| fprimex (~fprimex@fl-67-233-144-228.dhcp.embarqhsd.net) left irc: Quit: Leaving | 00:00 | |
| thinker` (~thinker@123-194-52-90.dynamic.kbronet.com.tw) left irc: Ping timeout: 240 seconds | 00:05 | |
| mfoord (~anonymous@python/psf/voidspace) left irc: Ping timeout: 240 seconds | 00:05 | |
| mfoord (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy. | 00:07 | |
| mfoord (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host | 00:07 | |
| mfoord (~anonymous@python/psf/voidspace) joined #pypy. | 00:07 | |
| asabil (~asabil@131.246.16.62.customer.cdi.no) left irc: Quit: Ex-Chat | 00:10 | |
| JaredW (~jared@061093200075.ctinets.com) joined #pypy. | 00:15 | |
| tlynn_ (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) left irc: Ping timeout: 248 seconds | 00:16 | |
| kenaan | 12ctismer win64-stage1 11d3f5435ca3b0 15/: manual big review of changes | 00:32 |
|---|---|---|
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 00:33 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 244 seconds | 00:40 | |
| Nick change: Gulaway -> Gulopine | 00:42 | |
| Nick change: Gulopine -> Gulaway | 00:47 | |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 00:47 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 248 seconds | 00:47 | |
| chronitis (~chronitis@88-104-120-223.dynamic.dsl.as9105.com) left irc: Quit: Lost terminal | 00:48 | |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Ping timeout: 244 seconds | 00:54 | |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 252 seconds | 00:55 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 00:58 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/146 | 01:01 |
| bgola (~bgola@189.100.46.89) joined #pypy. | 01:05 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 248 seconds | 01:18 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/720 | 01:26 |
| mfoord (~anonymous@python/psf/voidspace) left irc: Quit: mfoord | 01:30 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 01:30 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 01:38 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 01:39 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 01:40 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1854 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/583 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/619 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1117 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/278 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/804 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1446 | 01:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/715 | 01:42 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 258 seconds | 01:43 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 01:43 | |
| justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) joined #pypy. | 01:43 | |
| JaredW (~jared@061093200075.ctinets.com) left irc: Read error: Operation timed out | 01:44 | |
| jonathan__ (~jonathan@bas1-montreal29-1242341904.dsl.bell.ca) left irc: Remote host closed the connection | 01:44 | |
| justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) left irc: Quit: Leaving | 01:54 | |
| rokujyouhitoma (~rokujyouh@122x220x64x254.ap122.ftth.ucom.ne.jp) joined #pypy. | 01:55 | |
| Nick change: Gulaway -> Gulopine | 01:58 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection timed out | 02:02 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 02:03 | |
| rokujyouhitoma | hi! | 02:05 |
| rokujyouhitoma | I make a plan PyPy Advent Calendar in Japan and want the article of pypy-project member at the final day(Dec. 25). | 02:05 |
| rokujyouhitoma | I'd like to ask anyone to give a article. | 02:06 |
| rokujyouhitoma | Do anyone write it? | 02:06 |
| __class__ (~class@99-105-56-217.lightspeed.sntcca.sbcglobal.net) joined #pypy. | 02:09 | |
| _pyre_ (~stian@137.241.189.109.customer.cdi.no) joined #pypy. | 02:13 | |
| _pyre_ | Hey guys | 02:13 |
| _pyre_ | I teached the raw python pickle.py to use struct instead of marshal | 02:13 |
| _pyre_ | and some other minor things | 02:14 |
| _pyre_ | result in a small test, upto 2x improvement | 02:14 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 02:14 | |
| _pyre_ | (tested for dumps) | 02:14 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 02:15 | |
| _pyre_ | Something that might be push into pypy? :) | 02:17 |
| _pyre_ | for dict pickling it outperforms cPickle by the way | 02:18 |
| _pyre_ | 1000 dict entries with string keys and string values, pickled a thousand times. | 02:21 |
| _pyre_ | CPython cPickle did it in 1.15s while pypy nightly did it in 0.71s. Using default pickle pypy used 1.61s in the same test. And cpython using pickle used 6.11s | 02:21 |
| _pyre_ | and no, i didn't warm up the JIT properly either, just took the executation time directly so you guys might find more improvement | 02:22 |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 02:23 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 240 seconds | 02:23 | |
| _pyre_ | list with 1000 string entries also did about twice as good as cPickle | 02:24 |
| _pyre_ | cPickle still outperforms my own pickler with tuple's and int lists tho. | 02:24 |
| _pyre_ | with 5000 loops and 5000 entries, it's still 2x for string dicts compared to cPickle. | 02:26 |
| stakkars | ctismer | 02:27 |
| kenaan | 12ctismer win64-stage1 1177ee30370cd7 15/pytest.py: forgot this patch for sys.maxint, which has bad consequences. Note that this patch must stay until the end o... | 02:28 |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 252 seconds | 02:29 | |
| stakkars | fixed, works on (default, win32), (win64-stage1, win32), (win64-stage1, win64) | 02:30 |
| stakkars | arigatoÖ now, after 14 hours, this is reallz true. Waah | 02:31 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 02:31 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 240 seconds | 02:36 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-win-x86-64/builds/50 [12ctismer, win64-stage1] | 02:36 |
| _pyre_ (~stian@137.241.189.109.customer.cdi.no) left irc: Quit: Leaving. | 02:38 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 02:39 | |
| ericflo (~ericflo@75.103.8.110) left irc: Quit: ericflo | 02:40 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/949 | 02:40 |
| Shanita (~John@osbk-4db14086.pool.mediaWays.net) joined #pypy. | 02:46 | |
| Kaskuka (~John@osbk-4db1653f.pool.mediaWays.net) left irc: Ping timeout: 240 seconds | 02:46 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 02:47 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 260 seconds | 02:53 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) left irc: Quit: etrepum | 02:58 | |
| jimbaker (~jbaker@canonical/jimbaker) left irc: Quit: Coyote finally caught me | 03:15 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) joined #pypy. | 03:16 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) left irc: Changing host | 03:16 | |
| jimbaker (~jbaker@canonical/jimbaker) joined #pypy. | 03:16 | |
| jimbaker (~jbaker@canonical/jimbaker) left irc: Quit: Coyote finally caught me | 03:21 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) left irc: Remote host closed the connection | 03:22 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy. | 03:22 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) joined #pypy. | 03:24 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) left irc: Changing host | 03:24 | |
| jimbaker (~jbaker@canonical/jimbaker) joined #pypy. | 03:24 | |
| Ademan (~dan@adsl-71-141-224-79.dsl.snfc21.pacbell.net) joined #pypy. | 03:26 | |
| ernop (~ernie@173.233.132.94) joined #pypy. | 03:28 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 03:29 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1854 | 03:39 |
| ita (ita@kde/developer/tnagy) left #pypy. | 03:42 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 03:56 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) left irc: Ping timeout: 244 seconds | 03:56 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 04:01 | |
| ernop (~ernie@173.233.132.94) left irc: Ping timeout: 252 seconds | 04:14 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 258 seconds | 04:26 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 04:27 | |
| CIA-68 | 03justinpeel 07roundup * 10#927/Task takes more time on nightly than CPython 2.7: | 04:30 |
| CIA-68 | I think that a small card ref size for sets/dicts would help (I'll just talk | 04:30 |
| CIA-68 | about dicts from now on, but it applies to sets). However, I also d ... * 14https://bugs.pypy.org/issue927 | 04:30 |
| ders (~ders@202.72.107.130) joined #pypy. | 04:30 | |
| ders | Hello all. It seems though I'm having trouble importing cairo (specifically _cairo) under pypy 1.7 on MacOS X. Any suggestions? | 04:33 |
| ders | It imports smoothly enough on CPython 2.7 | 04:33 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1446 | 04:35 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Operation timed out | 04:36 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/804 | 04:37 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 04:39 | |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 04:42 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 04:43 | |
| kenaan | 12alex_gaynor default 11bb940ca20b4d 15/pypy/module/pypyjit/policy.py: allow inlining into marshal, which is used in cPickle | 04:43 |
| kenaan | 12alex_gaynor default 11bdb51cc58e95 15/pypy/: merged upstream | 04:43 |
| svenstaro (~svenstaro@archlinux/trusteduser/svenstaro) joined #pypy. | 04:44 | |
| svenstaro | hi, pypy 1.7 doesn't make for me :( http://pastie.org/2901611 | 04:45 |
| svenstaro | all I do it python2 translate.py -Ojit | 04:45 |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Ping timeout: 240 seconds | 04:46 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1117 | 04:50 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/619 | 04:50 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 05:00 | |
| canta (~canta@77-20-123-240-dynip.superkabel.de) joined #pypy. | 05:02 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-win-x86-64/builds/50 [12ctismer, win64-stage1] | 05:03 |
| Ademan | hi svenstaro what OS and architecture are you on? | 05:04 |
| svenstaro | arch linux x86_64 | 05:04 |
| svenstaro | I'm the arch packager for pypy | 05:04 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/583 | 05:05 |
| svenstaro | oh actually, the build that failed was in a 32bit build chroot | 05:06 |
| svenstaro | so technically 32bit is my arch | 05:06 |
| svenstaro | x86_64 build is still going | 05:06 |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 05:07 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 05:07 | |
| svenstaro | yeah x86_64 is now compiling | 05:09 |
| svenstaro | i688 already failed during translation | 05:09 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 05:12 | |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 05:12 | |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 05:13 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 05:13 | |
| davisagli (~davisagli@davisagli.com) left irc: Excess Flood | 05:35 | |
| davisagli (~davisagli@davisagli.com) joined #pypy. | 05:36 | |
| DanKluev (~root@93-88-2-081-xdsl.vntc.ru) left irc: Ping timeout: 244 seconds | 05:39 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/278 | 05:41 |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy. | 05:53 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) left irc: Client Quit | 05:53 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy. | 05:55 | |
| Ademan | svenstaro: sorry, any luck? How did you get the source? | 05:56 |
| Nick change: Gulopine -> Gulaway | 05:57 | |
| DanKluev (~root@93-88-1-161-xdsl.vntc.ru) joined #pypy. | 05:58 | |
| __name__ (~name@sburn/devel/name) left irc: Ping timeout: 248 seconds | 05:59 | |
| fijal (~fijal@197.107.25.33) joined #pypy. | 05:59 | |
| fijal | hi | 06:00 |
| __name__ (~name@sburn/devel/name) joined #pypy. | 06:01 | |
| svenstaro | from the bitbucket official download | 06:02 |
| svenstaro | Ademan: also, no luck | 06:02 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/715 | 06:04 |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 06:06 | |
| fijal | svenstaro: can you paste the crash? | 06:07 |
| fijal | and why are you using cpython to translate | 06:07 |
| fijal | it takes aaaaages on 2.6 | 06:07 |
| svenstaro | I got the same results using pypy | 06:07 |
| svenstaro | and I just pasted it, didn't I? | 06:07 |
| fijal | yes yes | 06:08 |
| Ademan | svenstaro: fijal wasn't in here | 06:08 |
| svenstaro | oh | 06:08 |
| Action: fijal can read the logs though | 06:08 | |
| svenstaro | http://pastie.org/2901611 | 06:08 |
| fijal | svenstaro: do you think I can get access to this pdb somehow? | 06:08 |
| fijal | and - do you have any local modifications? | 06:08 |
| Ademan | it'd be really nice if the default translation failure told you a bit about the graph where it exploded | 06:10 |
| Alex_Gaynor | it does | 06:10 |
| svenstaro | alright, shall I redirect the whole make process to a log for you guys? | 06:10 |
| svenstaro | this is the PKGBUILD: | 06:10 |
| svenstaro | http://projects.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD?h=packages/pypy | 06:10 |
| svenstaro | it's what we use on arch for the buildscript and it is really simple | 06:11 |
| fijal | no | 06:12 |
| fijal | I need pdb | 06:12 |
| Ademan | Alex_Gaynor: it's been a while since I translated anything ( :-( ) but I certainly don't see anything in svenstaro's output that describes that | 06:13 |
| svenstaro | fijal: tell me how and you will receive | 06:14 |
| fijal | svenstaro: preferably by ssh connection :) | 06:16 |
| fijal | but let me think for a second | 06:16 |
| fijal | your 32bit chroot must be peculiar | 06:17 |
| fijal | you sure, you don't have any local modifications? | 06:17 |
| svenstaro | yessir, it is installed from scratch each time we build | 06:17 |
| svenstaro | this is how we build all our packages, it just installs a new base system | 06:17 |
| fijal | ok | 06:17 |
| fijal | svenstaro: or you can try to fish where the staticmethod come from | 06:18 |
| fijal | svenstaro: but you need an interactive translation | 06:18 |
| fijal | without --batch | 06:18 |
| fijal | svenstaro: sorry I have to go some time soon, but I would like to investigate some more | 06:21 |
| fijal | will you be online say in an hour or two? | 06:21 |
| fijal | if you can get a pdb by that time it would be awesome | 06:21 |
| svenstaro | I will try to reproduce the error on other machines | 06:21 |
| svenstaro | and I will need your ssh pubkey then or something | 06:22 |
| fijal | cool | 06:22 |
| fijal | http://paste.pocoo.org/show/511269/ | 06:23 |
| fijal | sorry for spammy PM | 06:23 |
| svenstaro | alright | 06:23 |
| fijal | I should get more coffee I was thinking I pasted my private one for a sec.... | 06:25 |
| svenstaro | those start with fairly obvious comments though :P | 06:25 |
| svenstaro | ------ OMG THIS IS YOUR PRIVAT KEY DONT PASTE THIS ------- | 06:26 |
| fijal | never underestimate the lack of caffeine | 06:29 |
| canta (~canta@77-20-123-240-dynip.superkabel.de) left irc: Quit: Leaving. | 06:32 | |
| [mat^2] (~mathias@212.130.113.35) left irc: | 06:33 | |
| fijal (~fijal@197.107.25.33) left irc: Read error: Connection reset by peer | 06:39 | |
| svenstaro | http://pastie.org/private/lmqz4lzwbd20l2pjxh7iyw | 06:43 |
| svenstaro | new error | 06:43 |
| svenstaro | I used pypy this time and didn't use the no-sse2 instruction | 06:43 |
| Garen | svenstaro: Looks like you ran out of memory | 06:47 |
| svenstaro | what's the user space limit on 32bit? 3GB right? | 06:47 |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 06:48 | |
| svenstaro | and 2GB for applications or something | 06:49 |
| svenstaro | well then, how do I make pypy when the process itself terminates? :D | 06:49 |
| Garen | svenstaro: Are you building the 32-bit target on x64? Then you can get a full 4GB | 06:52 |
| Alex_Gaynor | svenstaro: that praticular memoryerror is a bug in pypy 1.6, you can't use it to translate 1.7 | 06:52 |
| Alex_Gaynor | the release notes really should have said that | 06:53 |
| svenstaro | Alex_Gaynor: oh | 06:54 |
| svenstaro | oh... :) | 06:54 |
| svenstaro | Garen: yes, the kernel is 64bit but I build in a 32bit chroot so I'm not technically cross-building | 06:55 |
| svenstaro | doesn't that still impose the restrictions since all my userland stuff is then striclty 32bit? | 06:56 |
| Garen | no | 06:56 |
| Garen | svenstaro: your kernel isn't reserving 1GB or so from the smaller 32-bit address space, since you really have 64-bit | 06:57 |
| svenstaro | off again with python2 I go! | 07:00 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 07:01 | |
| stakkars__ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 07:01 | |
| svenstaro | why did you guys take out the mandelbrot? | 07:04 |
| svenstaro | it was the only thing to keep me entertained | 07:04 |
| svenstaro | now I need to do something productive in the mean time | 07:05 |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Read error: Connection reset by peer | 07:19 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) joined #pypy. | 07:24 | |
| verte (~verte@python/site-packages/verte) joined #pypy. | 07:38 | |
| Ademan (~dan@adsl-71-141-224-79.dsl.snfc21.pacbell.net) left irc: Quit: leaving | 07:42 | |
| asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer | 07:48 | |
| verte (~verte@python/site-packages/verte) left irc: Ping timeout: 252 seconds | 07:49 | |
| fijal (~fijal@196-210-43-166.dynamic.isadsl.co.za) joined #pypy. | 07:50 | |
| derdon (~derdon@p5DE89237.dip.t-dialin.net) joined #pypy. | 07:53 | |
| stakkars_ (~tismer@dslb-088-072-224-185.pools.arcor-ip.net) left irc: Quit: schnarch | 07:54 | |
| fijal | svenstaro: if you use python2.7 over python2.6 it's much faster | 07:55 |
| rokujyouhitoma (~rokujyouh@122x220x64x254.ap122.ftth.ucom.ne.jp) left irc: Remote host closed the connection | 07:55 | |
| svenstaro | well this is arch :) python 2.6 was dropped from it 50 years go | 07:55 |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy. | 07:55 | |
| fijal | ah ok | 07:55 |
| fijal | yes, obviously :) | 07:56 |
| svenstaro | doesnt solve the missing mandelbrot now does it | 07:57 |
| fijal | nope | 07:58 |
| fijal | mandelbrot should be there | 07:58 |
| fijal | as far as I know | 07:58 |
| fijal | it does check if it's a terminal | 07:58 |
| fijal | are you having a weird terminal | 07:58 |
| fijal | ? | 07:58 |
| Action: fijal ponders what to do | 07:59 | |
| svenstaro | oh right, it can't get my tty | 07:59 |
| svenstaro | ye it's arlgiht | 07:59 |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 08:01 | |
| Trundle (~andy@p578bfdcf.dip0.t-ipconnect.de) joined #pypy. | 08:02 | |
| Trundle (~andy@p578bfdcf.dip0.t-ipconnect.de) left irc: Changing host | 08:02 | |
| Trundle (~andy@python/site-packages/trundle) joined #pypy. | 08:02 | |
| svenstaro | on different machine I was just able to make i686 using pypy 1.6 | 08:03 |
| svenstaro | huh | 08:03 |
| fijal | something is broken with your machine... | 08:04 |
| kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/ | 08:06 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy. | 08:10 | |
| svenstaro | oh actually that was a mistake, I used pypy | 08:10 |
| svenstaro | sorry, getting confused with too many remote terminals | 08:10 |
| svenstaro | I gotta make the hostnames marquee or something | 08:10 |
| [Arfreve1] (~Arfrever@minotaur.apache.org) joined #pypy. | 08:11 | |
| [Arfrever] (~Arfrever@apache/committer/Arfrever) left irc: Ping timeout: 248 seconds | 08:12 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) joined #pypy. | 08:15 | |
| fijal_ (~fijal@41.52.68.96) joined #pypy. | 08:15 | |
| fijal (~fijal@196-210-43-166.dynamic.isadsl.co.za) left irc: Disconnected by services | 08:15 | |
| Nick change: fijal_ -> fijal | 08:16 | |
| fijal | svenstaro: different colors | 08:16 |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) left irc: Ping timeout: 240 seconds | 08:18 | |
| tito (~tito@pdpc/supporter/bronze/tito) left irc: Remote host closed the connection | 08:25 | |
| tito (~tito@81-67-113-37.rev.numericable.fr) joined #pypy. | 08:27 | |
| tito (~tito@81-67-113-37.rev.numericable.fr) left irc: Changing host | 08:27 | |
| tito (~tito@pdpc/supporter/bronze/tito) joined #pypy. | 08:27 | |
| senyai (~senyai@62.173.135.205) joined #pypy. | 08:27 | |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Ping timeout: 260 seconds | 08:44 | |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 08:45 | |
| ders (~ders@202.72.107.130) left irc: Ping timeout: 260 seconds | 08:47 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy. | 08:59 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 245 seconds | 09:02 | |
| fijal | where is armin when you need one | 09:03 |
| Alex_Gaynor | fijal: on the mailing list apparently. | 09:04 |
| fijal | right | 09:04 |
| Alex_Gaynor | fijal: btw, hilarious bug: http://ci.django-cms.org/job/Django/501/database=sqlite3,python=pypy1.7/testReport/django.contrib.humanize.tests/HumanizeTests/test_naturaltime/ | 09:04 |
| fijal | justin's observation is kinda cool | 09:04 |
| Alex_Gaynor | which one? | 09:04 |
| fijal | on some recent issue, I lost the link | 09:05 |
| fijal | https://bugs.pypy.org/issue927 | 09:05 |
| fijal | this one | 09:05 |
| fijal | Alex_Gaynor: pretty hilarious | 09:05 |
| Alex_Gaynor | fijal: do you see the cause? it took me a minute :) | 09:06 |
| fijal | no :) | 09:07 |
| Alex_Gaynor | fijal: someone monkey patches datetime.datetime, and datetime.py has isinstance(x, datetime) whereas in CPython it just has a check using C struct, not module name | 09:07 |
| Alex_Gaynor | s/module name/global name/ | 09:07 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 09:08 | |
| Alex_Gaynor | I fixed the bug in django, because it would berak under a CPython with a pure python datetime.py as well | 09:08 |
| fijal | ah | 09:08 |
| timonator | oh, ouch | 09:10 |
| timonator | that's mean | 09:10 |
| iori (~iori@110-133-45-54.rev.home.ne.jp) left irc: Remote host closed the connection | 09:18 | |
| ders (~ders@202.72.107.133) joined #pypy. | 09:18 | |
| G2P (~G2P@fw-asn1.ornis.com) joined #pypy. | 09:21 | |
| svenstaro | alright I narrowed this a little, it explodes with --jit-backend=x86-without-sse2 | 09:21 |
| svenstaro | please confirm | 09:21 |
| svenstaro | it works fine without that | 09:21 |
| fijal | ah | 09:21 |
| fijal | ! | 09:21 |
| fijal | why do you want to do that? | 09:21 |
| fijal | we don't have such a buildbot and indeed it might crash | 09:21 |
| Alex_Gaynor | perhaps he doesn't have sse2 /hides | 09:22 |
| svenstaro | I'm fairly sure a core7 does :) | 09:22 |
| fijal | he's on arch, everything needs sse2 since 1992 | 09:22 |
| svenstaro | we're building it like that because i686 doesn't always have sse2 | 09:22 |
| fijal | I would strongly suggest to not ship a version of pypy with this option | 09:22 |
| fijal | it makes all float operations super-slow | 09:22 |
| svenstaro | frankly if you just give me "that's not supported" I'd be glad to just take that out and be on my way :P | 09:22 |
| Alex_Gaynor | we should either get a buildbot, or remove it | 09:22 |
| fijal | yes | 09:23 |
| Alex_Gaynor | I guess I claim this should be true for everything. | 09:23 |
| fijal | svenstaro: I would strongly suggest not to use it and indeed it's unsupported because we don't have a buildbot | 09:23 |
| svenstaro | that's good enough for me | 09:23 |
| Alex_Gaynor | why do we have it on our site then :/ | 09:23 |
| kenaan | 12arigo default 110261999420cf 15/pypy/jit/metainterp/optimizeopt/test/test_optimizeopt.py: Test for the case of really two distinct constant immutable objects pointing to each other. (works) | 09:24 |
| verte (~verte@python/site-packages/verte) joined #pypy. | 09:39 | |
| ders (~ders@202.72.107.133) left irc: Ping timeout: 252 seconds | 09:41 | |
| h0h0 (c010b8ae@gateway/web/freenode/ip.192.16.184.174) joined #pypy. | 09:45 | |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) joined #pypy. | 09:46 | |
| zk (~is@whatit.is) left irc: Ping timeout: 255 seconds | 09:49 | |
| zk (~is@whatit.is) joined #pypy. | 09:49 | |
| ojii (~ojii@40-34.60-188.cust.bluewin.ch) joined #pypy. | 09:53 | |
| canta (~canta@77-20-123-240-dynip.superkabel.de) joined #pypy. | 10:07 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 260 seconds | 10:08 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 10:11 | |
| kenaan | 12hager ppc-jit-backend 1110f8da3b0282 15/pypy/jit/backend/ppc/ppcgen/codebuilder.py: Added code that verifies sane use of scratch register. | 10:17 |
| fijal (~fijal@41.52.68.96) left irc: Ping timeout: 252 seconds | 10:20 | |
| amaury_ (~amaury_@74.125.57.34) joined #pypy. | 10:21 | |
| kenaan | 12hager ppc-jit-backend 1174db0eef66fe 15/pypy/jit/backend/ppc/ppcgen/codebuilder.py: Allow allocation without setting the scratch register value | 10:29 |
| fijal (~fijal@196-210-43-166.dynamic.isadsl.co.za) joined #pypy. | 10:32 | |
| ders (~ders@202.72.107.130) joined #pypy. | 10:33 | |
| lahwran | why is pypy-c-sandbox not provided as a version that can be downloaded? | 10:43 |
| lahwran | is there something about it that prevents such distribution? | 10:43 |
| lahwran | or is it just that not enough perceived interest has resulted in nobody spending the time to build it on a regular basis | 10:44 |
| fijal (~fijal@196-210-43-166.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds | 10:45 | |
| ders (~ders@202.72.107.130) left irc: Ping timeout: 244 seconds | 10:45 | |
| amaury_ (~amaury_@74.125.57.34) left irc: Ping timeout: 276 seconds | 10:50 | |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 10:50 | |
| amaury_ (~amaury_@74.125.57.34) joined #pypy. | 10:51 | |
| ronny | lahwran: there would be need for extra work to make a usefull distribution of it, since it needs the controller process and the libs for communication with the sandboxed slave | 10:51 |
| lahwran | I see | 10:51 |
| lahwran | so if I build a copy, and then write my own interaction code, it'd be perfectly fine for me to distribute the pypy-c-sandbox binary? | 10:52 |
| witulski (stupsi@fwstups.cs.uni-duesseldorf.de) left #pypy. | 10:52 | |
| ronny | lahwran: as far as i understood, yes | 10:53 |
| mfoord (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy. | 10:53 | |
| mfoord (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host | 10:53 | |
| mfoord (~anonymous@python/psf/voidspace) joined #pypy. | 10:53 | |
| mfoord (~anonymous@python/psf/voidspace) left irc: Client Quit | 10:54 | |
| lahwran | alright, sweet | 10:55 |
| mvt (~mvantelli@87.213.45.85) joined #pypy. | 10:57 | |
| amaury_ (~amaury_@74.125.57.34) left irc: Ping timeout: 240 seconds | 11:09 | |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy | 11:11 | |
| h0h0 (c010b8ae@gateway/web/freenode/ip.192.16.184.174) left irc: Ping timeout: 265 seconds | 11:19 | |
| berdario (~Tiibiidii@host118-77-dynamic.10-79-r.retail.telecomitalia.it) joined #pypy. | 11:21 | |
| berdario (~Tiibiidii@host118-77-dynamic.10-79-r.retail.telecomitalia.it) left irc: Client Quit | 11:22 | |
| berdario (~Tiibiidii@host118-77-dynamic.10-79-r.retail.telecomitalia.it) joined #pypy. | 11:24 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) left irc: Remote host closed the connection | 11:28 | |
| Sho_ (~EHS1@kde/hein) joined #pypy. | 11:36 | |
| lahwran | why is --withoutmod-unicodedata the default? | 11:40 |
| lahwran | ditto for --withoutmod-time | 11:41 |
| lahwran | what effects will enabling these have | 11:41 |
| lahwran | again for thread and bz2 | 11:41 |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) left irc: Read error: Connection reset by peer | 11:44 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) joined #pypy. | 11:44 | |
| fijal (~fijal@197.171.34.193) joined #pypy. | 11:46 | |
| fijal | meh | 11:59 |
| fijal | so disabling sse does not work | 11:59 |
| fijal | what do we do? | 11:59 |
| lahwran | hm? | 12:03 |
| fijal | lahwran: hi :) | 12:04 |
| lahwran | greetings | 12:04 |
| fijal | lahwran: default is --withmod-unicodedata | 12:04 |
| fijal | to be precise, default is --allworkingmodules | 12:04 |
| fijal | oh, you mean the sandbox version? | 12:05 |
| lahwran | how in the world are you responding to what i said before you joined | 12:05 |
| lahwran (lahwran@unaffiliated/lahwran) left #pypy. | 12:05 | |
| lahwran (lahwran@unaffiliated/lahwran) joined #pypy. | 12:05 | |
| lahwran | are you really looking in the logs just for that? ._. | 12:05 |
| fijal | yes | 12:05 |
| fijal | I read pypy logs | 12:05 |
| lahwran | okay. I don't feel creeped out anymore./ | 12:06 |
| lahwran | anyway, yeah, I'm building the sandbox version and I was looking through the translate.py options | 12:06 |
| lahwran | and I was curious why those default to off | 12:06 |
| fijal | sandbox has modules turned off | 12:07 |
| fijal | because well, it makes sense :) | 12:07 |
| chronitis (~chronitis@79-68-125-122.dynamic.dsl.as9105.com) joined #pypy. | 12:07 | |
| lahwran | not really | 12:07 |
| fijal | but unicodedata and time are almost always off | 12:07 |
| lahwran | why? | 12:07 |
| lahwran | time is one of the most essential modules | 12:07 |
| fijal | er | 12:08 |
| fijal | time and unicodedata are always on :) | 12:09 |
| fijal | I think we have two versions of time thought | 12:09 |
| fijal | time and rctime | 12:09 |
| lahwran | err | 12:09 |
| lahwran | lemme change my question: with --withoutmod-time, will 'import time' fail? | 12:09 |
| fijal | I don't think you can build without time, but I'm unsure | 12:09 |
| fijal | you certainly can't build without unicodedata | 12:10 |
| amaury_ (~amaury_@74.125.57.34) joined #pypy. | 12:13 | |
| amaury_ | Bad Cython optimization: if(PyList_CheckExact(L)) Py_SIZE(L) -= 1; | 12:17 |
| amaury_ | (used when translating x.pop()) | 12:17 |
| fijal | ups | 12:18 |
| fijal | you can create a very bad benchmark | 12:18 |
| amaury_ | it also checks for overallocation | 12:19 |
| fijal (~fijal@197.171.34.193) left irc: Read error: Connection reset by peer | 12:36 | |
| berdario (~Tiibiidii@host118-77-dynamic.10-79-r.retail.telecomitalia.it) left irc: Quit: Ex-Chat | 12:40 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 240 seconds | 12:43 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 12:43 | |
| stakkars_ (~tismer@i59F72840.versanet.de) joined #pypy. | 12:44 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 244 seconds | 12:46 | |
| verte (~verte@python/site-packages/verte) left irc: Quit: ~~~ Crash in JIT! | 12:47 | |
| jonvk (~jonathan@bas1-montreal29-1242341904.dsl.bell.ca) joined #pypy. | 12:54 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 13:00 | |
| lahwran | drat | 13:02 |
| lahwran | it seems to have been overusing /tmp, which was low on space | 13:03 |
| lahwran | and now it's crashed with a "no space left on device" | 13:03 |
| JaredW (~jared@061093200075.ctinets.com) joined #pypy. | 13:03 | |
| Rhy0lite (dje@nat/ibm/x-rpxbvgucbnivsokf) joined #pypy. | 13:06 | |
| lahwran | https://gist.github.com/ab8431c4000ca7afc2de | 13:07 |
| lahwran | I was running the translator in an old partition I hadn't used in a while | 13:07 |
| lahwran | because my root partition is low on space | 13:07 |
| lahwran | can I safely resume translating by running the command again? | 13:07 |
| lahwran | the command I used was: | 13:07 |
| lahwran | ../../../../pypy-1.7/bin/pypy translate.py -O2 --sandbox targetpypystandalone.py | 13:07 |
| lahwran | bah I'm just going to run the command again | 13:08 |
| lahwran | hopefully it resumes where it left off | 13:08 |
| lahwran | nope | 13:09 |
| lahwran | grr _ | 13:09 |
| chronitis (~chronitis@79-68-125-122.dynamic.dsl.as9105.com) left irc: Ping timeout: 252 seconds | 13:12 | |
| fijal (~fijal@41.55.152.102) joined #pypy. | 13:12 | |
| lahwran | well, there. I mount --bind'ed /tmp onto the spare partition | 13:14 |
| lizardo (~lizardo@189.2.128.130) joined #pypy. | 13:14 | |
| fijal | lahwran: no, you can't resume | 13:17 |
| fijal | it's a bit of a pity | 13:17 |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Read error: Operation timed out | 13:18 | |
| chronitis (~chronitis@88-104-124-49.dynamic.dsl.as9105.com) joined #pypy. | 13:26 | |
| kenaan | 12ctismer win64-stage1 11d40a0bdc27aa 15/pypy/: Merge with default | 13:32 |
| lahwran | suddenly, builds | 13:33 |
| chronitis | do pull requests to bitbucket get checked or do they need to be notified somewhere else for review? | 13:34 |
| exarkun | they're not automatically merged to default, if that's what you mean | 13:35 |
| chronitis | of course - I meant is it necessary to open a bug or similar at the same time? | 13:36 |
| fijal | chronitis: no | 13:42 |
| fijal | I've seen your request | 13:42 |
| fijal | amaury_: feel like reviewing it? it goes to py3k branch | 13:43 |
| fijal | also bitbucket crashed on me when I tried to look there :/ | 13:43 |
| chronitis | ...useful of it | 13:43 |
| fijal | chronitis: pester amaury, he's the py3k master | 13:44 |
| chronitis | will do | 13:44 |
| arigato (~arigo@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 13:48 | |
| fijal | Arach: hi | 13:48 |
| fijal | arigato: hi | 13:48 |
| arigato | hi | 13:48 |
| fijal | good release I would say :) | 13:49 |
| arigato | :-) | 13:49 |
| arigato | why? | 13:49 |
| fijal | because not too many complaints | 13:49 |
| fijal | because of | 13:50 |
| fijal | arigato: have you seen justin's finds about slow large dict perf? | 13:50 |
| arigato | (1) indeed, good. (2) no? on the bug report? | 13:51 |
| fijal | (2) yes | 13:51 |
| fijal | I mean noone reported crashes that are seriously bad | 13:51 |
| arigato | yes | 13:53 |
| arigato | apart from › on the blog | 13:53 |
| arigato | of course we're all going to spend hours trying to reproduce the segfault he doesn't want to tell us anything about | 13:53 |
| fijal | I would not count that one as a bug report | 13:56 |
| fijal | I suppose | 13:57 |
| Nick change: Gulaway -> Gulopine | 13:58 | |
| thinker (~thinker@123-194-52-90.dynamic.kbronet.com.tw) joined #pypy. | 14:00 | |
| Rhy0lite | arigato, fijal: Is there any detailed analysis of which PyPy optimizations contribute to its performance advantage for richards benchmark? | 14:05 |
| fijal | Rhy0lite: not particularly, because a lot of optimizations are mutually benefiting each other | 14:05 |
| fijal | if you do inlining, your stuff can stay virtual | 14:06 |
| fijal | etc. etc. | 14:06 |
| fijal | you can try playing with various --jit enable_opts=... | 14:06 |
| fijal | but that won't let you disable say virtualizables | 14:06 |
| Rhy0lite | ok, thanks | 14:06 |
| fijal | why do you ask? | 14:06 |
| fijal | as far as I know carl did some analysis for papers, but they were usually about simpler languages than python | 14:07 |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 14:07 | |
| cfbolz | hello | 14:07 |
| Rhy0lite | hello | 14:07 |
| Rhy0lite | speaking of | 14:07 |
| Rhy0lite | can we do this all the time? | 14:07 |
| fijal | measure impact of various optimizations? | 14:07 |
| fijal | hi carl | 14:07 |
| Rhy0lite | no, talk about Carl and have him appear | 14:08 |
| Rhy0lite | :-) | 14:08 |
| fijal | haha :) | 14:08 |
| cfbolz | :-) | 14:08 |
| cfbolz | how can I be of service? | 14:08 |
| Rhy0lite | fijal: some of my colleagues still are trying to understand why IBM's JIT is not as fast as PyPy for some simple benchmarks | 14:08 |
| fijal | Rhy0lite: IBM JIT being in this example .... ? | 14:09 |
| Rhy0lite | yes | 14:09 |
| fijal | I mean what sort of architecture we're talking about | 14:09 |
| Rhy0lite | cfbolz: the question is which pypy optimizations contribute to richards performance | 14:09 |
| cfbolz | Rhy0lite: that's peng et al, I assume? | 14:09 |
| Rhy0lite | cfbolz: yes, but more Jose | 14:10 |
| fijal | Rhy0lite: you can always show them compiled assembler using jitviewer | 14:10 |
| fijal | it'll show up why | 14:10 |
| fijal | cfbolz: can I merge default to list strategies? | 14:10 |
| fijal | so I have a cleaner review situation | 14:10 |
| cfbolz | fijal: sure | 14:10 |
| arigato | cfbolz: hi | 14:10 |
| cfbolz | arigato: hi, will review your code soon | 14:10 |
| fijal | cfbolz: set-strategies is also ready? | 14:11 |
| arigato | cfbolz: ok :-) | 14:11 |
| Rhy0lite | fijal: they are trying to perform some optimizations manually in their framework | 14:11 |
| Rhy0lite | for simple testcases | 14:11 |
| cfbolz | fijal: yes, reviews welcome | 14:11 |
| Action: arigato tried to explain in a blog comment how open source works | 14:11 | |
| cfbolz | fijal: but I would like to merge them after the other, and have a buildbot run in between | 14:11 |
| fijal | cfbolz: sure | 14:11 |
| cfbolz | Rhy0lite: I guess that's easy enough to find out by disabling the optimizations and see | 14:12 |
| Rhy0lite | arigato: good luck | 14:12 |
| fijal | Rhy0lite: did they got rid of frames and boxing? | 14:12 |
| cfbolz | Rhy0lite: but I think that it's very much a combination of many things | 14:12 |
| fijal | that's a good first step | 14:12 |
| cfbolz | Rhy0lite: anyway, are they running the JIT on jython? or on the java version of richards? | 14:12 |
| Rhy0lite | cfbolz: this is Fiorano, using IBM's JIT | 14:13 |
| cfbolz | Rhy0lite: yes, I guessed that | 14:13 |
| cfbolz | but which version of richards is it? | 14:14 |
| Rhy0lite | cfbolz: I have tried to explain PyPy's virtuals and escape analysis a couple of times | 14:14 |
| fijal | haha | 14:14 |
| Rhy0lite | but I don't think they quite get it | 14:14 |
| fijal | cfbolz: it's totally unmergeable automatically | 14:14 |
| Action: fijal fights | 14:14 | |
| cfbolz | fijal: huh? why? | 14:14 |
| Rhy0lite | cfbolz: Python version of richards, from U-S benchmarks | 14:14 |
| cfbolz | Rhy0lite: ok, on jython then | 14:14 |
| fijal | cfbolz: I wonder as well | 14:15 |
| Rhy0lite | Fiorano vs PyPy vs CPython | 14:15 |
| Rhy0lite | I don't understand why you keep saying jython | 14:15 |
| Action: arigato googles Fiorano | 14:15 | |
| Rhy0lite | arigato: I don't think you will find anything | 14:15 |
| cfbolz | Rhy0lite: I think I forgot what fiorano is, could you remind me please? | 14:15 |
| fijal | Rhy0lite: can you brief me a gain on the architecture of Fiorano? I remember you did that, but I don't remember details :) | 14:15 |
| fijal | hahaha :] | 14:16 |
| arigato | indeed | 14:16 |
| arigato | a secret closed source project? bah :-/ | 14:16 |
| Rhy0lite | Think Unladen-Swallow design (CPython as VM) connected to IBM's Testarossa JIT instead of LLVM | 14:16 |
| kenaan | 12fijal list-strategies 11424f89e54a68 15/: merge default | 14:16 |
| fijal | Rhy0lite: well ok, then how much of runtime is CPython and how much it is exposed? | 14:17 |
| Rhy0lite | but more invasive changes in CPython than Collin wanted to make | 14:17 |
| cfbolz | ok, now I remember, sorry | 14:17 |
| fijal | because on of the pypy strengths is about not calling into a runtime with fixed API | 14:17 |
| fijal | http://www.trl.ibm.com/projects/jit/index_e.htm | 14:17 |
| fijal | I love the animated gif | 14:18 |
| fijal | with all the broken fonts | 14:18 |
| Rhy0lite | :-) | 14:18 |
| arigato | from Psyco experience, reference counting slows things down in a JIT | 14:18 |
| Rhy0lite | we are specializing some parts of the runtime | 14:18 |
| fijal | Rhy0lite: some parts (as we have learned) is by far not good enough for anything but the most simple stuff | 14:18 |
| Rhy0lite | fijal: I agree | 14:18 |
| fijal | if you have an API that accepts boxes, it's more costly to have an optimizing JIT | 14:19 |
| Rhy0lite | I'm not trying to defend this approach | 14:19 |
| fijal | because you keep unboxing and boxing stuff | 14:19 |
| Rhy0lite | as Carl well knows | 14:19 |
| fijal | I know as well | 14:19 |
| Rhy0lite | :-) | 14:19 |
| fijal | I'm giving you arguments for the debate | 14:19 |
| Rhy0lite | yes, thanks | 14:19 |
| fijal | which surely goes on :) | 14:19 |
| Action: Rhy0lite is tired of debates with his colleagues | 14:19 | |
| cfbolz | Rhy0lite: my totally uneducated guess why PyPy wins is: virtuals + frames + faster lookups on classes | 14:19 |
| fijal | do you have the equivalent of jitviewer? | 14:19 |
| arigato | the technologies might be different, but it sounds similar to the HotPy VM | 14:19 |
| cfbolz | arigato: yes | 14:20 |
| arigato | "good luck" | 14:20 |
| Rhy0lite | arigato: to me or to Fiorano project? | 14:20 |
| arigato | to the Fiorano project | 14:21 |
| Rhy0lite | yeah ;-) | 14:21 |
| fijal | Rhy0lite: don't know about you, but such "good luck" from armin would be a good reason alone to reconsider your approach :) | 14:21 |
| Rhy0lite | fijal: this is why I am helping with PyPy | 14:21 |
| arigato | no no, I'm not saying e.g. HotPy will fail | 14:21 |
| cfbolz | it's just too much work | 14:21 |
| arigato | it may work, but it will be heavy, and getting more and more heavy as Python versions evolve | 14:22 |
| Rhy0lite | fijal: a colleagues made a number of manual transformations to benchmarks like nbody and float | 14:22 |
| Rhy0lite | and found a few, key optimizations that were important | 14:22 |
| fijal | sure | 14:22 |
| Rhy0lite | I believe he expects the same for richards | 14:22 |
| fijal | but those are "easy" cases | 14:22 |
| fijal | richards is "moderately easy" | 14:22 |
| Rhy0lite | fijal: shh! | 14:22 |
| fijal | ;] | 14:23 |
| arigato | no, richards is easy too | 14:23 |
| cfbolz | the big difference between those easy cases and richards is really better class handling | 14:23 |
| cfbolz | richards is much more object-oriented | 14:23 |
| Rhy0lite | yes, I think richards is more pervasive | 14:23 |
| arigato | yes, richards is testing method calls, and attribute gets and sets (not object creation) | 14:23 |
| cfbolz | arigato: exactly | 14:23 |
| fijal | arigato: I would like to point out that our first JIT didn't get fast richards | 14:23 |
| fijal | so it's "easy" | 14:23 |
| cfbolz | fijal: no, wait | 14:24 |
| fijal | no? | 14:24 |
| cfbolz | the first jit could have gotten a fast richard, if we had had mapdicts/version tags at the time | 14:24 |
| Rhy0lite | so richards mostly needs something like hidden classes? | 14:24 |
| cfbolz | Rhy0lite: in addition to the things float/nbody need, yes | 14:24 |
| arigato | I'm just saying it's an "easy" case because it's not using any deep Python tricks, to the point that there is little to manually optimize | 14:24 |
| Nick change: ojii -> ojii|desertbus | 14:24 | |
| cfbolz | Rhy0lite: I guess the right pointer is this years icooolps paper :-) http://dl.acm.org/citation.cfm?id=2069181 | 14:25 |
| Rhy0lite | thanks | 14:29 |
| Rhy0lite | and thanks for the feedback | 14:29 |
| Rhy0lite | need to get ready for my next meeting | 14:29 |
| kenaan | 12fijal list-strategies 114566b449070c 15/pypy/module/__builtin__/functional.py: remove dead/commented out code | 14:31 |
| fijal | cfbolz: do you have a student who want to make an efficient bytearray type? | 14:33 |
| kenaan | 12fijal default 11acfe60970d50 15/pypy/doc/project-ideas.rst: add a potential project idea | 14:35 |
| cfbolz | fijal: efficient in what sense? | 14:36 |
| santagada (~leonardo@187.113.196.61) joined #pypy. | 14:36 | |
| fijal | for example so file.readinto would not make a copy | 14:36 |
| kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy. | 14:37 | |
| cfbolz | fijal: ah. we can see how long lukas takes with instances | 14:37 |
| fijal | and a lot of operations delegate to string | 14:37 |
| fijal | like __contains__ | 14:37 |
| cfbolz | ah, right | 14:38 |
| cfbolz | that's very annoying to fix, of course | 14:38 |
| fijal | of course | 14:38 |
| fijal | we could try to share code between string/unicode and bytearray | 14:39 |
| fijal | or I don't know | 14:39 |
| fijal | but it's interesting to experiment with | 14:39 |
| fijal | or annoying | 14:39 |
| fijal | a bit more interesting would be a better buffer interface (maybe?) | 14:40 |
| cfbolz | or something | 14:40 |
| kenaan | 12fijal default 1162e07d175dd7 15/pypy/doc/project-ideas.rst: add a potential project idea | 14:41 |
| kenaan | 12fijal default 11ffe6c1bffaa0 15/: dumb merge | 14:41 |
| fijal | either way we lack a consistent view on "how do you use an efficient mutable buffer in pypy" | 14:41 |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy. | 14:41 | |
| Nick change: Gulopine -> Gulaway | 14:42 | |
| chronitis | given that a lot of bytearray methods call the stringobject equivalents, is there a good reason not to automatically delegate bytearray to bytes and only write implementations for the cases where we'd end up iterating over a list of characters anyway? | 14:52 |
| kenaan | 12l.diekmann type-specialized-instances 11c31774e1542e 15/pypy/objspace/std/: added tests and fixes for overwriting attributes with another type and for untaggable ints | 14:52 |
| chronitis | (this might apply more in py3k than default) | 14:55 |
| fijal | chronitis: bytearray is a disaster mostly in pypy | 14:55 |
| fijal | since a lot of those methods require conversion and coping | 14:55 |
| fijal | copying even | 14:56 |
| fijal | assert isinstance(wrappeditems, list) | 14:59 |
| fijal | cfbolz: this is in W_ListObject.__init__ | 14:59 |
| fijal | what does that do? | 14:59 |
| chronitis | specialising some of them to avoid copying would require hacking rstring to accept arrays, I think | 14:59 |
| fijal | chronitis: wrong level I believe :) | 15:00 |
| chronitis | fair enough | 15:00 |
| amaury_ | fijal: I can review the pull requests later tonight | 15:01 |
| amaury_ | or are you discussing them at the moment? | 15:01 |
| fijal | amaury_: no | 15:01 |
| chronitis | (I have been playing with bytearray a bit, but that's not in the pull request) | 15:03 |
| fijal | chronitis: I suggest coming up with a solution to the current mess and discussing it here if you're interested | 15:04 |
| chronitis | that would require coming up with a solution ;) | 15:04 |
| fijal | we can discuss it still, but I'm out of opinions for now :) | 15:04 |
| antocuni (~antocuni@host142-66-static.94-94-b.business.telecomitalia.it) joined #pypy. | 15:06 | |
| fijal | cfbolz: also, why there is | 15:13 |
| fijal | except IndexError: raise | 15:13 |
| cfbolz | fijal: that's needed, because otherwise you don't get an IndexError | 15:16 |
| cfbolz | I guess a comment would help | 15:16 |
| fijal | ok | 15:16 |
| fijal | there is an untested path in setslice | 15:16 |
| cfbolz | about the assert: I guess we can kill it now, in the beginning we ran into it | 15:16 |
| cfbolz | untested path :-( | 15:16 |
| cfbolz | add an XXX | 15:16 |
| fijal | ok | 15:16 |
| fprimex (~fprimex@brent-macbook.sc.fsu.edu) joined #pypy. | 15:18 | |
| kenaan | 12fijal list-strategies 115f5ee7044ff7 15/pypy/objspace/std/listobject.py: untested paths | 15:18 |
| engla (~ulrik@wikipedia/Sverdrup) joined #pypy. | 15:19 | |
| whitelynx (~whitelynx@63.241.75.144) joined #pypy. | 15:22 | |
| PiotrSikora (~none@nginx/adept/piotrsikora) left irc: Excess Flood | 15:23 | |
| PiotrSikora (~none@nginx/adept/piotrsikora) joined #pypy. | 15:24 | |
| ahall (~ahall@2001:bd0:100::666) joined #pypy. | 15:28 | |
| kkris (~kris@80-123-47-149.adsl.highway.telekom.at) joined #pypy. | 15:29 | |
| fijal (~fijal@41.55.152.102) left irc: Read error: Connection reset by peer | 15:37 | |
| JaredW (~jared@061093200075.ctinets.com) left irc: Ping timeout: 252 seconds | 15:39 | |
| maxyz (~maxy@186.23.74.228) left irc: Ping timeout: 248 seconds | 15:44 | |
| kenaan | 12arigo default 11cecce64567b2 15/pypy/jit/metainterp/optimizeopt/virtualize.py: (cfbolz, arigo) Clarify the intention. | 15:46 |
| kenaan | 12cfbolz default 113b335c5116cf 15/pypy/jit/metainterp/optimizeopt/optimizer.py: add a (probably not too costly) sanity check to make sure that the traces the optimizer produces have the SSA prop... | 15:47 |
| JaredW (~jared@061093200075.ctinets.com) joined #pypy. | 15:52 | |
| JaredW (~jared@061093200075.ctinets.com) left irc: Remote host closed the connection | 15:57 | |
| JaredW (~jared@061093200075.ctinets.com) joined #pypy. | 15:57 | |
| mvt (~mvantelli@87.213.45.85) left irc: Quit: This computer has gone to sleep | 15:59 | |
| maxyz (~maxy@186.23.74.228) joined #pypy. | 15:59 | |
| svenstaro (svenstaro@archlinux/trusteduser/svenstaro) left #pypy ("WeeChat 0.3.6"). | 16:01 | |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) left irc: Quit: Leaving... | 16:05 | |
| chronitis (~chronitis@88-104-124-49.dynamic.dsl.as9105.com) left irc: Read error: Operation timed out | 16:07 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) joined #pypy. | 16:20 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) left irc: Quit: etrepum | 16:30 | |
| `fox` (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) joined #pypy. | 16:38 | |
| [Arfreve1] (~Arfrever@minotaur.apache.org) left irc: Quit: leaving | 16:43 | |
| [Arfrever] (~Arfrever@apache/committer/Arfrever) joined #pypy. | 16:45 | |
| jnoller (~jnoller@vpn.nasuni.com) joined #pypy. | 16:49 | |
| Trundle (~andy@python/site-packages/trundle) left irc: Remote host closed the connection | 16:55 | |
| asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy. | 17:02 | |
| G2P (~G2P@fw-asn1.ornis.com) left irc: Quit: Leaving. | 17:08 | |
| voidspace (~voidspace@python/psf/voidspace) left irc: Read error: Connection reset by peer | 17:09 | |
| voidspace (~voidspace@87-194-212-65.bethere.co.uk) joined #pypy. | 17:10 | |
| voidspace (~voidspace@87-194-212-65.bethere.co.uk) left irc: Changing host | 17:10 | |
| voidspace (~voidspace@python/psf/voidspace) joined #pypy. | 17:10 | |
| arigato (~arigo@fwstups.cs.uni-duesseldorf.de) left irc: Quit: See you | 17:10 | |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy. | 17:10 | |
| voidspace (~voidspace@python/psf/voidspace) left irc: Client Quit | 17:10 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Ping timeout: 240 seconds | 17:11 | |
| antocuni (~antocuni@host142-66-static.94-94-b.business.telecomitalia.it) left irc: Ping timeout: 252 seconds | 17:13 | |
| DasIch (~dasich@p3E99064C.dip.t-dialin.net) joined #pypy. | 17:20 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) joined #pypy. | 17:21 | |
| fschwidom (~fschwidom@46.115.21.179) joined #pypy. | 17:25 | |
| chronitis (~chronitis@88-104-124-215.dynamic.dsl.as9105.com) joined #pypy. | 17:27 | |
| fox__ (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) joined #pypy. | 17:28 | |
| `fox` (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) left irc: Read error: Connection reset by peer | 17:28 | |
| stakkars_ (~tismer@i59F72840.versanet.de) left irc: Quit: schnarch | 17:40 | |
| stakkars_ (~tismer@i59F72840.versanet.de) joined #pypy. | 17:43 | |
| jnoller (~jnoller@vpn.nasuni.com) left irc: | 17:44 | |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) left irc: Quit: Leaving | 17:44 | |
| kenaan | 12ctismer win64-stage1 118a6f3e90b28c 15/pypy/__init__.py: test_zpy works on win64 | 17:47 |
| stakkars_ (~tismer@i59F72840.versanet.de) left irc: Quit: schnarch | 17:49 | |
| kenaan | 12hager ppc-jit-backend 11c489a73eaf5d 15/pypy/jit/backend/ppc/ppcgen/: Started implementation of CALL_ASSEMBLER | 17:51 |
| senyai (~senyai@62.173.135.205) left irc: Quit: senyai | 17:55 | |
| freenizzle (~crazy@pool-71-98-178-103.tampfl.dsl-w.verizon.net) joined #pypy. | 17:56 | |
| freenizzle (~crazy@pool-71-98-178-103.tampfl.dsl-w.verizon.net) left irc: Remote host closed the connection | 18:01 | |
| JaredW (~jared@061093200075.ctinets.com) left irc: Ping timeout: 276 seconds | 18:01 | |
| srin (~srin@257WIRELESS.CNS.NYU.EDU) joined #pypy. | 18:05 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) left irc: Remote host closed the connection | 18:05 | |
| stakkars_ (~tismer@g225044163.adsl.alicedsl.de) joined #pypy. | 18:15 | |
| amaury_ (~amaury_@74.125.57.34) left irc: Read error: Operation timed out | 18:20 | |
| fmilo (~mist0@cpe-76-168-60-20.socal.res.rr.com) joined #pypy. | 18:20 | |
| dmalcolm (david@nat/redhat/x-msmgylgwalntazsi) joined #pypy. | 18:21 | |
| srin (~srin@257WIRELESS.CNS.NYU.EDU) left irc: Ping timeout: 244 seconds | 18:26 | |
| fmilo (~mist0@cpe-76-168-60-20.socal.res.rr.com) left irc: Quit: fmilo | 18:26 | |
| fox__ (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) left irc: Ping timeout: 260 seconds | 18:29 | |
| fijal (~fijal@41.55.74.39) joined #pypy. | 18:31 | |
| engla (~ulrik@wikipedia/Sverdrup) left irc: Ping timeout: 252 seconds | 18:33 | |
| ojii|desertbus (~ojii@40-34.60-188.cust.bluewin.ch) left irc: Quit: Leaving | 18:37 | |
| tilgovi (~randall@c-98-210-155-124.hsd1.ca.comcast.net) joined #pypy. | 18:38 | |
| tilgovi (~randall@c-98-210-155-124.hsd1.ca.comcast.net) left irc: Changing host | 18:38 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 18:38 | |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer | 18:38 | |
| voidspace (~voidspace@87-194-212-65.bethere.co.uk) joined #pypy. | 18:41 | |
| voidspace (~voidspace@87-194-212-65.bethere.co.uk) left irc: Changing host | 18:41 | |
| voidspace (~voidspace@python/psf/voidspace) joined #pypy. | 18:41 | |
| stakkars__ (~tismer@g225044163.adsl.alicedsl.de) joined #pypy. | 18:52 | |
| stakkars_ (~tismer@g225044163.adsl.alicedsl.de) left irc: Read error: Connection reset by peer | 18:52 | |
| `fox` (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) joined #pypy. | 18:57 | |
| fijal | Rhy0lite: any luck persuaiding anyone? | 19:03 |
| nickik (~nick@217.150.239.67) joined #pypy. | 19:04 | |
| Rhy0lite | persuading anyone of what? | 19:05 |
| santagada (~leonardo@187.113.196.61) left irc: Ping timeout: 240 seconds | 19:05 | |
| kenaan | 12edelsohn ppc-jit-backend 11ecee5b13c5e2 15/pypy/jit/backend/ppc/ppcgen/opassembler.py: Make PPC64 _emit_call abi offset clearer. | 19:06 |
| srin (~srin@257WIRELESS.CNS.NYU.EDU) joined #pypy. | 19:06 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy. | 19:19 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Remote host closed the connection | 19:20 | |
| mcdonc_ (~mcdonc@ip72-209-213-54.dc.dc.cox.net) left irc: Remote host closed the connection | 19:23 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 19:25 | |
| mcdonc (~mcdonc@ip72-209-213-54.dc.dc.cox.net) joined #pypy. | 19:29 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) joined #pypy. | 19:39 | |
| fijal (~fijal@41.55.74.39) left irc: Read error: Connection reset by peer | 19:39 | |
| durin42 (~durin@adium/durin42) left irc: Ping timeout: 245 seconds | 19:44 | |
| srin (~srin@257WIRELESS.CNS.NYU.EDU) left irc: Ping timeout: 252 seconds | 19:44 | |
| Ademan (~dan@78.sub-97-152-91.myvzw.com) joined #pypy. | 19:46 | |
| durin42 (~durin@173-15-14-174-Illinois.hfc.comcastbusiness.net) joined #pypy. | 19:46 | |
| durin42 (~durin@173-15-14-174-Illinois.hfc.comcastbusiness.net) left irc: Changing host | 19:46 | |
| durin42 (~durin@adium/durin42) joined #pypy. | 19:46 | |
| etrepum (~bob@207.239.114.206) joined #pypy. | 19:46 | |
| ojii (~ojii@47-78.78-83.cust.bluewin.ch) joined #pypy. | 19:47 | |
| ojii (~ojii@47-78.78-83.cust.bluewin.ch) left irc: Remote host closed the connection | 19:47 | |
| DasIch_ (~dasich@p4FFDE8AC.dip.t-dialin.net) joined #pypy. | 19:53 | |
| DasIch (~dasich@p3E99064C.dip.t-dialin.net) left irc: Ping timeout: 258 seconds | 19:57 | |
| fschwidom (~fschwidom@46.115.21.179) left irc: Remote host closed the connection | 20:00 | |
| thoughtpolice (~thoughtpo@li111-206.members.linode.com) left irc: Read error: Operation timed out | 20:03 | |
| jml (~jml@canonical/launchpad/jml) left irc: Ping timeout: 240 seconds | 20:04 | |
| srin (~srin@NYUFGA-WLESSAUTHCLIENTS-02.NATPOOL.NYU.EDU) joined #pypy. | 20:05 | |
| thoughtpolice (~thoughtpo@li111-206.members.linode.com) joined #pypy. | 20:09 | |
| jml (~jml@mumak.net) joined #pypy. | 20:09 | |
| jml (~jml@mumak.net) left irc: Changing host | 20:09 | |
| jml (~jml@canonical/launchpad/jml) joined #pypy. | 20:09 | |
| Shanita (~John@osbk-4db14086.pool.mediaWays.net) left irc: Read error: Connection reset by peer | 20:10 | |
| Kaskuka (~John@osbk-4db14086.pool.mediaWays.net) joined #pypy. | 20:13 | |
| Ademan (~dan@78.sub-97-152-91.myvzw.com) left irc: Quit: leaving | 20:17 | |
| camara (8c9d2633@gateway/web/freenode/ip.140.157.38.51) joined #pypy. | 20:18 | |
| davisagli (~davisagli@davisagli.com) left irc: Excess Flood | 20:22 | |
| davisagli (~davisagli@davisagli.com) joined #pypy. | 20:23 | |
| ericflo (~ericflo@75.103.8.110) joined #pypy. | 20:28 | |
| camara (8c9d2633@gateway/web/freenode/ip.140.157.38.51) left #pypy. | 20:29 | |
| srin (~srin@NYUFGA-WLESSAUTHCLIENTS-02.NATPOOL.NYU.EDU) left irc: Ping timeout: 260 seconds | 20:37 | |
| srin (~srin@NYUFGA-WLESSAUTHCLIENTS-02.NATPOOL.NYU.EDU) joined #pypy. | 20:39 | |
| kenaan | 12alex_gaynor default 11b38fa1e72083 15/lib_pypy/_sqlite3.py: Added sqlite3.Connection.enable_load_extension, untested because it requires your libsqlite to have been comp... | 20:39 |
| kenaan | 12alex_gaynor default 11d550918b20a6 15/pypy/: merged upstream | 20:39 |
| kkris (~kris@80-123-47-149.adsl.highway.telekom.at) left irc: Quit: Leaving. | 20:50 | |
| lizardo (~lizardo@189.2.128.130) left irc: Quit: Leaving | 20:51 | |
| stakkars__ (~tismer@g225044163.adsl.alicedsl.de) left irc: Quit: schnarch | 20:52 | |
| `fox` (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) left irc: Read error: Connection reset by peer | 20:52 | |
| `fox` (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) joined #pypy. | 20:53 | |
| chronitis (~chronitis@88-104-124-215.dynamic.dsl.as9105.com) left irc: Quit: Lost terminal | 20:53 | |
| Rhy0lite (dje@nat/ibm/x-rpxbvgucbnivsokf) left irc: Quit: Leaving | 20:56 | |
| Alex_Gaynor | amaury_: ping | 20:58 |
| amaury_ | pong | 20:58 |
| Alex_Gaynor | amaury_: Do you know what exactly is supposed to be passed to an errcheck function in ctypes? | 20:58 |
| Alex_Gaynor | CPython seems to be passing something with an ._obj (a CArgObject) and we're passing the value directly. | 20:59 |
| amaury_ | IIRC, a callable that takes the return value as argument | 20:59 |
| amaury_ | ah | 20:59 |
| Alex_Gaynor | I mean it takes (result, func, cargs) | 21:00 |
| Alex_Gaynor | but what types are cargs | 21:00 |
| kenaan | 12amauryfa py3k 1132c718fe8261 15/pypy/module/_io/interp_bufferedio.py: Buffered IO should catch EINTR and retry. | 21:00 |
| kenaan | 12amauryfa py3k 114a4118ac402e 15/: hg merge default | 21:00 |
| amaury_ | in ctypes, all arguments are wrapped in CArgObject | 21:00 |
| amaury_ | (don't be confused by a class of the same name in pypy) | 21:00 |
| Alex_Gaynor | amaury_: I can't imagine why I'd be confused :) | 21:01 |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Ping timeout: 258 seconds | 21:01 | |
| Alex_Gaynor | amaury_: it causes an issue with geodjango, it tries to get an ._obj attribute since it's a CargObject, but we pass the underlying obj | 21:03 |
| amaury_ | hum | 21:03 |
| amaury_ | what is the function restype? | 21:04 |
| Alex_Gaynor | amaury_: grr, no --pdb on this system :) | 21:04 |
| Alex_Gaynor | amaury_: http://paste.pocoo.org/show/511613/ is what I get under pypy, I can't actually run this code under CPython :/ | 21:08 |
| dmalcolm (david@nat/redhat/x-msmgylgwalntazsi) left irc: Quit: Leaving | 21:09 | |
| amaury_ | "CFuncPtrFast" | 21:10 |
| amaury_ | it's possible that pypy's "fast ctypes" has some shortcuts | 21:10 |
| amaury_ | maybe Antonio knows better | 21:11 |
| Alex_Gaynor | with errcheck you mean? | 21:11 |
| amaury_ | yes, and other things | 21:11 |
| Alex_Gaynor | No, this doesn't appear to be related to fastctypes | 21:12 |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 21:13 | |
| fijal (~fijal@41.52.144.224) joined #pypy. | 21:17 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) joined #pypy. | 21:17 | |
| amaury_ | Then it's an issue with our _rawffi | 21:17 |
| amaury_ | which is not _raw enough | 21:17 |
| amaury_ | since it already returns real ints or floats | 21:18 |
| Alex_Gaynor | we should get the ctypes tests passing, so we can blame cpython | 21:18 |
| amaury_ | there is only one test about errcheck | 21:18 |
| Alex_Gaynor | why am I not surprised :( | 21:18 |
| fijal | hey | 21:18 |
| amaury_ | something you have to remember about ctypes: | 21:18 |
| etrepum (~bob@207.239.114.206) left irc: Read error: Operation timed out | 21:18 | |
| amaury_ | it was first developed on Windows | 21:19 |
| Alex_Gaynor | amaury_: based on the docs http://docs.python.org/library/ctypes.html#ctypes._FuncPtr.errcheck I claim CPython's behavior is broken | 21:19 |
| aboudreault (~alanb@osgeo/member/aboudreault) left irc: Quit: Leaving | 21:19 | |
| Alex_Gaynor | since it doesn't pass teh original *args, but the warpped ones | 21:19 |
| amaury_ | but you don't know the best part: | 21:19 |
| amaury_ | if errchek returns the "arguments" tuple that were passed | 21:20 |
| amaury_ | "then continue normal processing" | 21:20 |
| Alex_Gaynor | yes I saw that in the code :) | 21:20 |
| Alex_Gaynor | that makes sense | 21:20 |
| amaury_ | ah really? | 21:20 |
| amaury_ | it makes sense only when there are *several* return values | 21:20 |
| amaury_ | i.e. INOUT parameters | 21:21 |
| Alex_Gaynor | amaury_: no, I was being sarcastic :) | 21:21 |
| amaury_ | It makes sense in a Windows COM world | 21:21 |
| amaury_ | a bit | 21:21 |
| amaury_ | where INOUT parameters are used by some functions | 21:21 |
| amaury_ | that can also explain why there are so few unit tests for advanced features | 21:22 |
| fijal | amaury_: how does it explain so? | 21:22 |
| amaury_ | if there were some, they were written for win32 COM interfaces | 21:22 |
| amaury_ | and removed when ctypes was shipped in CPython | 21:23 |
| Alex_Gaynor | amaury_: here's what CPython has: http://dpaste.com/660487/ | 21:23 |
| Alex_Gaynor | so the last arg is the wrong type for some reason | 21:24 |
| etrepum (~bob@207.239.114.206) joined #pypy. | 21:25 | |
| amaury_ | How is this last param constructed? | 21:26 |
| amaury_ | a by_ref? | 21:26 |
| bgola (~bgola@189.100.46.89) left irc: Ping timeout: 260 seconds | 21:27 | |
| Alex_Gaynor | amaury_: yes | 21:27 |
| amaury_ | so fix by_ref :-) | 21:27 |
| Alex_Gaynor | so, it's our bug afterall :/ | 21:27 |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 21:28 | |
| amaury_ | yes | 21:28 |
| bgola (~bgola@189.100.42.16) joined #pypy. | 21:28 | |
| Alex_Gaynor | amaury_: :/ I'm not sure I understand what the expected behavior is even | 21:31 |
| amaury_ | byref(c_int(3))._obj | 21:32 |
| amaury_ | should work | 21:32 |
| amaury_ | in pypy the equivalent is byref(c_int(3)).contents | 21:33 |
| Alex_Gaynor | sigh | 21:33 |
| amaury_ | it's because we were lazy | 21:33 |
| Alex_Gaynor | we shouldn't be lazy, I guess | 21:33 |
| amaury_ | and implemented byref as a pointer | 21:33 |
| amaury_ | when it's a reference | 21:33 |
| amaury_ | it's a bit like C++ | 21:34 |
| amaury_ | they generate the same code | 21:34 |
| amaury_ | but are used with a different syntax | 21:34 |
| Alex_Gaynor | sigh | 21:34 |
| amaury_ | so it's not an errcheck issue after all | 21:38 |
| fijal (~fijal@41.52.144.224) left irc: Read error: Connection reset by peer | 21:38 | |
| amaury_ | it's just that normally byref() is directly passed as a function parameter | 21:38 |
| Alex_Gaynor | yup | 21:38 |
| amaury_ | so nobody sees the exact thing this returns | 21:38 |
| srin (~srin@NYUFGA-WLESSAUTHCLIENTS-02.NATPOOL.NYU.EDU) left irc: Read error: Connection reset by peer | 21:39 | |
| etrepum (~bob@207.239.114.206) left irc: Quit: etrepum | 21:41 | |
| lotrpy (~lotrpy@202.120.58.116) left irc: Ping timeout: 244 seconds | 21:46 | |
| lotrpy (~lotrpy@202.120.58.116) joined #pypy. | 21:46 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection | 21:54 | |
| stakkars_ (~tismer@p5DC46903.dip0.t-ipconnect.de) joined #pypy. | 21:55 | |
| amaury_ | float.__hash__ uses W_LongObject.fromfloat, it may be slow | 21:55 |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 21:55 | |
| amaury_ | I hope 1e32 is not often used as a dict key | 21:55 |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 240 seconds | 21:59 | |
| fprimex (~fprimex@brent-macbook.sc.fsu.edu) left irc: Quit: http://www.fprimex.com | 22:01 | |
| amaury_ | hmm, CPython defines sys.hash_info.modulus | 22:07 |
| amaury_ | (1<<61)-1 on 64bit platforms | 22:07 |
| amaury_ | what's the equivalent for pypy? | 22:07 |
| amaury_ | (I noticed that hash(somefloat) is always around 1<<32 | 22:08 |
| Alex_Gaynor | I would expect we copied CPython. gutworth did that work IIRC. | 22:13 |
| lotrpy (~lotrpy@202.120.58.116) left irc: Ping timeout: 260 seconds | 22:14 | |
| lotrpy (~lotrpy@202.120.58.116) joined #pypy. | 22:15 | |
| `fox` (~fox@host42-105-dynamic.51-79-r.retail.telecomitalia.it) left irc: Ping timeout: 244 seconds | 22:17 | |
| amaury_ | should I change pypy's float.__hash__ then? | 22:20 |
| nickik (~nick@217.150.239.67) left irc: Quit: Verlassend | 22:22 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) left irc: Remote host closed the connection | 22:26 | |
| antocuni (~antocuni@88-149-186-50.staticnet.ngi.it) joined #pypy. | 22:31 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) joined #pypy. | 22:34 | |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy. | 22:38 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/721 | 22:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/147 | 22:41 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/950 | 22:41 |
| etrepum (~bob@12.130.118.18) joined #pypy. | 22:51 | |
| mat^2 (~mathias@212.130.113.35) joined #pypy. | 22:55 | |
| kenaan | 12amauryfa py3k 11aa1d34b5a927 15/pypy/: Implement sys.hash_info. Initial patch by chronitis, thanks! | 23:00 |
| srin (~srin@NYUFGA-WLESSAUTHCLIENTS-02.NATPOOL.NYU.EDU) joined #pypy. | 23:10 | |
| srin (~srin@NYUFGA-WLESSAUTHCLIENTS-02.NATPOOL.NYU.EDU) left irc: Client Quit | 23:10 | |
| stakkars_ | amaury_: hi! | 23:15 |
| amaury_ | hi | 23:15 |
| stakkars_ | amaury_: do you happen to know which test modules load the boehm gc? | 23:16 |
| amaury_ | some translation tests IIRC | 23:16 |
| amaury_ | test_asmgcc for example | 23:16 |
| stakkars_ | it is almost impossible to find out with a text search, since "gc" comes so often, and "gc.dll" does not occur :-( | 23:17 |
| amaury_ | sure | 23:17 |
| stakkars_ | ah, thank you. | 23:17 |
| amaury_ | "boehm" should be the keyword | 23:17 |
| stakkars_ | I'd be happier if such things were searchable... | 23:17 |
| stakkars_ | I need to test the point where "gc64_dll.dll" will need to be loaded, but it doesn't work in any obvious way | 23:19 |
| stakkars_ | trying test_asmgcc (why &gcc, what does it mean?) | 23:20 |
| amaury_ | sorry, it's unrelated to boehm | 23:20 |
| amaury_ | asmgcc is a technique to get root pointers for gc | 23:21 |
| amaury_ | the one that are live on the stack | 23:21 |
| amaury_ | instead of a "shadow stack" i.e. you push your local variables when calling another function | 23:21 |
| stakkars_ | aha!?! | 23:21 |
| stakkars_ | ah yes | 23:21 |
| amaury_ | it makes a map of the register that can hold gc pointers | 23:21 |
| amaury_ | analysis of the generated assembler | 23:22 |
| amaury_ | it was first developed for gcc | 23:22 |
| amaury_ | but I made it work with Visual Studio | 23:22 |
| amaury_ | very fragile though | 23:23 |
| stakkars_ | I found test_asmgcroot instead | 23:24 |
| amaury_ | yes, that's the correct name | 23:24 |
| stakkars_ | ok, great, thank you again! | 23:24 |
| amaury_ | I think it's not used anymore | 23:26 |
| amaury_ | Armin wrote a version of the shadow stack that is fast enough, and compatible with the JIT | 23:27 |
| amaury_ | and generates normal C code | 23:27 |
| stakkars_ | hmm. I simply want to see the call to -- what? types? that causes the all to be loaded. | 23:28 |
| stakkars_ | types, ffi? It seems to no longer use "gc.h". I hacked on boehm and windows in 2005 ;-) | 23:29 |
| stakkars_ | (arg, there must be something that corrects my typing. wanted to type "types", twice) | 23:30 |
| stakkars_ | arch, again | 23:31 |
| stakkars_ | ctypes | 23:31 |
| voidspace (~voidspace@python/psf/voidspace) left irc: Quit: Uhm... gotta go | 23:32 | |
| amaury_ | stakkars: you can follow "configure_boehm" calls | 23:33 |
| amaury_ | it returns an ExternalCompilationInfo object | 23:33 |
| amaury_ | which represents the boehm library | 23:34 |
| stakkars_ | yes, I just found that. thanks! | 23:34 |
| amaury_ | but I don't think boehm is used in interpreted tests | 23:35 |
| amaury_ | (with ll2ctypes &co) | 23:35 |
| amaury_ | only in translated tests | 23:35 |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) left irc: Ping timeout: 240 seconds | 23:39 | |
| stakkars_ | jit/backend/llsupport/gc.py seems to do something, right? | 23:39 |
| stakkars_ | I will trace that. Obviously the only way to learn what pypy does. Tracing. | 23:40 |
| amaury_ | stakkars: this is for the JIT, of course | 23:41 |
| amaury_ | when it needs to allocate memory and such, it has to integrate to the gc | 23:41 |
| voidspace (~anonymous@python/psf/voidspace) joined #pypy. | 23:42 | |
| Action: amaury_ ->bed | 23:45 | |
| stakkars_ | thank you, good night | 23:46 |
| gnuvince (~vince@ip-96-43-237-62.dsl.netrevolution.com) joined #pypy. | 23:48 | |
| stakkars_ | ha, found what I searched in rffi_platform.configure_boehm | 23:48 |
| antocuni (~antocuni@88-149-186-50.staticnet.ngi.it) left irc: Ping timeout: 252 seconds | 23:49 | |
| gnuvince | I'm trying to run nosetests with PyPy, but I get an exception when it tries to import site.py. I've pasted the stack trace: http://www.ideone.com/NWBIN. Anyone knows what I can do to fix that? | 23:49 |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 252 seconds | 23:50 | |
| timonator | oh | 23:52 |
| timonator | you shouldn't set the pythonpath to anything from cpython | 23:53 |
| timonator | that's bad for pypy | 23:53 |
| timonator | if you need to install nose for it, the easiest way is probably to use a virtualenv, but if you have the right permissions to the pypy folder, you can just as well pypy setup.py install it | 23:53 |
| timonator | gnuvince: also a highlight for you | 23:55 |
| --- Wed Nov 23 2011 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!