| timonator | that's great! | 00:01 |
|---|---|---|
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 00:39 | |
| jterrace (~jterrace@pu116187.student.Princeton.EDU) left irc: Quit: Leaving. | 00:41 | |
| Guest81860 (~jared@061093200075.ctinets.com) joined #pypy. | 00:45 | |
| nedbat | Alex_Gaynor: hmm, I get a different but similar error with your build line: undefined reference to `SSLv2_method' | 00:48 |
| Alex_Gaynor | nedbat: grumble, I believe that's a bug in our ssl lib that was fixed between 1.6 and tip | 00:48 |
| Alex_Gaynor | :/ | 00:48 |
| nedbat | Alex_Gaynor: ok, I can try again with the latest stable nightly. | 00:49 |
| nedbat | Alex_Gaynor: when you get a chance, can you try my build line? ./pypy/translator/goal/translate.py -O2 --sandbox | 00:49 |
| Alex_Gaynor | k, I'm not booted into my linux partition ATM, but I'll try later this evening | 00:50 |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 00:52 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Client Quit | 00:53 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 00:53 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Client Quit | 00:53 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Quit: ZNC - http://znc.sourceforge.net | 00:55 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 01:10 | |
| stakkars | fijal, arigo: where exactly does the file "ffitarget.h" come from, and why are there changes? | 01:13 |
| stakkars | there are different versions out there, and I'd like to know more. | 01:15 |
| derdon (~derdon@pD9E1C4FD.dip.t-dialin.net) left irc: Remote host closed the connection | 01:18 | |
| Guest81860 (~jared@061093200075.ctinets.com) left irc: Ping timeout: 252 seconds | 01:18 | |
| derdon (~derdon@pD9E1C4FD.dip.t-dialin.net) joined #pypy. | 01:19 | |
| derdon (~derdon@pD9E1C4FD.dip.t-dialin.net) left irc: Remote host closed the connection | 01:19 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1840 [12alex, jit-dynamic-getarrayitem] | 01:20 |
| lameiro (~lameiro@173.120-240-81.adsl-dyn.isp.belgacom.be) left irc: Quit: Ex-Chat | 01:27 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/711 | 01:27 |
| stakkars | fijal: solved, thanks | 01:39 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1841 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/567 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/264 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/611 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1103 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/796 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1438 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/701 | 01:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/261 | 01:44 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/261 | 01:44 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/264 | 01:53 |
| arkx (arkx@hilla.kapsi.fi) left irc: Ping timeout: 276 seconds | 02:16 | |
| arkx (arkx@hilla.kapsi.fi) joined #pypy. | 02:16 | |
| nedbat | Alex_Gaynor: your build command works fine for me in 48894-56079dacea00. Trying --sandbox there now. | 02:35 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1841 | 02:42 |
| asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Ping timeout: 248 seconds | 02:44 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1438 | 02:45 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1103 | 02:45 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/611 | 02:45 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/567 | 02:45 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/796 | 02:45 |
| Shinata (~John@osbk-4db069c2.pool.mediaWays.net) left irc: Ping timeout: 240 seconds | 02:48 | |
| Moku (~John@osbk-4db06ce4.pool.mediaWays.net) joined #pypy. | 02:48 | |
| Nick change: Moku -> Guest77977 | 02:48 | |
| crakdmirror (~crakdmirr@174.127.114.26) left irc: Quit: ZNC - http://znc.sourceforge.net | 02:50 | |
| crakdmirror (~crakdmirr@174.127.114.26) joined #pypy. | 02:52 | |
| crakdmirror (~crakdmirr@174.127.114.26) left irc: Client Quit | 02:52 | |
| crakdmirror (~crakdmirr@174.127.114.26) joined #pypy. | 02:53 | |
| nedbat | Alex_Gaynor: OK, my command works fine in 48894-56079dacea0 also! I guess it isn;t an issue. fijal's changesets about this should be reverted I guess. Sorry for the false alarm. | 02:57 |
| nedbat | Alex_Gaynor: although, when I run pypy-c, I get: sll_os.ll_os_getenv(sPYPY_GENERATIONGC_NURSERY | 02:58 |
| nedbat | Fatal error during initialization: out of memory | 02:58 |
| nedbat | Aborted | 02:58 |
| asmeurer (~asmeurer@dhcp-altamirano-203.resnet.nmt.edu) joined #pypy. | 03:01 | |
| Nick change: crakdmirror -> basepi_ | 03:03 | |
| Nick change: basepi_ -> basepi | 03:03 | |
| Nick change: basepi -> basepi_ | 03:07 | |
| Nick change: basepi_ -> basepi | 03:07 | |
| basepi (~crakdmirr@174.127.114.26) left irc: Quit: ZNC - http://znc.sourceforge.net | 03:08 | |
| basepi (~basepi@174.127.114.26) joined #pypy. | 03:12 | |
| nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 258 seconds | 03:28 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) joined #pypy. | 03:37 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/701 | 03:42 |
| asmeurer (~asmeurer@dhcp-altamirano-203.resnet.nmt.edu) left irc: Ping timeout: 248 seconds | 03:42 | |
| asmeurer___ (~asmeurer@dhcp-altamirano-203.resnet.nmt.edu) joined #pypy. | 03:42 | |
| canta (~canta@77-20-123-240-dynip.superkabel.de) left irc: Quit: Leaving. | 03:57 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 04:30 | |
| chickamade (~chickamad@118.69.67.203) joined #pypy. | 04:34 | |
| apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) joined #pypy. | 04:40 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) joined #pypy. | 04:49 | |
| Sho_ (~EHS1@kde/hein) joined #pypy. | 04:57 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) left irc: Quit: Leaving. | 05:19 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) joined #pypy. | 05:27 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 05:30 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 05:34 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) left irc: Quit: Leaving. | 05:35 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 05:53 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 05:59 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Read error: Connection reset by peer | 06:01 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) joined #pypy. | 06:04 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) left irc: Quit: Leaving. | 06:09 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy. | 06:13 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) joined #pypy. | 06:21 | |
| Nick change: Gulopine -> Gulaway | 06:27 | |
| mat^2 (~mathias@212.130.113.35) left irc: | 06:35 | |
| Guest77977 (~John@osbk-4db06ce4.pool.mediaWays.net) left irc: Remote host closed the connection | 06:37 | |
| asmeurer___ (~asmeurer@dhcp-altamirano-203.resnet.nmt.edu) left irc: Quit: asmeurer___ | 06:37 | |
| Moku (~John@osbk-4db06ce4.pool.mediaWays.net) joined #pypy. | 06:39 | |
| Nick change: Moku -> Shinata | 06:40 | |
| asmeurer_ (~asmeurer@71.216.240.168) joined #pypy. | 06:54 | |
| Nick change: stargaming- -> stargaming | 07:06 | |
| kkris (~kris@80-123-43-97.adsl.highway.telekom.at) joined #pypy. | 07:07 | |
| asmeurer_ (~asmeurer@71.216.240.168) left irc: Quit: asmeurer_ | 07:08 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 240 seconds | 07:10 | |
| EnCuKou (~encukou@a91-156-243-246.elisa-laajakaista.fi) left irc: Remote host closed the connection | 07:18 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 07:24 | |
| amaury (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Read error: Operation timed out | 07:25 | |
| asmeurer_ (~asmeurer@71.216.240.168) joined #pypy. | 07:25 | |
| Nick change: Gulaway -> Gulopine | 07:27 | |
| ronny | moin | 07:30 |
| fijal (~fijal@77-254-63-221.adsl.inetia.pl) left irc: Ping timeout: 248 seconds | 07:31 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 240 seconds | 07:31 | |
| kkris (~kris@80-123-43-97.adsl.highway.telekom.at) left irc: Quit: Leaving. | 07:32 | |
| elipsion | mornin' | 07:38 |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 07:51 | |
| dgl (~dgl@109.86.165.231) joined #pypy. | 07:52 | |
| dgl (~dgl@109.86.165.231) left irc: Client Quit | 07:52 | |
| Nick change: Gulopine -> Gulaway | 07:55 | |
| EnCuKou (~encukou@193.167.65.94) joined #pypy. | 07:57 | |
| Ademan (~dan@adsl-71-141-252-21.dsl.snfc21.pacbell.net) left irc: Quit: leaving | 08:01 | |
| chickamade_ (~chickamad@118.69.67.203) joined #pypy. | 08:07 | |
| bobbyz (~bobbyz@c-67-184-43-39.hsd1.il.comcast.net) joined #pypy. | 08:09 | |
| chickamade (~chickamad@118.69.67.203) left irc: Ping timeout: 260 seconds | 08:10 | |
| Nick change: chickamade_ -> chickamade | 08:10 | |
| Trundle (~andy@p578bfdcf.dip0.t-ipconnect.de) joined #pypy. | 08:12 | |
| Trundle (~andy@p578bfdcf.dip0.t-ipconnect.de) left irc: Changing host | 08:12 | |
| Trundle (~andy@python/site-packages/trundle) joined #pypy. | 08:12 | |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 08:41 | |
| asmeurer_ (~asmeurer@71.216.240.168) left irc: Quit: asmeurer_ | 08:44 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy. | 08:51 | |
| lac (~quassel@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 08:52 | |
| lac (quassel@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left #pypy. | 08:52 | |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 08:54 | |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) joined #pypy. | 08:58 | |
| kenaan | 12fijal numpy-multidim-shards 11e49090902cb5 15/pypy/module/micronumpy/: one more test and start working on repr | 09:02 |
| arigato (~arigo@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 09:24 | |
| amaury_ (~amaury_@74.125.57.33) joined #pypy. | 09:25 | |
| fijal | arigato: hi | 09:26 |
| arigato | fijal: hi | 09:26 |
| fijal | we have a new pypy website for the release btw | 09:26 |
| arigato | are you fixing the clock_gettime failures or should I revert ad8b93cf993c for now? | 09:26 |
| fijal | I kinda forgot | 09:26 |
| Action: fijal is not fixing anything | 09:26 | |
| fijal | arigato: I'm flying today, so I'm unusable for two days probably | 09:27 |
| arigato | ok | 09:27 |
| fijal | it's a bit hard to fix, because it works for me :/ | 09:27 |
| arigato | :-/ | 09:27 |
| kenaan | 12arigo default 11f718100d7782 15/pypy/translator/platform/: Backout ad8b93cf993c and 38f173ee998a, which fail on tannit. | 09:29 |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) left irc: Quit: Leaving | 09:29 | |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) joined #pypy. | 09:29 | |
| G2P (~G2P@fw-asn1.ornis.com) joined #pypy. | 09:29 | |
| Action: arigato wonders how you know it works, given that translation is broken anyway | 09:29 | |
| arigato | (my fault) | 09:29 |
| kenaan | 12bivab arm-backend-2 1102cab8c5480f 15/: merge default | 09:29 |
| kenaan | 12bivab arm-backend-2 11f6952466347e 15/pypy/jit/backend/arm/test/: make sure arm backend tests are only executed when running on ARM | 09:30 |
| kenaan | 12bivab ppc-jit-backend 113e3ed29ed104 15/: merge upstream | 09:30 |
| fijal | arigato: I wrote a genc test | 09:30 |
| fijal | test_standalone | 09:30 |
| fijal | and poked a bit | 09:30 |
| fijal | but among other things, my gcc does not require -lrt to be at the end | 09:30 |
| fijal | so it's nothing new, the bug is there forever | 09:30 |
| fijal | but simply people came with new gcc that is more strict | 09:31 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-ppc-32/builds/1 [12David Schneider, ppc-jit-backend] | 09:31 |
| arigato | yes, but that doesn't explain how the tests fail on tannit | 09:31 |
| fijal | ah, that's yet-another-issue | 09:32 |
| fijal | no, maybe the same one | 09:32 |
| fijal | no idea, maybe older gcc wants -lrt to be at the beginning? | 09:32 |
| fijal | or we forgot to pass it at all or so? | 09:32 |
| fijal | yes, I think we simply don't write it to makefile | 09:33 |
| fijal | arigato: want me to fix it? | 09:33 |
| arigato | feel free | 09:33 |
| arigato | but please store the class attributes in tuples, not in lists | 09:33 |
| arigato | we had issues with exactly this at some point | 09:34 |
| fijal | ok | 09:38 |
| kenaan | 12arigo default 113ed133d0ce83 15/pypy/objspace/std/: Fix annotation issues. | 09:38 |
| kenaan | 12fijal default 113cbd94863224 15/pypy/translator/platform/: a test and a fix | 09:40 |
| kenaan | 12fijal default 11654821ec2610 15/pypy/translator/platform/: use tuple | 09:40 |
| kenaan | 12fijal default 11e36e879eb1fe 15/pypy/: merge | 09:40 |
| fijal | arigato: done | 09:40 |
| kenaan | 12arigo default 111931f105c85b 15/pypy/objspace/std/test/test_sliceobject.py: Skip an assert in a test that fails on Python 2.5. | 09:41 |
| kenaan | 12arigo default 111be28eed2789 15/pypy/translator/platform/: merge heads | 09:41 |
| fijal | arigato: if we're using Håkan all over the place, then also please use fijaB | 09:41 |
| fijal | :-) | 09:41 |
| arigato | fijaB: but then you don't get beeped on irc | 09:42 |
| fijal | indeed, but that's about blog posts and not irc | 09:43 |
| fijal | I suppose more "official" things | 09:44 |
| Sho_ (~EHS1@kde/hein) left irc: Quit: Stop leaking memory like it's going out of fashion. | 09:45 | |
| bivab (david@fwstups.cs.uni-duesseldorf.de) left #pypy. | 09:46 | |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 09:47 | |
| arigato | fijal: thanks! now the tests pass on tannit too :-) | 09:49 |
| fijal | good | 09:49 |
| fijal | this was a simple fix that broke everything :/ | 09:49 |
| fijal | and was supposed to only take one commit | 09:49 |
| arigato | of course | 09:49 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 09:49 | |
| fijal | when is/was FSCons? | 09:50 |
| arigato | I just talked with jacob who suggests that we try (again) to avoid doing small changes directly to "default" | 09:50 |
| antocuni (~antocuni@host70-121-dynamic.11-79-r.retail.telecomitalia.it) joined #pypy. | 09:50 | |
| arigato | antocuni: hi | 09:50 |
| arigato | I just talked with jacob who suggests that we try (again) to avoid doing small changes directly to "default" | 09:50 |
| antocuni | hi | 09:51 |
| antocuni | because we break it too often? | 09:51 |
| Sho_ (~EHS1@kde/hein) joined #pypy. | 09:51 | |
| arigato | yes, even if "too often" is only "once in a while it's broken for two days" | 09:51 |
| fijal | antocuni: hi | 09:51 |
| fijal | antocuni: I have opinions about the checkpointing work | 09:51 |
| antocuni | am I the culprit by chance? :-) | 09:51 |
| antocuni | fijal: yes? | 09:51 |
| fijal | well, the stuff you tried does not address the actual issue | 09:52 |
| fijal | the actual issue being that it's hard to write python applications that clean up the state | 09:52 |
| fijal | I tried a bit, but a lot of stuff is reading configurations at import time etc. | 09:52 |
| fijal | it's technically possible, but it would not work out of the box on pretty much anything | 09:52 |
| antocuni | yes, I fear you are right | 09:52 |
| antocuni | this is a problem of both the checkpointing approach and the fork approach | 09:53 |
| witulski (stupsi@fwstups.cs.uni-duesseldorf.de) left #pypy. | 09:53 | |
| fijal | yes, precisely | 09:54 |
| fijal | that's why advanced application checkpointing does not address anything really | 09:54 |
| antocuni | well, it was more for trying whether it works | 09:54 |
| Action: arigato writes a mail instead | 09:55 | |
| fijal | arigato: instead of? | 09:55 |
| arigato | instead of fighting to have a chance to expose the whole problem without being interrupted? | 09:55 |
| fijal | arigato: sorry :) | 09:56 |
| antocuni | did I jump in the middle of a discussion by chance? :-) | 09:56 |
| fijal | arigato: I can discuss that a bit, mostly because I'm against that rule | 09:56 |
| arigato | antocuni: no, I only posted one line | 09:57 |
| arigato | so you arrived at the aborted start of the discussion | 09:57 |
| ojii (~ojii@40-34.60-188.cust.bluewin.ch) joined #pypy. | 09:59 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1104 [12arigo] | 09:59 |
| arigato | fijal: sorry, seems that I'm in a bad mood this morning :-( please ignore me... | 10:00 |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) left irc: Ping timeout: 240 seconds | 10:02 | |
| antocuni | pff, two days of "upgrading" ubuntu and I'm still not 100% done | 10:05 |
| kenaan | 12arigo default 111368405e24e1 15/pypy/module/test_lib_pypy/test_pwd.py: skip on win32. | 10:12 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 256 seconds | 10:13 | |
| derdon (~derdon@p5DE89701.dip.t-dialin.net) joined #pypy. | 10:21 | |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) joined #pypy. | 10:21 | |
| fijal | arigato: meh, got kicked out | 10:21 |
| fijal | arigato: so, what is your opinion? | 10:22 |
| fijal | or you simply don't want to discuss? | 10:22 |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) left irc: Quit: unbit | 10:22 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) joined #pypy. | 10:22 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) left irc: Read error: Connection reset by peer | 10:23 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) joined #pypy. | 10:23 | |
| chickamade (~chickamad@118.69.67.203) left irc: Quit: chickamade | 10:25 | |
| antocuni | Alex_Gaynor: I'm looking at jit-dynamic-getarrayitem: so, IIUC, the {get,set}items which are done through libffi are always emitted as {get,set}interiorfield_raw? | 10:26 |
| fijal | antocuni: ok, cool, you could have told me before you would look there ;-) | 10:28 |
| antocuni | fijal: uh? | 10:31 |
| fijal | well, alex told me to review this | 10:31 |
| fijal | if you're reviewing, then it does not make sense to repeat no? | 10:31 |
| antocuni | I just skimmed over the commits | 10:32 |
| fijal | if you didn't plan to review, then please do it :) | 10:32 |
| antocuni | they look fine, it's the same kind of code which I had to write for ffistruct | 10:32 |
| fijal | maybe it should be the same code? | 10:32 |
| antocuni | fijal: ok, I review everything but the x86 part, and you review that one :-) | 10:32 |
| fijal | ok | 10:32 |
| kenaan | 12arigo default 11335d74f2d6e3 15/pypy/interpreter/test/test_executioncontext.py: This number of quotes suddenly stopped working on Windows, but with one level of quotes less it seems to work. I h... | 10:33 |
| antocuni | fijal: no, most of it it's really a long chain of functions which are almost empty and almost the same but need to be different for the annotator, the codewriter, etc. | 10:33 |
| fijal | right | 10:33 |
| fijal | the x86 part looks fine | 10:34 |
| fijal | (also very small) | 10:34 |
| fijal | then there is a question why do we need the static version at all | 10:34 |
| arigato | fijal: sorry, seems that I'm in a bad mood this morning :-( please ignore me... | 10:34 |
| fijal | (does it buy us anything really?) | 10:34 |
| fijal | arigato: :] | 10:34 |
| antocuni | fijal: sorry, I'm going afk | 10:34 |
| fijal | antocuni: I suppose that's more a question to alex | 10:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/262 [12arigo] | 10:36 |
| fijal | http://www.jobcareerdb.com/2011/job-vacancy-devopsinfrastructure-engineer-palo-alto.html | 10:37 |
| fijal | pypy in job ad | 10:37 |
| fijal | in a job ad | 10:38 |
| stakkars (~tismer@213-65-251-34-no35.business.telia.com) left irc: Quit: schnarch | 10:39 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Ping timeout: 240 seconds | 10:41 | |
| arigato | :-) | 10:42 |
| fijal | a technology choice seems a bit arbitrary, but would be cool to know if surveymonkey is using pypy | 10:43 |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy. | 10:44 | |
| lucian_ (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy. | 10:45 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Read error: Connection reset by peer | 10:45 | |
| Nick change: lucian_ -> lucian | 10:46 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1104 [12arigo] | 10:50 |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) left irc: Quit: Leaving | 10:52 | |
| fijal_ (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) joined #pypy. | 10:52 | |
| Nick change: fijal_ -> fijal | 10:52 | |
| Nick change: ojii -> ojii|lunch | 10:57 | |
| antocuni | fijal: wow, that's cool | 11:03 |
| antocuni | maybe we should write to them asking whether they are actually using pypy? | 11:03 |
| fijal | feel free | 11:03 |
| Action: antocuni does | 11:03 | |
| antocuni | eh, the hard part is to find the right email addres | 11:04 |
| antocuni | *address | 11:04 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-ppc-32/builds/1 [12David Schneider, ppc-jit-backend] | 11:09 |
| Squarepy (~Squarepy@ip4da70b11.direct-adsl.nl) joined #pypy. | 11:14 | |
| kenaan | 12antocuni extradoc 11bba5526b4eb5 15/blog/draft/2011-11-gborg-sprint-report.rst: twaks | 11:18 |
| Squarepy (~Squarepy@ip4da70b11.direct-adsl.nl) left irc: Changing host | 11:22 | |
| Squarepy (~Squarepy@unaffiliated/squarepy) joined #pypy. | 11:22 | |
| kenaan | 12arigo extradoc 1137cc0ee778a9 15/talk/fscons2011/: latest version | 11:22 |
| kenaan | 12arigo extradoc 11614c01f2ba2b 15/blog/draft/2011-11-gborg-sprint-report.rst: Typo | 11:22 |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection | 11:22 | |
| fijal (~fijal@188.147.230.166.nat.umts.dynamic.t-mobile.pl) left irc: Ping timeout: 260 seconds | 11:23 | |
| arigato | yay, "that will be the perfect excuse to eat another cake :-)" | 11:23 |
| antocuni | yep :-) | 11:24 |
| antocuni | if nobody objects, I'm posting it | 11:24 |
| JaRoel (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 11:24 | |
| kenaan | 12bivab ppc-jit-backend 1108426f22e5d5 15/pypy/jit/codewriter/heaptracker.py: fix for tests | 11:25 |
| antocuni | http://morepypy.blogspot.com/2011/11/gothenburg-sprint-report.html | 11:26 |
| Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy. | 11:32 | |
| kenaan | 12bivab ppc-jit-backend 11cdd52796a997 15/pypy/jit/backend/x86/test/test_assembler.py: another fix for the ppc tests | 11:36 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-ppc-32/builds/2 [12David Schneider, ppc-jit-backend] | 11:37 |
| amaury_ (~amaury_@74.125.57.33) left irc: Ping timeout: 244 seconds | 11:44 | |
| Fluxid | antocuni: "This needs a lot of work because a lot of PyPy is written under the assumption than assumption that the long type in C has the same bit size than void*," | 11:56 |
| Fluxid | assumption than assumption that? | 11:57 |
| kirma | there should be empty line between paragraphs | 11:58 |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 12:06 | |
| Nick change: ojii|lunch -> ojii | 12:16 | |
| antocuni | Fluxid: thank you, I fixed it | 12:24 |
| Fluxid | np | 12:26 |
| ltratt (~ltratt@87.115.8.1) joined #pypy. | 12:34 | |
| ltratt | Hi. Does anyone have a dummys explanation of how yield is implemented in PyPy? In the RPython Converge VM, I'm using stacklets to implement generators, but a bug (#907) prevents that working on --opt>1. | 12:37 |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 12:40 | |
| cfbolz | hello | 12:40 |
| cfbolz | I am seeing typos in the sprint report, ok to fix them? | 12:41 |
| Arfrever (~Arfrever@apache/committer/Arfrever) joined #pypy. | 12:42 | |
| arigato | ltratt: hi | 12:44 |
| arigato | ltratt: sorry I didn't answer your mail yet, I was busy at a sprint and a conference | 12:44 |
| ltratt | arigato: No problem! I thought it might be something that others were better placed to answer - I have very little idea who's responsible for what bits in PyPy yet :) | 12:45 |
| arigato | :-) | 12:46 |
| ltratt | The new Converge VM is getting to the point where I'm struggling to answer questions just by grep'ing, which has been a remarkably successful tactic upto now. | 12:46 |
| cfbolz | :-) | 12:47 |
| cfbolz | ltratt: at the latest you can pester me with question in january :-) | 12:47 |
| ltratt | cfbolz: oh, you can bet your bottom euro I will ;) | 12:47 |
| arigato | basically, for generators, you need to have a "frame" class, but I think you do in the "cf" variable | 12:48 |
| antocuni | cfbolz: I think you can just click on the small pencil at the bottom of the page | 12:48 |
| cfbolz | antocuni: I was more wondering whether somebody else was editing at the moment | 12:48 |
| antocuni | I'm not | 12:48 |
| arigato | ltratt: that means that to execute a "yield", you do it like a "return", except that the "cf" is stored around | 12:48 |
| cfbolz | antocuni: you "bake" a cake, you don't "cook" it :-) | 12:49 |
| arigato | ltratt: later, to resume the generator, you call bc_loop again with the same "cf" | 12:49 |
| ltratt | arigato: ok, that's what I guessed. It means that RPython generators need to manually maintain state over yields. | 12:49 |
| arigato | yes | 12:49 |
| antocuni | cfbolz: yes, but the important thing is that you *eat* it :-) | 12:49 |
| cfbolz | eh | 12:49 |
| cfbolz | yes, indeed | 12:49 |
| arigato | ltratt: that's a bit more work than just using stacklet, but it's more explicit, and avoids needing stacklets | 12:50 |
| ltratt | arigato: I tried that in one C VM years ago, and it made many things very icky. Maybe because of the relative terseness of RPython it won't be so bad. | 12:50 |
| cfbolz | ltratt: can your generators recurse? or are they really like python's? | 12:51 |
| ltratt | arigato: i like the stacklets approach to generators, though I'm unclear what the performance penalty is. | 12:51 |
| ltratt | cfbolz: they are mostly like Python's, except they're much more ubiquitous. | 12:51 |
| cfbolz | rights | 12:51 |
| ltratt | the bytecode generators are handled nicely, but its the ones implemented in RPython that are a pain. Storing / unpacking state can be very tedious. | 12:52 |
| basepi (~basepi@174.127.114.26) left irc: Quit: ZNC - http://znc.sourceforge.net | 12:52 | |
| arigato | ah, I see | 12:52 |
| arigato | indeed, Python (being CPython or PyPy) doesn't have generators at the level of the interpreter | 12:53 |
| arigato | it's implemented explicitly by storing and unpacking state | 12:54 |
| basepi (~basepi@server.cmyers.net) joined #pypy. | 12:54 | |
| ltratt | arigato: stacklets get around that very neatly :) | 12:55 |
| arigato | yes | 12:55 |
| ltratt | arigato: in the old C VM I used setjmp / longjmp and stack copying for the same purpose. Evil, but surprisingly portable. | 12:56 |
| arigato | :-/ | 12:56 |
| arigato | what you may really need is generator support in RPython | 12:56 |
| arigato | it is something we are vaguely thinking of adding to RPython every now and then | 12:57 |
| ltratt | arigato: that would be even easier i guess | 12:57 |
| ltratt | one advantage i have is that in Converge I know when called functions might possibly be resumed, so often i can dynamically optimise away expensive techniques (e.g. stacklets) | 12:58 |
| mat^2 (~mathias@212.130.113.35) joined #pypy. | 12:59 | |
| stargaming (stargaming@server3.raumopol.de) left irc: Changing host | 13:01 | |
| stargaming (stargaming@unaffiliated/stargaming) joined #pypy. | 13:01 | |
| ronny | arigato: isnt it possible to transform a generator to a iterator with statemachine? | 13:02 |
| arigato | ronny: yes | 13:04 |
| arigato | ideally we'd like an automatic transformation of this kind done very early during translation | 13:04 |
| arigato | it would allow us to re-enable at least pypy/interpreter/pyparser/pytokenizer.py | 13:05 |
| arigato | it was originally a big generator, but has been converted to accumulate and return a list of results | 13:06 |
| mat^2 | I see that during the sprint, some Windows bugs were fixed - does that include https://bugs.pypy.org/issue849? | 13:06 |
| arigato | it creates issues like requiring tons of memory to compile large files, or not really supporting incremental compilation (e.g. when the input is a pipe) | 13:07 |
| arigato | mat^2: yes, I think so | 13:07 |
| mat^2 | I'm translating myself, so I'll make sure to report back | 13:08 |
| mat^2 | arigato: nice | 13:08 |
| arigato | thank you :-) | 13:08 |
| ronny | arigato: btw, are there any possibilities for significantly speeding up translation? | 13:12 |
| arigato | not that I can think up | 13:12 |
| ronny | arigato: im tinkering with the idea to use execnet to do the actual translation on a other box and streaming back the results | 13:13 |
| arigato | what is the advantage over ssh'ing there? | 13:14 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-ppc-32/builds/2 [12David Schneider, ppc-jit-backend] | 13:14 |
| ronny | platform checks on the local box, so gcc would run local | 13:14 |
| ronny | also it could be used to kill the memory usag right before invoking gcc | 13:14 |
| ronny | so it would help to automate doing the expensive stuff on a remote box and help to avoid one "swap of death" issue | 13:15 |
| arigato | note that this is not a "swap of death" issue | 13:15 |
| ronny | arigato: on my box it kind of is | 13:16 |
| arigato | it's just a "swap a lot for a bit but then proceed normally" issue | 13:16 |
| k_bx (~k_bx@195.20.130.1) joined #pypy. | 13:16 | |
| arigato | you can already work around it, e.g. by setting "translate.py --source" | 13:16 |
| ronny | btw, can we switch the default mode of translate, to invoke itself with --source, then start the compile | 13:17 |
| arigato | feel free to hack something, preferrably based on the existing pypy/tool/runsubprocess.py | 13:19 |
| arigato | e.g. the final call to run_subprocess() could pass a flag "final=True" that kills the current process by doing an execv instead of a popen | 13:20 |
| arigato | well, it's not really related to runsubprocess.py, actually | 13:21 |
| arigato | probably just do a execv to run the final compilation, and that's it | 13:22 |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) left irc: Quit: Leaving. | 13:23 | |
| lizardo (~lizardo@201009105166.user.veloxzone.com.br) joined #pypy. | 13:29 | |
| kenaan | 12bivab ppc-jit-backend 11deb6836af9be 15/pypy/jit/: more test fixes | 13:34 |
| Guest81860 (~jared@061093200075.ctinets.com) joined #pypy. | 13:36 | |
| Alex_Gaynor | antocuni: yes I'm using interiorfield everywhere, I added an optimization to the x86 backend so if it's equivilant to getarrayitem_raw there's no penalty | 13:37 |
| antocuni | ok, cool | 13:37 |
| dimazest (~dimazest@balticom-202-59.balticom.lv) joined #pypy. | 13:38 | |
| Rhy0lite (dje@nat/ibm/x-vtoqopwdtgiwgqdy) joined #pypy. | 13:40 | |
| jared (~jared@061093200075.ctinets.com) joined #pypy. | 13:41 | |
| Guest81860 (~jared@061093200075.ctinets.com) left irc: Read error: Connection reset by peer | 13:41 | |
| Nick change: jared -> Guest35710 | 13:41 | |
| Nick change: Hong_MinHee -> hongminhee | 13:45 | |
| Nick change: Gulaway -> Gulopine | 13:54 | |
| lucian | this nuitka thing might be interesting | 13:56 |
| cfbolz | lucian: nuitka? | 13:56 |
| lucian | this thing http://www.nuitka.net/blog/nuitka-a-python-compiler/ | 13:57 |
| lucian | it seems to actually work | 13:57 |
| cfbolz | ah, yet another static compiler for python | 13:57 |
| cfbolz | gives a speedup of 2-3 at most | 13:57 |
| Alex_Gaynor | cfbolz: Not Python, a subset of Python :) | 13:57 |
| cfbolz | even worse | 13:58 |
| lucian | yeah. it's GPLv3 and has an odd project structure, so unlikely to get too far | 13:58 |
| lucian | Alex_Gaynor: it's a surprisingly large subset | 13:58 |
| lucian | it falls back on libpython's eval if needed, afaict | 13:58 |
| kenaan | 12cfbolz set-strategies 11c3ed604fcfb5 15/: merge | 14:00 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1105 [12Carl Friedrich Bolz, set-strategies] | 14:00 |
| Alex_Gaynor | cfbolz: Is llgraph's Descr supposed to be seen in translation (even in tests) ever? | 14:02 |
| cfbolz | don't think so | 14:02 |
| cfbolz | Alex_Gaynor: btw, it seems lukas will start on type-specializing instances soonish | 14:02 |
| Action: Alex_Gaynor wonders what he did to break this on his branch | 14:02 | |
| Alex_Gaynor | cfbolz: Cool, dynamic specialization, or a static set of subclasses? | 14:03 |
| cfbolz | Alex_Gaynor: in a minimal way, ie with tagging and not GC support | 14:03 |
| cfbolz | dynamic | 14:03 |
| Alex_Gaynor | cfbolz: how is his tagging different from the translator option? | 14:04 |
| cfbolz | completely unrelated | 14:04 |
| cfbolz | the tagging is only used to not have to implement GC support | 14:05 |
| cfbolz | arigato: should I still get memory errors in _shape_to_ffi_type? | 14:05 |
| ltratt (~ltratt@87.115.8.1) left irc: Ping timeout: 252 seconds | 14:05 | |
| ltratt (~ltratt@87.115.8.1) joined #pypy. | 14:05 | |
| cfbolz | Alex_Gaynor: the idea is still to have the type of an instance's fields encoded in the map | 14:06 |
| cfbolz | and then store unboxedly | 14:06 |
| Alex_Gaynor | cfbolz: so in theory someone coul combine his work on encoding it in the map with GC integration :) | 14:08 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 14:08 | |
| cfbolz | Alex_Gaynor: yes | 14:08 |
| Alex_Gaynor | cool | 14:08 |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) joined #pypy. | 14:09 | |
| Action: Alex_Gaynor -> class | 14:11 | |
| arigato | cfbolz: re memory errors: | 14:17 |
| EnCuKou (~encukou@193.167.65.94) left irc: Remote host closed the connection | 14:17 | |
| arigato | cfbolz: yes, I fear so, unless the pypy that you're *using* to run translate.py is very recent | 14:17 |
| arigato | unsure if it's worth writing a hack to make it work on older pypy's | 14:18 |
| Rhy0lite (dje@nat/ibm/x-vtoqopwdtgiwgqdy) left irc: Quit: Leaving | 14:30 | |
| lizardo_ (~lizardo@201008201030.user.veloxzone.com.br) joined #pypy. | 14:30 | |
| lizardo (~lizardo@201009105166.user.veloxzone.com.br) left irc: Ping timeout: 258 seconds | 14:30 | |
| kenaan | 12arigo default 11913f736ff114 15/pypy/module/_rawffi/test/test__rawffi.py: Add a test for ed83fd7b7ec1. | 14:32 |
| Rhy0lite (dje@nat/ibm/x-sjkqybftuqlceyri) joined #pypy. | 14:32 | |
| whitelynx (~whitelynx@li117-47.members.linode.com) joined #pypy. | 14:38 | |
| cfbolz | arigato: ah, I see | 14:43 |
| cfbolz | arigato: ok, will update my pypy | 14:43 |
| EnCuKou (~encukou@a91-156-243-246.elisa-laajakaista.fi) joined #pypy. | 14:43 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) joined #pypy. | 14:45 | |
| cfbolz | arigato: i've been playing with the slf-in-slf meta-interpreter again | 14:46 |
| arigato | :-) | 14:46 |
| cfbolz | our optimizations have gotten a lot better | 14:46 |
| arigato | good | 14:46 |
| cfbolz | so now we can remove the meta-interpreter almost fully | 14:46 |
| arigato | what is left? :-) | 14:47 |
| cfbolz | some allocation of integers | 14:47 |
| cfbolz | because the same integer exists in several places | 14:47 |
| cfbolz | ie is passed twice along the jump | 14:47 |
| canta (~canta@77-20-123-240-dynip.superkabel.de) joined #pypy. | 14:48 | |
| arigato | ah, which is also a (minor) issue with pypy (running python) | 14:48 |
| cfbolz | it is? | 14:49 |
| arigato | yes | 14:50 |
| arigato | http://paste.pocoo.org/show/507693/ | 14:51 |
| arigato | this is not as optimized as without the "j = i" | 14:51 |
| arigato | (at least it was the case in pypy 1.6) | 14:51 |
| cfbolz | ah, bah | 14:51 |
| cfbolz | I see | 14:51 |
| cfbolz | arigato: also the meta-slf stuff is a great benchmark for optimizer performance :-) | 14:52 |
| arigato | :-) | 14:52 |
| aboudreault | Is there a PPA for pypy with ubuntu packages? | 14:52 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1105 [12Carl Friedrich Bolz, set-strategies] | 14:52 |
| arigato | aboudreault: I can't definitely answer this question, but I think so | 14:53 |
| aboudreault | arigato, ok | 14:54 |
| arigato | but you might get better results anyway by downloading the trunk at http://buildbot.pypy.org/nightly/trunk/ | 14:54 |
| aboudreault | ok, just checking to replace CPython on a server. | 14:54 |
| bfirsh (u1308@gateway/web/irccloud.com/x-vqqljpezvqvklqug) joined #pypy. | 14:59 | |
| Nick change: Gulopine -> Gulaway | 14:59 | |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy. | 15:00 | |
| gutworth (~benjamin@dhcp-128-36-89-186.central.yale.edu) joined #pypy. | 15:05 | |
| ltratt | arigato: do exceptions propagate over stacklet switches? I'm guessing not. | 15:05 |
| arigato | if there is an exception, you cannot call the switch() method at all | 15:06 |
| arigato | but if a non-main stacklet terminates with an exception, then yes | 15:07 |
| masquerade (~rdeaton@galactica.lug.udel.edu) left irc: Read error: Operation timed out | 15:07 | |
| arigato | ah, no | 15:07 |
| arigato | indeed, you crash if a non-main stacklet terminates with an exception | 15:08 |
| ltratt | arigato: ok, that's what I'm seeing :) | 15:08 |
| arigato | there is code in module/_continuation that catches the exception and propagates it to the parent | 15:08 |
| masquerade (~rdeaton@galactica.lug.udel.edu) joined #pypy. | 15:08 | |
| arigato | basically a non-main stacklet *must* terminate by returning a valid stacklet to jump to | 15:08 |
| Nick change: masquerade -> Guest58894 | 15:09 | |
| ltratt | OK, I was wondering if I might need to do that. | 15:09 |
| fijal (~fijal@195.175.230.246) joined #pypy. | 15:10 | |
| ltratt | Although I'm not sure if I should keep on using stacklets or not, anway! Decisions, decisions.. | 15:10 |
| arigato | :-) | 15:11 |
| ltratt | I think I probably will keep using them, as they keep the VM nice and clean. If they're fixed, brilliant. If RPython grows generators, it's a simple search and replace to move thing to generators. If I move to the "store state and restore", it's more of a one-way ticket, and it's pretty icky... | 15:12 |
| Action: fijal greets from istanbul airport | 15:19 | |
| Rhy0lite | heh. hi, fijal! | 15:20 |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 15:20 | |
| Squarepy (~Squarepy@unaffiliated/squarepy) left irc: Quit: Leaving | 15:22 | |
| arigato | hi :-) | 15:22 |
| fijal | ronny: remote machine translation is a very likely non-feature | 15:23 |
| fijal | Alex_Gaynor: ping | 15:23 |
| espes (espes@2001:470:1f05:ca1::) left irc: Ping timeout: 244 seconds | 15:25 | |
| exarkun | fijal: You mean PyPy will never translate super fast on my beowulf cluster? | 15:27 |
| fijal | exarkun: this question is so wrong, I don't know how to answer | 15:28 |
| xorAxAx | exarkun: it will probably some day, but not remotely if you follow fijals point :) | 15:29 |
| fijal | exarkun: I guess my point was that ssh+screen is a very good remote compilation protocol | 15:29 |
| cfbolz | exarkun: you still have one of those? | 15:30 |
| exarkun | cfbolz: Heh. No, I never had one. :) | 15:31 |
| cfbolz | beowulf clusters are so 90ies | 15:31 |
| exarkun | indeed :) | 15:31 |
| fijal | cfbolz: cool about specialized instances | 15:32 |
| fijal | cfbolz: we should start merging lukas branches though ;-) | 15:32 |
| cfbolz | will take some time | 15:33 |
| cfbolz | yes | 15:33 |
| fijal | did alex review list-strategies? | 15:33 |
| cfbolz | fijal: I think list-strategies should be mergeable after the release | 15:33 |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) joined #pypy. | 15:33 | |
| fijal | ok | 15:33 |
| fijal | we have the release branch | 15:33 |
| fijal | so feel free | 15:33 |
| arigato | well | 15:34 |
| cfbolz | no, that makes merging more annoying | 15:34 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1842 [12fijal, release-1.7.x] | 15:34 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/702 [12fijal, release-1.7.x] | 15:34 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-win-x86-32/builds/537 [12fijal, release-1.7.x] | 15:34 |
| arigato | I honestly think that we should just merge everything on the release branch soon | 15:34 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1106 [12fijal, release-1.7.x] | 15:34 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/568 [12fijal, release-1.7.x] | 15:34 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/265 [12fijal, release-1.7.x] | 15:34 |
| fijal | arigato: did we fix bugs? | 15:34 |
| arigato | yes, tons, on Windows, for example | 15:34 |
| fijal | ok | 15:35 |
| fijal | should I stop the buildbots then? | 15:35 |
| arigato | the Windows situation improved from "crashes randomly" to "seems to work" | 15:35 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/265 [12fijal, release-1.7.x] | 15:35 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/568 [12fijal, release-1.7.x] | 15:35 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1106 [12fijal, release-1.7.x] | 15:35 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/pypy-c-app-level-win-x86-32/builds/537 [12fijal, release-1.7.x] | 15:35 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/702 [12fijal, release-1.7.x] | 15:35 |
| bbot2 | 13Exception: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1842 [12fijal, release-1.7.x] | 15:36 |
| fijal | no point in overloading slaves then | 15:36 |
| arigato | I'm still trying to run the Windows buildbot and to figure out which tests are really fixed and which ones need more attention | 15:36 |
| fijal | ok | 15:36 |
| fijal | I'll take care about the new website and renaming numpy then | 15:36 |
| arigato | (please do not stop my windows build :-) i'm waiting for it since 6 hours) | 15:37 |
| fijal | I only stopped mine | 15:37 |
| kenaan | 12edelsohn ppc-jit-backend 113db474e494aa 15/pypy/jit/backend/ppc/ppcgen/codebuilder.py: refactor load_from_addr. clarify register usage in store_reg. | 15:40 |
| kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/ | 15:41 | |
| Rhy0lite | bivab: ping | 15:41 |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) left irc: Remote host closed the connection | 15:43 | |
| kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy. | 15:44 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) left irc: Quit: unbit | 15:45 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) joined #pypy. | 15:45 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) left irc: Read error: Connection reset by peer | 15:46 | |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 15:51 | |
| cfbolz | Rhy0lite: david and sven are both busy right now, sven might come online a while | 15:52 |
| Rhy0lite | ok | 15:52 |
| Action: arigato -> ferry | 15:52 | |
| Action: cfbolz -> home | 15:55 | |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) left irc: Client Quit | 15:55 | |
| gutworth (benjamin@dhcp-128-36-89-186.central.yale.edu) left #pypy. | 15:56 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Read error: Connection reset by peer | 15:58 | |
| nedbat (nedbat@python/psf/nedbat) left #pypy ("Leaving"). | 16:02 | |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 16:02 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 16:03 | |
| nedbat | fijal: btw: using the latest stable nightly code, I didn't have that same build problem. I think you should probably revert the change you made for me. | 16:05 |
| nedbat | fijal: sorry for the noise. | 16:05 |
| fijal | nedbat: well, latest nightly fixed that | 16:06 |
| fijal | I added extra small change | 16:06 |
| fijal | (for makefiles) | 16:06 |
| nedbat | ok, thanks. | 16:07 |
| Nick change: Guest58894 -> masquerade | 16:07 | |
| arigato | as usual, no clue whatsoever what detail of the windows configuration changed | 16:12 |
| fijal | arigato: are you on a ferry? | 16:12 |
| arigato | now the test run fails to locate libexpat.dll and gives a nice dialog box saying so | 16:12 |
| arigato | fijal: no, about to leave | 16:12 |
| arigato | I will be tonight | 16:12 |
| arigato | ...well, as usual I suppose I'll give up understanding what could have changed, and hack, e.g. by copying libexpat.dll around | 16:13 |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Quit: bivab | 16:14 | |
| arigato | fwiw, on win64 christian used a tool that is able to figure out that a dialog box has been opened and click itself on the "Ok" button | 16:16 |
| arigato | advanced technology | 16:16 |
| fijal | I've seen such tools for windows | 16:16 |
| sven_hager (~sven@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 16:18 | |
| arigato | :-/ | 16:18 |
| arigato (~arigo@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Quit: See you | 16:18 | |
| fijal | https://bitbucket.org/fijal/talk/src/default/blog/analysing-performance.rst | 16:19 |
| fijal | anyone feel like reviewing? | 16:19 |
| JaRoel (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection | 16:19 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 16:20 | |
| Rhy0lite | fijal: I'll take a loot | 16:20 |
| Rhy0lite | look | 16:20 |
| fijal | don't take my loot please | 16:20 |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 16:23 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 240 seconds | 16:24 | |
| Nick change: lizardo_ -> lizardo | 16:27 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) joined #pypy. | 16:27 | |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) left irc: Quit: Leaving... | 16:28 | |
| AndrewHarrison (~Adium@66-215-59-147.dhcp.atsc.ca.charter.com) left irc: Client Quit | 16:29 | |
| fijal (~fijal@195.175.230.246) left irc: Ping timeout: 244 seconds | 16:29 | |
| Alex_Gaynor | fijal: pong | 16:34 |
| Rhy0lite | I hope they didn't take him to a Turkish prison | 16:38 |
| nedbat (nedbat@python/psf/nedbat) left #pypy ("Leaving"). | 16:38 | |
| Alex_Gaynor | antocuni: ping | 16:40 |
| fijal (~fijal@195.175.230.246) joined #pypy. | 16:41 | |
| kenaan | 12hager ppc-jit-backend 11a1d07117b3f1 15/pypy/jit/backend/ppc/ppcgen/ppc_assembler.py: Fixed assertion that made multiple tests fail due to side effect | 16:44 |
| Alex_Gaynor | fijal: pong again? | 16:45 |
| mattip (~chatzilla@109.64.118.235) joined #pypy. | 16:45 | |
| fijal | Alex_Gaynor: why do we need static descrs now that we have dynamic ones btw? | 16:45 |
| Alex_Gaynor | fijal: you mean have all descrs be purely integer based? | 16:46 |
| fijal | yes | 16:46 |
| fijal | I guess | 16:46 |
| Alex_Gaynor | not sure, I don't know the original design decisions for descrs | 16:46 |
| fijal | you mean why creating descrs is not rpython? | 16:47 |
| fijal | I think caching plays a role | 16:47 |
| fijal | but we can cache on rpython level as well | 16:47 |
| fijal | btw, do you make sure that identical descrs are cached? | 16:48 |
| Alex_Gaynor | Uh, I guess, I just mean I don't know what other purposes descrs might serve, the dynamic info seems sufficient for what the backend does | 16:48 |
| Trundle (~andy@python/site-packages/trundle) left irc: Remote host closed the connection | 16:48 | |
| Alex_Gaynor | fijal: they aren't with llmodel ATM, that should be a WeakValueDict for a cache I suppose? | 16:49 |
| fijal | then well, let's make all descrs RPython | 16:49 |
| fijal | why Weak? | 16:49 |
| Alex_Gaynor | if you deallocate assembler you should lose descr, I suppose | 16:49 |
| Alex_Gaynor | fijal: on that note, what's the point of llgraph's Descr class, instead of using the llsupport ones? | 16:50 |
| fijal | who's using llgraph Descr? | 16:50 |
| fijal | it should not be seen by the translation | 16:50 |
| Alex_Gaynor | llgraph | 16:50 |
| `fox` (~fox@host150-111-dynamic.17-79-r.retail.telecomitalia.it) joined #pypy. | 16:50 | |
| Alex_Gaynor | it's seen in my branch's test_ztranslation | 16:50 |
| Alex_Gaynor | I don't know what I did to break that | 16:50 |
| fijal | test_ztranslation runs on llgraph no? | 16:51 |
| Alex_Gaynor | I have no idea | 16:51 |
| fijal | yes | 16:51 |
| fijal | metainterp/test/test_ztranslation.py | 16:51 |
| fijal | this one | 16:51 |
| Alex_Gaynor | yes | 16:52 |
| stakkars (~tismer@89.204.154.73) joined #pypy. | 16:53 | |
| Alex_Gaynor | fijal: it fails on my branch because getdescr is seen, I guess because it can in theory be called at runtime | 16:53 |
| fijal | ah | 16:53 |
| fijal | so you changed that :-P | 16:53 |
| fijal | you have to do something about it then | 16:53 |
| Alex_Gaynor | how else should I create a descr at runtime! | 16:53 |
| fijal | make it rpython then | 16:53 |
| Alex_Gaynor | well ISTM it's fixed automatically if you just switch to using llsupport's descrs | 16:53 |
| apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) left irc: Remote host closed the connection | 16:54 | |
| fijal | does that even work on llgraph? | 16:55 |
| fijal | if so maybe we can move descrs to model and just ignore llgraph.Decr | 16:55 |
| fijal | or remove it entirely? | 16:55 |
| Sho_ (~EHS1@kde/hein) left irc: Quit: Stop leaking memory like it's going out of fashion. | 16:55 | |
| mattip (chatzilla@109.64.118.235) left #pypy. | 16:55 | |
| fijal | I mean - this is the point where it's worth simplifying things | 16:55 |
| fijal | without adding new stuff to the mess | 16:55 |
| Alex_Gaynor | fijal: I dunno, I was trying to have one branch that didn't involve 3 layers of yak shaving | 16:55 |
| fijal | how about having jit-simplify-descrs branch first then? | 16:56 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/262 [12arigo] | 16:56 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/263 [12fijal, release-1.7.x] | 16:56 |
| Alex_Gaynor | Yes I suppose. /me wishes he had a lacky who would do this work for him. | 16:56 |
| fijal | I'm busy with other stuff :-P | 16:57 |
| antocuni_ (~antocuni@host216-125-dynamic.6-79-r.retail.telecomitalia.it) joined #pypy. | 16:57 | |
| antocuni (~antocuni@host70-121-dynamic.11-79-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 17:00 | |
| witulski (~stupsi@134.99.16.23) joined #pypy. | 17:02 | |
| mattip (~chatzilla@bzq-109-64-118-235.red.bezeqint.net) joined #pypy. | 17:05 | |
| mattip | Alex_Gaynor: ping | 17:07 |
| Alex_Gaynor | mattip: pong | 17:07 |
| JaRoel|4d (~jaroel|4d@2001:980:7133:1:6c40:286b:dd77:2045) joined #pypy. | 17:08 | |
| mattip | I am available for small well defined tasks | 17:08 |
| fijal | mattip: hi | 17:08 |
| mattip | Fijal: hi | 17:09 |
| mattip | fijal: I have been playing with multidimensional arrays | 17:09 |
| fijal | cool | 17:09 |
| fijal | I've rewritten a large chunk of them | 17:09 |
| mattip | and I think I see an easy way to do transpose and matrixmultiply | 17:10 |
| mattip | using your shards | 17:10 |
| mattip | So I'm just kind of hanging out waiting for the branch to mature | 17:10 |
| fijal | cool | 17:10 |
| k_bx (~k_bx@195.20.130.1) left irc: Remote host closed the connection | 17:11 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Ping timeout: 240 seconds | 17:13 | |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy. | 17:14 | |
| ahall (~ahall@2001:bd0:100::666) joined #pypy. | 17:19 | |
| Action: fijal found a crash in split_jit_merge_points | 17:19 | |
| fijal | nice :/ | 17:20 |
| fprimex (~fprimex@fl-67-233-144-228.dhcp.embarqhsd.net) joined #pypy. | 17:20 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) joined #pypy. | 17:20 | |
| kenaan | 12fijal numpy-multidim-shards 119121efb3b83f 15/pypy/module/micronumpy/interp_numarray.py: move to_str to the base class | 17:22 |
| kenaan | 12fijal numpy-multidim-shards 11c857554d0c6c 15/pypy/module/micronumpy/: rpythonize a bit. Disable _immutable_fields_ for now | 17:22 |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Ping timeout: 240 seconds | 17:23 | |
| ixokai (~ixokai@rrcs-24-199-8-246.west.biz.rr.com) joined #pypy. | 17:25 | |
| ixokai (~ixokai@rrcs-24-199-8-246.west.biz.rr.com) left irc: Changing host | 17:25 | |
| ixokai (~ixokai@pdpc/supporter/student/ixokai) joined #pypy. | 17:25 | |
| G2P (~G2P@fw-asn1.ornis.com) left irc: Quit: Leaving. | 17:26 | |
| fijal | good | 17:32 |
| fijal | Alex_Gaynor: I have an obscure case of arraycopy not optimized correctly | 17:33 |
| fijal | want to have a look? | 17:33 |
| fijal | mattip: making descr_repr work on numpy-multidim-shards is a good idea? | 17:33 |
| Alex_Gaynor | fijal: sure | 17:33 |
| kenaan | 12fijal numpy-multidim-shards 1114f8da6a95bf 15/pypy/module/micronumpy/: Make array iterators a once-off immutable things. | 17:34 |
| fijal | Alex_Gaynor: run test_zjit -k test_multidim_slice --viewloops | 17:35 |
| Alex_Gaynor | fijal: on that branch? | 17:35 |
| fijal | yes | 17:35 |
| fijal | guards are there to stay I believe | 17:35 |
| fijal | store sinking does not work across boundaries, otherwise the above commit would make no sense | 17:36 |
| Alex_Gaynor | fijal: uh, segfaults for me | 17:36 |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy | 17:37 | |
| antocuni__ (~antocuni@host86-122-dynamic.6-79-r.retail.telecomitalia.it) joined #pypy. | 17:38 | |
| fijal | Alex_Gaynor: yes, but after it shows the loop ;-) | 17:39 |
| fijal | not sure what segfaults | 17:39 |
| Alex_Gaynor | fijal: no,it just segfaults and dies :) | 17:39 |
| fijal | nice | 17:39 |
| mattip | fijal: sorry, went away for a minute. I'm not sure what the question is. | 17:39 |
| mattip | fijal: it's a good idea, do you want me to do it? | 17:40 |
| fijal | mattip: you were looking for a simple task right? | 17:40 |
| ltratt (~ltratt@87.115.8.1) left irc: Quit: Bye all. | 17:40 | |
| fijal | if you feel like doing it that would be great | 17:40 |
| fijal | because I broke it | 17:40 |
| antocuni_ (~antocuni@host216-125-dynamic.6-79-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 17:40 | |
| fijal | while moving to shards | 17:40 |
| fijal | also, it would be nice if to_str receives a builder | 17:40 |
| fijal | instead of creating a new one locally | 17:40 |
| mattip | OK, I have to run, but I will take a look later tonight. Have a good flight. | 17:40 |
| stakkars (~tismer@89.204.154.73) left irc: Quit: schnarch | 17:41 | |
| fijal | mattip: cool | 17:41 |
| mattip | fijal: builder? Like a StringBuilder? | 17:41 |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) joined #pypy. | 17:42 | |
| fijal | yes | 17:42 |
| mattip | OK, I'll poke around and see if I "get" the concept. | 17:43 |
| mattip (chatzilla@bzq-109-64-118-235.red.bezeqint.net) left #pypy ("late"). | 17:43 | |
| fijal | Alex_Gaynor: ups | 17:43 |
| kenaan | 12fijal numpy-multidim-shards 11ba1a3becc049 15/pypy/module/micronumpy/test/test_numarray.py: one more test and make the other test failing without problems with repr | 17:43 |
| fijal | ups, indeed I broke it | 17:43 |
| fijal | Alex_Gaynor: I'll fix it first | 17:44 |
| Alex_Gaynor | fijal: thanks | 17:45 |
| asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy. | 17:47 | |
| jacob22_ | fijal: The line " PYPYLOG=log ./test.py" needs a bit of explaining. Is "log" the name of the logfile you are creating? Is ./test.py the program you are profiling, and if so, I guess it needs a #/usr/bin/env pypy line at the beginning. | 17:50 |
| sven_hager (~sven@fwstups.cs.uni-duesseldorf.de) left irc: Quit: Leaving | 17:51 | |
| jacob22_ | fijal: The rest of the post is very nice and clear. | 17:51 |
| fijal | yes, you have to run it under pypy | 17:51 |
| fijal | whatever way | 17:51 |
| fijal | I guess I'm officially dead by now | 17:54 |
| fzzzy (~donovan@nat/mozilla/x-auzwlaoobuwscgrg) joined #pypy. | 17:56 | |
| Alex_Gaynor | fijal: llsupport's descr things need gc_ll_descr, which llgraph doesn't have | 17:56 |
| fijal | see? | 17:57 |
| Alex_Gaynor | ? | 17:57 |
| fijal | it's mess | 17:58 |
| fijal | a mess :) | 17:58 |
| Alex_Gaynor | yes | 17:58 |
| fijal | Call1 has values None | 18:00 |
| fijal | how that can happen? | 18:01 |
| fijal | uh | 18:02 |
| fijal | Alex_Gaynor: consider this | 18:02 |
| fijal | a = numpy.array([...]) | 18:02 |
| fijal | b = a + a | 18:02 |
| fijal | c = b + b | 18:03 |
| fijal | and I force b | 18:03 |
| fijal | now I can't force c | 18:03 |
| fijal | any more | 18:03 |
| Alex_Gaynor | that's a bug | 18:03 |
| Alex_Gaynor | we even have a test for this, I claim | 18:03 |
| kenaan | 12hager ppc-jit-backend 11c76ce86e60ff 15/pypy/jit/backend/ppc/ppcgen/: Implemented COPYSTRCONTENT | 18:04 |
| fijal | anyway, I'm down in the yak-shaving territory | 18:06 |
| Alex_Gaynor | fijal: pff, me too, descrs are a mess, Idon't feel like refactoring the whole world | 18:06 |
| fijal | we don | 18:06 |
| fijal | 't have a test for this | 18:06 |
| fijal | not directly I think | 18:06 |
| Alex_Gaynor | I don't understand how they translate at all though, since some descrs have an int attribute and others have a EffectINfo one | 18:07 |
| lizardo_ (~lizardo@201009096007.user.veloxzone.com.br) joined #pypy. | 18:07 | |
| Alex_Gaynor | fijal: test_numarry:test_auto_force | 18:07 |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 18:07 | |
| fijal | I think this is a different combination of forces | 18:08 |
| fijal | anyway, fixed | 18:08 |
| lizardo (~lizardo@201008201030.user.veloxzone.com.br) left irc: Ping timeout: 258 seconds | 18:08 | |
| Alex_Gaynor | fijal: want to fix my test_ztranslation issue now :) | 18:08 |
| fijal | no :) | 18:10 |
| Alex_Gaynor | me neither | 18:10 |
| kenaan | 12fijal numpy-multidim-shards 118521a920ed05 15/pypy/module/micronumpy/: provide some sort of descr_repr (a broken one) and a fix | 18:10 |
| Tobu (~Tobu@unaffiliated/tobu) left irc: Ping timeout: 240 seconds | 18:14 | |
| Alex_Gaynor | fijal: can you explain how llgraph translate at all ATM? | 18:14 |
| fprimex (~fprimex@fl-67-233-144-228.dhcp.embarqhsd.net) left irc: Quit: Leaving | 18:14 | |
| fijal | with black magic | 18:16 |
| fijal | it has a boundary | 18:16 |
| fijal | rpython/not rpython | 18:16 |
| fijal | look how the runner.py interacts with llimpl | 18:17 |
| Alex_Gaynor | fijal: I'm talking about the Descrs themselves, some have an int extra_info, others have an EffecInfo | 18:17 |
| fijal | but they get compiled | 18:18 |
| fijal | separately into stuff that's in llimpl | 18:18 |
| fijal | no? | 18:18 |
| Alex_Gaynor | I don't know, I'm asking :) | 18:18 |
| ojii (~ojii@40-34.60-188.cust.bluewin.ch) left irc: Quit: Leaving | 18:18 | |
| fijal | Alex_Gaynor: I'm sorry I'm dead | 18:20 |
| fijal | I'll go off maybe read a book | 18:20 |
| Alex_Gaynor | ok, I'll hack hack hack I guess | 18:20 |
| kenaan | 12fijal numpy-multidim-shards 11f7cb08c0396d 15/pypy/module/micronumpy/: some partial fixes, will continue later | 18:20 |
| fijal | read the blog post I posted here | 18:21 |
| kenaan | 12hager ppc-jit-backend 1134e139792020 15/pypy/jit/backend/ppc/ppcgen/: Implemented COPYUNICODECONTENT | 18:22 |
| Alex_Gaynor | fijal: you keep changing these tests, maybe make sure they're correct before you commit them in the first place ;) | 18:23 |
| fijal | well | 18:23 |
| fijal | that's the whole problem | 18:23 |
| fijal | I don't quite know what they should be upfront | 18:23 |
| fijal | I'll change them again | 18:23 |
| fijal | (they fail now) | 18:23 |
| mat^2 | I can confirm that https://bugs.pypy.org/issue849 is fixed in trunk | 18:24 |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) left irc: Quit: etrepum | 18:25 | |
| bobbyz (~bobbyz@c-67-184-43-39.hsd1.il.comcast.net) left irc: Quit: Leaving | 18:26 | |
| fijal (~fijal@195.175.230.246) left irc: Ping timeout: 260 seconds | 18:30 | |
| witulski (stupsi@134.99.16.23) left #pypy. | 18:30 | |
| mat^2 | running pyglet in pypy trunk on win32 raises this exception: http://paste.pocoo.org/show/jyAabBpSG1kocwpkPGWs/ | 18:32 |
| Nick change: lizardo_ -> lizardo | 18:32 | |
| CIA-68 | 03matpow2 07roundup * 10#849/pypy 1.6, pyglet crash on win32: | 18:35 |
| CIA-68 | The segfault seems to have been fixed in PyPy trunk. Now I get exceptions and | 18:35 |
| CIA-68 | warnings like this, though: http://paste.pocoo.org/show/rPNQr4baBys ... * 14https://bugs.pypy.org/issue849 | 18:35 |
| pedronis_ (~pedronis@73-53.195-178.cust.bluewin.ch) joined #pypy. | 18:41 | |
| pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) left irc: Ping timeout: 260 seconds | 18:41 | |
| Nick change: antocuni__ -> antocuni | 18:45 | |
| antocuni | Alex_Gaynor: pong | 18:45 |
| Alex_Gaynor | antocuni: do you know if metainterp/test_ztranslation passed on ffistruct? | 18:45 |
| antocuni | I could translate until the most recent commits, so I suppose that it passed | 18:46 |
| antocuni | but I never actually tried, I think | 18:46 |
| Alex_Gaynor | well, my branch translates as well, but fails that one because llgraph is different from x86 | 18:47 |
| antocuni | uhm | 18:47 |
| Alex_Gaynor | but looking at your diffa gain, my issue wouldn't apply to you, so I guess I'll just hack some more | 18:47 |
| antocuni | (I'm reading the logs) if the translator sees the llgraph descrs, it's probably because you are doing too much in runner.py | 18:48 |
| antocuni | you should move everything which is not rpython inside llimpl.py | 18:48 |
| Alex_Gaynor | yes, I guess so, thanks | 18:48 |
| antocuni | btw, I'm not sure about fijal's proposal about descrs | 18:48 |
| Guest35710 (~jared@061093200075.ctinets.com) left irc: Ping timeout: 245 seconds | 18:48 | |
| Alex_Gaynor | I gave up on simplifying descrs,someone else can try with that mess | 18:48 |
| antocuni | so far they are prebuilt because we can cache them | 18:49 |
| antocuni | and for dynamic ones, we don't case about creating new ones because they are not too many anyway | 18:49 |
| Alex_Gaynor | ok, I'll hack some more | 18:50 |
| etrepum (~bob@accessnat4.mochimedia.net) joined #pypy. | 18:50 | |
| etrepum (~bob@accessnat4.mochimedia.net) left irc: Client Quit | 18:54 | |
| etrepum (~bob@accessnat4.mochimedia.net) joined #pypy. | 18:55 | |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 19:08 | |
| kenaan | 12edelsohn ppc-jit-backend 113a6600bf032a 15/pypy/jit/backend/ppc/ppcgen/opassembler.py: setarrayitem and getarrayitem offsets are immediate values. | 19:11 |
| Nick change: mcdonc_ -> mcdonc | 19:14 | |
| kenaan | 12edelsohn ppc-jit-backend 11b6a18d1530bf 15/pypy/jit/backend/ppc/ppcgen/opassembler.py: setarrayitem and getarrayitem cannot add offset with addi. | 19:19 |
| hakanardo | how do I debug this: http://paste.pocoo.org/show/507882/ | 19:20 |
| hakanardo | any way I can dig out the name offending attribute? | 19:20 |
| [mat^2] (~mathias@212.130.113.35) joined #pypy. | 19:20 | |
| Alex_Gaynor | hakanardo: change that raise Exception to just be `raise`, and you'll probably have to do it in 1 other place, but eventually you'll get a useful location for pdb | 19:21 |
| Alex_Gaynor | hakanardo: I did this exact thing the other day :/ | 19:21 |
| mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 260 seconds | 19:21 | |
| hakanardo | ok, I'll try | 19:21 |
| hakanardo | thanx | 19:21 |
| aisipos (~aisipos@wc.lax.truecarcorp.com) joined #pypy. | 19:22 | |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy. | 19:25 | |
| nickik (~nick@217.150.239.67) joined #pypy. | 19:28 | |
| nickik (~nick@217.150.239.67) left irc: Client Quit | 19:29 | |
| kenaan | 12hakanardo jit-targets 1136b8365a8bba 15/pypy/jit/metainterp/optimizeopt/unroll.py: translation fix | 19:32 |
| Shinata (~John@osbk-4db06ce4.pool.mediaWays.net) left irc: Read error: Connection reset by peer | 19:33 | |
| hakanardo | that gave me two more lines of mandelbrot and a new crash :) | 19:35 |
| Alex_Gaynor | progress! | 19:35 |
| Moku (~John@osbk-4db06ce4.pool.mediaWays.net) joined #pypy. | 19:35 | |
| Nick change: Moku -> Guest78224 | 19:36 | |
| hakanardo | any idea on how to tackle this one: http://paste.pocoo.org/show/507894/ | 19:36 |
| hakanardo | Alex_Gaynor: indeed! | 19:37 |
| Alex_Gaynor | no idea what tha tone means | 19:38 |
| hakanardo | me neither... | 19:38 |
| hakanardo_ (~hakanardo@h59ec009e.seluldx.dyn.perspektivbredband.net) joined #pypy. | 19:43 | |
| Alex_Gaynor | hakanardo: any idea what http://paste.pocoo.org/show/507895/ means? | 19:45 |
| kkris (~kris@80-123-36-150.adsl.highway.telekom.at) joined #pypy. | 19:48 | |
| hakanardo | not realy | 19:48 |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 260 seconds | 19:48 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 19:48 | |
| hakanardo | a wild guess would be something about mixing *args, **kwyards and/or default argments | 19:49 |
| hakanardo | differently on different implementations of the same method | 19:49 |
| kenaan | 12alex_gaynor jit-dynamic-getarrayitem 1146c3f314a1e7 15/pypy/jit/backend/model.py: added these to base model | 19:52 |
| kenaan | 12alex_gaynor jit-dynamic-getarrayitem 1199efe7320cf1 15/pypy/jit/backend/llgraph/runner.py: don't cache dynamic descrs, progress maybe | 19:52 |
| kenaan | 12alex_gaynor jit-dynamic-getarrayitem 1169c9793c54c0 15/pypy/jit/backend/model.py: remove the dupe | 19:53 |
| aat (~aat@c-67-188-9-79.hsd1.ca.comcast.net) joined #pypy. | 19:58 | |
| tlynn (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) joined #pypy. | 19:59 | |
| fox__ (~fox@host239-108-dynamic.244-95-r.retail.telecomitalia.it) joined #pypy. | 20:07 | |
| `fox` (~fox@host150-111-dynamic.17-79-r.retail.telecomitalia.it) left irc: Ping timeout: 240 seconds | 20:07 | |
| Squarepy (~Squarepy@ip4da70b11.direct-adsl.nl) joined #pypy. | 20:07 | |
| Squarepy (~Squarepy@ip4da70b11.direct-adsl.nl) left irc: Changing host | 20:08 | |
| Squarepy (~Squarepy@unaffiliated/squarepy) joined #pypy. | 20:08 | |
| stakkars (~tismer@p5DDB5E98.dip.t-dialin.net) joined #pypy. | 20:08 | |
| aboudreault (~alanb@osgeo/member/aboudreault) left irc: Quit: Leaving | 20:10 | |
| etrepum (~bob@accessnat4.mochimedia.net) left irc: Quit: etrepum | 20:17 | |
| Squarepy (~Squarepy@unaffiliated/squarepy) left irc: Quit: Leaving | 20:23 | |
| Trundle (~andy@89.204.155.169) joined #pypy. | 20:28 | |
| Trundle (~andy@89.204.155.169) left irc: Changing host | 20:28 | |
| Trundle (~andy@python/site-packages/trundle) joined #pypy. | 20:28 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Ping timeout: 252 seconds | 20:32 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy. | 20:38 | |
| Nick change: Guest78224 -> Shinita | 20:39 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 20:43 | |
| aisipos (~aisipos@wc.lax.truecarcorp.com) left irc: Read error: Operation timed out | 20:46 | |
| asmeurer_ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer_ | 20:48 | |
| stakkars_ (~tismer@dslb-088-072-232-162.pools.arcor-ip.net) joined #pypy. | 20:51 | |
| antocuni (~antocuni@host86-122-dynamic.6-79-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 20:54 | |
| ixokai (~ixokai@pdpc/supporter/student/ixokai) left irc: Quit: Bye! | 20:55 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) joined #pypy. | 20:56 | |
| stakkars_ (~tismer@dslb-088-072-232-162.pools.arcor-ip.net) left irc: Quit: schnarch | 20:57 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy. | 20:57 | |
| __main__ (~main@adsl-99-173-15-158.dsl.pltn13.sbcglobal.net) left irc: Read error: Connection reset by peer | 20:59 | |
| basepi (~basepi@server.cmyers.net) left irc: Quit: ZNC - http://znc.sourceforge.net | 21:02 | |
| basepi (~basepi@server.cmyers.net) joined #pypy. | 21:03 | |
| __main__ (~main@adsl-99-173-15-158.dsl.pltn13.sbcglobal.net) joined #pypy. | 21:09 | |
| kkris (~kris@80-123-36-150.adsl.highway.telekom.at) left irc: Quit: Leaving. | 21:13 | |
| stakkars_ (~tismer@dslb-088-072-232-162.pools.arcor-ip.net) joined #pypy. | 21:13 | |
| tilgovi (~randall@199.36.158.16) joined #pypy. | 21:16 | |
| tilgovi (~randall@199.36.158.16) left irc: Changing host | 21:16 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 21:16 | |
| aisipos (~aisipos@wc.lax.truecarcorp.com) joined #pypy. | 21:18 | |
| stakkars__ (~tismer@p5DDB5E98.dip.t-dialin.net) joined #pypy. | 21:19 | |
| stakkars (~tismer@p5DDB5E98.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 21:19 | |
| Nick change: stakkars__ -> stakkars | 21:19 | |
| kenaan | 12boemmels lang-scheme 1184d83d4e7639 15/scheme/execution.py: naming consistency | 21:28 |
| Nick change: pedronis_ -> pedronis | 21:29 | |
| jaw^ (~jaw@karl1-149.cust.blixtvik.net) joined #pypy. | 21:32 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 21:34 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 21:40 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 21:41 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Ping timeout: 245 seconds | 21:45 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/263 [12fijal, release-1.7.x] | 21:53 |
| rekamso (~textual@67.51.82.66) joined #pypy. | 21:57 | |
| fzzzy (~donovan@nat/mozilla/x-auzwlaoobuwscgrg) left irc: Quit: fzzzy | 21:58 | |
| fzzzy (~donovan@nat/mozilla/x-immkjxtsvtnduvxe) joined #pypy. | 22:00 | |
| Rhy0lite (dje@nat/ibm/x-sjkqybftuqlceyri) left irc: Quit: Leaving | 22:04 | |
| fox__ (~fox@host239-108-dynamic.244-95-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 22:08 | |
| kenaan | 12ctismer win64-stage1 11bb40c355d38d 15/pypy/rlib/clibffi.py: fixed clibffi.py, but win32.c must be replaced, and I don't know yes by what. | 22:11 |
| kenaan | 12ctismer win64-stage1 114b214639c276 15/: Merge with default | 22:11 |
| kenaan | 12ctismer win64-stage1 110b37384fbaad 15/pypy/translator/c/src/libffi_msvc/win64.asm: added the win64.asm source from cpython | 22:11 |
| kenaan | 12ctismer win64-stage1 11e1834e63c531 15/pypy/rlib/clibffi.py: added win64.asm | 22:11 |
| stakkars_ (~tismer@dslb-088-072-232-162.pools.arcor-ip.net) left irc: Quit: schnarch | 22:12 | |
| asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy. | 22:23 | |
| fzzzy (~donovan@nat/mozilla/x-immkjxtsvtnduvxe) left irc: Quit: fzzzy | 22:25 | |
| fzzzy (~donovan@nat/mozilla/x-zalwykhlbyeqkhii) joined #pypy. | 22:27 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 22:28 | |
| stakkars_ (~tismer@p5DC4696F.dip0.t-ipconnect.de) joined #pypy. | 22:30 | |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Remote host closed the connection | 22:31 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 22:33 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 22:33 | |
| stakkars__ (~tismer@p5DC4696F.dip0.t-ipconnect.de) joined #pypy. | 22:34 | |
| stakkars_ (~tismer@p5DC4696F.dip0.t-ipconnect.de) left irc: Read error: Connection reset by peer | 22:34 | |
| stakkars_ (~tismer@p5DC4696F.dip0.t-ipconnect.de) joined #pypy. | 22:36 | |
| stakkars__ (~tismer@p5DC4696F.dip0.t-ipconnect.de) left irc: Read error: Connection reset by peer | 22:36 | |
| zain2 (~zain@69.38.221.130) joined #pypy. | 22:37 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Ping timeout: 260 seconds | 22:38 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 22:43 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/138 | 22:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/712 | 22:44 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/941 | 22:44 |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 22:50 | |
| Rhyolite (~dje@pool-108-6-25-71.nycmny.fios.verizon.net) left irc: Quit: Leaving | 22:51 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 22:51 | |
| stakkars__ (~tismer@p5DC4696F.dip0.t-ipconnect.de) joined #pypy. | 22:53 | |
| stakkars_ (~tismer@p5DC4696F.dip0.t-ipconnect.de) left irc: Read error: Connection reset by peer | 22:53 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Read error: Connection reset by peer | 22:53 | |
| Rhyolite (~dje@pool-108-6-25-71.nycmny.fios.verizon.net) joined #pypy. | 22:53 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 22:53 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 22:56 | |
| lizardo (~lizardo@201009096007.user.veloxzone.com.br) left irc: Quit: Leaving | 22:57 | |
| antocuni (~antocuni@host86-122-dynamic.6-79-r.retail.telecomitalia.it) joined #pypy. | 23:02 | |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 23:05 | |
| JaRoel|4d (~jaroel|4d@2001:980:7133:1:6c40:286b:dd77:2045) left irc: Remote host closed the connection | 23:10 | |
| stakkars__ (~tismer@p5DC4696F.dip0.t-ipconnect.de) left irc: Quit: schnarch | 23:10 | |
| JaRoel|4d (~jaroel|4d@2001:980:7133:1:6c40:286b:dd77:2045) joined #pypy. | 23:10 | |
| JaRoel|4d (~jaroel|4d@2001:980:7133:1:6c40:286b:dd77:2045) left irc: Ping timeout: 244 seconds | 23:15 | |
| mattip (~chatzilla@bzq-109-64-118-235.red.bezeqint.net) joined #pypy. | 23:16 | |
| mattip | amaury: ping | 23:17 |
| mattip | amaury_: ping | 23:18 |
| amaury_ | pong | 23:18 |
| mattip | Hi. I saw you commented on wxPython-dev about recreating your patch | 23:19 |
| amaury_ | :-) | 23:19 |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 23:19 | |
| mattip | I applied it to the wtk files, built it and tried to run, but something went wrong | 23:19 |
| amaury_ | traceback, or gdb stack? | 23:20 |
| whitelynx (~whitelynx@li117-47.members.linode.com) left irc: Quit: Ex-Chat | 23:20 | |
| amaury_ | ah, with a recent pypy you'll get weird errors | 23:20 |
| amaury_ | like: a wx.Window is not a wx.EvtHandler | 23:21 |
| mattip | In short, _PyObject_GetDictPtr gave an empty pointer | 23:21 |
| amaury_ | mattip: that's expected in pypy | 23:21 |
| mattip | Where under cPython it gave a proper pointer to a dictionary. | 23:21 |
| amaury_ | yes, it cannot be fixed | 23:21 |
| mattip | Hmm. | 23:22 |
| amaury_ | pypy objects are not PyObjects | 23:22 |
| amaury_ | so how can I give the address of the pointer to x.__dict__? | 23:22 |
| Shinita (~John@osbk-4db06ce4.pool.mediaWays.net) left irc: Ping timeout: 240 seconds | 23:23 | |
| amaury_ | at this point, x.__dict__ is still a pypy reference, | 23:23 |
| mattip | Don't the swig wrappers need this to work? | 23:23 |
| amaury_ | no | 23:23 |
| amaury_ | in every swig generated file there is #define SWIG_PYTHON_SLOW_GETSET_THIS | 23:24 |
| amaury_ | by default this is commented out | 23:24 |
| amaury_ | but pypy needs it | 23:24 |
| mattip | Ahah. | 23:24 |
| amaury_ | mattip: didn't you apply the patch? | 23:24 |
| amaury_ | it contains a similar fix | 23:25 |
| mattip | I guess I missed some of the chunks, I had to do it by hand. | 23:25 |
| amaury_ | I did the same | 23:25 |
| mattip | I will take a closer look. | 23:25 |
| amaury_ | enabling SWIG_PYTHON_SLOW_GETSET_THIS is easier | 23:25 |
| mattip | OK, thnks. | 23:26 |
| amaury_ | but the current pypy is a bit broken | 23:27 |
| amaury_ | the typesystems of different source files are not correctly shared | 23:27 |
| amaury_ | and you'll get type errors like: self is not of the expected type | 23:28 |
| mattip | Well, it's a start anyway... | 23:29 |
| maxyz (~maxy@186.23.74.228) left irc: Ping timeout: 276 seconds | 23:30 | |
| maxyz (~maxy@186.23.74.228) joined #pypy. | 23:33 | |
| mattip | amaury_: It seems that the last half of the patch is already in the sources. | 23:37 |
| amaury_ | possible | 23:37 |
| amaury_ | not entirely though | 23:38 |
| amaury_ | ah bitbucket is *very slow* here | 23:38 |
| kenaan | 12amauryfa default 11c074503990bb 15/pypy/module/cpyext/: cpyext: turn some parameters into const char* | 23:40 |
| kenaan | 12amauryfa default 1105ed38d2c537 15/pypy/module/cpyext/modsupport.py: Fix usages of Py_InitModule outside of the module init() function. SWIG for example uses it to share the type sy... | 23:41 |
| amaury_ | mattip: fixed | 23:41 |
| mattip | amaury_: cool. Hopefullly it will translate and tomorow I can give it a spin. | 23:42 |
| amaury_ | well, it did translate here :-) | 23:43 |
| amaury_ | mattip: I also fixed a couple of reference counting issues in wxPython | 23:44 |
| mattip | amaury_: beyond the ones in the patch? | 23:46 |
| apanda (~apanda@dhcp-45-162.EECS.Berkeley.EDU) joined #pypy. | 23:46 | |
| amaury_ | mattip: yes | 23:46 |
| amaury_ | for example, in helpers.cpp, in __wxPySetDictionary | 23:47 |
| amaury_ | it would be good to Py_INCREF(wxPython_dict) | 23:47 |
| amaury_ | otherwise this stores a borrowed reference in a global variable | 23:48 |
| amaury_ | i.e. cpyext will delete it as soon as the function returns | 23:48 |
| mattip | Right. | 23:49 |
| amaury_ | (more exactly: when the args tuple is reclaimed) | 23:49 |
| bbot2 | 4Finished ??: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/138 | 23:50 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/139 | 23:50 |
| Alex_Gaynor | amaury_: perhaps you should speak with dmalcolm about getting some detectors into his GCC thingy for common refcounting issues | 23:50 |
| amaury_ | Note that it's never an issue with CPython | 23:51 |
| Alex_Gaynor | amaury_: so is it a bug on our end? | 23:54 |
| amaury_ | not really | 23:54 |
| amaury_ | I'll try to explain | 23:54 |
| amaury_ | suppose you create a new object: obj = PyString_FromString("hello") | 23:55 |
| amaury_ | its refcount is 1 | 23:55 |
| amaury_ | now put it into a tuple: t = PyTuple_New(), PyTuple_SetItem(t, 0, obj); | 23:55 |
| amaury_ | PyTuple_SetItem *steals* the reference | 23:56 |
| Alex_Gaynor | lowering the effective refcount to 0? | 23:56 |
| amaury_ | no | 23:56 |
| amaury_ | not with CPython | 23:56 |
| amaury_ | because the tuple has a reference to the object | 23:56 |
| amaury_ | but with cpyext, | 23:56 |
| amaury_ | the tuple is a pypy object | 23:56 |
| amaury_ | which stores pypy objects | 23:57 |
| amaury_ | so refcount falls to zero | 23:57 |
| amaury_ | now, if the next lines happens to use obj again | 23:57 |
| amaury_ | it will work with CPython | 23:57 |
| amaury_ | and crash pypy | 23:57 |
| Alex_Gaynor | hmm | 23:58 |
| amaury_ | strictly speaking, pypy is "right" to crash | 23:58 |
| Alex_Gaynor | good enough for me | 23:58 |
| amaury_ | because obj does not own the reference anymore | 23:58 |
| amaury_ | ownership was transferred to the tuple | 23:59 |
| Alex_Gaynor | sounds like it's fair for dave's tool to complain them | 23:59 |
| Alex_Gaynor | then* | 23:59 |
| amaury_ | it's tricky sometimes | 23:59 |
| --- Tue Nov 15 2011 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!