice behavior.13:23
jnoller (~jnoller@vpn.nasuni.com) joined #pypy.13:24
mfoord (~anonymous@python/psf/voidspace) left irc: Quit: mfoord13:31
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/288 [12mattip, default]13:37
mattipfirst time I did that.13:37
stakkars_your first bot start?13:39
mattipyep.13:39
stakkars_well, I'm just 5 weeks ahead13:39
mattipBut what a 5 weeks its been!13:40
lmoura (~lmoura@186.212.130.54) left irc: Ping timeout: 240 seconds13:40
stakkars_because of the first win64 build? hee13:41
chrisq (chrisq@193.69.165.197) left irc: Quit: leaving13:41
mattip (4fb46e04@gateway/web/freenode/ip.79.180.110.4) left irc: Quit: Page closed13:41
mkerrin (~mkerrin@wg1-nat.fwgal01.deri.ie) joined #pypy.13:44
Arach (~arach@2.95.139.234) left irc: Read error: Connection reset by peer13:47
fenrrir (~fenrrir@187.40.145.179) joined #pypy.13:47
kushal (~kdas@fedora/kushal) joined #pypy.13:53
lmoura (~lmoura@187.112.24.204) joined #pypy.13:54
camara_ (4c1c615e@gateway/web/freenode/ip.76.28.97.94) joined #pypy.13:54
camara_stakkars_: you will find that Xeon will be faster for building pypy as it has larger cache, higher bandwidth to the system memory, and tends to have additional optimizations13:54
stakkars_ok. It is Eur 69 vs Eur 49. Is it 30 % faster? :-)13:56
camara_no likely about 15% if comparing similar clock speeds13:57
stakkars_per month, that really counts, and I'm coming from core i7 960, Eur. 8913:57
kirmaeuros for what?13:58
stakkars_so the change to the Eur 49 offer is still more attractive. (buying the 2nd best is a good strategy)13:58
stakkars_for my server, per month!13:59
kirmahow about EC2 instances for the amount of time you actually need13:59
stakkars_I'm again upgrading my snakepit machine, which has stackless.com, python.net and a few others, and my build slave13:59
camara_ (4c1c615e@gateway/web/freenode/ip.76.28.97.94) left irc: Quit: Page closed14:00
stakkars_because I cannot go without speed14:00
Nick change: Gulaway -> Gulopine14:01
kushal (~kdas@fedora/kushal) left irc: Read error: Connection reset by peer14:10
tos9 (~tos9@unaffiliated/tos9) left irc: Quit: leaving14:10
tos9 (~tos9@unaffiliated/tos9) joined #pypy.14:10
mfoord (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy.14:14
mfoord (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host14:14
mfoord (~anonymous@python/psf/voidspace) joined #pypy.14:14
harrison (~quassel@adsl-69-209-208-236.dsl.chcgil.sbcglobal.net) joined #pypy.14:18
cpets (c358b60a@gateway/web/freenode/ip.195.88.182.10) joined #pypy.14:26
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 248 seconds14:35
whitelynx (~whitelynx@63.241.75.144) joined #pypy.14:35
khs (~khs@2001:700:300:2120:725a:b6ff:fee5:a44) left irc: Remote host closed the connection14:37
fijal (~fijal@196.209.233.36) joined #pypy.14:38
fijalexarkun: what buffer? I can probably help14:39
gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy.14:39
fijalstakkars: hey14:39
fijalstakkars: I'm sorry for breaking windows14:39
fijalkirma: EC2 does not really work well with stuff like benchmarks14:40
fijalbesides, our usage is quite high14:40
Nick change: Gulopine -> Gulaway14:41
exarkunfijal: __builtin__.buffer14:44
fijalto support it in cpyext?14:45
exarkunyes14:45
exarkunin getargs.c14:45
fijaloh14:45
fijalfor pyopenssl?14:45
exarkunright14:46
fijaluh aliens stole my laptop14:46
Action: fijal looks for excuses14:46
fijalexarkun: ok, I'll have a look14:46
exarkunOpenSSL/ssl/connection.c, ssl_Connection_send14:46
fijaldid you write a test?14:46
exarkunOne of the pyOpenSSL tests should cover it14:46
exarkunLet me check14:46
fijalno, as in cpyext test14:47
fijalwould be a good start14:47
exarkunBlech.  PyPy looks in ~/.local/lib/python2.7/site-packages before looking in its own site-packages14:48
bbot24Failure: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/970 [12arigo, temp2]14:48
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/971 [12arigo, SpecialisedTuples]14:48
exarkunAdding something to test_getargs.py might be easy I guess14:51
exarkunmaybe easy enough that I should just let you do it :)  but let me know if you'd rather I.14:51
kenaan12bivab arm-backend-2 11a94098254217 15/pypy/jit/backend/arm/opassembler.py: (bivab, arigo) use constants here14:52
kenaan12bivab arm-backend-2 110cde4d93fe16 15/pypy/jit/backend/arm/runner.py: (arigo, bivab): write the correct value at the force index in the frame14:52
fenrrir (~fenrrir@187.40.145.179) left irc: Quit: Leaving14:55
kenaan12ctismer win64-stage1 11ffe08320d6bc 15/pypy/: Quite a simplification and improvement: removed the extra mess for signed_defn.h and put it into the standar...14:56
fijalexarkun: are you sure?14:57
fijalexarkun: I would definitely be happy if you do it :)14:57
Action: fijal trying not very hard to make slides14:57
stakkars_ (~tismer@89.204.139.170) left irc: Quit: schnarch15:03
abki_ (~Amirouche@hyper.pilotsystems.net) joined #pypy.15:03
abki_hi15:03
abki_is there an api similar to gevent on pypy 1.7 ?15:03
thinker (~thinker@123-194-52-90.dynamic.kbronet.com.tw) joined #pypy.15:04
abki_the documentation at http://doc.pypy.org/en/latest/index.html still seem to reference 1.6 code15:05
fijalno15:05
fijalI'll fix the docs15:05
fijalgreenlets work though15:05
fijaland eventlet is reported to work15:06
abki_thx15:08
Insectoid (~Q@12.237.80.67) joined #pypy.15:12
fijalAlex_Gaynor: ping15:14
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer15:16
ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) left irc: Quit: Leaving...15:16
kkris (~kris@80-123-41-118.adsl.highway.telekom.at) joined #pypy.15:18
jerith (~jerith@chthon.slipgate.za.net) left irc: Ping timeout: 248 seconds15:19
ronan (~ronan@host86-142-249-87.range86-142.btcentralplus.com) joined #pypy.15:24
fijal (~fijal@196.209.233.36) joined #pypy.15:33
fijalAlex_Gaynor: ping15:33
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer15:34
overminder (~overminde@wf107-175.ust.hk) joined #pypy.15:34
exarkunabki_: Plus, there's Twisted.15:39
exarkunHey, what is the C struct for the __builtin__.buffer type?15:40
exarkundoes cpyext even have it?15:40
Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy.15:43
exarkunBah, I can't run pypy's tests anymore15:44
arigatoit's PyBuffer_Type15:44
exarkunfijal would have been done with the whole thing by now :/15:44
arigatoyes, it is in cpyext15:44
arigatoah no, confusion15:45
arigatoit is called PyBufferObject, but even CPython 2.7 doesn't expose it15:45
exarkunpypy/module/cpyext/test/test_getargs.py ......Segmentation fault15:46
exarkun:(15:46
fijal (~fijal@196.209.233.36) joined #pypy.15:50
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer15:52
K-oshi! when running python py.py it does some gcc-compilation stuff at the beginning.15:52
K-oscan anybody tell me what it is?15:52
timotimoplatform-checks15:52
timotimothings like "how big is a pointer" and such15:52
MostAwesomeDudeK-os: It's checking your toolchain capabilities.15:52
MostAwesomeDudeKinda like those checks that happen during an autoconf.15:53
jterrace (~jterrace@aegis.CS.Princeton.EDU) joined #pypy.15:53
K-osok, but when it's running on top of python, why does it need to know this low-level stuff?15:53
K-osor is there actually a JIT involved?15:53
apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) left irc: Remote host closed the connection15:54
timotimopypy uses foreign-function-interface things15:54
timotimoctypes, for example, or its own _ffi module15:54
apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) joined #pypy.15:54
K-osoh ok15:54
K-osand the cpython interpreter can not provide this information?15:55
MostAwesomeDudeOh, this is on top of CPython, interpreted? In that case, I'm not sure why gcc is invoked.15:55
K-osi thought so15:55
`fox` (~fox@84.18.157.166) joined #pypy.15:56
fijal (~fijal@196.209.233.36) joined #pypy.15:56
K-osin the pypy checkout i'm in the pypy/bin/ directory15:56
K-osand just doing "python py.py" where python is my standard distribution supplied cpython interpreter15:57
fijalexarkun: :(15:57
exarkunfijal: I'm stuck now.  I don't know why it segfaults.  When I try to add debug info, it fails with C compiler errors that make no sense.15:57
santagadaK-os, it need to know the same thing than when running standalone15:58
exarkunhttp://codepad.org/XG9qsg3K15:58
K-osok15:58
santagadaK-os, because ctypes don't know this stuff15:58
fijalexarkun: you added a test?15:58
fijalthat segfaults?15:59
exarkunfijal: yes15:59
fijalok15:59
apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) left irc: Ping timeout: 248 seconds15:59
Rhy0litefijal: we have Alex in custody15:59
stakkars (~tismer@p5DDB7E68.dip.t-dialin.net) left irc: Read error: Connection reset by peer15:59
fijalRhy0lite: heh :)15:59
K-oshm. funny. i would think a module for interfacing with native code would provide that info :-)15:59
fijalRhy0lite: is he telling you "all that travis said is obsolete"?16:00
exarkunK-os: How would it provide you with the info without running a C compiler?16:00
stakkars (~tismer@p5DDB7E68.dip.t-dialin.net) joined #pypy.16:00
Rhy0litewhat did travis say?16:00
fijalexarkun: is there a reason why PyBuffer_Release is to the right so much?16:00
fijalRhy0lite: that you should use C/cython I presume16:00
exarkunfijal: I just copied it from a similar test in the module.16:00
fijalok16:01
Rhy0liteoh, that16:01
exarkunI don't even remember what it does.16:01
Rhy0liteTravis didn't say anything16:01
K-osctypes could do that on 'installation'16:01
exarkunfijal: Oh.  In the paste.  I don't know.  Copy/paste buffers are weird I guess.16:01
Insectoid (~Q@12.237.80.67) left irc: Ping timeout: 240 seconds16:01
exarkunfijal: It doesn't look like that in the source file.16:01
Rhy0liteTravis was a marketing pitch about NumPy/SciPy16:01
Rhy0lite"look at the pretty pictures"16:01
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer16:02
K-osor on compilation of the cpython interpreter. but that doesn't have to do with pypy any more... :-)16:02
bivab_ (~david@fwstups.cs.uni-duesseldorf.de) left irc: Quit: bivab_16:03
exarkunK-os: That's not a practical idea.16:03
santagadaK-os, yep. probably some form of platform detection should be part of ctypes16:04
santagadaexarkun, why do you think so?16:04
K-osthat's what i would think of a foreign function module...16:04
exarkunsantagada: Why is impractical to gather information about every possible ABI at installation time?  I can think of many reasons.16:05
exarkunsantagada: Any one should be sufficient to convince you.  For example, perhaps some of the libraries you'll eventually be interested in haven't been written yet.16:05
santagadaexarkun, I mean the same thing platform detection code in pypy does, discover pointer sizes and other simple stuff from std c16:06
santagadaexarkun, the same module that pypy uses could be part of ctypes.16:07
exarkunsantagada: The module PyPy uses invokes gcc16:07
exarkunAm I completely turned around about what K-os is talking about?  I thought he was saying it should be possible to gather the information without running gcc.16:08
K-osyes, that's what i thought16:08
santagadano there is no way to do that16:09
exarkunAnd you can't do it without running gcc, and you can't run gcc on _everything in the world_ up front.16:09
santagadaI mean, if you implement gcc, then yes16:09
K-osin case it's only pointer and int and long sizes, it should be easy for the runtime to provide that16:09
santagadaexarkun, probably most of the simple stuff everyone needs could come pre compiled/defined16:09
exarkunK-os: And arbitrary struct alignment, and macro "resolution"16:10
santagadaK-os, every macro on every library also has to be introspected16:10
exarkunK-os: Can't do it.16:10
K-osok, that's another thing16:10
K-osi get it16:10
asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy.16:12
K-osanother question. why does py.py not run on top of pypy-c. is it just not of interest or are there more fundamental problems?16:12
fijal (~fijal@196.209.233.36) joined #pypy.16:12
fijalK-os: yes, but there are more complex things that sizes of ints16:13
fijalK-os: I failed to provide "common sizes" in svn16:13
fijalRhy0lite:  I kind of hoped he would tell a bit about architecture, why it's good, python productivity16:13
fijalor something16:13
K-osfijal: yes i understand the problem. i just didn't think about structs and stuff...16:14
kennethreitz (~kennethre@adsl-99-191-194-225.dsl.pltn13.sbcglobal.net) joined #pypy.16:14
exarkunK-os: py.py does run on top of pypy-c.16:16
K-oshm, not with my version16:16
exarkunGet a newer version, or file a bug.16:17
fijaler no16:17
fijal--no-faking16:17
fijalor --allworkingmodules16:17
fijalonly runs16:17
fijalexarkun: ok, next step would be killing some #if 0 in getargs.c16:17
exarkunfijal: To fix the segfault?16:18
fijaldo you want me to do it or do you feel adventerous?16:18
fijalyes16:18
exarkunfijal: The failure I encountered with the Twisted website was the NotImplementedError case, not a segfault.16:18
fijalunforuntately segfault is a perfectly fine failing test for cpyext :(16:18
exarkunfijal: So I guess my unit test does not correctly reflect the use case I am interested in.16:18
fijalit's commented by #if 016:18
fijalit might lead to various things16:18
fijalit's not commented by "if this, raise error"16:18
Rhy0litefijal: Alex is in my office. do you need him for something?16:19
exarkunfijal: When I mean that I encountered the NotImplementedError case, I mean that NotImplementedError got raised by the code.16:19
fijalyes16:19
exarkuns/mean/say16:19
fijalRhy0lite: no, not really16:19
fijalexarkun: yes ok16:19
exarkunfijal: I don't get why the next step isn't writing a unit test that triggers NotImplementedError to be raised.16:20
fijalexarkun: but still I don't think it proves anything16:20
exarkunI'm really stupid today though.16:20
fijalbecause you might get a segfault before exception16:20
fijalas far as I can see in the code16:20
abki_I'm wondering whether postgresql use a btree to store tuples, and table information16:20
K-osah! with --nofaking it works16:20
exarkunabki_: #postgres16:21
abki_oups16:21
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer16:24
Action: exarkun wonders if he can buy fijal a better internet connection16:25
fijal (~fijal@196.209.233.36) joined #pypy.16:25
fijalexarkun: indeed, it should explode with NotImplementedError16:26
fijalI have no idea why it does not16:26
exarkunDo you know why it fails to compile when I add `printf("OH NO %s %d\n", buf.buf, buf.len);´?16:27
K-osbye16:27
K-os (~gabriel@dhcp-205.zbh.uni-hamburg.de) left irc: Quit: Leaving16:27
fijalno16:28
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer16:28
exarkunah right, I can read the generated c in /tmp16:29
exarkunAnd of course, "\n" is a stupid thing to put in a string literal in a C program embedded in a Python string16:29
exarkun`-1075718480` looks like a pretty awful value for buf.len16:33
jerith (~jerith@chthon.slipgate.za.net) joined #pypy.16:33
exarkunUgh man.  PyString_Check can _fail_?16:34
exarkunThat's no kind of right16:35
G2P (~G2P@fw-asn1.ornis.com) left irc: Quit: Leaving.16:35
santagada (~leonardo@189.27.166.216.dynamic.adsl.gvt.net.br) left irc: Quit: Ex-Chat16:35
nephics (~nephics@78-72-71-194-no33.tbcn.telia.com) left irc: Quit: Leaving.16:42
exarkunOkay... something about error handling in the s* non-string case is wrong16:43
exarkunI don't know what, but it's setting NotImplementedError but not causing PyArg_ParseTuple to fail16:44
exarkunAnd behaving differently untranslated vs translated16:44
fijal (~fijal@196.209.233.36) joined #pypy.16:45
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer16:46
exarkunif I replace the NotImplementedError with a converterr() call then it fails appropriately16:46
exarkunmaybe returning NULL from convertsimple is how you signal _success_?16:47
nephics (~nephics@78-72-71-194-no33.tbcn.telia.com) joined #pypy.16:50
fijal (~fijal@196.209.233.36) joined #pypy.16:50
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer16:52
mkerrin (~mkerrin@wg1-nat.fwgal01.deri.ie) left irc: Quit: mkerrin16:52
exarkunWhat is cpyext/stubs.py?16:53
exarkun(why won't you write documentation :()16:53
kenaan12arigo default 11a72429e0e0ed 15/pypy/: Write a hack to distinguish between "true built-in modules" and "pseudo-extension built-in modules".  The latters a...16:55
fijal (~fijal@196.209.233.36) joined #pypy.16:56
fijalthat was probably me who wrote the code and I might not be paying attention16:57
fijalexarkun: maybe16:57
CIA-1303arigo 07roundup * 10#945/import shadowing module from __init__.py as __main__ fails: [resolved] Checked in as a straightforward hack: a72429e0e0ed * 14https://bugs.pypy.org/issue94516:57
exarkunWhere does PyBuffer_IsContiguous belong?16:57
exarkunin CPython it's in Objects/abstract.c16:57
exarkunShould I create a new cpyext/abstract.py?16:57
fijalexarkun: isn't it a thing that always returns True anyway?16:58
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer16:58
exarkunI don't know yet.16:58
exarkunThere's a sequence.py, so I guess I'll create buffer.py16:58
lmoura (~lmoura@187.112.24.204) left irc: Ping timeout: 240 seconds16:59
fijal (~fijal@196.209.233.36) joined #pypy.17:01
Action: fijal tries to stabilize his connection17:02
exarkunOh, there's already a Py_buffer defined in api.py, in addition to the one defined in stubs.py17:02
exarkunSo the one in stubs.py should clearly be deleted17:03
fijalI think stubs.py is not even imported17:03
fijalbut yes17:03
tos9 (~tos9@unaffiliated/tos9) left irc: Read error: Operation timed out17:03
exarkunOkay, so for now I'll try to say that contiguous buffers are the only ones pypy supports17:04
exarkunwhich means the implementation is `return not fortran´17:04
tos9 (~tos9@unaffiliated/tos9) joined #pypy.17:04
exarkunI mean `return fortran == 'C'´17:04
exarkunnnng what17:05
exarkun`return fortran == 'C' or fortran == 'A'´17:05
fijalI think it's the only one that people support anyway17:05
rokujyouhitoma (~rokujyouh@173.227.61.3) joined #pypy.17:05
fijalif you try to pass a non-contiguous buffer to cpython internals it'll explode in so many places17:05
exarkunI hate this interface17:05
exarkunIt's a total catastrophe17:05
fijalwelcome to the club17:06
arigato:-)17:06
fijalI'm actually amazed it somehow works17:06
exarkunWhat does this mean?   @cpython_api([Py_buffer, lltype.Char], ...)17:07
ronnynot 100% broken is practically working17:07
ronnyexarkun: parameter types cpyext will have in c17:08
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy.17:08
exarkunBlech, obviously.17:08
exarkunsorry, I'm probably going to be stupid all day.17:08
antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds17:08
fijal (~fijal@196.209.233.36) left irc: Ping timeout: 240 seconds17:11
kenaan12arigo buildbot 1183d6d146e64f 15/bot2/pypybuildbot/builds.py: Give it a bit more time before timing out.  Useful for very slow final "make" runs.17:11
lmoura (lmoura@gateway/indt/session) joined #pypy.17:12
Action: arigato off17:13
arigato (~arigo@fwstups.cs.uni-duesseldorf.de) left irc: Quit: See you17:13
`fox` (~fox@84.18.157.166) left irc: Ping timeout: 255 seconds17:18
`fox` (~fox@84.18.157.166) joined #pypy.17:18
cpets (c358b60a@gateway/web/freenode/ip.195.88.182.10) left irc: Quit: Page closed17:20
lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Ping timeout: 252 seconds17:21
bbot24Failure: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/971 [12arigo, SpecialisedTuples]17:26
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/972 [12arigo, temp2]17:26
fijal (~fijal@196.209.233.36) joined #pypy.17:27
rokujyouhitoma (~rokujyouh@173.227.61.3) left irc: Remote host closed the connection17:34
rokujyouhitoma (~rokujyouh@173.227.61.3) joined #pypy.17:35
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy.17:36
rokujyouhitoma (~rokujyouh@173.227.61.3) left irc: Remote host closed the connection17:37
jterraceAlex_Gaynor: you spelled W_UnsignedIntgerBox wrong :)17:38
fijalexarkun: cpython_api has a docstring!17:42
exarkunfijal: I moved on17:42
exarkunI don't remember how to set a breakpoint for gdb now.17:43
fijalb lineno17:44
fijalor b funcname17:44
fijalor b filename:lineno17:44
Insectoid (~Q@c-98-230-8-37.hsd1.fl.comcast.net) joined #pypy.17:45
exarkunnone of those seems to work for cpyext17:45
fijalI fear you need to hack cpyext to pass -g to gcc17:45
nephics (~nephics@78-72-71-194-no33.tbcn.telia.com) left irc: Quit: Leaving.17:46
exarkunI dunno, I might give up17:46
kennethreitz (~kennethre@adsl-99-191-194-225.dsl.pltn13.sbcglobal.net) left irc: Quit: Computer has gone to sleep.17:47
fijal:/17:47
exarkunwrite more docs, get rid of the redundant dead code sitting around like traps, make debugging possible without sticking in one-time hacks deep in the internals17:48
exarkunmaybe I'll succeed next time17:48
exarkun(I know, you have plenty of other things to do instead)17:49
fijalyeah, cpyext is not the sexiest thing to work on17:49
exarkunembedding cpython in pypy was probably fun, and you got an awesome blog post out of it17:50
exarkunbut cpyext is probably the thing that's most useful to actual adoption of pypy, for people who have not yet adopted it17:51
fijalyes or no17:51
fijalI can't see cpyext ever dealing with matplotlib17:52
kenaan12fijal default 11be6e2485ed4a 15/pypy/module/cpyext/api.py: add -g so debugger can be used17:52
exarkuncan you see cpython embedded in pypy ever being practically useful to someone?  is it ever going to work with a release of pypy, without a checkout of your hacks directory?17:52
fijalexarkun: here you go17:52
fijalgood question17:53
exarkunhg won't let me update to that revision17:53
fijalwhy?17:53
exarkunbecause I have uncommitted outstanding changes I guess17:53
fijalno, update is fine17:54
fijalhg is happy with that17:54
exarkunabort: outstanding uncommitted changes17:54
fijalwhen you try to ....17:54
fijal?17:54
fijalhg pull -u works for me17:54
fijal?17:54
exarkunuh ok that worked17:54
fijalyeah17:54
fijalit's not the same17:54
fijal(surprise, surprise)17:54
fijalexarkun: I believe my hacks can be useful, I believe it's possible to make it work with the release17:55
fijaland without checkout of my hack dir17:55
fijalwhat you would do would be pip install pypy_matplotlib_wrapper17:55
fijalthat would pull my hacks for you17:55
fijalI got something useful (barely) in day and a half17:55
fijalthat might be deceiving but it's better than cpyext17:55
exarkunokay17:56
exarkunand then when I want to use matplotlib to serve up graphs with twisted over https?17:56
fijalthis is the only reasonable path I see for matplotlib/scipy17:56
fijalhow do you do that now?17:56
bbot213Exception: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/972 [12arigo, temp2]17:56
exarkunwell, I can't, because pypy and twisted/https don't work (that's the feature I'm trying to implement)17:56
exarkunbut if I ever manage to implement it, then twisted/https will work on pypy with cpyext17:57
exarkunwill matplotlib work with pypy with cpyext?17:57
exarkunI still can't set a breakpoint by filename:lineno btw17:57
exarkunI mean, I can set it, but it doesn't break17:58
fijal:(17:58
exarkunoh there we go, using an absolute path worked17:58
`fox` (~fox@84.18.157.166) left irc: Ping timeout: 244 seconds17:58
exarkunwell, almost17:59
fijalexarkun: it should be possible to mix cpyext and my hacks17:59
exarkunoff by line error in the line number handling17:59
fijalwith deliberate #defines, like we do when testing17:59
fijalwe have deliberate hacks17:59
exarkunugh, no it's not just an off by one error18:00
exarkunsomething about optimization I guess18:00
exarkunfijal: well, okay18:01
fijal-O3 is completely unnecessary here18:01
fijalgood question how do I remove it18:01
exarkunI got what I needed, just needed to set the breakpoint far earlier than it seemed18:02
exarkunbut turning off O3 would be nice for the future18:02
fijalremove -O3 from translator/platform/linux.py18:02
exarkunand okay, obviously it's broken, because tp_as_buffer is full of null function pointers18:02
fijaldunno how to do it nicer18:02
exarkunthough I'm pretty sure I saw that those were initialized18:03
exarkunI remember there being something like PyType_Ready somewhere, doing a ton of special-case checks that could maybe be related to this18:04
exarkunbut I don't remember where that is18:04
exarkunmaybe _type_realize18:04
fijalexarkun: note that I have little experience with cpyext18:04
fijalexarkun: to summarize, I believe my hacks are the right way for some stuff18:05
fijalespecially for scientific balls of mud18:05
fijalif we ever want to run them18:05
exarkunpartly I'm just annoyed that because people went and wrote all that awful, awful code in matplotlib, it means that you're spending time supporting their crappy stupid code instead of supporting my code18:06
exarkunand also I'm really grumpy today for totally unrelated reasons18:06
fijalwell18:07
fijalit won't make you happier, but I'm very unlikely to develop my hacks more, unless someone pays me to do that18:07
exarkunis cpyext/src/bufferobject.c used at all in pypy now?18:07
fijalit's compiled18:08
fijalif it's not full of #if 0, it's used18:08
exarkunokay, I don't see any of those18:08
exarkunit initializes PyBuffer_Type's tp_as_buffer slot with &buffer_as_buffer, which has lots of nice function pointers in it18:08
fijalprobably some parts are ignored18:08
fijallike this one18:08
exarkunbut the buffer object I get in getargs.c has all nulls18:09
fijalhm18:09
fijalmaybe not?18:09
Action: fijal is confused18:09
exarkunam I supposed to add special support for buffer objects to type_attach in typeobject.py?18:09
exarkunis bufferobject.c even implementing the __builtin__.buffer, or is it some other kind of buffer?18:09
fijalgood question18:10
fijalI think it's some other buffer18:10
fijalyou can look at test_bufferobject18:11
fijalit *seems* to be supported to some degree18:11
Action: fijal is confues18:11
exarkunPyBuffer_Type does seem properly initialized, so I guess it isn't being used in this case18:16
exarkunThough I think it _should_ be used in this case18:16
exarkunInstead some other less-fully-defined buffer type is used18:17
jterracefijal: how can i skip a numpy test? i cant seem to import py so i can use py.test.skip18:23
fijaljterrace: skip("test")18:23
fijalit "magically" appears in your namespace18:23
jterraceoh18:23
jterracethanks18:23
fijalit predates my tenure in the project :)18:23
jterracehehe18:23
jterracei have more patches in the queue18:24
jterracei exposed uint8, uint16, and uint32 and made the fromstring function work for other types and parameters18:24
jterraceuint64 is broken so i didnt expose it18:25
rokujyouhitoma (~rokujyouh@173.227.61.3) joined #pypy.18:25
fijalcool18:25
jterracenot sure if alex got to review my tolist pull request18:26
exarkunthere's probably no way to go from a _typeobject to its ... definition, is there?18:27
exarkunThis type really sucks: http://codepad.org/hxtk1nQP18:28
etrepum (~bob@75.94.138.216) joined #pypy.18:30
kennethreitz (~kennethre@204.14.152.118) joined #pypy.18:31
fijaltp_bases = <tuple at remote 0xaf63998>18:32
fijal?18:32
exarkunit's (object,)18:33
exarkunand even though it has tp_as_numeric and a few others, the function pointers inside those structs are all null18:33
exarkunI'm trying to look at the boxing now, since it seems like that's where things go wrong18:34
voidspace (~voidspace@python/psf/voidspace) left irc: Quit: Uhm... gotta go18:35
Nick change: mfoord -> voidspace18:35
DasIch_ (~dasich@p4FFDCD7B.dip.t-dialin.net) joined #pypy.18:37
exarkunis W_TupleObject(<pypy.interpreter.buffer.StringBuffer object at 0xb74f70c>) a tuple with a StringBuffer as its only element?18:40
mtigas (~mtigas@users.spokesman.com) joined #pypy.18:40
DasIch (~dasich@p4FFDE31D.dip.t-dialin.net) left irc: Ping timeout: 248 seconds18:40
rokujyouhitoma (~rokujyouh@173.227.61.3) left irc: Remote host closed the connection18:42
rokujyouhitoma (~rokujyouh@173.227.61.3) joined #pypy.18:42
overminder (~overminde@wf107-175.ust.hk) left irc: Ping timeout: 252 seconds18:43
voidspace (~anonymous@python/psf/voidspace) left irc: Quit: voidspace18:43
rokujyouhitoma (~rokujyouh@173.227.61.3) left irc: Remote host closed the connection18:44
rokujyouhitoma (~rokujyouh@173.227.61.3) joined #pypy.18:44
exarkunughhh oh my god, lltype2ctypes18:45
ericflo (~ericflo@75.103.8.110) joined #pypy.18:45
nephics (~nephics@78-72-71-194-no33.tbcn.telia.com) joined #pypy.18:51
kennethreitz (~kennethre@204.14.152.118) left irc: Ping timeout: 240 seconds18:53
exarkunthere's no way I'll understand this function today18:56
kennethreitz (~kennethre@204.14.152.118) joined #pypy.18:56
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) joined #pypy.19:03
`fox` (~fox@host91-171-dynamic.245-95-r.retail.telecomitalia.it) left irc: Ping timeout: 255 seconds19:23
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.19:27
exarkunAlex_Gaynor: When is PyPy going to be usable, in practice, by the lay person, with a postgresql driver?19:32
jterracei'm a lay person and i approve this message19:33
gsneddersI'm a lay person and don't want to sound American.19:34
fijalexarkun: have you tried the ctypes-based drivers?19:39
exarkunfijal: no19:39
fijalalso pg8000 should work19:39
fijalpeople reported such stuff working :)19:39
exarkunHow "should"?19:39
fijalas in someone said it worked for him19:39
fijalI did not try myself19:40
kennethr_ (~kennethre@204.14.152.118) joined #pypy.19:40
kennethreitz (~kennethre@204.14.152.118) left irc: Ping timeout: 244 seconds19:41
lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy.19:42
kennethr_ (~kennethre@204.14.152.118) left irc: Ping timeout: 248 seconds19:44
lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) left irc: Ping timeout: 240 seconds19:47
rokujyouhitomaHi, all19:51
rokujyouhitomaI have a question.19:51
rokujyouhitomaI want RPyton interpreter(or RPython implements) is it exist?19:52
masqueraderokujyouhitoma: any python interpreter is also an RPython interpreter19:52
rokujyouhitomamasquerade: yes, I see. it is means RPython is a row layer Python language too.19:55
rokujyouhitomamasquerade: thanks.19:55
dmalcolm (~david@c-24-61-12-82.hsd1.ma.comcast.net) joined #pypy.19:58
gsneddersIs there any automated regression testing done of major Python projects, to ask a random question?19:59
exarkungsnedders: http://buildbot.pypy.org/20:00
etrepum (~bob@75.94.138.216) left irc: Quit: etrepum20:00
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Remote host closed the connection20:12
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy.20:13
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: Leaving20:14
fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.20:15
fijal (~fijal@196.209.233.36) left irc: Ping timeout: 240 seconds20:15
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Ping timeout: 240 seconds20:17
jterracewhen i try to pass sep=" " parameter to this function (http://pastebin.com/EY54MBJt) i get (application-level) TypeError: unsupported operand type for int(): 'str'20:18
jterracei dont understand the error20:18
jterracethe function works otherwise, but only when i try to set "sep" does it fail20:19
hazmat (~hazmat@204-232-198-49.static.cloud-ips.com) left irc: Changing host20:25
hazmat (~hazmat@plone/hazmat) joined #pypy.20:25
aboudreault (~alanb@osgeo/member/aboudreault) left irc: Quit: Leaving20:26
harrison (~quassel@adsl-69-209-208-236.dsl.chcgil.sbcglobal.net) left irc: Ping timeout: 252 seconds20:26
lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy.20:26
jterraceoh i see20:27
jterracedoes space.int(w_item) work if type(w_item) == str?20:27
jterraceseems like it doesnt20:27
lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) left irc: Ping timeout: 252 seconds20:31
nephics (~nephics@78-72-71-194-no33.tbcn.telia.com) left irc: Quit: Leaving.20:33
kennethreitz (~kennethre@204.14.152.118) joined #pypy.20:37
kennethreitz (~kennethre@204.14.152.118) left irc: Client Quit20:40
jnoller (~jnoller@vpn.nasuni.com) left irc: Quit: Leaving...20:40
kennethreitz (~kennethre@204.14.152.118) joined #pypy.20:41
e1ven (~E1ven@SQ7/ProjectLead/E1ven) left irc: Quit: Computer has gone to sleep.20:44
norox (~norox@2a00:801::2d0:b7ff:fea0:c753) left irc: Ping timeout: 240 seconds21:02
apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy.21:04
kennethreitz (~kennethre@204.14.152.118) left irc: Quit: Computer has gone to sleep.21:05
kennethreitz (~kennethre@204.14.152.118) joined #pypy.21:07
jterraceAlex_Gaynor: ping21:11
ojii (~ojii@40-34.60-188.cust.bluewin.ch) left irc: Remote host closed the connection21:15
Rhy0litejterrace: we have him trapped21:16
asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Ping timeout: 252 seconds21:16
jterracetrying to figure out how to convert an object-level string to an interpreter-level int21:16
jterracespace.int(somestr) throws a valueerror21:17
Rhy0liteI mentioned this lack of documentation during the Q&A session after his talk this morning ;-)21:18
fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy21:18
kennethreitz (~kennethre@204.14.152.118) left irc: Quit: Computer has gone to sleep.21:29
ronnyjterrace: whats the type of somestr21:33
legogris (~legogris@c80-217-207-157.bredband.comhem.se) left irc: Read error: Connection reset by peer21:33
jterracestr21:33
jterracee.g., space.int(space.wrap('3')) raises ValueError21:34
jterracewhat's the correct way to do that?21:34
jterraceint(space.unwrap(w_somestr)) ?21:34
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection21:36
legogris (~legogris@c80-217-207-157.bredband.comhem.se) joined #pypy.21:37
ronnyhmm21:37
ronnyok, im just lost in that code as well21:38
jterraceim talking about the object space functions: http://codespeak.net/pypy/dist/pypy/doc/objspace.html#operations-on-objects-in-the-object-space21:38
jterracespace.int(space.wrap(3)) returns 321:38
fijal (~fijal@196.209.233.36) joined #pypy.21:39
fijaljterrace: what do you want to do?21:40
jterracethis is for numpy.fromstring()21:40
jterraceif i use dtype's coerce method, it throws an error when i give it a string21:40
fijalyou have W_StringObject and you want rpython int?21:40
jterraceyes21:40
jterraceor a W_IntObject21:41
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer21:42
derdon (~derdon@pD9E1CE32.dip.t-dialin.net) joined #pypy.21:44
nedbat (~nedbat@python/psf/nedbat) joined #pypy.21:44
fijal (~fijal@196.209.233.36) joined #pypy.21:45
fijalgrrrr21:46
fijalspace.call_function(space.w_int, space.wrap('3'))21:46
fijalor21:46
fijalint(space.str_w(space.wrap('3')))21:46
fijaljterrace: ^^^21:46
jterraceah21:46
Action: fijal off21:46
jterracethanks21:46
fijal (~fijal@196.209.233.36) left irc: Read error: Connection reset by peer21:48
pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) left irc: Ping timeout: 252 seconds21:53
davisagli (~davisagli@davisagli.com) left irc: Excess Flood21:58
davisagli (~davisagli@davisagli.com) joined #pypy.21:58
danchr (~danchr@2001:4978:f:34f::2) left irc: Quit: ZNC - http://znc.sourceforge.net22:03
danchr (~danchr@cl-848.chi-02.us.sixxs.net) joined #pypy.22:03
stakkars_ (~tismer@88.130.164.164) joined #pypy.22:04
stakkars_hi22:11
Rhy0lite (dje@nat/ibm/x-pnagwetamayeztft) left irc: Quit: Leaving22:14
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 255 seconds22:14
CIA-1303r.lobb 07roundup * 10#949/1.7 Sandbox problems: 22:14
CIA-13[chatting] I'd be grateful if someone could either confirm or deny whether the sandbox is22:14
CIA-13broken in the 1.7 release. I've used the 1.4 version i ... * 14https://bugs.pypy.org/issue94922:14
dmalcolm (~david@c-24-61-12-82.hsd1.ma.comcast.net) left irc: Quit: Leaving22:19
kenaan12ctismer win64-stage1 1191bdfa626831 15/pypy/rlib/test/test_libffi.py: small fix to test_libffi22:23
bbot2Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/73922:36
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/16522:36
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/97322:36
pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) joined #pypy.22:42
kenaan12arigo default 11eff9db9ac0c8 15/pypy/jit/metainterp/test/test_fficall.py: Python 2.5 compat22:46
kenaan12arigo default 119b45755e2c2b 15/pypy/jit/backend/: Improve the hack 062e9d06c908: revert the changes done in the RegisterManager, and instead write a more involved bu...22:46
kenaan12arigo default 1194737f156c30 15/pypy/jit/backend/llsupport/: test and fix22:46
kennethreitz (~kennethre@204.14.152.118) joined #pypy.22:46
nedbat (~nedbat@python/psf/nedbat) joined #pypy.22:50
kkris (~kris@80-123-41-118.adsl.highway.telekom.at) left irc: Ping timeout: 240 seconds22:57
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 248 seconds23:08
bbot2Started: 15http://buildbot.pypy.org/builders/own-win-x86-64/builds/81 [12ctismer, win64-stage1]23:09
asmeurer__ (~asmeurer@c-174-56-21-245.hsd1.nm.comcast.net) joined #pypy.23:10
papercrane (~papercran@75.101.111.82) joined #pypy.23:12
kkris (~kris@80-123-38-181.adsl.highway.telekom.at) joined #pypy.23:14
papercrane (~papercran@75.101.111.82) left irc: Client Quit23:15
papercrane (~papercran@75.101.111.82) joined #pypy.23:16
Action: mikefc 's mind boggles at the py2k and py3k running simultaneously in pypy23:16
papercrane (~papercran@75.101.111.82) left irc: Client Quit23:17
papercrane (~papercran@75.101.111.82) joined #pypy.23:18
dracman (~draco@212.255.16.24) left irc: Ping timeout: 252 seconds23:21
dracman (~draco@212.255.27.92) joined #pypy.23:23
kkris (~kris@80-123-38-181.adsl.highway.telekom.at) left irc: Quit: Leaving.23:25
whitelynx (~whitelynx@63.241.75.144) left irc: Quit: Ex-Chat23:35
stakkars_on testing:23:36
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy.23:36
stakkars_there are tests which are failing, and there are tests that are skipped.23:36
stakkars_many of the skipped tests are skipped because they do not apply, which is ok.23:37
stakkars_but a few tests are skipped because they don't work, and they are not left as failures, but just skipped23:38
stakkars_with a message "try to make it run, somehow".23:38
stakkars_I think this is a problem in the buildbot, because I doubt that someone reads the skipping reasons, but just skips over it.23:39
stakkars_one way out is to let these tests fail.23:39
stakkars_another way is to add a new marker that flags the problematic test in an other way as just skipping.23:40
stakkars_But this way, it is no good and gives the misleading feeling "all ok"!23:40
stakkars_ (~tismer@88.130.164.164) left irc: Read error: Operation timed out23:47
mikefc"The revised text handling design in Python 3 is definitely a case of the pursuit of correctness triumphing over convenience."23:52
mikefc"The efficient Unicode representation coming in Python 3.3..."23:54
etrepum (~bob@69-9-57-57.static.logl001.digis.net) joined #pypy.23:55
azanella (~azanella@189.6.80.207) left irc: Ping timeout: 252 seconds23:56
--- Sat Dec 10 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!