| hazmat (~hazmat@204-232-198-49.static.cloud-ips.com) joined #pypy. | 00:10 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/741 | 00:10 |
|---|---|---|
| jterrace (~jterrace@pu116187.student.Princeton.EDU) left irc: Quit: Leaving. | 00:19 | |
| derdon (~derdon@p5DE89B26.dip.t-dialin.net) left irc: Remote host closed the connection | 00:33 | |
| overminder (~overminde@wf113-028.ust.hk) joined #pypy. | 00:37 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/167 | 01:07 |
| xiaochen (xiaochen@nat/redhat/x-ygytqdeupiljirth) joined #pypy. | 01:27 | |
| overminder (~overminde@wf113-028.ust.hk) left irc: Ping timeout: 240 seconds | 01:28 | |
| Nick change: Gulaway -> Gulopine | 01:29 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Ping timeout: 252 seconds | 01:30 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) joined #pypy. | 01:34 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) left irc: Changing host | 01:34 | |
| jimbaker (~jbaker@canonical/jimbaker) joined #pypy. | 01:34 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1885 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/627 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/302 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/640 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1157 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/824 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1466 | 01:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/747 | 01:35 |
| mfoord (~anonymous@python/psf/voidspace) left irc: Quit: mfoord | 01:46 | |
| pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) left irc: Ping timeout: 255 seconds | 01:47 | |
| overminder (~overminde@143.89.167.3) joined #pypy. | 01:49 | |
| pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) joined #pypy. | 01:51 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Read error: Connection reset by peer | 02:20 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 02:21 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Ping timeout: 244 seconds | 02:25 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 02:26 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Remote host closed the connection | 02:26 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy. | 02:28 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 02:32 | |
| Shanita (~John@osbk-4db15917.pool.mediaWays.net) left irc: Ping timeout: 240 seconds | 02:40 | |
| Moku (~John@osbk-4db1409b.pool.mediaWays.net) joined #pypy. | 02:40 | |
| Nick change: Moku -> Guest66857 | 02:40 | |
| jacob22 (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 252 seconds | 02:43 | |
| Action: mikefc starts a .au pypy chapter so he'll have someone to banter with | 02:44 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/975 | 02:50 |
| thinker` (~thinker@218-167-129-150.dynamic.hinet.net) joined #pypy. | 03:00 | |
| Nick change: thinker` -> Sinquer | 03:02 | |
| _pyre_ (~stian@137.241.189.109.customer.cdi.no) joined #pypy. | 03:04 | |
| _pyre_ | No improvement in the speed for the tuple specialization :( | 03:05 |
| _pyre_ | nice improvement in nbody tho | 03:07 |
| harrison (~quassel@69.209.222.34) left irc: Ping timeout: 240 seconds | 03:08 | |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 03:10 | |
| _pyre_1 (~stian@235.183.9.46.customer.cdi.no) joined #pypy. | 03:11 | |
| _pyre_ (~stian@137.241.189.109.customer.cdi.no) left irc: Ping timeout: 240 seconds | 03:13 | |
| _pyre_ (~stian@137.241.189.109.customer.cdi.no) joined #pypy. | 03:13 | |
| _pyre_1 (~stian@235.183.9.46.customer.cdi.no) left irc: Ping timeout: 240 seconds | 03:16 | |
| nedbat (~nedbat@python/psf/nedbat) left irc: Quit: Leaving | 03:17 | |
| Sinquer (~thinker@218-167-129-150.dynamic.hinet.net) left irc: Remote host closed the connection | 03:17 | |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) joined #pypy. | 03:18 | |
| mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 252 seconds | 03:20 | |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 03:27 | |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 03:31 | |
| nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 255 seconds | 03:34 | |
| anish_ (anish@nat/redhat/x-blrujkfejymsmnxl) joined #pypy. | 03:39 | |
| thinker` (~thinker@218-167-129-150.dynamic.hinet.net) joined #pypy. | 03:44 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1885 | 03:51 |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Ping timeout: 252 seconds | 03:52 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 03:53 | |
| nirbheek | Hey folks, what's the pypy hg repo size? | 03:53 |
| Action: nirbheek is on a bad connection, and is wondering whether to clone it | 03:54 | |
| _pyre_ (~stian@137.241.189.109.customer.cdi.no) left irc: Quit: Leaving. | 03:55 | |
| Garen | nirbheek: On disk or over the wire size? | 03:55 |
| nirbheek | Garen, over the wire | 03:55 |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 03:56 | |
| Garen | nirbheek: I'd guess it's around 100-150MB over the wire and 350-400 on disk | 03:57 |
| nirbheek | Garen, okay, I guess I can wait for the dl to complete, thanks for the info! :D | 03:57 |
| Garen | nirbheek: What kind of connection do you have? | 03:57 |
| nirbheek | Garen, 256kbps | 03:58 |
| cheater (~cheater@ip-80-226-24-12.vodafone-net.de) left irc: Ping timeout: 248 seconds | 03:58 | |
| nirbheek | Garen, I don't use hg much, can I resume the clone if I interrupt it? | 03:59 |
| nirbheek | or will it restart from the beginning ala git | 03:59 |
| Garen | no resuming :) | 04:00 |
| nirbheek | darnit | 04:00 |
| Garen | it compresses better that way I'm sure | 04:00 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Quit: brb | 04:05 | |
| Nick change: Gulopine -> Gulaway | 04:09 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 04:10 | |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) joined #pypy. | 04:24 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Ping timeout: 252 seconds | 04:27 | |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) left irc: Client Quit | 04:28 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 04:36 | |
| gutworth (~benjamin@128.36.84.91) left irc: Ping timeout: 252 seconds | 04:43 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1466 | 04:54 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/824 | 04:55 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/640 | 04:59 |
| Nick change: Gulaway -> Gulopine | 05:01 | |
| mat^2 (~mathias@212.130.113.35) joined #pypy. | 05:16 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1157 | 05:17 |
| brutal_c1aos (~chaos@francium.electricgrey.com) left irc: Quit: leaving | 05:18 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Quit: Konversation terminated! | 05:19 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 05:20 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/627 | 05:29 |
| xiaochen (xiaochen@nat/redhat/x-ygytqdeupiljirth) left irc: Quit: leaving | 05:32 | |
| xiaochen (xiaochen@nat/redhat/x-qartjkdpdbseyokp) joined #pypy. | 05:32 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Quit: Konversation terminated! | 05:37 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 05:37 | |
| xiaochen (xiaochen@nat/redhat/x-qartjkdpdbseyokp) left #pypy. | 05:46 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Ping timeout: 276 seconds | 05:47 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 05:48 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/302 | 05:49 |
| overminder (~overminde@143.89.167.3) left irc: Read error: Connection reset by peer | 05:58 | |
| overminder (~overminde@143.89.167.3) joined #pypy. | 05:59 | |
| Nick change: thinker` -> Sinker | 06:08 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Quit: Konversation terminated! | 06:12 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) joined #pypy. | 06:26 | |
| anish_ (anish@nat/redhat/x-blrujkfejymsmnxl) left irc: Ping timeout: 252 seconds | 06:28 | |
| Nick change: Gulopine -> Gulaway | 06:28 | |
| Ademan (~dan@adsl-71-141-224-79.dsl.snfc21.pacbell.net) left irc: Quit: leaving | 06:28 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 240 seconds | 06:28 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/747 | 06:28 |
| nirbheek (~nirbheek@117.197.151.239) joined #pypy. | 06:44 | |
| nirbheek (~nirbheek@117.197.151.239) left irc: Changing host | 06:45 | |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 06:45 | |
| anish_ (anish@nat/redhat/x-apcuadzpqvlzytzw) joined #pypy. | 06:45 | |
| asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer__ | 06:45 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy. | 06:51 | |
| __main__ (~main@adsl-99-173-15-158.dsl.pltn13.sbcglobal.net) left irc: Read error: Connection reset by peer | 06:51 | |
| raydeo (~raydeo@merickel.org) left irc: Quit: Coyote finally caught me | 06:54 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 244 seconds | 06:56 | |
| Taggnostr (~quassel@dyn57-215.yok.fi) joined #pypy. | 07:00 | |
| __main__ (~main@adsl-99-173-15-158.dsl.pltn13.sbcglobal.net) joined #pypy. | 07:00 | |
| apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) left irc: Remote host closed the connection | 07:10 | |
| apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) joined #pypy. | 07:10 | |
| raydeo (~raydeo@merickel.org) joined #pypy. | 07:15 | |
| Sinker (~thinker@218-167-129-150.dynamic.hinet.net) left irc: Ping timeout: 252 seconds | 07:18 | |
| thinker` (~thinker@218-167-101-153.dynamic.hinet.net) joined #pypy. | 07:28 | |
| espes (~espes@123-243-221-248.tpgi.com.au) left irc: Read error: Connection reset by peer | 07:28 | |
| Nick change: thinker` -> Sinker | 07:29 | |
| espes (~espes@123-243-221-248.tpgi.com.au) joined #pypy. | 07:31 | |
| stakkars_ (~tismer@p5DDB72FD.dip.t-dialin.net) left irc: Quit: schnarch | 07:31 | |
| espes (~espes@123-243-221-248.tpgi.com.au) left irc: Read error: Connection reset by peer | 07:39 | |
| kenaan | 12hakanardo jit-targets 11073665c57a62 15/pypy/module/pypyjit/test_pypy_c/: Retracing does not fail in the same set of situations as before. In test_getattr_promote we previously got ... | 07:39 |
| kenaan | 12hakanardo jit-targets 11d3d7d07ab453 15/pypy/: hg merge default | 07:40 |
| kenaan | 12hakanardo jit-targets 11cccccb4a9c72 15/: closing to be merged branch | 07:40 |
| kenaan | 12hakanardo default 111a9a44331010 15/pypy/: Merge jit-targets. It introduces an label ResOperation which can be used as a target for jump operations. Curre... | 07:40 |
| espes (~espes@123-243-221-248.tpgi.com.au) joined #pypy. | 07:40 | |
| overminder (~overminde@143.89.167.3) left irc: Ping timeout: 248 seconds | 07:42 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 07:42 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection | 07:48 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 07:50 | |
| fijal (~fijal@197.172.190.65) joined #pypy. | 07:53 | |
| fijal (~fijal@197.172.190.65) left irc: Read error: Connection reset by peer | 07:57 | |
| fijal (~fijal@41.53.87.204) joined #pypy. | 07:58 | |
| arigato (~arigo@89.204.155.202) joined #pypy. | 08:00 | |
| mvt (~mvantelli@87.213.45.85) joined #pypy. | 08:00 | |
| anish_ (anish@nat/redhat/x-apcuadzpqvlzytzw) left irc: Remote host closed the connection | 08:06 | |
| fijal | arigato: morning | 08:08 |
| fijal | http://pyppet.blogspot.com/2011/12/rpythonic-044.html | 08:09 |
| fijal | have you seen this? | 08:09 |
| fijal | the project name is really obscure | 08:09 |
| arigato | hi | 08:10 |
| squiddy (~squiddy@g224196119.adsl.alicedsl.de) joined #pypy. | 08:13 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 08:13 | |
| anish_ (anish@nat/redhat/x-obhcbhtmvbfwdlom) joined #pypy. | 08:14 | |
| jacob22 (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 08:16 | |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 08:20 | |
| kenaan | 12arigo jit-simplify-backendintf 11793032bf57e5 15/pypy/jit/backend/x86/assembler.py: Fix for StackLocs with a negative 'position' (corresponding to a positive offset from %ebp). | 08:37 |
| kenaan | 12arigo jit-simplify-backendintf 117ae9b2f7c938 15/pypy/jit/backend/x86/regalloc.py: 64-bit support. | 08:37 |
| kenaan | 12arigo jit-simplify-backendintf 113a5a45bf3039 15/pypy/jit/backend/x86/assembler.py: Fix call_assembler. | 08:37 |
| kenaan | 12arigo jit-simplify-backendintf 11dbe437dbf368 15/pypy/jit/backend/x86/: redirect_call_assembler() fix. | 08:37 |
| kenaan | 12arigo jit-simplify-backendintf 11b9daf080154c 15/pypy/jit/backend/: Fixes. | 08:37 |
| arigato | fijal: ah | 08:38 |
| fijal | arigato: ah? | 08:39 |
| arigato | rpythonic | 08:39 |
| fijal | it's a very confusing name | 08:41 |
| fijal | but if it really works the way it's advertised, it's great | 08:41 |
| fijal | (I doubt however, ctypes are tricky) | 08:42 |
| arigato | not that horribly tricky | 08:42 |
| fijal (~fijal@41.53.87.204) left irc: Read error: Connection reset by peer | 08:42 | |
| fijal (~fijal@41.54.214.136) joined #pypy. | 08:43 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy. | 08:53 | |
| antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) joined #pypy. | 09:00 | |
| kenaan | 12arigo jit-simplify-backendintf 118564866c0e68 15/pypy/jit/backend/x86/test/test_assembler.py: Fix test | 09:03 |
| kenaan | 12arigo jit-simplify-backendintf 1173f9a638b3f9 15/pypy/jit/backend/test/calling_convention_test.py: Fix test | 09:03 |
| kenaan | 12arigo jit-simplify-backendintf 1170bc31a4262f 15/pypy/jit/backend/: Fix tests | 09:03 |
| kenaan | 12arigo jit-simplify-backendintf 11a7c8b3b608d5 15/pypy/jit/backend/x86/test/test_recompilation.py: Fix test | 09:03 |
| kenaan | 12arigo jit-simplify-backendintf 11bdc61f3c13cb 15/pypy/jit/: Yay. More code is killed. Killing this was actually the goal of the branch, somehow. | 09:13 |
| G2P (~G2P@fw-asn1.ornis.com) joined #pypy. | 09:15 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Ping timeout: 268 seconds | 09:17 | |
| ojii (~ojii@40-34.60-188.cust.bluewin.ch) joined #pypy. | 09:25 | |
| JaRoel (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 09:27 | |
| fijal | arigato: are you doing work based on jit-targets? | 09:27 |
| arigato | fijal: yes | 09:28 |
| fijal | cool stuff | 09:29 |
| Action: fijal does not feel very productive | 09:29 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 240 seconds | 09:30 | |
| Sinker (~thinker@218-167-101-153.dynamic.hinet.net) left irc: Remote host closed the connection | 09:38 | |
| mattip (c1a9466d@gateway/web/freenode/ip.193.169.70.109) joined #pypy. | 09:49 | |
| mattip | fijal: hi | 09:51 |
| kkris (~kris@80-123-36-219.adsl.highway.telekom.at) joined #pypy. | 09:53 | |
| mattip (c1a9466d@gateway/web/freenode/ip.193.169.70.109) left irc: Quit: Page closed | 09:55 | |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy. | 09:56 | |
| sunoano (~sa@mail.heartinternet.co.uk) joined #pypy. | 10:17 | |
| sunoano (~sa@mail.heartinternet.co.uk) left irc: Changing host | 10:17 | |
| sunoano (~sa@unaffiliated/sunoano) joined #pypy. | 10:17 | |
| mat^2 (~mathias@212.130.113.35) left irc: | 10:17 | |
| mwhudson (~mwh@linaro/mwhudson) left irc: Ping timeout: 245 seconds | 10:19 | |
| anish_ (anish@nat/redhat/x-obhcbhtmvbfwdlom) left irc: Ping timeout: 252 seconds | 10:22 | |
| kenaan | 12arigo jit-simplify-backendintf 115960a81ae802 15/pypy/jit/metainterp/test/: Fix tests | 10:28 |
| kenaan | 12arigo jit-simplify-backendintf 11cd85bd047d43 15/pypy/jit/backend/llgraph/: Add an extra assertion in the llgraph backend. | 10:28 |
| kenaan | 12arigo jit-simplify-backendintf 11bcdbd1bc9618 15/pypy/jit/backend/test/test_random.py: Fix | 10:28 |
| kenaan | 12arigo jit-simplify-backendintf 116e71614fa5d3 15/pypy/jit/metainterp/history.py: Expand the comment. | 10:28 |
| kenaan | 12arigo jit-simplify-backendintf 11ee8cf5779f0c 15/pypy/jit/backend/: Fix test_random. | 10:29 |
| fijal (~fijal@41.54.214.136) left irc: Read error: Connection reset by peer | 10:35 | |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) joined #pypy. | 10:36 | |
| kenaan | 12hager ppc-jit-backend 11f4618a1f60ba 15/pypy/jit/backend/ppc/ppcgen/ppc_assembler.py: fix wrong computation of stack offset in gen_direct_bootstrap_code | 10:41 |
| kenaan | 12hager ppc-jit-backend 119130b26b5bad 15/pypy/jit/backend/ppc/ppcgen/ppc_assembler.py: remove comments and debug stuff | 10:41 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Quit: Leaving | 10:47 | |
| CIA-13 | 03mikefc 07roundup * 10#959/numpypy array operation improperly jitted away: (log message trimmed) | 10:49 |
| CIA-13 | simplified the example a bit more. | 10:49 |
| CIA-13 | ========================= | 10:49 |
| CIA-13 | import numpypy as numpy | 10:49 |
| CIA-13 | import random * 14https://bugs.pypy.org/issue959 | 10:49 |
| CIA-13 | 03mikefc 07roundup * 10#959/numpypy array operation improperly jitted away: | 10:51 |
| CIA-13 | My output. JIT must kick in at 2079. Everything is TRUE up to there. | 10:51 |
| CIA-13 | ================================ | 10:51 |
| CIA-13 | 2075 [2075] [True] | 10:51 |
| CIA-13 | 2076 [2076] [True ... * 14https://bugs.pypy.org/issue959 | 10:51 |
| mikefc | anyone tried ctopy? | 10:56 |
| arigato (~arigo@89.204.155.202) left irc: Quit: See you | 10:56 | |
| kkris (~kris@80-123-36-219.adsl.highway.telekom.at) left irc: Quit: Leaving. | 11:15 | |
| azanella (~azanella@189.6.80.207) joined #pypy. | 11:24 | |
| mkerrin (~mkerrin@wg1-nat.fwgal01.deri.ie) joined #pypy. | 12:01 | |
| lmoura (lmoura@gateway/indt/session) left irc: Ping timeout: 240 seconds | 12:01 | |
| lmoura (~lmoura@187.113.98.156) joined #pypy. | 12:14 | |
| ericholscher (~eric@173.203.223.63) left irc: Ping timeout: 240 seconds | 12:15 | |
| pjenvey (~pjenvey@underboss.org) left irc: Read error: Operation timed out | 12:16 | |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 12:18 | |
| pjenvey (~pjenvey@underboss.org) joined #pypy. | 12:20 | |
| ericholscher (~eric@173.203.223.63) joined #pypy. | 12:22 | |
| derdon (~derdon@pD9E1CB1E.dip.t-dialin.net) joined #pypy. | 12:33 | |
| gutworth (~benjamin@dhcp-128-36-84-91.central.yale.edu) joined #pypy. | 12:54 | |
| linq (~ident@24-246-25-39.cable.teksavvy.com) left irc: Ping timeout: 248 seconds | 12:56 | |
| arigato (~arigo@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 13:03 | |
| kenaan | 12arigo jit-simplify-backendintf 117e98ea1a1c5e 15/pypy/jit/backend/x86/assembler.py: Fix. This is also a bug on trunk, actually, but because the stack checking is only done after CAL... | 13:04 |
| khs (~khs@2001:700:300:2120:725a:b6ff:fee5:a44) joined #pypy. | 13:04 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jitonly-own-linux-x86-32/builds/45 [12arigo, jit-simplify-backendintf] | 13:04 |
| xenru (~Adium@matrix-217-146-248-10.matrix.farlep.net) joined #pypy. | 13:16 | |
| arigato | "in real-address mode, if the ESP register is 1 when the PUSH instruction is executed, ..." | 13:19 |
| voidspace_ (~voidspace@87-194-212-65.bethere.co.uk) joined #pypy. | 13:19 | |
| voidspace_ (~voidspace@87-194-212-65.bethere.co.uk) left irc: Changing host | 13:19 | |
| voidspace_ (~voidspace@python/psf/voidspace) joined #pypy. | 13:19 | |
| voidspace (~voidspace@python/psf/voidspace) left irc: Ping timeout: 240 seconds | 13:22 | |
| Nick change: voidspace_ -> voidspace | 13:22 | |
| Insectoid (~Q@c-98-230-8-37.hsd1.fl.comcast.net) left irc: Ping timeout: 252 seconds | 13:32 | |
| e1ven (~E1ven@SQ7/ProjectLead/E1ven) left irc: Quit: Computer has gone to sleep. | 13:34 | |
| nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 276 seconds | 13:35 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/jitonly-own-linux-x86-32/builds/45 [12arigo, jit-simplify-backendintf] | 13:35 |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 13:38 | |
| jnoller (~jnoller@vpn.nasuni.com) joined #pypy. | 13:47 | |
| exarkun | Does anyone want to help with s* buffer yet | 13:52 |
| Nick change: Gulaway -> Gulopine | 13:59 | |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy. | 14:03 | |
| derdon (~derdon@pD9E1CB1E.dip.t-dialin.net) left irc: Remote host closed the connection | 14:04 | |
| lmoura (~lmoura@187.113.98.156) left irc: Read error: Connection reset by peer | 14:09 | |
| kenaan | 12arigo jit-simplify-backendintf 11cc11be2ea8a9 15/pypy/jit/tool/: Fix test | 14:10 |
| kenaan | 12arigo jit-simplify-backendintf 11b43f4ac19a1b 15/pypy/jit/backend/llgraph/llimpl.py: Translation fix. | 14:10 |
| kenaan | 12arigo default 115ba9f567b515 15/pypy/rpython/: Tests and fix for ``for c in string'' when the string turns out to be always a single character. | 14:10 |
| kenaan | 12arigo jit-simplify-backendintf 11aff416c812ea 15/pypy/: hg merge default | 14:10 |
| lmoura (lmoura@gateway/indt/session) joined #pypy. | 14:26 | |
| Rhy0lite (dje@nat/ibm/x-iqvummkaxufizbro) joined #pypy. | 14:26 | |
| aboudreault (~alanb@osgeo/member/aboudreault) left irc: Quit: Leaving | 14:27 | |
| aboudreault (~alanb@osgeo/member/aboudreault) joined #pypy. | 14:30 | |
| harrison (~quassel@adsl-76-217-39-2.dsl.chcgil.sbcglobal.net) joined #pypy. | 14:30 | |
| harrison_ (~quassel@adsl-76-217-39-2.dsl.chcgil.sbcglobal.net) joined #pypy. | 14:34 | |
| whitelynx (~whitelynx@63.241.75.144) joined #pypy. | 14:34 | |
| harrison (~quassel@adsl-76-217-39-2.dsl.chcgil.sbcglobal.net) left irc: Ping timeout: 244 seconds | 14:34 | |
| fijal (~fijal@41.49.241.166) joined #pypy. | 14:35 | |
| Nick change: Gulopine -> Gulaway | 14:40 | |
| bfirsh (u1308@gateway/web/irccloud.com/x-vertixxxctpxdezm) left irc: Remote host closed the connection | 14:58 | |
| hongminhee (u295@gateway/web/irccloud.com/x-juuehcrfxwjrnmtl) left irc: Remote host closed the connection | 14:58 | |
| oal (u4126@gateway/web/irccloud.com/x-nmdnucfmlpdbesxp) left irc: Read error: Connection reset by peer | 14:58 | |
| Alex_Gaynor (u1246@gateway/web/irccloud.com/x-rgtzudjklhivhojf) left irc: Remote host closed the connection | 14:58 | |
| oal (u4126@gateway/web/irccloud.com/x-mrrmdxqibenxbofl) joined #pypy. | 14:58 | |
| bfirsh (u1308@gateway/web/irccloud.com/x-bxagoggwardmynye) joined #pypy. | 15:00 | |
| fijal | exarkun: yes | 15:02 |
| Action: arigato debugs libffi on ppc32 | 15:07 | |
| exarkun | fijal: Okay. Do you want to hear what trivial, probably misleading, discoveries I did manage to make before I was defeated last week? | 15:09 |
| exarkun | (do you want to help now? later?) | 15:09 |
| hongminhee (u295@gateway/web/irccloud.com/x-xwvmgozxcovfnhfl) joined #pypy. | 15:10 | |
| fijal | I can help now | 15:11 |
| fijal | and yes, I want to hear where did you get | 15:11 |
| fijal | would you mind checking whatever you have to a branch? | 15:11 |
| fijal | I don't care how broken it is | 15:11 |
| exarkun | I'm happy to do that, yes, but I don't remember how hg branching works, so I'll need some hints | 15:11 |
| fijal | you have a copy with modification right? | 15:11 |
| exarkun | yep | 15:11 |
| fijal | you do hg branch <some name> && hg ci -m 'some message' && hg push | 15:12 |
| fijal | as far as I remember | 15:12 |
| kenaan | 12exarkun pyarg-parsetuple-s-star-buffer 110561684806ca 15/pypy/module/cpyext/: fiddle around a bit - add a unit test that might exercise the desired functionality; play ... | 15:13 |
| fijal | exarkun: I'm trying to write a blogpost why am I unemployable and a conference abstract submission for a web scaling conference | 15:14 |
| fijal | seems like parsing s* beats that | 15:14 |
| exarkun | Heh :) Glad to provide a useful distraction. | 15:14 |
| exarkun | As far as I managed to follow the behavior, a brand new type with zero defined behavior and a name "buffer" is synthesized when the unit test passes a buffer instance to the C function it defines. | 15:15 |
| exarkun | in gdb I can _see_ the other buffer type struct, and it is full of useful stuff | 15:16 |
| fijal | does not sound too good | 15:16 |
| exarkun | but it doesn't seem to get used, and I don't know where the code that would decide to use it or not lives | 15:16 |
| fijal | exarkun: any reason why you checked in import pdb;pdb.set_trace()? | 15:16 |
| fijal | or can I just remove it? | 15:16 |
| exarkun | feel free to do anything to it | 15:17 |
| fijal | ok | 15:17 |
| exarkun | I was breaking there so I could step through the Python side of the function call logic | 15:17 |
| fijal | right | 15:18 |
| fijal | have you seen pdb/gdb thing? | 15:18 |
| exarkun | I don't think so | 15:19 |
| fijal | it's a set of gdb scripts | 15:19 |
| exarkun | But I couldn't even understand the Python code by itself, so I don't think I'm advanced enough to need that tool yet. | 15:19 |
| fijal | ok | 15:19 |
| fijal | ImportError: No module named buffer | 15:20 |
| fijal | exarkun: you forgot to hg add something | 15:20 |
| exarkun | hm | 15:20 |
| fijal | import pypy.module.cpyext.buffer | 15:20 |
| fijal | I don't have it | 15:20 |
| fijal | do you? | 15:20 |
| exarkun | yes. adding now. | 15:21 |
| kenaan | 12exarkun pyarg-parsetuple-s-star-buffer 11421fe5b137ea 15/pypy/module/cpyext/buffer.py: Add trivial buffer implementation module | 15:21 |
| fijal | exarkun: https://bitbucket.org/antocuni/pygdb2/overview | 15:22 |
| Action: exarkun takes a look | 15:22 | |
| e1ven (~E1ven@SQ7/ProjectLead/E1ven) joined #pypy. | 15:23 | |
| exarkun | The more I think about that branch, the more I think I must just be overlooking some necessary step to map a particular python type to a particular cpyext type. | 15:23 |
| fijal | I think there is a missing call to make_typedescr | 15:26 |
| fijal | look at init_stringobject in stringobject.py | 15:26 |
| fijal | you need something like this for buffer | 15:27 |
| fijal | exarkun: essentially you're missing how to move data python <-> cpyext | 15:28 |
| fijal | or how to map types precisely | 15:28 |
| fijal | I *think* | 15:28 |
| exarkun | api.py does something | 15:30 |
| exarkun | There's an init_bufferobject | 15:30 |
| exarkun | Hm, but it doesn't make_typedescr though, it does something else. | 15:31 |
| fijal | yes | 15:31 |
| fijal | it maps interface to C functions | 15:31 |
| fijal | but you still need to be able to somehow map that to pypy types | 15:31 |
| fijal | there is a type called 'buffer' that has behaviors | 15:32 |
| fijal | but it's not connected to pypy buffer type at all | 15:32 |
| bivab_ (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 15:34 | |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Read error: Connection reset by peer | 15:34 | |
| Nick change: bivab_ -> bivab | 15:34 | |
| kenaan | 12arigo ppc-jit-backend 117ef8482ed95b 15/pypy/jit/backend/ppc/ppcgen/ppc_assembler.py: (sven, arigo) Fix a bug, and add a print that is useful for debugging with gdb. | 15:34 |
| berdario (~Tiibiidii@217.220.118.217) joined #pypy. | 15:35 | |
| jterrace (~jterrace@aegis.CS.Princeton.EDU) joined #pypy. | 15:45 | |
| fijal | require | 15:47 |
| fijal | this is how you write it? | 15:47 |
| exarkun | yes | 15:47 |
| kenaan | 12hager ppc-jit-backend 1155b2c881ec01 15/pypy/jit/backend/ppc/ppcgen/ppc_assembler.py: (bivab, hager): do stack pointer alignment | 15:48 |
| fijal | exarkun: do I make any sense? | 15:49 |
| exarkun | fijal: Abstractly. I'm trying to get some other things out of my way so I can focus on just this for a while. | 15:50 |
| exarkun | So I haven't really started trying yet. Maybe in 15 minutes. | 15:50 |
| fijal | ok | 15:50 |
| arigato | hakanardo: is jit-targets ready to merge? | 15:56 |
| arigato | ah | 15:57 |
| arigato | you merged today :-) | 15:57 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jitonly-own-linux-x86-32/builds/46 [12arigo, jit-simplify-backendintf] | 15:59 |
| fijal | anyone feel like reviewing a personal rant? | 16:01 |
| arigato | sure | 16:01 |
| Alex_Gaynor (u1246@gateway/web/irccloud.com/x-unwydkcmqavjigvx) joined #pypy. | 16:01 | |
| #pypy: mode change '+o Alex_Gaynor' by ChanServ!ChanServ@services. | 16:01 | |
| jterrace | is it your rant about python 3? | 16:01 |
| jterrace | i liked that one | 16:01 |
| fijal | no, that was mitsuhiko's | 16:01 |
| jterrace | oh | 16:01 |
| capncapacitor (~capncapac@72.1.196.82) joined #pypy. | 16:02 | |
| xenru (~Adium@matrix-217-146-248-10.matrix.farlep.net) left irc: Quit: Leaving. | 16:02 | |
| fijal | https://bitbucket.org/fijal/talk/src/default/blog/talent-shortage.rst | 16:02 |
| nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy. | 16:02 | |
| fijal | arigato: ^^^ | 16:02 |
| Action: arigato reads | 16:02 | |
| nirbheek | Hey guys, aren't the instructions for compiling pypy on this page: http://pypy.org/download.html  outdated? | 16:03 |
| fijal | arigato: feel free to say "it's a frustrated rant" | 16:03 |
| nirbheek | It mentions --stackless which is no longer present in 1.7 | 16:03 |
| Action: fijal cleans up | 16:03 | |
| fijal | nirbheek: thanks | 16:03 |
| jterrace | Alex_Gaynor: i tried adding pi, min, max, sum to numpy, but it causes one of the zjit tests to fail - any idea why? the errors dont make sense to me | 16:03 |
| capncapacitor | Has anyone had any issues/luck getting PyOpenSSL working with Pypy? 1.7 is locking up our entire Twisted process | 16:03 |
| nirbheek | fijal, that was fast, thanks :D | 16:04 |
| fijal | jterrace: paste them | 16:04 |
| jterrace | http://paste.pocoo.org/show/519799/ | 16:05 |
| exarkun | capncapacitor: I haven't seen "locking up" | 16:05 |
| exarkun | capncapacitor: There's at least one fairly serious issue that will cause an unhandled exception when trying to send some data, though. | 16:06 |
| kenaan | 12fijal pypy.org[extradoc] 11ba6be317c363 15/: mention the default build contains stackless, remove the stackless build instruction | 16:06 |
| kenaan | 12fijal pypy.org[extradoc] 1186a7edc8b73e 15/: merge | 16:06 |
| fijal | nirbheek: thanks! | 16:06 |
| fijal | antocuni: ping | 16:06 |
| exarkun | Maybe if you're looking at things from the perspective of a peer connected over a socket, it could look like a lock up. | 16:06 |
| antocuni | pong | 16:06 |
| fijal | antocuni: can you reply to the dude on mailing list about rawffi/_ffi? | 16:07 |
| fijal | nirbheek: should be fixed | 16:07 |
| fijal | jterrace: well, you broke stuff :0 | 16:07 |
| antocuni | yes, I'll do it | 16:07 |
| capncapacitor | exarkun, Trying to send an email using SMTP/SSL results in the entire process locking up. Twisted stops responding, can only kill -9 the process, not quite sure how to debug it. | 16:07 |
| fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy. | 16:07 | |
| jterrace | but i dont understand why - all i did was add some top-level, simple functions | 16:07 |
| exarkun | capncapacitor: Hm. gdb, I suppose. | 16:07 |
| exarkun | capncapacitor: Or https://bitbucket.org/antocuni/pygdb2/overview | 16:07 |
| jterrace | the diff: https://bitbucket.org/jterrace/pypy/changeset/d4d7f068865e | 16:08 |
| fijal | jterrace: diff? | 16:08 |
| nirbheek | fijal, is the information about stackless under "Other versions" still accurate? | 16:08 |
| nirbheek | http://pypy.org/download.html#other-versions | 16:08 |
| fijal | didn't I remove it? | 16:08 |
| nirbheek | refreshing the page still shows it, will it take some time to show up? | 16:09 |
| fijal | yop | 16:09 |
| fijal | no, I just had to hg pull | 16:09 |
| nirbheek | oh, I got confused because the option removal showed up immediately | 16:09 |
| nirbheek | ah, okay, it shows up now, sorry for the confusion :) | 16:09 |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Quit: bivab | 16:09 | |
| jterrace | fijal: there really is a talent shortage | 16:10 |
| fijal | jterrace: then I must be super-dumb | 16:10 |
| jterrace | most people don't have your constraints | 16:10 |
| arigato | fijal: excellent "rant" | 16:11 |
| jterrace | have you tried http://careers.stackoverflow.com/ ? | 16:11 |
| jterrace | quality there is usually better | 16:11 |
| fijal | jterrace: I interviewed with various companies | 16:11 |
| fijal | and got contacted by a few recruiters | 16:11 |
| fijal | all of them got blasted away by my list | 16:11 |
| jterrace | yeah | 16:11 |
| fijal | startups won't even bother replying to my mail, because H1B is a mess | 16:12 |
| jterrace | they dont understand what productivity actually is | 16:12 |
| fijal | hence I declare talent shortage a myth | 16:12 |
| fijal | arigato: good to go? | 16:12 |
| arigato | from me point of view, yes | 16:12 |
| jterrace | the H1B mess is one of the reasons there is a shortage | 16:12 |
| fijal | note if it's not clear - this is my personal blog not a pypy blog (obviously) | 16:12 |
| fijal | jterrace: my point is the need for H1B is the first problem | 16:12 |
| fijal | why not work remotely? | 16:12 |
| jterrace | because most companies don't understand what productivity really is | 16:13 |
| jterrace | they should read The Lean Startup | 16:13 |
| timotimo | can you expand H1B for me? | 16:13 |
| timotimo | Hire 1 Badass? | 16:14 |
| timotimo | Hour 1 Business? | 16:14 |
| fijal | jterrace: then they have no idea and their complaints should not be valid | 16:14 |
| timotimo | oh, it's some immigration thing | 16:14 |
| fijal | timotimo: an immigration visa | 16:14 |
| mvt (~mvantelli@87.213.45.85) left irc: Quit: This computer has gone to sleep | 16:15 | |
| fijal | jterrace: so how the talent shortage is a real thing? | 16:15 |
| antocuni | fijal: nice blog post :-) | 16:15 |
| jterrace | i mean it definitely varies by location, but there are more people hiring than there are qualified applicants | 16:15 |
| jterrace | silicon valley for example | 16:15 |
| fijal | jterrace: but they don't want to hire good people | 16:15 |
| fijal | there is no such place as silicon valley any more | 16:16 |
| fijal | there is just internet | 16:16 |
| jterrace | more companies need this: http://zachholman.com/posts/how-github-works-asynchronous/ | 16:17 |
| jterrace | but the old mantra of needing people to be in the office is still widespread unfortunately | 16:17 |
| fijal | well | 16:17 |
| fijal | but that only entrenches my point | 16:17 |
| fijal | there is no talent shortage, just incompetent management | 16:17 |
| Rhy0lite | :-) | 16:18 |
| jterrace | i think they are separate issues | 16:18 |
| fijal | are they? | 16:18 |
| jterrace | you're right that a lot of companies can't recognize talent | 16:18 |
| jterrace | but there still is a shortage of talent, even if you account for it | 16:18 |
| jterrace | there aren't many people like you | 16:18 |
| fijal | I know quite a few | 16:18 |
| jterrace | that's bias because of what you do :) | 16:18 |
| fijal | I don't have enough evidence to say that that there is a shortage of talent even if you account for it | 16:19 |
| fijal | and I doubt you do | 16:19 |
| fijal | and if you do I want references so I can read some more :) | 16:19 |
| Rhy0lite | fijal: the problem is it takes time and effort to determine if a person is talented / competent | 16:19 |
| Rhy0lite | and it's easier to use some trivial binning process | 16:20 |
| fijal | sure, but we have tools | 16:20 |
| Rhy0lite | than to actually understand the applicant | 16:20 |
| fijal | OS is a good tool | 16:20 |
| Rhy0lite | or "the company hired me and I'm doing well, so clearly what the company needs is more people exactly like me" | 16:21 |
| Rhy0lite | I definitely see the divide between interviewers who are able to talk to a person and determine if the person is smart and competent versus interviewers who want candidates to jump through hoops | 16:24 |
| mattip (~chatzilla@bzq-79-177-112-183.red.bezeqint.net) joined #pypy. | 16:24 | |
| fijal | mattip: hi | 16:24 |
| fijal | jterrace: I think the spotify recruiter was a tipping point for me | 16:24 |
| jterrace | did you ask him what their business model was and he started crying? | 16:25 |
| mattip | hi | 16:25 |
| fijal | no, I told him stockholm is not nearly as cool as cape town | 16:25 |
| capncapacitor | exarkun, Looks like the lockup is happening in OpenSSL's ssl_Connection_dealloc () | 16:25 |
| Rhy0lite | fijal: I have found personal contacts much more useful than recruiters | 16:26 |
| jterrace | agreed | 16:26 |
| fijal | Rhy0lite: yes, but the problem is our personal contacts tend to go to people who have no budget | 16:26 |
| fijal | that is my usual problem | 16:26 |
| xenru (~Adium@31.31.97.244) joined #pypy. | 16:26 | |
| exarkun | capncapacitor: Well, that won't be fun, then. There are no loops or anything there, so it's presumably a problem down in the memory management system, maybe due to memory having been corrupted already at some earlier point. | 16:27 |
| exarkun | Maybe someone else can point out some cpyext debugging tips that are better than bare gdb | 16:27 |
| capncapacitor | exarkun, here's the thread stack at the point of locking up: http://paste.pocoo.org/show/519807/ | 16:28 |
| xenru (~Adium@31.31.97.244) left irc: Client Quit | 16:28 | |
| stakkars_ (~tismer@i59F71C2E.versanet.de) joined #pypy. | 16:29 | |
| rguillebert | hi | 16:29 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jitonly-own-linux-x86-32/builds/46 [12arigo, jit-simplify-backendintf] | 16:29 |
| fijal | rguillebert: hi | 16:30 |
| exarkun | capncapacitor: Do you have a simple example that reproduces the problem? | 16:30 |
| jterrace | fijal: any idea on that zjit error? | 16:30 |
| fijal | jterrace: I asked for the diff | 16:30 |
| jterrace | oh i pasted it | 16:31 |
| fijal | and then I stopped caring :) | 16:31 |
| fijal | so I missed it | 16:31 |
| fijal | can you repaste? | 16:31 |
| capncapacitor | I'll come up with something | 16:31 |
| jterrace | https://bitbucket.org/jterrace/pypy/changeset/d4d7f068865e | 16:31 |
| mattip | jterrace: did the test work before you make your changes? I tried to check out your branch and run the test there and it failed for me with dlopen errors | 16:34 |
| jterrace | it worked before yes | 16:35 |
| jterrace | it's based off default, and it works there | 16:35 |
| fijal | jterrace: ok, I agree it should not break | 16:35 |
| fijal | jterrace: why are you not developing on the pypy repo btw? | 16:35 |
| jterrace | i dont have access | 16:35 |
| fijal (~fijal@41.49.241.166) left irc: Read error: Connection reset by peer | 16:36 | |
| jterrace | no wait, i think it's broken on default too | 16:36 |
| fijal (~fijal@197.169.119.41) joined #pypy. | 16:37 | |
| jterrace | fijal: i think test_zjit might be broken on default too | 16:37 |
| fijal | jterrace: you want some? | 16:37 |
| jterrace | sure, i can push to branches | 16:37 |
| fijal | username? jterrace? | 16:37 |
| jterrace | yeah | 16:37 |
| berdario_ (~Tiibiidii@217.220.118.217) joined #pypy. | 16:38 | |
| berdario (~Tiibiidii@217.220.118.217) left irc: Read error: Connection reset by peer | 16:39 | |
| mattip | fijal: Monday is my pypy evening. I would like to see you a) get a job, b) write the performance blog on the "laplace" iterations. Seeing as I can't really help with a), can I do anything for b) ? | 16:41 |
| fijal | jterrace: done | 16:41 |
| jterrace | thanks | 16:42 |
| fijal | mattip: I'm working on sharing iterators | 16:42 |
| jterrace | can someone confirm that test_zjit is failing on default? | 16:42 |
| mattip | Yes, the cache idea is cool. | 16:42 |
| arigato | fijal: if you're not too productive, maybe you can review jit-simplify-backendintf? | 16:42 |
| arigato | it passes all the jitonly tests | 16:43 |
| arigato | on 32-bit linux at least | 16:43 |
| fijal | ok | 16:43 |
| fijal | mattip: not sure how you can help | 16:43 |
| mattip | jterrace: it seems to fail for me too, I am also on 32 bit linux. | 16:43 |
| fijal | with that | 16:43 |
| jterrace | ok | 16:43 |
| fijal | arigato: looking at test_zjit, will look at that later | 16:45 |
| mattip | fijal: ok. | 16:45 |
| fijal | mattip: I don't want to post "we're >2x slower than C" if we don't have to | 16:46 |
| fijal | hum | 16:47 |
| fijal | arigato: jit-targets broke check_simple_loop | 16:47 |
| fijal | that' | 16:47 |
| fijal | s why test_zjit is failing | 16:47 |
| fijal | can you run test_zjit -k test_slice2? | 16:47 |
| apanda (~apanda@c-24-7-101-62.hsd1.ca.comcast.net) left irc: Remote host closed the connection | 16:48 | |
| dracman (~draco@212.255.38.116) joined #pypy. | 16:49 | |
| kenaan | 12jterrace numpy-pi-sum-min-max 11d4d7f068865e 15/pypy/module/micronumpy/: Added pi, sum, min, and max | 16:50 |
| dmalcolm (david@nat/redhat/x-fbwfkuafatpwdkyv) joined #pypy. | 16:50 | |
| jterrace | that branch should be fine to merge | 16:50 |
| fijal | jterrace: looks good to me, can you merge it? | 16:51 |
| jterrace | push it to default you mean? | 16:51 |
| fijal | yes | 16:51 |
| jterrace | sure | 16:51 |
| fijal | as in hg merge numpy-pi-sum-min-max | 16:51 |
| fijal | (assuming it passes all tests except test_zjit) | 16:52 |
| fijal | like, test_ztranslation | 16:52 |
| jterrace | yeah they do | 16:52 |
| jterrace | actually my default branch might not be clean | 16:52 |
| jterrace | im going to nuke it | 16:52 |
| fijal | do whatever fu is necessary | 16:52 |
| fijal | arigato: ? | 16:52 |
| G2P (~G2P@fw-asn1.ornis.com) left irc: Quit: Leaving. | 16:53 | |
| thinker (~thinker@123-194-52-90.dynamic.kbronet.com.tw) left irc: Ping timeout: 255 seconds | 16:54 | |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) joined #pypy. | 16:55 | |
| thinker (~thinker@123-194-52-90.dynamic.kbronet.com.tw) joined #pypy. | 16:56 | |
| fijal | arigato: can you make compile_started_vars more-obviously rpython? | 16:56 |
| kenaan | 12arigo jit-simplify-backendintf 113c385108e3da 15/pypy/jit/backend/x86/regalloc.py: Another simplification: _x86_arglocs can be just one list, instead of a tuple of two lists. | 16:57 |
| arigato | fijal: sorry ,back | 16:57 |
| capncapacitor | exarkun, Twisted's SSL echoserver and echoclient illustrate the problem. Lockup occurs only on the client | 16:57 |
| arigato | fijal: any code in llimpl.py is not RPython | 16:58 |
| fijal | except AttributeError: # when translated | 16:58 |
| fijal | then what does it mean? | 16:58 |
| fijal | I got confused by the comment | 16:58 |
| thrain42 (~durin@adium/durin42) joined #pypy. | 16:59 | |
| arigato | it's when "the world around" is translated | 16:59 |
| fijal | arigato: ok | 16:59 |
| fijal | arigato: can you look at test_zjit failure? | 16:59 |
| arigato | i.e. in metainterp/test/test_ztranslated.py (and in no other test) | 16:59 |
| `fox` (~fox@95.233.109.194) joined #pypy. | 16:59 | |
| arigato | yes | 16:59 |
| durin42 (~durin@adium/durin42) left irc: Ping timeout: 255 seconds | 17:00 | |
| kenaan | 12arigo jit-simplify-backendintf 11d56a7f15aeba 15/pypy/jit/backend/llgraph/llimpl.py: fix comment (thanks fijal) | 17:00 |
| arigato | fijal: about micronumpy: | 17:00 |
| arigato | it's great that there are contributions from various people, | 17:00 |
| arigato | but it also means it's a bit annoying e.g. for people that work on unrelated things | 17:00 |
| fijal | arigato: ok, I would like to be a bit nitpicky and say that random changes to regalloc strategies should not *really* be a part of this branch | 17:00 |
| fijal | right | 17:00 |
| arigato | they do "hg merge default" and sometimes pypy doesn't translate any more | 17:01 |
| fijal | that's why we have test_ztranslations these days :) | 17:01 |
| arigato | well, that's not enough | 17:01 |
| fijal | I promise to look how to solve stuff next time numpy causes breakage of translation | 17:01 |
| fijal | did it cause anything recently? | 17:01 |
| arigato | yesterday or two days ago, hakanardo got a crash of a whole translation, even though test_ztranslation worked | 17:02 |
| fijal | note that test_zjit failure is because check_simple_loop stopped working | 17:02 |
| fijal | ok, sorry | 17:02 |
| arigato | np | 17:02 |
| arigato | just saying, it would make it easier for you too | 17:02 |
| arigato | e.g. you wouldn't have to fix test_zjit absolutely now | 17:02 |
| papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep. | 17:03 | |
| fijal | you mean to move numpy development outside of default? | 17:03 |
| arigato | but now you do, because we don't want to be left with failures on default :-) | 17:03 |
| fijal | I'm fine with that | 17:03 |
| arigato | I don't know if it's necessary, but it may help, simply because a lot of people commit to it | 17:03 |
| fijal | I'm fine with dealing with problems | 17:03 |
| fijal | I don't want numpy to be a problem for everyone, but it's cool numpy comes in the nightlies | 17:04 |
| arigato | you could e.g. make one long-lived branch, and do merges in both directions when appropriate | 17:04 |
| exarkun | capncapacitor: Hm. Yea. I see a problem with that, though it happens earlier than the one you're describing. | 17:04 |
| exarkun | But I'm going to focus on one thing at a time (which I can finally start doing now) | 17:04 |
| thrain42 (~durin@adium/durin42) left irc: Read error: Operation timed out | 17:05 | |
| fijal | arigato: and also release contains it | 17:05 |
| fijal | I can think about some mid-ground scenarios | 17:05 |
| arigato | I'm not saying it should be disabled, or it should be done only in a branch | 17:05 |
| fijal | like we have a branch that got merged nightly to default when it translates or so | 17:05 |
| arigato | yes, something like that | 17:05 |
| __main__ (~main@adsl-99-173-15-158.dsl.pltn13.sbcglobal.net) left irc: Read error: Connection reset by peer | 17:06 | |
| arigato | either automatic or manual | 17:06 |
| fijal | I would vastly prefer automtic | 17:07 |
| fijal | automatic even | 17:07 |
| fijal | I'll look into that | 17:07 |
| fijal | once I clean up my todo list | 17:07 |
| arigato | :-) | 17:07 |
| arigato | the alternative was to have a "default-translates" branch | 17:08 |
| durin42 (~durin@adium/durin42) joined #pypy. | 17:08 | |
| arigato | it would contain the latest known-to-translate version of "default" | 17:08 |
| arigato | branch or tag or something | 17:08 |
| fijal | we had this idea for a while | 17:08 |
| arigato | test_zjit: looks like a mess, because the test doesn't fail if run alone | 17:08 |
| fijal | arigato: it totally does | 17:08 |
| fijal | test_zjit -k test_slice2 | 17:09 |
| fijal | did you run the correct one? | 17:09 |
| arigato | yes | 17:09 |
| fijal | it fails for me | 17:09 |
| fijal | on 64 bit | 17:09 |
| arigato | passes when run alone, on 32 bit :-( | 17:10 |
| arigato | on 64 bit too :-( | 17:10 |
| mattip | ahah! It fails when run as pypy pytest.py ... but succeeds when run as python pytest.py ... | 17:12 |
| arigato | iow it's random... | 17:12 |
| jterrace | i've had it pass and fail | 17:13 |
| jterrace | so i think it's also random | 17:13 |
| fijal | arigato: can you explain to me changes in register allocation btw? | 17:13 |
| fijal | because it seems it would produce different asm now | 17:13 |
| arigato | yes | 17:13 |
| arigato | about jit-simplify-backendintf specifically, or about jit-targets? | 17:14 |
| fijal | jit-simplify | 17:14 |
| arigato | ok | 17:14 |
| fijal | because you precisely modified a test that was there for numpy | 17:14 |
| fijal | (and made it 10% faster) | 17:14 |
| arigato | which is...? | 17:15 |
| fijal | notably things that are passed around in a loop but are unused should not be in registers | 17:15 |
| fijal | and you killed two more | 17:15 |
| fijal | that were testing variants of it | 17:15 |
| arigato | yes | 17:15 |
| __main__ (~main@adsl-99-173-15-158.dsl.pltn13.sbcglobal.net) joined #pypy. | 17:15 | |
| arigato | the "unused" computation was killed | 17:15 |
| arigato | along with the rest of the related code, actually | 17:16 |
| fijal | well ok | 17:16 |
| arigato | because there is no more place where we have to decide out of thin air "this box goes there" | 17:16 |
| fijal | because you have a continuity between the preamble and the main loop? | 17:17 |
| timotimo | http://pypy.org/features.html � this still claims that the stackless build is separate | 17:17 |
| arigato | no | 17:17 |
| fijal | so why? | 17:17 |
| fijal | does the loop still go in two pieces into regalloc now? | 17:18 |
| arigato | no | 17:18 |
| arigato | that's a different issue | 17:18 |
| arigato | the main piece of code that jit-simplify kills is the one that makes arglocs for the start of the loop | 17:18 |
| fijal | which is replaced by... | 17:18 |
| fijal | ? | 17:18 |
| arigato | and the piece of generated code that came before, which loads stuff from fail_args | 17:18 |
| arigato | there is only the piece of generated code that was used for CALL_ASSEMBLERs | 17:19 |
| fijal | there is still _set_initial_bindings | 17:19 |
| arigato | and if there is no LABEL operation at the start of the loop, there is no way to JUMP to it later | 17:19 |
| arigato | yes, but note that _set_initial_bindings() doesn't produce any machine code at all | 17:20 |
| fijal | yes sure | 17:20 |
| fijal | but it was the case before as well IMO | 17:20 |
| arigato | it initializes the bindings to be exactly where the calling convention will put them | 17:20 |
| fijal | ok | 17:20 |
| fijal | so this is at the beginning of the preamble? | 17:20 |
| fijal | or at the beginning of the loop? | 17:20 |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) left irc: Quit: etrepum | 17:21 | |
| arigato | I think there is some confusion: "premable" and "loop" don't make sense for the backend | 17:21 |
| fijal | or both? | 17:21 |
| fijal | ok | 17:21 |
| fijal | so they come in glued with a LABEL in between? | 17:21 |
| arigato | yes | 17:21 |
| fijal | ok, I asked the same question before and the answer was "no" :) | 17:22 |
| arigato | the backend is sent loops and bridges making a tree, like before, but the difference is that they end in a JUMP going always to a LABEL, which is typically in the middle of some other generated code | 17:22 |
| fijal | but that explains a lot | 17:22 |
| arigato | uh, did I? sorry | 17:22 |
| fijal | nevermind :) | 17:22 |
| fijal | I guess that's fine | 17:22 |
| fijal | then | 17:23 |
| arigato | so in this model you cannot JUMP to the start of a loop, unless there is also a LABEL there | 17:23 |
| fijal | because we have slightly less arbitrary set of moves, right? | 17:23 |
| arigato | yes | 17:23 |
| fijal | I mean typically a JUMP in the end would jump to LABEL in the middle | 17:23 |
| fijal | which would have roughly-the-same setup | 17:23 |
| arigato | yes | 17:23 |
| arigato | at least I hope | 17:23 |
| fijal | my main point of contention is that we usually also have things like move const NULL to this place on valuestack | 17:24 |
| fijal | etc. | 17:24 |
| lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Ping timeout: 252 seconds | 17:24 | |
| fijal | so the variable is unused at all | 17:24 |
| fijal | but it's not passed around | 17:24 |
| fijal | at the entrance there is a var and at the end there is a const | 17:24 |
| arigato | a "not" too many?... | 17:24 |
| arigato | but yes | 17:24 |
| kenaan | 12jterrace default 11863190739032 15/pypy/module/micronumpy/: Merged numpy-pi-sum-min-max, adding numpypy.pi, .sum, .min, and .max | 17:24 |
| fijal | I think it's the right amount, but confusing :) | 17:25 |
| fijal | typically | 17:25 |
| fijal | [i0] | 17:25 |
| fijal | jump(0) | 17:25 |
| arigato | yes | 17:25 |
| fijal | should not generate assembler | 17:25 |
| fijal | but indeed, I think it's already better than before | 17:25 |
| Alex_Gaynor | jterrace: can you close the merged branch as well? | 17:25 |
| arigato | well, that's hard and a different issue, I suppose | 17:25 |
| Action: jterrace is looking up how to do that | 17:25 | |
| fijal | yes | 17:25 |
| Alex_Gaynor | jterrace: on the branch you merged hg ci --close-branch (or something like that) | 17:26 |
| fijal | arigato: and what are the various changes to register allocator doing? | 17:26 |
| fijal | like the new flag | 17:26 |
| fijal | arigato: it's a cool cleanup, feel free to merge | 17:27 |
| arigato | _hint_dont_reuse_quickly? | 17:27 |
| fijal | yes | 17:27 |
| arigato | just a hack, you can see the effect in the source | 17:27 |
| kenaan | 12jterrace numpy-pi-sum-min-max 1153d2c2028ef3 15/: Close merged branch | 17:27 |
| arigato | a hack of unclear benefits I suppose, but useful in some tests | 17:28 |
| fijal | :] | 17:29 |
| davidpk (~davidpk@5e06d080.bb.sky.com) joined #pypy. | 17:29 | |
| arigato | used in consider_label(), to avoid that the possibly_free_var(tmpvar) we do early ends up putting eax as the next register to use | 17:29 |
| arigato | e.g. if we are generating a tight loop, | 17:29 |
| fijal | ok | 17:30 |
| arigato | then without this hack it would then use eax in preference, | 17:30 |
| mkerrin (~mkerrin@wg1-nat.fwgal01.deri.ie) left irc: Quit: mkerrin | 17:30 | |
| arigato | and then spill it in consider_jump() | 17:30 |
| stakkars_ (~tismer@i59F71C2E.versanet.de) left irc: Quit: schnarch | 17:31 | |
| arigato | test_zjit is a bit unpredictable because it translates one version and reuses it for all tests | 17:33 |
| fijal | yes | 17:34 |
| fijal | well, it's a speed hack | 17:34 |
| arigato | in this case I get test_slice2 that ends up generating a bridge out a loop produced by test_slice | 17:34 |
| fijal | pffff | 17:34 |
| fijal | pfff | 17:34 |
| fijal | pffffff | 17:34 |
| fijal | ok | 17:34 |
| fijal | it totally should not | 17:34 |
| fijal | arigato: so it's testing a bit extra as well | 17:34 |
| fijal | :) | 17:34 |
| arigato | :-) | 17:34 |
| fijal | arigato: it makes sense to do it - this is a real scenario | 17:35 |
| fijal | I think slices are kind of broken on default anyway | 17:35 |
| thrain42 (~durin@adium/durin42) joined #pypy. | 17:35 | |
| durin42 (~durin@adium/durin42) left irc: Ping timeout: 240 seconds | 17:36 | |
| fijal (~fijal@197.169.119.41) left irc: Read error: Connection reset by peer | 17:36 | |
| mattip | I have a translated pypy from 315da4da5ab0, just starting it up and calling "import ctypes" throws an exception, something about dlopen. | 17:41 |
| fijal (~fijal@41.52.213.72) joined #pypy. | 17:42 | |
| Alex_Gaynor | mattip: basically you need a newer translated pypy, something in teh stdlib was changed | 17:44 |
| Action: fijal changed something in the stdlib | 17:44 | |
| arigato | fijal: I'm a bit unsure, but I think it's basically expected to reuse a piece of code between one test and the next one | 17:44 |
| fijal | yes | 17:45 |
| fijal | but not if two expressions differ | 17:45 |
| Alex_Gaynor | fijal: test_slice2 is sometimes working and sometimes not for me | 17:45 |
| arigato | yes, but in this case, they don't differ, or do they= | 17:45 |
| fijal | if it's a bridge they do, no? | 17:46 |
| arigato | no | 17:46 |
| fijal | ah | 17:46 |
| arigato | they are indeed the same, according to the test run | 17:46 |
| fijal | indeed | 17:46 |
| fijal | so why this test is even there? | 17:47 |
| arigato | so you get nonsense because slice2 immediately jumps to the existing assembler | 17:47 |
| fijal | checking two identical expressions is not giving you much | 17:47 |
| arigato | but then it fails a guard | 17:47 |
| arigato | I think that's the finish guard anyway | 17:47 |
| fijal | right | 17:47 |
| fijal | ok | 17:47 |
| fijal | I gonna kill this test then | 17:47 |
| arigato | it could be fixed by making reset_stats() more aggressive | 17:48 |
| kenaan | 12fijal default 11770cb2b481e2 15/pypy/module/micronumpy/test/test_zjit.py: kill redundant test | 17:49 |
| etrepum (~bob@accessnat4.mochimedia.net) joined #pypy. | 17:49 | |
| Alex_Gaynor | fijal: well in theory it tests something different | 17:49 |
| fijal | how different? | 17:49 |
| Alex_Gaynor | in practice not | 17:49 |
| Alex_Gaynor | it's a 2-arg slice vs. a 3-arg slice | 17:50 |
| fijal | then it should go to test_compile | 17:50 |
| fijal | or so | 17:50 |
| arigato | yay, fixed | 17:51 |
| fijal | arigato: I killed the test? | 17:52 |
| arigato | yes, but I fixed it by fixing reset_stats() | 17:52 |
| nedbat (~nedbat@python/psf/nedbat) left irc: Disconnected by services | 17:52 | |
| arigato | so I'm going to check in anyway, I suppose | 17:52 |
| fijal | cool | 17:52 |
| fijal | yes | 17:52 |
| fijal | Alex_Gaynor: what we don't have is a slice of a ufunc | 17:53 |
| fijal | or some other operation | 17:53 |
| Alex_Gaynor | write one, I suppos | 17:53 |
| kenaan | 12arigo default 112edda8cbb1b0 15/pypy/jit/metainterp/: Fix reset_stats() to clear a bit more things. It's still not perfect, but better... | 17:53 |
| kenaan | 12arigo default 11c805becb84aa 15/pypy/module/micronumpy/: merge heads | 17:53 |
| thinker (~thinker@123-194-52-90.dynamic.kbronet.com.tw) left irc: Ping timeout: 244 seconds | 17:54 | |
| khs (~khs@2001:700:300:2120:725a:b6ff:fee5:a44) left irc: Quit: Leaving | 17:54 | |
| fijal | Alex_Gaynor: why is getarrayitem_raw getinteriorfield_raw these days? | 17:57 |
| fijal | did it get unified? | 17:57 |
| Alex_Gaynor | fijal: because of how Primitive.read() works, and how composite/complex dtypes will work | 17:58 |
| fijal | ok | 17:58 |
| arigato (~arigo@fwstups.cs.uni-duesseldorf.de) left irc: Ping timeout: 252 seconds | 17:59 | |
| kenaan | 12alex_gaynor extradoc 11160ba3f22a89 15/planning/micronumpy.txt: anotehr item | 18:02 |
| PiotrSikora (~none@nginx/adept/piotrsikora) left irc: Excess Flood | 18:04 | |
| PiotrSikora (~none@nginx/adept/piotrsikora) joined #pypy. | 18:05 | |
| xenru (~Adium@31.31.97.244) joined #pypy. | 18:07 | |
| Alex_Gaynor | fijal: it occurs to me, I think we've finally introduced enough complexity so that constants for greenkeys aren't happening in numpy tests :) | 18:08 |
| xenru (~Adium@31.31.97.244) left irc: Client Quit | 18:09 | |
| exarkun | in api.py, is SYMBOLS_C a list of the subset of the cpyext exported C API which is implemented in .c files instead of in rpython? | 18:10 |
| apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy. | 18:13 | |
| fijal | exarkun: I think so | 18:14 |
| exarkun | I guess it doesn't help that I'm not all that clear on the difference between the different buffer types in CPython :/ | 18:15 |
| Action: exarkun tries to sort that out | 18:15 | |
| Alex_Gaynor | exarkun: if it makes you feel better, no one understand teh different buffer types | 18:16 |
| exarkun | Heh :) | 18:17 |
| exarkun | It looks like maybe I should try ignoring Py_buffer for a while. At least until I can pass a buffer object from Python into C, anyway. | 18:17 |
| exarkun | PyPy currently defines PyBuffer_Type in bufferobject.c. Do I need to translate that into RPython, or can it somehow stay where it is? | 18:18 |
| exarkun | structseq looks like it's in a similar state to bufferobject, but I don't know if structseq works on pypy either ;) | 18:21 |
| Action: exarkun creates a bufferobject.py and starts cargo-culting from stringobject.py | 18:22 | |
| Rhy0lite | alex_gaynor: ping | 18:23 |
| Alex_Gaynor | Rhy0lite: pong | 18:23 |
| Nick change: thrain42 -> durin42 | 18:34 | |
| etrepum_ (~bob@accessnat4.mochimedia.net) joined #pypy. | 18:35 | |
| nickik (~nick@217.150.239.67) joined #pypy. | 18:36 | |
| fijal | exarkun: I didn't know you can make a verb out of that | 18:37 |
| pjenvey | you can verb just about anything if you really want to | 18:38 |
| arigato (~arigo@89.204.153.160) joined #pypy. | 18:38 | |
| etrepum (~bob@accessnat4.mochimedia.net) left irc: Ping timeout: 240 seconds | 18:38 | |
| Nick change: etrepum_ -> etrepum | 18:38 | |
| capncapacitor | exarkun: The same bug as before exists in PyOpenSSL trunk as well, FYI | 18:39 |
| DasIch_ (~dasich@p4FFDF269.dip.t-dialin.net) left irc: Ping timeout: 240 seconds | 18:40 | |
| exarkun | What kind of a thing is space.w_str? | 18:40 |
| Alex_Gaynor | space.w_str is W_TypeObject | 18:41 |
| exarkun | There's probably no space.w_buffer, right? | 18:41 |
| Alex_Gaynor | I think there is, but I won't swear to it | 18:41 |
| mwhudson (~mwh@linaro/mwhudson) joined #pypy. | 18:42 | |
| Alex_Gaynor | no, there isn't | 18:42 |
| Alex_Gaynor | exarkun: the equivilant is space.gettypefor(Buffer) with from pypy.interpreter.buffer import Buffer | 18:43 |
| exarkun | The abstractness of Buffer makes me feel like I'm about to run into trouble. | 18:43 |
| jterrace (~jterrace@aegis.CS.Princeton.EDU) left irc: Quit: Leaving. | 18:43 | |
| Alex_Gaynor | exarkun: what do you mean? | 18:44 |
| Guest__ (~textual@216.190.27.14) joined #pypy. | 18:44 | |
| exarkun | eg PyBufferObject has a field `b_base´ that I think is a pointer to the PyObject* the buffer is exposing | 18:44 |
| exarkun | Buffer doesn't have that data because its up to a specialized subclass to know how to be a buffer for any particular type | 18:45 |
| exarkun | So if I try to implement buffer_attach, then I don't know where to get a value to use to populate b_base | 18:45 |
| berdario_ (~Tiibiidii@217.220.118.217) left irc: Ping timeout: 276 seconds | 18:45 | |
| Rhy0lite | Alex_Gaynor: where should Priya initiate the discussion about contributing the codespeed benchmark? | 18:45 |
| exarkun | Maybe I have to do this once per Buffer subclass that is to be supported in cpyext? | 18:45 |
| Alex_Gaynor | Rhy0lite: pypy-dev mailing list is as good as anyplace | 18:46 |
| JaRoel (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection | 18:46 | |
| Rhy0lite | okay | 18:46 |
| Alex_Gaynor | exarkun: hmm, yes that would be a problem | 18:46 |
| exarkun | make_typedescr a StringBuffer, make_typedescr a ... StringLikeBuffer? make_typedescr a ... SubBuffer? | 18:46 |
| exarkun | otoh today I don't care about support for anything except StringBuffer | 18:47 |
| exarkun | Building the cpyext bindings for each concrete Buffer type is at least a feasible strategy, if boring, because it's not like new Buffer types can appear at runtime. | 18:47 |
| exarkun | Right? | 18:47 |
| exarkun | On the other hand, the reverse mapping is harder. | 18:48 |
| exarkun | If someone uses the CPython API to throw a PyBufferObject of their own design at PyPy, then what? | 18:49 |
| exarkun | (not that I care about that case either) | 18:49 |
| Alex_Gaynor | I don't understand the buffer protocol, excedpt to say it appears to be designed around a) non moving memory, b) using a char* for everything | 18:50 |
| Alex_Gaynor | both of which are a little unfortunate | 18:50 |
| exarkun | But I don't know that the cpyext infrastructure, eg make_typedescr, can support multiple PyPy type objects being associated with a single CPython type struct. | 18:50 |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) joined #pypy. | 18:50 | |
| jterrace (~jterrace@aegis.CS.Princeton.EDU) joined #pypy. | 18:50 | |
| exarkun | Yea, it's not a very good protocol. And its implementation is vastly overly complicated since the new buffer protocol was introduced (which in practice has no new features). | 18:51 |
| exarkun | But those are the easy problems, I think. | 18:51 |
| Nick change: Guest__ -> rekamso | 18:51 | |
| exarkun | I guess I'll try only supporting StringBuffer for starters | 18:51 |
| Alex_Gaynor | well, don't forget that the new protocol isn't totally implemented | 18:51 |
| DasIch_ (~dasich@p4FFDC921.dip.t-dialin.net) joined #pypy. | 18:53 | |
| exarkun | Oh. Supporting StringBuffer is not sufficient. Because buffer() can also created SubBuffers depending on the value of its arguments. | 18:54 |
| exarkun | So another idea is that I could go via the apparent Buffer interface, which has things such as as_str()... but that doesn't help with b_base, I guess. | 18:55 |
| exarkun | There's always isinstance... | 18:56 |
| exarkun | Or will I get in trouble with that? | 18:56 |
| Alex_Gaynor | I mean, isinstance works | 18:57 |
| Alex_Gaynor | and it's RPython | 18:57 |
| Action: exarkun finds out | 18:58 | |
| exarkun | I'm still missing something | 19:05 |
| exarkun | I wrote a bunch of RPython but I don't see how this will become associated with the stuff in src/bufferobject.c | 19:05 |
| exarkun | >:O | 19:14 |
| danchr (~danchr@cl-848.chi-02.us.sixxs.net) left irc: Quit: ZNC - http://znc.sourceforge.net | 19:14 | |
| exarkun | But of course none of this can work now, because PyBufferObject isn't actually exposed at the C level either. | 19:14 |
| exarkun | What does cpython_struct actually do? | 19:15 |
| danchr (~danchr@cl-848.chi-02.us.sixxs.net) joined #pypy. | 19:15 | |
| hruske (~Gasper@188-230-156-183.dynamic.t-2.net) left irc: Quit: Leaving | 19:15 | |
| JaRoel|4d (~jaroel|4d@2001:980:7133:1:d45d:8c46:9aee:b3eb) joined #pypy. | 19:16 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy. | 19:20 | |
| exarkun | pytest --pdb doesn't actually stop on the exception that my test encounters :( | 19:23 |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 19:24 | |
| lmoura (lmoura@gateway/indt/session) left irc: Ping timeout: 240 seconds | 19:26 | |
| amaury_ | exarkun: because it fails in setup or teardown? | 19:26 |
| exarkun | I don't think it does, but it's hard to tell. | 19:26 |
| mattip | are you running pytest under pypy or python? | 19:26 |
| exarkun | python | 19:26 |
| lmoura (lmoura@gateway/indt/session) joined #pypy. | 19:28 | |
| apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) left irc: Ping timeout: 248 seconds | 19:30 | |
| exarkun | I guess it's because errors originate somewhere inside cpyext | 19:34 |
| stakkars | hi! | 19:34 |
| exarkun | it needs to be correct enough to not segfault at least until after it comes back out of cpyext or pdb doesn't have a chance | 19:34 |
| stakkars | so do we have decided about default-translating or else? | 19:34 |
| stakkars | I think something to rely on would be good. | 19:35 |
| exarkun | My changes still seem to have no effect (despite the fact that they are able to cause segfaults :) - my unit test still fails like `TypeError: argument 1 must be convertible to a buffer, not buffer´ | 19:36 |
| fijal (~fijal@41.52.213.72) left irc: Read error: Connection reset by peer | 19:37 | |
| fijal (~fijal@197.108.108.70) joined #pypy. | 19:38 | |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy. | 19:42 | |
| kkris (~kris@80-123-34-214.adsl.highway.telekom.at) joined #pypy. | 19:44 | |
| mattip | stakkars: hi. | 19:49 |
| mattip | I have a windows 7 machine, i5 processor with 8GB RAM. I am trying to convince my partner to allow us to run a slave buildbot on it. | 19:50 |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) left irc: Ping timeout: 240 seconds | 19:51 | |
| mattip | But until I get the 24/7 thing worked out, I am happy to help with the win32/64 port | 19:51 |
| etrepum (~bob@accessnat4.mochimedia.net) left irc: Remote host closed the connection | 19:52 | |
| etrepum (~bob@accessnat4.mochimedia.net) joined #pypy. | 19:52 | |
| exarkun | Yea, weird. My code runs, but the wrong buffer type is still used. | 19:57 |
| amaury_ | exarkun: it is committed in pyarg-parsetuple-s-star-buffer branch? | 19:57 |
| exarkun | it can be in a moment | 19:57 |
| exarkun | maybe | 19:58 |
| amaury_ | :) | 19:58 |
| kenaan | 12exarkun pyarg-parsetuple-s-star-buffer 11339b179f2f51 15/pypy/module/cpyext/: Make PyBufferObject public so cpython_struct can work | 20:00 |
| kenaan | 12exarkun pyarg-parsetuple-s-star-buffer 111396a5482b12 15/pypy/module/cpyext/: Try to teach cpyext about buffer objects | 20:00 |
| exarkun | that's it | 20:01 |
| kenaan | 12hakanardo jit-multilabel 11667ace56adf5 15/pypy/jit/metainterp/optimizeopt/unroll.py: kill constant_inputargs and aliases from the exported state and encode it all in inputarg_setup_ops instead | 20:04 |
| kenaan | 12hakanardo jit-multilabel 11faa7d41a5f50 15/pypy/jit/metainterp/optimizeopt/optimizer.py: fix | 20:04 |
| asmeurer__ (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy. | 20:04 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/628 [12hakanardo, jit-multilabel] | 20:05 |
| Alex_Gaynor | hakanardo: is multilabel so something like max(a, b) has a very short bridge which jumps back into the middle of the main loop? | 20:06 |
| hakanardo | well it's more general | 20:07 |
| fijal | of course, it's pypy after all | 20:08 |
| Alex_Gaynor | :) | 20:08 |
| fijal | mattip, jterrace: I remove the offending zjit test | 20:08 |
| hakanardo | to suport multiple intermediate labels | 20:08 |
| jterrace | ok | 20:08 |
| jterrace | Alex_Gaynor: can #18 be merged? | 20:09 |
| Alex_Gaynor | jterrace: have you explored if the call_function(w_int) approach works? | 20:09 |
| exarkun | amaury_: Maybe I really need to implement alloc? | 20:09 |
| jterrace | no ill try that | 20:09 |
| kenaan | 12fijal default 1104116407532d 15/pypy/jit/: (arigo) merge jit-simplify-backendintf branch. This branch kills set_future_value_xxx and replaces it with a functi... | 20:09 |
| kenaan | 12fijal jit-simplify-backendintf 11539833042097 15/: close merged branch | 20:09 |
| fijal | Alex_Gaynor: what would space.call_function(space.w_int) do? | 20:11 |
| Alex_Gaynor | fijal: well, with an argument of course | 20:11 |
| Alex_Gaynor | fijal: well, it returns space.wrpa(0) with no arg | 20:11 |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy. | 20:11 | |
| fijal | why? | 20:11 |
| jterrace | it would make it so you don't have to special case strings | 20:11 |
| amaury_ | exarkun: try to run with the "-s" option | 20:12 |
| amaury_ | I get unraisable exceptions | 20:12 |
| jterrace | space.float_w(space.float(w_item)) does not work if w_item is a string, but space.call_function(space.w_int, w_item) might | 20:12 |
| Alex_Gaynor | fijal: for strings | 20:12 |
| Alex_Gaynor | jterrace: it will works | 20:12 |
| fijal | do you want applevel strings or rpython level strings? | 20:13 |
| Alex_Gaynor | you'll have to fix up fake space somehow though | 20:13 |
| fijal | because if you want rpython level | 20:13 |
| Alex_Gaynor | app-level | 20:13 |
| fijal | you want to do int(space.str_w) | 20:13 |
| fijal | or float(space.str_w) | 20:13 |
| amaury_ | exarkun: first error in buffer_attach(), no field b_readonly | 20:13 |
| fijal | maybe you want space.wrap(int(space.str_w(w_obj)) | 20:13 |
| fijal | then? | 20:13 |
| exarkun | amaury_: the -s option to what? | 20:13 |
| amaury_ | to pytest | 20:13 |
| jterrace | float(space.str_w(space.str(w_item))) is what i do | 20:13 |
| amaury_ | exarkun: the issue is that PyTuple_GET_ITEM may fail in pypy | 20:14 |
| exarkun | I don't get unraisable exceptions with or without -s | 20:14 |
| amaury_ | because it needs to convert the pypy object into a PyObject* | 20:15 |
| amaury_ | exarkun: I run | 20:16 |
| exarkun | Oh. I was already passing --capture=no | 20:16 |
| amaury_ | ../pytest.py module/cpyext/test/test_getargs.py -s | 20:16 |
| amaury_ | probably the same? | 20:16 |
| exarkun | right | 20:17 |
| exarkun | There's a lot of output, but I don't see anything about unraisable exceptions (but I might not know what you mean) | 20:17 |
| exarkun | I'm only running one test, fwiw. Maybe I broke some other code paths. | 20:17 |
| Alex_Gaynor | fijal: no, nonoe of those are what we want. | 20:17 |
| exarkun | /usr/bin/python ./pytest.py --pdb -k old_buffer --capture=no --verbose pypy/module/cpyext/test/test_getargs.py | 20:18 |
| exarkun | That's what I'm doing | 20:18 |
| Alex_Gaynor | trust me, space.call_function(space.w_float, w_item) is what we want | 20:18 |
| fijal | okey | 20:18 |
| kenaan | 12fijal extradoc 11a2c53a086126 15/talk/sea2012/abstract.rst: Add an abstract | 20:19 |
| fijal | anyone feel free to review | 20:22 |
| amaury_ | exarkun: anyway, here is what I get | 20:22 |
| amaury_ | http://paste.pocoo.org/show/519925/ | 20:22 |
| amaury_ | I suggest to use rffi.setintfield(py_buf, 'b_readonly', 1) | 20:23 |
| jterrace | Alex_Gaynor: i don't think space.call_function(space.w_float, w_item) works either | 20:23 |
| Alex_Gaynor | jterrace: why not? | 20:23 |
| jterrace | it's expecting a W_FloatObject | 20:23 |
| Alex_Gaynor | jterrace: space.float_w(<that>) | 20:23 |
| jterrace | so it barfs on a string | 20:23 |
| antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 20:24 | |
| Alex_Gaynor | jterrace: wtf, it shouldn't, give me a few minutes in a meeting, and Ill look | 20:24 |
| jterrace | k | 20:24 |
| amaury_ | jterrace: space.w_float is the Python 'float' | 20:24 |
| Rhy0lite | fijal: want me to edit in place? | 20:24 |
| fijal | Rhy0lite: please | 20:24 |
| Alex_Gaynor | amaury_: so calling it should be the same as float("12.3") right? | 20:24 |
| fijal | Rhy0lite: http://sea.ucar.edu/conference/2012 | 20:24 |
| amaury_ | yes | 20:24 |
| fijal | this is the conference call, not very detailed | 20:24 |
| exarkun | amaury_: I have no idea why you get that behavior and I don't. Since I don't have a problem with setting that int field, I suspect that using setintfield won't help me at all. :/ | 20:24 |
| amaury_ | ah, you may use a 32bit platform already | 20:25 |
| amaury_ | where Signed === INT | 20:25 |
| exarkun | yes | 20:25 |
| amaury_ | (sorry for the javascript operator) | 20:25 |
| Alex_Gaynor | amaury_: we accept your apology | 20:26 |
| Alex_Gaynor | :) | 20:26 |
| fijal | we should add automatic casts one day... | 20:26 |
| exarkun | add documentation for everything else first | 20:26 |
| fijal | like cpyext? | 20:27 |
| fijal | documentation easily falls out of sync... | 20:27 |
| exarkun | yes, including cpyext :) | 20:27 |
| exarkun | I never said it was easy | 20:27 |
| fijal | I think pypy is seriously being developed too fast for that | 20:28 |
| lambacck (~chris@d24-150-124-118.home.cgocable.net) left irc: Quit: lambacck | 20:28 | |
| fijal | we kinda have trouble with just keeping a little bit that we have up to date | 20:28 |
| exarkun | in another 2 hours though I'm going to give up on this round of pypy development, whether I have implemented this feature or not | 20:28 |
| exarkun | and once again it has been so unproductive that I'll wait at least another year before I try again | 20:28 |
| fijal | :/ | 20:28 |
| fijal | there is a reason why everyone hates the idea of developing cpyext | 20:28 |
| fijal | or say a cpyext-friendly swig backend | 20:29 |
| fijal | or cython | 20:29 |
| fijal | amaury_: can you look into helping exarkun a bit? | 20:29 |
| exarkun | I haven't spend hours staring at core files or corrupt C stacks | 20:29 |
| amaury_ | I'm doing it atm | 20:29 |
| fijal | cause I have no clue how cpyext works | 20:29 |
| fijal | cool, thanks :) | 20:29 |
| exarkun | I've spent hours staring at PyPy code that makes my eyes bleed | 20:29 |
| exarkun | I grant that it is a hard problem, and I think that what is there is very impressive | 20:29 |
| fijal | I think this is what you get with tons of metaprogramming | 20:30 |
| exarkun | But I wonder who can actually understand it well enough to do anything with it | 20:30 |
| fijal | and this tons is kind of necessary to wrap your head around tons and tons of apis | 20:30 |
| Action: fijal can't | 20:30 | |
| fijal | amaury_: documenting cpyext would be very cool btw | 20:30 |
| exarkun | Maybe having people understand it is not the point, but considering how much of getargs.c is still commented out, I think a case can be made that _more_ people need to understand it so it will actually get finished | 20:31 |
| fijal | exarkun: you know what's the problem? every single problem in computer science is either hard or trivial | 20:32 |
| fijal | (well, sometimes it's also impossible) | 20:32 |
| fijal | once you understood it's trivial | 20:32 |
| fijal | and you have to understand it to document it | 20:32 |
| capncapacitor | exarkun: I know how you feel. Today was the first day I looked at Pypy's source and I don't really remember what happened after that. When I came to there were a few sacrificial goats missing from the pen. | 20:32 |
| fijal | and once you did, there is no longer any point | 20:32 |
| Alex_Gaynor | capncapacitor: can we step back a few steps and ask why you have a goat pen? | 20:33 |
| fijal | Alex_Gaynor: we don't want to know | 20:33 |
| Alex_Gaynor | fijal: probably, why did you merge another branch making me compile again | 20:33 |
| fijal | I'm sorry | 20:33 |
| jterrace | Alex_Gaynor: nevermind, it works. I was trying to box an object-space value instead of interpreter-space value | 20:33 |
| capncapacitor | I work with Twisted, EC2 and Cassandra. A lot of deals have to be made. | 20:34 |
| exarkun | fijal: I think these are all excuses | 20:34 |
| fijal | capncapacitor: is your stuff working on pypy? | 20:34 |
| Alex_Gaynor | jterrace: that'll do it :) | 20:34 |
| fijal | exarkun: yes, I don't understand cpyext either | 20:34 |
| fijal | exarkun: I try documenting a bit stuff *I* touch | 20:34 |
| fijal | sometimes | 20:34 |
| exarkun | fijal: It is hard to write good documentation, but software for any domain can be documented. | 20:34 |
| capncapacitor | fijal: Not entirely. We're seeing massive, massive improvements everywhere that it is working though | 20:34 |
| kenaan | 12amauryfa pyarg-parsetuple-s-star-buffer 11ab4be8e09a58 15/pypy/module/cpyext/bufferobject.py: Fix for 64bit platforms | 20:34 |
| exarkun | regardless of how complicated the domain itself is | 20:35 |
| kenaan | 12edelsohn extradoc 11b91e6cab9b06 15/talk/sea2012/abstract.rst: Improve language in abstract. | 20:35 |
| amaury_ | exarkun: ok, now I see your error :-) | 20:35 |
| fijal | capncapacitor: good :) feel like writing a blog post/post to mailing list how it did work for you? | 20:35 |
| amaury_ | well, this means I get the same message | 20:35 |
| Rhy0lite | fijal: now for the next paragraph :-) | 20:35 |
| fijal | Rhy0lite: it does have only one? | 20:35 |
| Rhy0lite | yes | 20:35 |
| Rhy0lite | when you write the next one | 20:35 |
| fijal | should it have more? | 20:35 |
| fijal | I was thinking that's it | 20:36 |
| Rhy0lite | for the abstract, yes | 20:36 |
| fijal | yes | 20:36 |
| fijal | thanks :) | 20:36 |
| Action: arigato randomly thinks what other thread of conversation he could think about | 20:36 | |
| capncapacitor | fijal: Once we're closer to replacing CPython in production with Pypy, sure. The last big problem was openssl, but we have a temporary workaround for that. | 20:36 |
| fijal | capncapacitor: cool | 20:36 |
| jnoller (~jnoller@vpn.nasuni.com) left irc: | 20:36 | |
| capncapacitor | Huh, didn't know Jesse hung out here | 20:40 |
| fijal | sometimes | 20:40 |
| fijal | he's lurking mostly | 20:40 |
| amaury_ | exarkun: pb->bf_getreadbuffer is NULL | 20:40 |
| exarkun | amaury_: Yes | 20:40 |
| exarkun | amaury_: The type of the PyObject* that gets pushed into the C code is wrong | 20:41 |
| exarkun | amaury_: It's a type named "buffer" with no other fields initialized | 20:41 |
| exarkun | I can't understand the code that selects a typedescr to use when crossing the bridge from PyPy to C, so I dunno why the type is wrong | 20:41 |
| amaury_ | maybe the tp_as_buffer structure is not correctly filled | 20:42 |
| exarkun | no | 20:42 |
| exarkun | I mean, I have no clue, none of this stuff makes any sense to me, but that explanation doesn't seem correct based on what I've observed | 20:43 |
| exarkun | There is PyBuffer_Type, because of src/bufferobject.c | 20:43 |
| Rhy0lite (dje@nat/ibm/x-iqvummkaxufizbro) left irc: Quit: Leaving | 20:43 | |
| exarkun | It is all correctly initialized | 20:43 |
| exarkun | And when a __builtin__.buffer instance is passed from PyPy to C, the resulting PyObject* is _not_ an instance of PyBuffer_Type | 20:43 |
| amaury_ | ah | 20:44 |
| fijal | there is a missing make_typedescr or so | 20:44 |
| exarkun | I'm sure I've failed to glue PyBuffer_Type to pypy.interpreter.buffer.Buffer properly (because I have no idea how to do that) | 20:45 |
| amaury_ | so we need a way to link space.w_buffer to PyBuffer_Type | 20:45 |
| amaury_ | eh, track_reference() in pyobject.py | 20:46 |
| amaury_ | it even has a docstring | 20:46 |
| papercrane (~papercran@75.101.111.82) joined #pypy. | 20:48 | |
| exarkun | I stepped through that function a bunch of times in pdb. I'm not sure I see the connection. | 20:49 |
| exarkun | It's about instances, not types, isn't it? | 20:49 |
| amaury_ | we need to correctly fill ob_type | 20:50 |
| amaury_ | when exposing to C a pypy.interpreter.buffer.Buffer instance, | 20:50 |
| amaury_ | we would like ob_type to be PyBuffer_Type | 20:50 |
| amaury_ | the one in bufferobject.c | 20:50 |
| Action: exarkun nods | 20:50 | |
| papercrane (~papercran@75.101.111.82) left irc: Read error: Connection reset by peer | 20:50 | |
| amaury_ | and not some other struct autogenerated by cpyext | 20:51 |
| exarkun | Should buffer_attach be responsible for that? Or is this a new thing that needs support somewhere else in cpyext? | 20:51 |
| mattip (chatzilla@bzq-79-177-112-183.red.bezeqint.net) left #pypy. | 20:51 | |
| amaury_ | it can be done in buffer_attach | 20:52 |
| xenru (~Adium@31.31.97.244) joined #pypy. | 20:52 | |
| papercrane (~papercran@75.101.111.82) joined #pypy. | 20:53 | |
| amaury_ | but I would like space.type(someBuffer) return the same thing as someBuffer->ob_type | 20:53 |
| exarkun | I don't think I understand that. isn't space.type(someBuffer) a PyPy representation of the type, and someBuffer->ob_type a CPython representation of the type? | 20:55 |
| e1ven (~E1ven@SQ7/ProjectLead/E1ven) left irc: Ping timeout: 255 seconds | 20:56 | |
| mat^2 (~mathias@212.130.113.35) joined #pypy. | 20:58 | |
| amaury_ | yes, but I'd like one to be convertible to the other | 20:58 |
| exarkun | get_typedescr-style convertible, or something else? | 20:59 |
| fijal | Rhyolite: thx for the abstract btw | 21:06 |
| amaury_ | exarkun: I wonder what's the use of this src/bufferobject.c | 21:06 |
| Alex_Gaynor | jterrace: is the call_function approach working out? | 21:06 |
| jterrace | yeah just updated the pull request | 21:06 |
| Alex_Gaynor | cool, tkaing a look now | 21:06 |
| jterrace | all tests pass, and it looks much nicer: https://bitbucket.org/pypy/pypy/pull-request/18 | 21:06 |
| jterrace | 8 lines -> 1 | 21:07 |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) left irc: Ping timeout: 248 seconds | 21:07 | |
| Alex_Gaynor | jterrace: some of that StringObject code is dead now right? | 21:07 |
| exarkun | amaury_: Beats me. | 21:07 |
| jterrace | yes, but i thought id keep it since might have a use for it in the future? | 21:08 |
| exarkun | amaury_: Should it be deleted and the type and APIs be reimplemented in Rpython? | 21:08 |
| Alex_Gaynor | jterrace: we can always add it back later | 21:08 |
| jterrace | sure | 21:08 |
| jterrace | want me to remove? | 21:08 |
| Alex_Gaynor | please | 21:08 |
| amaury_ | well, bufferobject.c is the implementation of an object methods | 21:08 |
| Alex_Gaynor | then I'll press the merge button :) | 21:08 |
| amaury_ | which we already have in pypy | 21:08 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/628 [12hakanardo, jit-multilabel] | 21:10 |
| tilgovi (~randall@75-149-58-169-SFBA.hfc.comcastbusiness.net) joined #pypy. | 21:11 | |
| tilgovi (~randall@75-149-58-169-SFBA.hfc.comcastbusiness.net) left irc: Changing host | 21:11 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 21:11 | |
| fijal (~fijal@197.108.108.70) left irc: Read error: Connection reset by peer | 21:14 | |
| bgola (~bgola@c9518082.virtua.com.br) joined #pypy. | 21:14 | |
| fijal (~fijal@197.105.44.141) joined #pypy. | 21:14 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) joined #pypy. | 21:14 | |
| mwhudson (~mwh@linaro/mwhudson) left irc: Ping timeout: 245 seconds | 21:15 | |
| kenaan | 12arigo extradoc 11ec0b4da1f5e8 15/talk/sea2012/abstract.rst: Confusing first usage of 'our' this late in the abstract | 21:16 |
| arigato | fijal: added a comment ^^^ | 21:16 |
| fijal (~fijal@197.105.44.141) left irc: Read error: Connection reset by peer | 21:16 | |
| capncapacitor (~capncapac@72.1.196.82) left irc: Quit: Leaving | 21:18 | |
| GDorn (~gdorn@li223-147.members.linode.com) left irc: Ping timeout: 252 seconds | 21:19 | |
| amaury_ | exarkun: ok, now I remember why I did not implement buffers before :-/ | 21:19 |
| davidpk (~davidpk@5e06d080.bb.sky.com) left irc: Quit: davidpk | 21:20 | |
| ojii (~ojii@40-34.60-188.cust.bluewin.ch) left irc: Quit: Leaving | 21:21 | |
| jterrace | Alex_Gaynor: updated: https://bitbucket.org/pypy/pypy/pull-request/18 | 21:26 |
| Alex_Gaynor | jterrace: merging, thanks! | 21:27 |
| jterrace | great thanks | 21:28 |
| jterrace | ill add fromstring on a branch from default now | 21:28 |
| exarkun | amaury_: Yea? | 21:28 |
| Alex_Gaynor | jterrace: sweet! | 21:28 |
| kenaan | 12jterrace numpy-dtype-strings 11aaa9b6a48bbb 15/pypy/module/micronumpy/: Updated int and float types to take strings in their constructors so things like int32('34') work | 21:28 |
| kenaan | 12jterrace numpy-dtype-strings 11d0fa1bba8dd6 15/pypy/module/micronumpy/compile.py: Add string to FakeSpace to fix failing compile tests | 21:28 |
| kenaan | 12jterrace numpy-dtype-strings 11ed24df405fd9 15/pypy/module/micronumpy/types.py: Change string special case to use call_function instead | 21:28 |
| kenaan | 12jterrace numpy-dtype-strings 11cf42a69ca368 15/pypy/module/micronumpy/compile.py: Remove FakeSpace string code no longer needed | 21:28 |
| Alex_Gaynor | jterrace: err, so apparently the merge button just merged it onto a branch, want to merge tha tinto default and close the branch? | 21:29 |
| jterrace | hah ok | 21:29 |
| kkris (~kris@80-123-34-214.adsl.highway.telekom.at) left irc: Quit: Leaving. | 21:31 | |
| jterrace | Alex_Gaynor: no i think it merged it | 21:31 |
| Alex_Gaynor | jterrace: into default? | 21:31 |
| jterrace | oh wait maybe not | 21:31 |
| Alex_Gaynor | yeah it did | 21:32 |
| Alex_Gaynor | weird | 21:32 |
| jterrace | yeah it did | 21:32 |
| amaury_ | exarkun: I'm trying something like this | 21:32 |
| amaury_ | http://paste.pocoo.org/show/519964/ | 21:32 |
| Alex_Gaynor | jterrace: I guess this new branch needs to be closed though | 21:33 |
| Alex_Gaynor | I'll get it | 21:33 |
| jterrace | is it still there? | 21:33 |
| jterrace | i dont see it | 21:33 |
| davidpk (~davidpk@5e06d080.bb.sky.com) joined #pypy. | 21:34 | |
| Alex_Gaynor | jterrace: weird, the pull request wasn't closed either | 21:34 |
| jterrace | i got an email saying it was accepted | 21:35 |
| mwhudson (~mwh@linaro/mwhudson) joined #pypy. | 21:35 | |
| Alex_Gaynor | ok so I'm not going to touch anything now | 21:35 |
| amaury_ | exarkun: I see "OH NO foo 11" | 21:35 |
| amaury_ | but the final comparison fails | 21:35 |
| jterrace | well f pull requests | 21:35 |
| exarkun | amaury_: I was afraid of that special case in the last hunk. | 21:35 |
| tumbleweed | is there an easy way to increase the verbosity of the make stage in translate? (So that our build system doesn't time it out for inactivity)? | 21:36 |
| Alex_Gaynor | jterrace: yeah you've got commit now, so we can use branches for review | 21:36 |
| exarkun | amaury_: But that sounds like progress. | 21:36 |
| jterrace | k | 21:36 |
| Alex_Gaynor | jterrace: so I got an email saying it was merged, but the site doesn't show this | 21:37 |
| Action: Alex_Gaynor goes to bother bitbucket people | 21:37 | |
| amaury_ | exarkun: in the first place, buffer('aaa') == 'aaa' is False | 21:37 |
| jterrace | yeah who knows | 21:37 |
| amaury_ | with pypy | 21:37 |
| amaury_ | with cpython as well | 21:37 |
| derdon (~derdon@p5DE8ACB8.dip.t-dialin.net) joined #pypy. | 21:39 | |
| mattip (~chatzilla@bzq-79-177-112-183.red.bezeqint.net) joined #pypy. | 21:39 | |
| exarkun | amaury_: I'm not that surprised that the test doesn't pass as written. | 21:39 |
| amaury_ | it would not pass with cpython :-) | 21:40 |
| exarkun | amaury_: I spent about 13 seconds writing it (and then 10 hours trying to make it do something other than segfault or raise a TypeError) | 21:40 |
| exarkun | I expect it will take about 13 more seconds to change it to assert something sensible? | 21:40 |
| exarkun | like str(buffer(...)) == str(f(buffer(...)))? | 21:40 |
| khs (~khs@151.43.202.84.customer.cdi.no) joined #pypy. | 21:40 | |
| amaury_ | yes | 21:41 |
| amaury_ | well, you can replace str(buffer('xxx')) by 'xxx' | 21:41 |
| exarkun | I expect so :) | 21:41 |
| mattip | jterrace: ping | 21:43 |
| jterrace | mattip: pong | 21:44 |
| mattip | In changeset 716c084b46a1b964 you/Alex added a bunch of tests for int32, uint32, ... | 21:44 |
| jterrace | yeah | 21:44 |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) left irc: Quit: Leaving... | 21:44 | |
| mattip | things like assert numpy.int32(2147483648) == -2147483648 | 21:45 |
| jterrace | yes | 21:45 |
| exarkun | amaury_: Are you interested in trying to debug and fix the pyOpenSSL issue capncapacitor pointed out today too (looks like something hangs inside a GC call acquiring a lock)? I could probably pay something for it. | 21:45 |
| reversefold (~papercran@75.101.111.82) joined #pypy. | 21:45 | |
| mattip | I am trying to pass these tests, on 32 bit ubuntu and they fail :( | 21:45 |
| papercrane (~papercran@75.101.111.82) left irc: Read error: Connection reset by peer | 21:45 | |
| Nick change: reversefold -> papercrane | 21:45 | |
| jterrace | hmm that's not good | 21:46 |
| jterrace | the int32 one fails? | 21:46 |
| mattip | Am I doing something wrong (again?) | 21:46 |
| jterrace | i wouldnt be too surprised if the int64's were failing on 32-bit | 21:46 |
| jterrace | but int32 failing is strange | 21:46 |
| mattip | Yes, the one with the negative number gives me an overflow error. | 21:46 |
| jterrace | does it work for you in cpython? | 21:47 |
| Action: mattip testing | 21:47 | |
| mattip | nope. OverflowError: Python int too large to convert to C long | 21:48 |
| jterrace | hmm | 21:48 |
| `fox` (~fox@95.233.109.194) left irc: Ping timeout: 240 seconds | 21:48 | |
| jterrace | does it get an overflowerror in pypy too? | 21:48 |
| jterrace | that might be the correct thing then | 21:48 |
| jterrace | and the test might have to change | 21:49 |
| e1ven (~E1ven@SQ7/ProjectLead/E1ven) joined #pypy. | 21:50 | |
| mattip | I'm playing away on a branch, I will put a skip on them in my branch. | 21:50 |
| fijal (~fijal@197.169.235.100) joined #pypy. | 21:51 | |
| fijal | eh | 21:53 |
| fijal | GPRS :/ | 21:53 |
| fijal | 3G seems to have died | 21:53 |
| jterrace | Alex_Gaynor: any way to call runpack for each data type with its format specifier? | 21:57 |
| jterrace | without putting a separate one in each type | 21:57 |
| amaury_ | fijal: IIRC, __del__ cannot do too many things | 21:57 |
| amaury_ | is it true only for subclasses of Wrappable? | 21:57 |
| fijal | amaury_: ? | 21:58 |
| fijal | that was armin's idea | 21:58 |
| fijal | but otherwise JIT is very confused | 21:58 |
| fijal | I think so | 21:58 |
| fijal | why? | 21:58 |
| amaury_ | sooner today someone reported an issue in ssl_Connection_dealloc() | 21:59 |
| amaury_ | (run by pypy with cpyext) | 21:59 |
| fijal | yes? | 21:59 |
| amaury_ | and this is a case when a __del__ can run arbitrary C code | 21:59 |
| Alex_Gaynor | jterrace: if the call is in a mixin you can have the format be a class attribute | 22:00 |
| jterrace | oh | 22:00 |
| GDorn (~gdorn@li223-147.members.linode.com) joined #pypy. | 22:01 | |
| Alex_Gaynor | jterrace: the buildbot shows the int32 issues also, if regular numpy is also throwing the OverflowError just the tests need to be updated | 22:02 |
| jterrace | yeah i can just add a conditional | 22:03 |
| Alex_Gaynor | windows64 also shows it | 22:03 |
| fijal (~fijal@197.169.235.100) left irc: Ping timeout: 248 seconds | 22:04 | |
| squiddy (~squiddy@g224196119.adsl.alicedsl.de) left irc: Quit: Leaving | 22:04 | |
| nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 255 seconds | 22:07 | |
| mattip | stupid question time: how does the "def descr__new__(space, w_subtype ..." inside a class called MyClass and "__new__ = interp2app(MyClass.descr__new__.im_func)" in its typedef interact? | 22:08 |
| mattip | where does the w_subtype come from? | 22:08 |
| gutworth (~benjamin@dhcp-128-36-84-91.central.yale.edu) left irc: Ping timeout: 240 seconds | 22:08 | |
| gutworth (~benjamin@dhcp-128-36-84-91.central.yale.edu) joined #pypy. | 22:09 | |
| Alex_Gaynor | mattip: it's because __new__ is a classmethod in python always | 22:09 |
| mattip | I'm confused as to the im_func and the w_subtype. Neither of those are familiar to me, is it easy to explain or should I look them up elsewhere? | 22:10 |
| Alex_Gaynor | im_func is an attribute on UnboundMethods | 22:11 |
| Alex_Gaynor | basically it returns the underlying function object | 22:11 |
| Alex_Gaynor | this way interp2app doesn't think it's a normal method | 22:11 |
| Alex_Gaynor | w_subtype is the equivilant of `cls` in a @classmethod | 22:11 |
| mattip | Thanks. That, suprisingly enough, helped. | 22:12 |
| jterrace | Alex_Gaynor: the 'char' attribute of W_Dtype i dont think is a valid argument for runpack | 22:14 |
| jterrace | even when i make it an immutable field, it has to be a constant somehow | 22:14 |
| Alex_Gaynor | jterrace: it has to be a classattribute, it should be on itemtype | 22:14 |
| Alex_Gaynor | anyways | 22:15 |
| jterrace | you're saying to move it to types.py? | 22:15 |
| Alex_Gaynor | yes, this should be a method on each tyep | 22:15 |
| jterrace | ok | 22:16 |
| jterrace | but "char" attribute is on W_Dtype level, so ill move it down a level | 22:16 |
| Alex_Gaynor | not sure those are always the same thing | 22:16 |
| Alex_Gaynor | can we keep them seperate for now | 22:16 |
| jterrace | ok | 22:16 |
| davidpk (~davidpk@5e06d080.bb.sky.com) left irc: Quit: davidpk | 22:17 | |
| Alex_Gaynor | I'm pretty sure in the cases of composite dtypes, complexes, and strings they'll be different things | 22:17 |
| jterrace | yes | 22:17 |
| arigato (~arigo@89.204.153.160) left irc: Quit: See you | 22:17 | |
| fprimex (~brent@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: quit | 22:23 | |
| jterrace | whoo it translates | 22:24 |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 248 seconds | 22:27 | |
| antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) joined #pypy. | 22:30 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/168 | 22:35 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/976 | 22:35 |
| Insectoid (~Q@c-98-230-8-37.hsd1.fl.comcast.net) joined #pypy. | 22:36 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/742 | 22:36 |
| mattip | So I could have done linspace, but starting frompyfunc was more of a challenge. | 22:36 |
| mattip | I still need to add docstrings, and sleep on the implementation cuz it is now not lazy at all. | 22:37 |
| mattip | And if the "hg push" ever finishes, you can critique. | 22:39 |
| jterrace | are you having trouble pushing too? | 22:39 |
| jterrace | mine has been sitting there | 22:39 |
| rekamso (~textual@216.190.27.14) left irc: Ping timeout: 240 seconds | 22:39 | |
| jterrace | mattip: is yours stuck? | 22:40 |
| mattip | yes. Someone complained that it takes a loooooong time. | 22:40 |
| jterrace | grr | 22:41 |
| jterrace | mine has been sitting for quite a while doing nothing | 22:41 |
| mattip | 10 minutes and counting. | 22:45 |
| jterrace | bitbucket fail | 22:46 |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Remote host closed the connection | 22:47 | |
| mattip | I give up, will try again tomorrow. | 22:48 |
| mattip | Anyhow, there is a frompyfunc on the way. | 22:49 |
| mattip (chatzilla@bzq-79-177-112-183.red.bezeqint.net) left #pypy. | 22:49 | |
| lucian (~lucian@cpc1-newc15-2-0-cust84.gate.cable.virginmedia.com) joined #pypy. | 22:49 | |
| jterrace | nice | 22:51 |
| jterrace | it looks like the repo is locked | 22:51 |
| tilgovi (~randall@75-149-58-169-SFBA.hfc.comcastbusiness.net) joined #pypy. | 22:51 | |
| tilgovi (~randall@75-149-58-169-SFBA.hfc.comcastbusiness.net) left irc: Changing host | 22:51 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 22:51 | |
| fzzzy (~donovan@nat/mozilla/x-ukedxtcujsjcpbmn) joined #pypy. | 22:54 | |
| apanda (~apanda@dhcp-44-187.EECS.Berkeley.EDU) joined #pypy. | 22:55 | |
| reversefold (~papercran@75.101.111.82) joined #pypy. | 22:57 | |
| papercrane (~papercran@75.101.111.82) left irc: Ping timeout: 244 seconds | 22:57 | |
| Nick change: reversefold -> papercrane | 22:57 | |
| JaRoel|4d (~jaroel|4d@2001:980:7133:1:d45d:8c46:9aee:b3eb) left irc: Remote host closed the connection | 23:04 | |
| nickik (~nick@217.150.239.67) left irc: Quit: Verlassend | 23:08 | |
| papercrane (~papercran@75.101.111.82) left irc: Read error: Connection reset by peer | 23:13 | |
| papercrane (~papercran@75.101.111.82) joined #pypy. | 23:15 | |
| whitelynx (~whitelynx@63.241.75.144) left irc: Quit: Ex-Chat | 23:21 | |
| bbot2 | 4Finished ??: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/742 | 23:22 |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Quit: Konversation terminated! | 23:22 | |
| bgola (~bgola@c9518082.virtua.com.br) left irc: Ping timeout: 276 seconds | 23:38 | |
| CIA-13 | 03stefanor 07roundup * 10#673/Linux L2 cache size: | 23:38 |
| CIA-13 | Here's a quick patch I whipped up for the architectures that I can determine the L2 | 23:38 |
| CIA-13 | cache size on (i386, amd64, alpha, ppc, ia64, pa-risc, and s ... * 14https://bugs.pypy.org/issue673 | 23:38 |
| mfoord (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy. | 23:39 | |
| mfoord (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host | 23:39 | |
| mfoord (~anonymous@python/psf/voidspace) joined #pypy. | 23:39 | |
| jaredw (~jared@122-59-91-107.jetstream.xtra.co.nz) left irc: Ping timeout: 252 seconds | 23:40 | |
| dmalcolm (david@nat/redhat/x-fbwfkuafatpwdkyv) left irc: Quit: Leaving | 23:42 | |
| bgola (~bgola@c9518082.virtua.com.br) joined #pypy. | 23:45 | |
| jterrace (~jterrace@aegis.CS.Princeton.EDU) left irc: Quit: Leaving. | 23:45 | |
| DanKluev (~root@78-24-225-196-dsl.vntc.ru) left irc: Ping timeout: 252 seconds | 23:47 | |
| fzzzy (~donovan@nat/mozilla/x-ukedxtcujsjcpbmn) left irc: Quit: fzzzy | 23:56 | |
| azanella (~azanella@189.6.80.207) left irc: Ping timeout: 268 seconds | 23:56 | |
| --- Tue Dec 13 2011 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!