| Alex_Gaynor | amaury_: the slice fix wasn't enough, there are apparently even more bugs :/ | 00:02 |
|---|---|---|
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) left irc: Quit: ZOMG BYE | 00:09 | |
| canta (~canta@77-20-123-240-dynip.superkabel.de) left irc: Quit: Leaving. | 00:14 | |
| rekamso (~textual@216.190.27.14) left irc: Quit: Computer has gone to sleep. | 00:17 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/685 [12alex, list-strategies] | 00:20 |
| k_bx (~k_bx@94.244.19.62) left irc: Ping timeout: 252 seconds | 00:20 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 260 seconds | 00:26 | |
| mcdonc (~mcdonc@cabana.palladion.com) joined #pypy. | 00:30 | |
| dmsk (~dmsk@62.61.145.89.generic-hostname.arrownet.dk) left irc: Quit: Leaving | 00:31 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/700 | 00:34 |
| ronan__ (~ronan@host86-141-80-27.range86-141.btcentralplus.com) left irc: Ping timeout: 240 seconds | 00:47 | |
| justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) joined #pypy. | 00:49 | |
| stakkars_ (~tismer@213-65-251-34-no35.business.telia.com) left irc: Quit: stakkars_ | 00:50 | |
| stakkars_ (~tismer@213-65-251-34-no35.business.telia.com) joined #pypy. | 00:53 | |
| fijal | __pv: yes, but the current representation will probably be constant folded by the JIT to strides at a compile time | 00:56 |
| fijal | it's unclear whether we want to use this explicitly (maybe) | 00:57 |
| ronan__ (~ronan@host86-172-63-105.range86-172.btcentralplus.com) joined #pypy. | 01:00 | |
| fprimex | fijal! I want to get to contributing soon but work has kept me very busy | 01:01 |
| fprimex | hopefully next week I can do some more. can you give me a crash course some time? | 01:02 |
| fprimex (~fprimex@fl-67-233-128-10.dhcp.embarqhsd.net) left irc: Quit: AFT DECK! | 01:12 | |
| Nick change: Gulaway -> Gulopine | 01:15 | |
| rguillebert (~rguillebe@46.7.250.249) joined #pypy. | 01:17 | |
| etrepum (~bob@accessnat4.mochimedia.net) left irc: Ping timeout: 248 seconds | 01:20 | |
| rekamso (~textual@c-71-197-238-154.hsd1.wa.comcast.net) joined #pypy. | 01:27 | |
| kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy. | 01:35 | |
| mcdonc (~mcdonc@cabana.palladion.com) left irc: Ping timeout: 258 seconds | 01:47 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1821 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/554 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/599 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1084 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/252 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/785 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1417 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/686 | 01:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/250 | 01:47 |
| Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Ping timeout: 240 seconds | 01:51 | |
| aisipos (~aisipos@wc.lax.truecarcorp.com) left irc: Ping timeout: 276 seconds | 01:52 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 248 seconds | 01:54 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) left irc: Ping timeout: 255 seconds | 02:01 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy. | 02:02 | |
| DasIch (~dasich@p4FFDF22B.dip.t-dialin.net) left irc: Read error: Operation timed out | 02:12 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/928 [12alex, list-strategies] | 02:17 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/929 | 02:17 |
| rguillebert (~rguillebe@46.7.250.249) left irc: Ping timeout: 260 seconds | 02:20 | |
| kushal (~kdas@114.143.165.240) joined #pypy. | 02:25 | |
| kushal (~kdas@114.143.165.240) left irc: Changing host | 02:25 | |
| kushal (~kdas@fedora/kushal) joined #pypy. | 02:25 | |
| DasIch (~dasich@p4FFDE996.dip.t-dialin.net) joined #pypy. | 02:27 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 02:29 | |
| ericflo (~ericflo@75.103.8.110) left irc: Quit: ericflo | 02:30 | |
| rguillebert (~rguillebe@46.7.250.249) joined #pypy. | 02:33 | |
| mwhudson (~mwh@linaro/mwhudson) joined #pypy. | 02:49 | |
| mwhudson (~mwh@linaro/mwhudson) left irc: Remote host closed the connection | 02:51 | |
| Guest60050 (~John@osbk-4db157fb.pool.mediaWays.net) left irc: Ping timeout: 240 seconds | 02:51 | |
| Guest60050 (~John@osbk-4db14547.pool.mediaWays.net) joined #pypy. | 02:52 | |
| kushal (~kdas@fedora/kushal) left irc: Quit: This computer has gone to sleep | 02:53 | |
| rguillebert (~rguillebe@46.7.250.249) left irc: Quit: Quitte | 02:53 | |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy. | 02:55 | |
| daniloaf (~daniloaf@187.114.211.191) left irc: Quit: Fui embora | 03:19 | |
| justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) left irc: Ping timeout: 252 seconds | 03:42 | |
| mcdonc (~mcdonc@cabana.palladion.com) joined #pypy. | 03:43 | |
| mcdonc (~mcdonc@cabana.palladion.com) left irc: Remote host closed the connection | 04:10 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/125 | 04:34 |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy | 04:36 | |
| berdario (~quassel@host66-14-dynamic.52-82-r.retail.telecomitalia.it) joined #pypy. | 04:37 | |
| kushal (kdas@nat/redhat/x-kewgjzfbwlindkbc) joined #pypy. | 04:56 | |
| kushal (kdas@nat/redhat/x-kewgjzfbwlindkbc) left irc: Changing host | 04:56 | |
| kushal (kdas@fedora/kushal) joined #pypy. | 04:56 | |
| mitchell1 (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 05:10 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Read error: Connection reset by peer | 05:10 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 05:11 | |
| mitchell1 (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Read error: Connection reset by peer | 05:11 | |
| mitchell1 (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 05:13 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Read error: Connection reset by peer | 05:13 | |
| mitchell1 (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Read error: Connection reset by peer | 05:14 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 05:15 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Read error: Connection reset by peer | 05:21 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) joined #pypy. | 05:21 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1417 | 05:31 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/785 | 05:31 |
| berdario (~quassel@host66-14-dynamic.52-82-r.retail.telecomitalia.it) left irc: Ping timeout: 260 seconds | 05:49 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1084 | 05:49 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/599 | 05:54 |
| asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) left irc: Quit: asmeurer | 05:57 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/554 | 06:00 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/252 | 06:17 |
| alexsuraci (u1962@gateway/web/irccloud.com/x-npxtfptmiuuxjgoa) left irc: Ping timeout: 248 seconds | 06:27 | |
| mitchellh (~mitchellh@c-71-202-125-40.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 06:33 | |
| rekamso (~textual@c-71-197-238-154.hsd1.wa.comcast.net) left irc: Quit: Computer has gone to sleep. | 06:43 | |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy. | 06:50 | |
| alexsuraci (u1962@gateway/web/irccloud.com/x-nhitmfzlkwvfcrby) joined #pypy. | 06:51 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy. | 06:54 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/250 | 06:58 |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 248 seconds | 07:05 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/686 | 07:09 |
| canta (~canta@77-20-123-240-dynip.superkabel.de) joined #pypy. | 07:19 | |
| mvt (~mvantelli@87.213.45.85) joined #pypy. | 07:21 | |
| k_bx (~k_bx@94.244.19.62) joined #pypy. | 07:35 | |
| antocuni_ (~antocuni@host68-126-dynamic.14-87-r.retail.telecomitalia.it) joined #pypy. | 07:37 | |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) left irc: Quit: fzzzy | 07:45 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy. | 07:53 | |
| Arfrever (~Arfrever@apache/committer/Arfrever) joined #pypy. | 07:59 | |
| EnCuKou (~encukou@a91-156-243-246.elisa-laajakaista.fi) left irc: Remote host closed the connection | 08:00 | |
| k_bx (~k_bx@94.244.19.62) left irc: Ping timeout: 260 seconds | 08:05 | |
| costy_petrisor (c358b60a@gateway/web/freenode/ip.195.88.182.10) joined #pypy. | 08:18 | |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 08:24 | |
| canta (~canta@77-20-123-240-dynip.superkabel.de) left irc: Quit: Leaving. | 08:36 | |
| fijal (~fijal@178-36-132-231.adsl.inetia.pl) left irc: Ping timeout: 258 seconds | 08:38 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/929 | 08:41 |
| pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) left irc: Ping timeout: 255 seconds | 08:43 | |
| pedronis (~pedronis@73-53.195-178.cust.bluewin.ch) joined #pypy. | 08:47 | |
| Trundle (~andy@p578bfdcf.dip0.t-ipconnect.de) joined #pypy. | 08:49 | |
| Trundle (~andy@p578bfdcf.dip0.t-ipconnect.de) left irc: Changing host | 08:49 | |
| Trundle (~andy@python/site-packages/trundle) joined #pypy. | 08:49 | |
| arigato (~arigo@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 09:11 | |
| lucian (~lucian@93-97-174-115.zone5.bethere.co.uk) joined #pypy. | 09:12 | |
| kenaan | 12arigo default 11eb27c44ca6ad 15/pypy/jit/backend/x86/: - add in the backend, for binary instructions, a memo function that returns True if there is any NAME_xy that cou... | 09:12 |
| kenaan | 12arigo default 115478d1f631fa 15/pypy/jit/backend/x86/: Add tests for two special cases of "MOV" in INSN(). | 09:12 |
| kenaan | 12arigo default 11665b14e5263a 15/pypy/jit/metainterp/optimizeopt/: merge heads | 09:12 |
| arigato | hakanardo: there is still one failure in test_optimizeopt.py | 09:13 |
| k_bx (~k_bx@195.20.130.1) joined #pypy. | 09:15 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 09:18 | |
| G2P (~G2P@fw-asn1.ornis.com) joined #pypy. | 09:22 | |
| kenaan | 12arigo stm 11755507f9382b 15/pypy/: Improve targetdemo. | 09:26 |
| Da_Blitz (~Da_Blitz@203.56.250.63) left irc: Read error: Connection reset by peer | 09:28 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy. | 09:29 | |
| stakkars_ | good morning, guys. I'm a bit late, | 09:30 |
| stakkars_ | will come around 11:15 | 09:30 |
| stakkars_ | arigato: hi, is Laura not there? | 09:31 |
| stakkars_ | jacob22_: good morning | 09:31 |
| asabil (~asabil@cm-84.208.74.38.getinternet.no) joined #pypy. | 09:34 | |
| arigato | stakkars_: hi | 09:34 |
| arigato | yes, she's around somewhere | 09:34 |
| stakkars_ | will do a phone call and then come over. | 09:35 |
| ojii (~ojii@40-34.60-188.cust.bluewin.ch) joined #pypy. | 09:40 | |
| k_bx (~k_bx@195.20.130.1) left irc: Remote host closed the connection | 09:41 | |
| jacob22_ | stakkars_: Goodmorning. | 09:44 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1821 | 09:50 |
| kenaan | 12hakanardo default 11d0466dedbb14 15/pypy/jit/metainterp/optimizeopt/optimizer.py: interning ints aswell | 09:51 |
| kenaan | 12hakanardo default 119f81b789732c 15/pypy/jit/metainterp/optimizeopt/test/test_optimizeopt.py: test short preamble and non constant case aswell | 09:51 |
| kenaan | 12hakanardo default 11ce8c2eb5ccba 15/pypy/jit/metainterp/optimizeopt/optimizer.py: corner case not handled very well | 09:51 |
| kenaan | 12hakanardo default 115fb2ee9b17b4 15/pypy/jit/metainterp/optimizeopt/optimizer.py: dissable for now, it makes test_convert_from_SmallFunctionSetPBCRepr_to_FunctionsPBCRep fail | 09:51 |
| kenaan | 12hakanardo default 11569f16f25b1b 15/pypy/jit/metainterp/optimizeopt/: alternative fix that does not rely on interning ints | 09:51 |
| kenaan | 12hakanardo default 118b75e3ece413 15/pypy/jit/metainterp/optimizeopt/: allow setarrayitem to update the cache exported from the preamble to the loop the same way setfield does | 09:51 |
| kenaan | 12hakanardo default 119aea5197a2d6 15/pypy/jit/backend/x86/: hg merge | 09:51 |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 09:54 | |
| cfbolz | hello | 09:54 |
| arigato | hi | 09:54 |
| cfbolz | arigato: more sprinters now? | 09:55 |
| arigato | hakanardo just showed up | 09:55 |
| arigato | still waiting for anto | 09:55 |
| arigato | and christian | 09:55 |
| antocuni_ | how is the temperature btw? | 09:55 |
| arigato | :-) | 09:57 |
| hakanardo_ (~hakan@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 09:57 | |
| arigato | marc replies: 11 +/- 2 ° | 09:57 |
| antocuni_ | not too bad :-) | 09:58 |
| antocuni_ | thanks | 09:59 |
| arigato | :-) | 09:59 |
| stakkars_ (~tismer@213-65-251-34-no35.business.telia.com) left irc: Quit: schnarch | 10:00 | |
| DanKluev (~root@78-24-231-218-udsl.vntc.ru) left irc: Quit: Leaving. | 10:03 | |
| DanKluev (~root@78-24-231-218-udsl.vntc.ru) joined #pypy. | 10:03 | |
| fijal (~fijal@94-75-88-124.home.aster.pl) joined #pypy. | 10:08 | |
| fijal | hi | 10:08 |
| arigato | hi | 10:08 |
| fijal | arigato: I'm not coming :( | 10:09 |
| arigato | ok | 10:09 |
| fijal | not completely ok | 10:09 |
| fijal | but letting you know | 10:09 |
| intgr (ack@zombie.life.ee) left #pypy. | 10:11 | |
| fijal | arigato: what's on the sprint so far? | 10:11 |
| kenaan | 12arigo stm 1165eb6e47e3b4 15/pypy/translator/stm/: Tweaks. | 10:11 |
| arigato | I suppose it's pointless to ask you for a justification so I just accept it | 10:12 |
| fijal | not that I have to justify it obviously :) | 10:12 |
| arigato | of course | 10:12 |
| fijal | but anyway, combination of issues really | 10:12 |
| fijal | a combination | 10:12 |
| antocuni_ | fijal: :-(( | 10:12 |
| fijal | I could come on Sat, but then it breaks some of my other plans | 10:13 |
| verte (~verte@python/site-packages/verte) joined #pypy. | 10:14 | |
| fijal | since originally I planned to live the sprint on 8th | 10:14 |
| fijal | anyway | 10:15 |
| fijal | I should work on multidim arrays and the release | 10:15 |
| kvda (~kvda@124-169-130-71.dyn.iinet.net.au) joined #pypy. | 10:18 | |
| kenaan | 12arigo stm 110524190818dc 15/pypy/translator/stm/: Yay, the first example of RPython program that runs successfully on multiple threads. | 10:18 |
| verte | arigato: well done | 10:18 |
| cfbolz | arigato: is it any faster? :-) | 10:18 |
| cfbolz | (just kidding) | 10:18 |
| kvda (~kvda@124-169-130-71.dyn.iinet.net.au) left irc: Client Quit | 10:19 | |
| kvda (~kvda@124-169-130-71.dyn.iinet.net.au) joined #pypy. | 10:20 | |
| fijal | arigato: do you feel like reviewing justin's branch about memory pressure? | 10:24 |
| xorAxAx | cfbolz: it runs in O(1) :) | 10:25 |
| xorAxAx | because of the sleep | 10:25 |
| xorAxAx | if you neglect the spawn time | 10:25 |
| arigato | ah right, I can fix this easily without using locks | 10:26 |
| fijal | arigato: ok, I'll review it, didn't know you're the only core dev on the sprint | 10:28 |
| kenaan | 12arigo stm 1165545adde075 15/pypy/translator/stm/test/targetdemo.py: A poor man's lock: just use a regular counter and check it every second. As it's all protected by STM it works nicely. | 10:28 |
| arigato | fijal: so far, antonio is not here | 10:28 |
| arigato | fijal: as usual there are live discussions going on, sorry for not being present (this sentence was interrupted twice while I was writing it too :-) | 10:30 |
| fijal | it's ok :) | 10:30 |
| fijal | arigato: feel free to offload some work that can be done remotely to me | 10:30 |
| xorAxAx | hehe, "collected" | 10:32 |
| xorAxAx | is an interesting swedish name | 10:32 |
| lmoura (~lmoura@177.40.185.9) joined #pypy. | 10:33 | |
| antocuni_ | arigato: just to be sure that you are not waiting for me today: I'm arriving tomorrow | 10:34 |
| kenaan | 12hakanardo default 11d9708bf78c40 15/pypy/jit/metainterp/optimizeopt/: break up circular dependencies among short_boxes and give up | 10:34 |
| arigato | ah sorry, misread the people.txt | 10:35 |
| antocuni_ | around 12:30 | 10:35 |
| arigato | ok | 10:35 |
| kenaan | 12arigo stm 11b30caa32b11c 15/pypy/translator/stm/: kill these two C functions. | 10:38 |
| kenaan | 12arigo stm 11b4212b951b97 15/pypy/translator/stm/src_stm/et.c: Break a line that is definitely too long in the log. | 10:38 |
| arigato | cfbolz: stm: I would expect this simple example to have the same complexity as the C examples, i.e. around 6-7 times slower | 10:39 |
| arigato | so in order to see any speed-up you'd need at least 8 real cores | 10:40 |
| cfbolz | arigato: so you only need an 8-core? | 10:40 |
| arigato | maybe | 10:40 |
| cfbolz | :-) | 10:40 |
| cfbolz | and something else than boehm | 10:40 |
| arigato | tannit is probably not enough, as it's an 8-core 4-cpus | 10:40 |
| arigato | true, although the pressure on the GC should be reasonable in this demo | 10:40 |
| Da_Blitz | do overheads go up as you add cores? | 10:41 |
| arigato | not that I know | 10:41 |
| fijal | arigato: have you seen a recent comment on pypy's blog about STM? | 10:41 |
| fijal | "We are actually working on implementing this directly into stailaOS." | 10:41 |
| fijal | and a link | 10:41 |
| fijal | http://blog.staila.com/?p=81 | 10:41 |
| cfbolz | bit obscure | 10:41 |
| cfbolz | also seems not open source | 10:41 |
| arigato | Da_Blitz: basically the main constant per-core overhead is the need to transfer the global "version number" between cores | 10:41 |
| arigato | Da_Blitz: which means that the current stm implementation does not really really scale to large numbers of cores | 10:42 |
| arigato | but the limit is probably large enough for now :-) | 10:42 |
| xorAxAx | "It is an ideal solution for applications like real-time analytics, high frequency trading systems and many more." | 10:43 |
| verte | hopefully, with the jit involved, it will be possible to step up the grain size | 10:45 |
| arigato | a bit no clue what "this" means in the sentence "We are actually working on implementing this directly into stailaOS." | 10:45 |
| arigato | verte: yes | 10:45 |
| santagada (~leonardo@189.27.169.123.dynamic.adsl.gvt.net.br) joined #pypy. | 10:46 | |
| fijal | arigato: I suppose "the topic of the blog post" | 10:46 |
| fijal | or "stm" | 10:46 |
| fijal | but well | 10:46 |
| fijal | what do I know | 10:46 |
| lizardo (~lizardo@189.2.128.130) joined #pypy. | 10:47 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 258 seconds | 10:53 | |
| jacob22_ (~jacob@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 10:54 | |
| hakanardo_ (~hakan@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 240 seconds | 10:54 | |
| stakkars_ (~tismer@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 10:57 | |
| fijal | arigato: what about a different approach - malloc would grow a flag add_memory_pressure that can be either a bool (True) or an int (size)? | 10:58 |
| arigato | why? | 10:59 |
| fijal | because we might do a different thing depending whether we allocate new stuff in nursery or not | 11:00 |
| fijal | hm | 11:00 |
| fijal | we can just check where the object is I suppose | 11:00 |
| fijal | I still wonder what to do with stuff that increases memory pressure, but lives in nursery | 11:00 |
| fijal | forcing major collection faster does not seem like a perfect solution | 11:01 |
| arigato | agreed | 11:01 |
| dalke1 (~Adium@c-2ec29d45-74736162.cust.telenor.se) joined #pypy. | 11:01 | |
| dalke1 | Hi! Had work this morning and figured that hour trip in to down and back for the sprint wouldn't be that helpful. I'll be working on ctypes bindings for blas/lapack/etc. | 11:03 |
| JaredW | hi, what's the status of "the STM Vinnova project" | 11:03 |
| fijal | dalke1: hi | 11:03 |
| arigato | dalke1: great | 11:04 |
| dalke (~Adium@c-2ec29d45-74736162.cust.telenor.se) joined #pypy. | 11:06 | |
| dalke1 (~Adium@c-2ec29d45-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 11:06 | |
| arigato | JaredW: hi | 11:08 |
| arigato | nothing is known so far | 11:09 |
| canta (~canta@77-20-123-240-dynip.superkabel.de) joined #pypy. | 11:17 | |
| fijal (~fijal@94-75-88-124.home.aster.pl) left irc: Ping timeout: 252 seconds | 11:29 | |
| unbit (~unbit@93-58-162-185.ip159.fastwebnet.it) left irc: Quit: unbit | 11:44 | |
| dalke1 (~Adium@c-2ec29d45-74736162.cust.telenor.se) joined #pypy. | 11:50 | |
| dalke (~Adium@c-2ec29d45-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 11:50 | |
| dalke1 (~Adium@c-2ec29d45-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 11:51 | |
| dalke (~Adium@c-2ec29d45-74736162.cust.telenor.se) joined #pypy. | 11:51 | |
| Nick change: ojii -> ojii|food | 11:54 | |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 11:57 | |
| kenaan | 12hakanardo default 11a27a481ec877 15/pypy/module/: only use a single counter in xrange iterators (should save a setitem) | 11:59 |
| witulski (stupsi@fwstups.cs.uni-duesseldorf.de) left #pypy. | 12:03 | |
| antocuni_ (~antocuni@host68-126-dynamic.14-87-r.retail.telecomitalia.it) left irc: Ping timeout: 260 seconds | 12:06 | |
| Nick change: ojii|food -> ojii|busy | 12:19 | |
| antocuni_ (~antocuni@host68-126-dynamic.14-87-r.retail.telecomitalia.it) joined #pypy. | 12:21 | |
| amaury_ (~amaury_@74.125.57.34) joined #pypy. | 12:21 | |
| Rhy0lite (dje@nat/ibm/x-loatbeckkpypsdko) joined #pypy. | 12:23 | |
| EnCuKou (~encukou@193.167.65.94) joined #pypy. | 12:25 | |
| kenaan | 12arigo stm 11e76f2b79fd27 15/pypy/translator/stm/src_stm/: Baaaaah. setjmp() cannot be called on a jmp_buf that belongs to a parent frame, because then longjmp()ing to it will n... | 12:26 |
| kenaan | 12hakanardo step-one-xrange 116cf1ae5ff5d6 15/pypy/module/__builtin__/functional.py: special case xrange without any step specified | 12:28 |
| kvda (~kvda@124-169-130-71.dyn.iinet.net.au) left irc: Quit: -___- | 12:30 | |
| aboudreault (~alanb@osgeo/member/aboudreault) joined #pypy. | 12:37 | |
| fijal (~fijal@178-36-132-231.adsl.inetia.pl) joined #pypy. | 12:39 | |
| mwhudson (mwh@linaro/mwhudson) joined #pypy. | 12:39 | |
| Alex_Gaynor | fijal: we aren't ready to release, a freshly translated pypy can't be used to translate again, it's another str bug | 12:54 |
| fijal | nice | 12:54 |
| espes (espes@2001:470:1f05:ca1::) left irc: Ping timeout: 240 seconds | 12:59 | |
| voidspace (~anonymous@python/psf/voidspace) joined #pypy. | 12:59 | |
| Alex_Gaynor | fijal: I've been using http://paste.pocoo.org/show/502233/ to see it, not my progress os far :/ | 13:00 |
| Action: Alex_Gaynor -> class | 13:00 | |
| G2P (~G2P@fw-asn1.ornis.com) left irc: Remote host closed the connection | 13:03 | |
| hakanardo_ (~hakan@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 13:04 | |
| cfbolz (~cfbolz@fwstups.cs.uni-duesseldorf.de) left irc: Ping timeout: 245 seconds | 13:04 | |
| G2P (~G2P@fw-asn1.ornis.com) joined #pypy. | 13:05 | |
| espes (espes@2001:470:1f05:ca1::) joined #pypy. | 13:14 | |
| mat^2 (~mathias@212.130.113.35) joined #pypy. | 13:18 | |
| dalke | I have been investigating what needs to be done to get the BLAS bindings working. | 13:26 |
| dalke | BLAS is a set of low-level routines for linear algebra. It's implemented by a number of packages. | 13:27 |
| dalke | It's always set up as a library interface, with a specific API. | 13:27 |
| dalke | I implement a ctypes interface to "level 1" BLAS functions yesterday, and looked for some way to validate them. | 13:28 |
| dalke | The validation suite is a FORTRAN program which calls the library. Having Fortran call pypy? Mmmm, not so much. | 13:28 |
| dalke | SciPy has some validation code for the blas interface. I pulled it out, and realized that they have a slightly higher-level interface to blas. For example, | 13:29 |
| dalke | srotg() in blas takes 4 arguments. Two are inputs (specified as pointers, since that's what FORTRAN does) and two are outputs. | 13:29 |
| dalke | SciPy makes this be a function which takes two parameters and returns two parameters. | 13:30 |
| dalke | So I dug into how SciPy does it. It's through the f2py program. One thing that f2py does is create a C extension from an interface file, written in the "pyf" format. | 13:31 |
| fijal | nice | 13:32 |
| dalke | Only, SciPy doesn't have the needed pyf files. Instead, it generates them from a set of templates. This is because blas defines a family of functions; srotg takes single precision floats, drotg takes double precision floats, and there are other prefixes for other data types. | 13:32 |
| fijal | yes, scipy has some obscure obscure templating systems | 13:33 |
| dalke | The template -> fpy translation is in scipy/linalg/interface_gen.py . There's a __main__ function, but it doesn't work. Remove the "fblas" from the process_all() function. | 13:33 |
| dalke | Now that that's in place, I'm looking to see if I can use the f2py parser to get the code needed to generate a Python + ctypes interface to the library. | 13:34 |
| dalke | Change "some" into "multiple" | 13:34 |
| fijal | dalke: good work btw :) | 13:35 |
| EnCuKou (~encukou@193.167.65.94) left irc: Remote host closed the connection | 13:39 | |
| __pv | porting f2py would be great, it's used quite a lot in wrapping ancient fortran codes. however, its implementation is a bit ... special, so good luck ;) | 13:39 |
| dalke | Thanks. BTW, for anyone who wants to generate the f2py Python extension code, here's the command-line: | 13:40 |
| dalke | f2py ~/ftps/scipy-0.10.0b2/scipy/linalg/cblas.pyf --build-dir blah | 13:40 |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) joined #pypy. | 13:40 | |
| dalke | This would not be a port; I'm looking into writing an alternate back-end. | 13:40 |
| whitelynx (~whitelynx@li117-47.members.linode.com) joined #pypy. | 13:40 | |
| dalke1 (~Adium@c-2ec29d45-74736162.cust.telenor.se) joined #pypy. | 13:43 | |
| dalke (~Adium@c-2ec29d45-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 13:43 | |
| PiotrSikora (~none@nginx/adept/piotrsikora) left irc: Excess Flood | 13:46 | |
| PiotrSikora (~none@nginx/adept/piotrsikora) joined #pypy. | 13:47 | |
| kenaan | 12fijal rgc-mem-pressure 1192885c7cf7b6 15/pypy/module/_hashlib/interp_hashlib.py: optimize it slightly. not look up the dictionary each time we see _digest_size | 13:50 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1822 [12fijal, rgc-mem-pressure] | 13:52 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/687 [12fijal, rgc-mem-pressure] | 13:52 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-win-x86-32/builds/536 [12fijal, rgc-mem-pressure] | 13:52 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1085 [12fijal, rgc-mem-pressure] | 13:52 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/555 [12fijal, rgc-mem-pressure] | 13:52 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/251 [12fijal, rgc-mem-pressure] | 13:52 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/253 [12fijal, rgc-mem-pressure] | 13:52 |
| Nick change: Gulopine -> Gulaway | 13:53 | |
| fijal | once buildbots are done the branch is ready to be merged | 13:55 |
| kenaan | 12arigo stm 117a7ae4b45135 15/pypy/translator/stm/transform.py: hum. | 14:00 |
| kenaan | 12arigo stm 1116ac40bcfc6e 15/: hg merge default | 14:00 |
| kenaan | 12arigo stm 11e6d9748a9589 15/pypy/: A few extra operations that are always allowed. | 14:00 |
| arigato | "ah" | 14:03 |
| arigato | my stm demo crashes (rarely) if run repeatedly for long enough | 14:03 |
| arigato | but now I've figured out that the segfault occurs in boehm's malloc | 14:03 |
| fijal | not thread safe? | 14:03 |
| fijal | eh? | 14:04 |
| arigato | well it's supposed to be thread safe | 14:04 |
| fijal | "supposed" | 14:04 |
| arigato | yes | 14:04 |
| arigato | but well, we know that we already had troubles with Boehm | 14:04 |
| fijal | yes :( | 14:05 |
| arigato | I'm still 50-50 about blaming Boehm or my program that could corrupt memory randomly | 14:05 |
| arigato | ah, I can also try using --gc=none | 14:05 |
| Action: fijal expects armin to quit IRC shortly | 14:05 | |
| arigato | :-) | 14:06 |
| fijal | arigato: what's stopping us from merging the IO branch? | 14:06 |
| hakanardo_ (~hakan@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 248 seconds | 14:06 | |
| fijal | did you follow up | 14:06 |
| fijal | ? | 14:06 |
| arigato | good, with --gc=none at least this small program doesn't explode my RAM | 14:07 |
| fijal | https://bitbucket.org/pypy/pypy/pull-request/12/dont-lose-data-when-doing-non-blocking-i-o | 14:07 |
| fijal | this one | 14:07 |
| fijal | arigato: or you have nothing to do with it at all? | 14:07 |
| arigato | no, I didn't follow this branch | 14:07 |
| fijal | ok | 14:08 |
| fijal | well, it does something last time I looked | 14:08 |
| fijal | it *might* be non-exhaustive | 14:08 |
| fijal | I would still go for merging it, since it fixes some bugs | 14:08 |
| kenaan | 12hakanardo step-one-xrange 113aaee477e4be 15/pypy/module/pypyjit/test_pypy_c/test_misc.py: test ensuring xrange iterator only produces a single setitem | 14:09 |
| EnCuKou (~encukou@a91-156-243-246.elisa-laajakaista.fi) joined #pypy. | 14:09 | |
| arigato | if you reviewed it then it's ok I suppose | 14:10 |
| Action: fijal will look again | 14:11 | |
| fijal | I certainly did not review the last revision | 14:12 |
| arigato | and yes, I agree that it fixes bugs so it's good :-) | 14:12 |
| arigato | I'm a bit confused though | 14:13 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1418 [12hakanardo, step-one-xrange] | 14:13 |
| fijal | with what? | 14:15 |
| fijal | with "why can't we deal with it in a more systematic manner"? | 14:15 |
| fijal | or "why do we have 17 ways to buffer files"? | 14:15 |
| arigato | yes, about the systematic manner | 14:18 |
| fijal | I agree | 14:18 |
| fijal | but I guess it boils down to the second question | 14:19 |
| arigato | not necessarily | 14:19 |
| rguillebert (~rguillebe@46.7.250.249) joined #pypy. | 14:19 | |
| arigato | how is file.read() with no argument defined to work in CPython if it gets EAGAIN? | 14:20 |
| mcdonc (~mcdonc@cabana.palladion.com) joined #pypy. | 14:20 | |
| fijal | I think it tries to repeat | 14:20 |
| arigato | ah no, it does the same as 30fe8615eee1 | 14:21 |
| arigato | but it checks for EWOULDBLOCK in addition to EAGAIN | 14:22 |
| lizardo (~lizardo@189.2.128.130) left irc: Quit: Leaving | 14:22 | |
| antocuni__ (~antocuni@host44-121-dynamic.10-79-r.retail.telecomitalia.it) joined #pypy. | 14:22 | |
| arigato | but CPython does it only here, instead of everywhere | 14:23 |
| fijal | nice | 14:23 |
| fijal | but it does the same on the say socket no? | 14:23 |
| fijal | ah | 14:23 |
| antocuni_ (~antocuni@host68-126-dynamic.14-87-r.retail.telecomitalia.it) left irc: Ping timeout: 260 seconds | 14:23 | |
| fijal | you mean EWOULDBLOCK is only here? | 14:23 |
| kushal (kdas@fedora/kushal) left irc: Quit: This computer has gone to sleep | 14:23 | |
| arigato | I have no clue about sockets for now | 14:24 |
| verte (~verte@python/site-packages/verte) left irc: Quit: ~~~ Crash in JIT! | 14:25 | |
| rguillebert (~rguillebe@46.7.250.249) left irc: Ping timeout: 240 seconds | 14:26 | |
| arigato | I don't see anything that would justify e.g. bf0c77b5c906, because CPython doesn't seem to do it | 14:26 |
| kkris (~kris@93-82-36-194.adsl.highway.telekom.at) joined #pypy. | 14:28 | |
| arigato | hum, or more precisely, there is nothing specific to EAGAIN or EWOULDBLOCK about it in CPython | 14:29 |
| arigato | I agree that it's careful to not loose data | 14:29 |
| arigato | but it does so for any exception | 14:29 |
| arigato | ...or I'm not sure :-) | 14:30 |
| arigato | I *think* that in CPython if in the middle of your readline() you get EAGAIN, then the start of the line is lost | 14:30 |
| lmoura (~lmoura@177.40.185.9) left irc: Ping timeout: 252 seconds | 14:31 | |
| lizardo (~lizardo@189.2.128.130) joined #pypy. | 14:31 | |
| fijal | well | 14:31 |
| arigato | but I have no real clue, it needs to be checked carefully | 14:31 |
| fijal | I would say it's a CPython hard-to-reproduce bug | 14:31 |
| exarkun | reading from what kind of file? | 14:31 |
| arigato | my point is only that I'm trying to figure out if PyPy handles all the cases that CPython does handle, | 14:32 |
| exarkun | presumably if you do socket.makefile().readline(), an EAGAIN doesn't lose the beginning of the line - otherwise people would lose half their lines all the time. | 14:32 |
| arigato | and possibly create more CPython bug reports while I'm at it :-) | 14:32 |
| arigato | socket.makefile() is a completely different beast anyway | 14:33 |
| arigato | it's not at all a file object | 14:33 |
| canta (~canta@77-20-123-240-dynip.superkabel.de) left irc: Quit: Leaving. | 14:34 | |
| exarkun | so only normal on-disk files are in question here? It's certainly not impossible to get EAGAIN on a read from one of those, but it's definitely not very common. | 14:34 |
| arigato | it's based on sock.recv(), which I guess is not raising EAGAIN but EINTR | 14:34 |
| fijal | exarkun: NFS? | 14:34 |
| exarkun | fijal: Possibly sometimes, yea. | 14:34 |
| exarkun | Also fuse. | 14:34 |
| fijal | indeed it is less common | 14:34 |
| exarkun | I think the common surprising NFS error is ENOTCONN | 14:35 |
| exarkun | Normal filesystems like ext4 and VFAT probably never give you EAGAIN on a read though. | 14:35 |
| arigato | ...no, I'm just confused: looking at the CPython's Lib/socket.py, if a recv(1) gets you a EWOULDBLOCK then the line is lost | 14:35 |
| fijal | meh :/ | 14:37 |
| arigato | a-bit-no-clue | 14:38 |
| fijal | non pep-8 commits are annoying | 14:39 |
| mcdonc (~mcdonc@cabana.palladion.com) left irc: Remote host closed the connection | 14:40 | |
| Action: whitelynx can agree with that | 14:40 | |
| whitelynx | fijal: do you use vim? | 14:41 |
| fijal | nope | 14:41 |
| fprimex (~fprimex@brent-macbook.sc.fsu.edu) joined #pypy. | 14:41 | |
| whitelynx | aah ok | 14:42 |
| fijal | it's more than just that | 14:42 |
| fijal | creating strings by concatenation is not good either | 14:42 |
| mwhudson (mwh@linaro/mwhudson) left irc: Ping timeout: 248 seconds | 14:42 | |
| whitelynx | aah yeah | 14:42 |
| lizardo (~lizardo@189.2.128.130) left irc: Ping timeout: 260 seconds | 14:43 | |
| whitelynx | well, if anyone here uses vim and wants to check code with pep8 and pyflakes, I put together a vim plugin a while ago based on pyflakes.vim: http://projects.g33xnexus.com/pychecks-vim/wiki/ | 14:44 |
| whitelynx | i think it'll be moving to bitbucket soon, when i get time to convert the repo | 14:44 |
| lmoura (~lmoura@186.215.206.130) joined #pypy. | 14:44 | |
| fijal | whitelynx: I use the emacs equivalent | 14:45 |
| fijal | I'm using even | 14:45 |
| whitelynx | cool :) | 14:45 |
| fijal | well | 14:45 |
| fijal | the thing is I can see where it's wrong, I just don't want to fix all the places myself | 14:45 |
| whitelynx | that's understandable | 14:46 |
| lizardo (~lizardo@189.2.128.130) joined #pypy. | 14:48 | |
| lmoura (~lmoura@186.215.206.130) left irc: Ping timeout: 240 seconds | 14:49 | |
| lmoura (~lmoura@187.113.99.115) joined #pypy. | 14:49 | |
| arigato | fijal: well, in summary, someone needs to look carefully at what CPython guarantees and make sure we cover the same guarantees, and possibly report back as a CPython bug report the missing parts | 14:55 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1822 [12fijal, rgc-mem-pressure] | 14:55 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1823 [12hakanardo, step-one-xrange] | 14:55 |
| arigato | fijal: from this 10-minutes look it seems to me that just 30fe8615eee1 is enough to cover the same cases as CPython does (but EWOULDBLOCK is missing) | 14:56 |
| exarkun | fwiw, `errno.EWOULDBLOCK == errno.EAGAIN` is true most of the time ;) | 14:58 |
| arigato | I see :-) | 14:58 |
| kenaan | 12ctismer win64 test 1119ea93d6b3ae 15/: merge default | 14:58 |
| arigato | also, it can be tested directly: | 14:59 |
| arigato | you need your own subclass of Stream with a method read() that returns characters or raises EAGAIN at the correct time | 14:59 |
| arigato | (I didn't look at how it's tested in stefano's branch but it seems different) | 15:00 |
| arigato | (feel free to paste any of my comments back to the pull request) | 15:01 |
| fijal | arigato: I think amaury complained about direct tests | 15:08 |
| fijal | is there a way to take a few checkins and park them on a branch? | 15:09 |
| arigato | in this case, you cannot successfully hope to refactor streamio.py without direct tests | 15:09 |
| kenaan | 12ctismer win64_gborg 115179e1483dfb 15/: continuing win64 | 15:15 |
| mwhudson (mwh@linaro/mwhudson) joined #pypy. | 15:15 | |
| kkris (~kris@93-82-36-194.adsl.highway.telekom.at) left irc: Ping timeout: 276 seconds | 15:16 | |
| fijal | arigato: I agree | 15:17 |
| kkris (~kris@80-123-45-230.adsl.highway.telekom.at) joined #pypy. | 15:18 | |
| Alex_Gaynor | hey all :) | 15:21 |
| fijal | Alex_Gaynor: hey | 15:21 |
| fijal | Alex_Gaynor: descr_str is full of nonsense :( | 15:21 |
| Alex_Gaynor | fijal: so talk to matti, perhaps ask him to move it to a branch? | 15:22 |
| fijal | yes, preferably | 15:22 |
| fijal | he can finish fixing it, but then he should move | 15:22 |
| asabil (~asabil@cm-84.208.74.38.getinternet.no) left irc: Ping timeout: 260 seconds | 15:23 | |
| dalke (~Adium@c-2ec29d45-74736162.cust.telenor.se) joined #pypy. | 15:27 | |
| dalke1 (~Adium@c-2ec29d45-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 15:27 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy. | 15:28 | |
| kenaan | 12fijal numpy-multidim 11fa14e6831e42 15/pypy/module/micronumpy/: kill a function that was only used by one test. a bit pep-8ify, not too much though | 15:29 |
| Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Disconnected by services | 15:36 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy. | 15:38 | |
| amaury_ (~amaury_@74.125.57.34) left irc: Ping timeout: 240 seconds | 15:42 | |
| mvt (~mvantelli@87.213.45.85) left irc: Quit: Leaving | 15:59 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/687 [12fijal, rgc-mem-pressure] | 16:01 |
| jimbaker (~jbaker@canonical/jimbaker) left irc: Quit: Coyote finally caught me | 16:02 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) joined #pypy. | 16:03 | |
| jimbaker (~jbaker@c-75-71-80-146.hsd1.co.comcast.net) left irc: Changing host | 16:03 | |
| jimbaker (~jbaker@canonical/jimbaker) joined #pypy. | 16:03 | |
| hakanardo_ (~hakan@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy. | 16:03 | |
| Nick change: ojii|busy -> ojii | 16:05 | |
| mwhudson (mwh@linaro/mwhudson) left irc: Ping timeout: 240 seconds | 16:06 | |
| asmeurer (~asmeurer@dhcp-baca-230.resnet.nmt.edu) joined #pypy. | 16:07 | |
| fijal | good | 16:10 |
| bivab (~david@fwstups.cs.uni-duesseldorf.de) left irc: Quit: bivab | 16:14 | |
| tilgovi (~randall@75.101.111.78) joined #pypy. | 16:22 | |
| tilgovi (~randall@75.101.111.78) left irc: Changing host | 16:22 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 16:22 | |
| Alex_Gaynor | fijal: so far the only way I have to reproduce this is to translate pypy itself, is there a saner way to get the JIT log than to urn with PYPYLOG=jit-log-opt:x.pypylog and pray my HDD doesn't explode? | 16:24 |
| fijal | no :( | 16:24 |
| fijal | we lack some reasonable API | 16:24 |
| fijal | well | 16:24 |
| fijal | you can try to reuse the hooks I put there on applevel? | 16:24 |
| rekamso (~textual@216.190.27.14) joined #pypy. | 16:25 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) joined #pypy. | 16:29 | |
| kenaan | 12hager ppc-jit-backend 11a82d9737ffe8 15/pypy/jit/backend/ppc/ppcgen/: (bivab, hager): Fixed enoying error which occurred at calls because of the backchain. | 16:35 |
| kenaan | 12hager ppc-jit-backend 110dbe1538b91a 15/pypy/jit/backend/ppc/ppcgen/: merge | 16:35 |
| fijal | can I use precompiled regexes in RPython? | 16:37 |
| Alex_Gaynor | fijal: you mean using rsre? | 16:37 |
| fijal | yeah | 16:37 |
| fijal | somehow | 16:37 |
| Alex_Gaynor | I think it uses the same code as the host CPython, but I doubt that's guarnteed | 16:38 |
| Alex_Gaynor | (or the host pypy) | 16:38 |
| fijal | er | 16:38 |
| fijal | no, in an RPython program | 16:38 |
| Alex_Gaynor | right, I'm saying if you compile it with the host I think you can still eval it with rsre | 16:39 |
| fzzzy (~donovan@76-198-130-19.lightspeed.mtvwca.sbcglobal.net) joined #pypy. | 16:39 | |
| Rhy0lite | Alex_Gaynor: see pvt | 16:39 |
| fijal | Rhy0lite: hi | 16:40 |
| Rhy0lite | hi | 16:40 |
| Alex_Gaynor | got it | 16:40 |
| mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 258 seconds | 16:45 | |
| JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection | 16:47 | |
| voidspace (~anonymous@python/psf/voidspace) left irc: Quit: voidspace | 16:48 | |
| ramusara (~ramusara@220.156.210.236.user.e-catv.ne.jp) left irc: Quit: Leaving... | 16:49 | |
| Alex_Gaynor | Anyone know how long it takes to do a `make bootstrap; make -k check` cycle on GCC? | 16:50 |
| fenrrir (~fenrrir@201009115189.user.veloxzone.com.br) joined #pypy. | 16:50 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1418 [12hakanardo, step-one-xrange] | 16:51 |
| Trundle (~andy@python/site-packages/trundle) left irc: Remote host closed the connection | 16:51 | |
| lucian (~lucian@93-97-174-115.zone5.bethere.co.uk) left irc: Read error: Connection reset by peer | 16:54 | |
| k_bx (~k_bx@195.20.130.1) joined #pypy. | 16:55 | |
| lmoura (~lmoura@187.113.99.115) left irc: Read error: Connection reset by peer | 16:59 | |
| lmoura (~lmoura@186.215.206.130) joined #pypy. | 17:00 | |
| Alex_Gaynor | a fresh GCC checkout of trunk is 1.5GB | 17:01 |
| G2P (~G2P@fw-asn1.ornis.com) left irc: Quit: Leaving. | 17:01 | |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 17:02 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1085 [12fijal, rgc-mem-pressure] | 17:03 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1086 [12hakanardo, step-one-xrange] | 17:03 |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) left irc: Read error: Connection reset by peer | 17:04 | |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy. | 17:04 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/555 [12fijal, rgc-mem-pressure] | 17:05 |
| kenaan | 12arigo stm 11497d967a02c3 15/pypy/translator/stm/: In-progress | 17:07 |
| kenaan | 12hakanardo jit-targets 11505538a47fdb 15/pypy/jit/: introduce targets that can be placed somewhere in a trace that can be used jump targets | 17:08 |
| daniloaf (~daniloaf@177.42.188.153) joined #pypy. | 17:13 | |
| stakkars_ (~tismer@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Quit: schnarch | 17:14 | |
| costy_petrisor (c358b60a@gateway/web/freenode/ip.195.88.182.10) left irc: Quit: Page closed | 17:15 | |
| Alex_Gaynor | hakanardo: the objective of this is to fold the tails of loops? effectively giving you diamond control flow | 17:17 |
| hakanardo_ | that's one objective yes | 17:18 |
| hakanardo_ | the other is a cleaner way to retrace loops | 17:18 |
| fijal | uuuu | 17:22 |
| fijal | scary | 17:22 |
| fijal | I think I have a bit too much fun with parsers | 17:23 |
| Alex_Gaynor | fijal: compiling GCC is scarier. | 17:23 |
| fijal | Alex_Gaynor: are you fixing it? | 17:23 |
| Alex_Gaynor | fijal: trying, I'm not even managing to compile it ATM though | 17:23 |
| fijal | http://buildbot.pypy.org/summary/longrepr?testname=modified&builder=pypy-c-jit-linux-x86-32&build=1085&mod=lib-python.modified-2.7.test.test_builtin | 17:24 |
| fijal | does that sound like some sort of an old bug? | 17:24 |
| fijal | I think I forgot to merge default | 17:24 |
| rguillebert (~rguillebe@46.7.250.249) joined #pypy. | 17:24 | |
| Alex_Gaynor | fijal: yes, that's an old bug, armin fixed it ages ago | 17:25 |
| fijal | good | 17:25 |
| fijal | seems rgc-mem-pressure is good to go | 17:26 |
| Alex_Gaynor | so hg merge it :) | 17:26 |
| Alex_Gaynor | cool, I now have a 1.6GB pypylog | 17:27 |
| Alex_Gaynor | fijal: eh, I think I might be killing tannit with a search in vim | 17:29 |
| hakanardo_ (~hakan@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 260 seconds | 17:30 | |
| apanda (~apanda@dhcp-45-222.EECS.Berkeley.EDU) joined #pypy. | 17:31 | |
| santagada (~leonardo@189.27.169.123.dynamic.adsl.gvt.net.br) left irc: Ping timeout: 240 seconds | 17:36 | |
| witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) left irc: Quit: Leaving. | 17:36 | |
| k_bx (~k_bx@195.20.130.1) left irc: Ping timeout: 252 seconds | 17:37 | |
| fenrrir (~fenrrir@201009115189.user.veloxzone.com.br) left irc: Ping timeout: 248 seconds | 17:41 | |
| mattip (~chatzilla@109.64.113.99) joined #pypy. | 17:44 | |
| mattip | fijal: ping | 17:44 |
| fijal | mattip: pong | 17:45 |
| fijal | I have few issues regarding the str changes | 17:45 |
| mattip | Hit me with it. | 17:45 |
| fijal | so | 17:46 |
| fijal | pep 8 is the first thing | 17:46 |
| fijal | he other thing is building strings with concatenation - it's not a great idea | 17:46 |
| fijal | either use str.join(list) | 17:46 |
| fijal | or StringBuilder | 17:46 |
| mattip | OK, would you like me to fix it? | 17:47 |
| fijal | yes please | 17:48 |
| mattip | looking at the code, what would be a better fit, StringBuilder vs join? | 17:49 |
| Alex_Gaynor | fijal: going through a 1.6GB logfile sucks | 17:49 |
| fijal | StringBuilder is better | 17:49 |
| mattip | OK, I will commit something a bit later. | 17:50 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1086 [12hakanardo, step-one-xrange] | 17:54 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-app-level-win-x86-32/builds/536 [12fijal, rgc-mem-pressure] | 17:57 |
| kushal (~kdas@fedora/kushal) joined #pypy. | 18:00 | |
| mattip (chatzilla@109.64.113.99) left #pypy. | 18:02 | |
| kenaan | 12arigo stm 11a45743e6ee4e 15/pypy/translator/stm/: Test and fix. | 18:03 |
| kenaan | 12arigo stm 11b27ec3dc59d2 15/pypy/translator/stm/: Yay! targetdemo is fixed and seems to be working. Added a test for it. | 18:03 |
| kenaan | 12hakanardo jit-targets 111860421891fe 15/pypy/jit/: use TargetToken to refere to a target | 18:06 |
| rguillebert (~rguillebe@46.7.250.249) left irc: Ping timeout: 276 seconds | 18:07 | |
| kushal (~kdas@fedora/kushal) left irc: Quit: This computer has gone to sleep | 18:07 | |
| fijal | arigato: ping | 18:08 |
| arigato | pong | 18:08 |
| fijal | how can I use regular expressions in RPython? | 18:09 |
| fijal | precompiled ones are fine | 18:09 |
| arigato | yes | 18:10 |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Remote host closed the connection | 18:10 | |
| arigato | you can precompile them using the 're' module | 18:10 |
| fijal | and then match on them should work? | 18:11 |
| arigato | see the tests | 18:11 |
| fijal | ok | 18:11 |
| fijal | I seem not to be able to put a bunch of them in a list | 18:11 |
| fijal | but hey, unrolling iterable | 18:11 |
| arigato | basically you need test_match.get_code(), which should ideally be moved outside the test file | 18:11 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/253 [12fijal, rgc-mem-pressure] | 18:12 |
| arigato | get_code() returns something RPython-ish that can be passed to rsre_core.match() | 18:12 |
| arigato | don't try to have real regexp objects in your RPython programs | 18:12 |
| arigato | I think that get_code() just returns a list of integers | 18:13 |
| fijal | a bit obscure | 18:13 |
| fijal | ok | 18:13 |
| arigato | no, it's just that the function should move somewhere else | 18:13 |
| arigato | so far it's only used in these tests | 18:14 |
| fijal | sre_compile_hacked is a very obscure hack | 18:14 |
| arigato | yes, completely | 18:14 |
| arigato | but a localized one :-) | 18:14 |
| arigato | I didn't find any other way to do that | 18:14 |
| fijal | rsre/__init__.py as a place? | 18:15 |
| arigato | ? | 18:16 |
| fijal | where should I put it | 18:16 |
| arigato | ah, no | 18:17 |
| arigato | our __init__.py files are empty as a rule | 18:17 |
| fijal | translator/platform/__init__.py is not | 18:18 |
| arigato | just make a new module | 18:18 |
| arigato | yes, it's the only exception and I'm not too happy about it :-) | 18:18 |
| Trundle (~andy@89.204.137.130) joined #pypy. | 18:19 | |
| Trundle (~andy@89.204.137.130) left irc: Changing host | 18:19 | |
| Trundle (~andy@python/site-packages/trundle) joined #pypy. | 18:19 | |
| fijal | :-) | 18:19 |
| fijal | it's also full of state | 18:19 |
| fijal | and I'm not too happy about that either | 18:19 |
| arigato | yes | 18:19 |
| fijal | I guess that was my idea :)\ | 18:19 |
| Alex_Gaynor | grrr, I still have no idea where the bug is :( | 18:20 |
| fijal | any idea about a name? | 18:20 |
| arigato | fijal: pick one :-) | 18:21 |
| rguillebert (~rguillebe@46.7.250.249) joined #pypy. | 18:21 | |
| fijal | pom pom pom | 18:21 |
| voidspace (anonymous@python/psf/voidspace) joined #pypy. | 18:22 | |
| costy_petrisor (~JDI@79.117.215.111) joined #pypy. | 18:24 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 18:25 | |
| fijal | arigato: rlib.rsre.rpy | 18:29 |
| fijal | ? | 18:29 |
| arigato | if you want | 18:30 |
| Alex_Gaynor | arigato: are you aware of the fact that translation doesn't bootstrap ATM? | 18:31 |
| stakkars_ (~tismer@213-65-251-34-no35.business.telia.com) joined #pypy. | 18:31 | |
| arigato | I remember reading it here, yes, but I didn't look. I think fijal said it's related to vstring stuff | 18:31 |
| Action: fijal didn't say that | 18:32 | |
| Alex_Gaynor | I think it's related to vstring, it looks like other vstring issues (i.e. characters being NUL for unknown reasons), but I haven't proved it | 18:32 |
| elipsion (~elipsion@81-232-46-166-no38.tbcn.telia.com) joined #pypy. | 18:35 | |
| arigato | try to disable vstring -- I think you can even do it by saying "pypy --jit something" | 18:35 |
| arigato | pypy --jit enable_opts=X | 18:36 |
| arigato | for X = the list of all optimizations apart from vstring | 18:36 |
| arigato | (helpfully) | 18:36 |
| Alex_Gaynor | good idea, unfortunately it takes ~45 minutes to reproduce the issue, the only thing I've found that shows it is translation so far | 18:37 |
| Alex_Gaynor | I'll try later this evening | 18:37 |
| arigato | even translating smaller things than pypy doesn't show the problem? | 18:37 |
| Alex_Gaynor | I'm not sure, I haven't tried yet :/ | 18:37 |
| arigato | ok | 18:38 |
| kenaan | 12fijal numpy-multidim 116d64103f1147 15/pypy/module/micronumpy/: make compile use a real tokenizer - breaks test_zjit for now | 18:38 |
| kenaan | 12fijal numpy-multidim 111582795d14f5 15/pypy/rlib/rsre/: shuffle code around. Now get_code lives outside of tests | 18:38 |
| ericflo (~ericflo@75.103.8.110) joined #pypy. | 18:40 | |
| mat^2 (~mathias@212.130.113.35) joined #pypy. | 18:41 | |
| pingveno (adevore@ruby.cat.pdx.edu) joined #pypy. | 18:41 | |
| jblaze_ipad (~jblazeipa@mobile-166-147-114-174.mycingular.net) joined #pypy. | 18:43 | |
| pingveno | Is it possible to cross compile from Solaris to Linux x86_64? I have a nice beefy server available to me for building for Linux. | 18:43 |
| fijal | arigato: why is group in rsre_core marked as NOT_RPYTHON? | 18:43 |
| pingveno | CS students don't take must processing power :) | 18:43 |
| fijal | hm | 18:48 |
| fijal | I nicely rendered it completely unusable | 18:48 |
| Rhy0lite | fijal: ping | 18:49 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1823 [12hakanardo, step-one-xrange] | 18:55 |
| fijal | Rhy0lite: pong | 18:55 |
| fijal | 6 minutes, not great | 18:55 |
| apanda (~apanda@dhcp-45-222.EECS.Berkeley.EDU) left irc: Ping timeout: 260 seconds | 18:57 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/251 [12fijal, rgc-mem-pressure] | 19:00 |
| Alex_Gaynor | Rhy0lite: pong | 19:03 |
| Alex_Gaynor | ping* | 19:03 |
| Sho_ (~EHS1@kde/hein) joined #pypy. | 19:05 | |
| elipsion | so, is this really where compilers are made or are you just making funny noises all day long? | 19:05 |
| MostAwesomeDude | elipsion: It can't be both? | 19:06 |
| apanda (~apanda@dhcp-45-222.EECS.Berkeley.EDU) joined #pypy. | 19:06 | |
| elipsion | MostAwesomeDude: that depends of the compiler/noise ratio | 19:06 |
| elipsion | after half an hour here, it's about 1:1 | 19:07 |
| MostAwesomeDude | elipsion: Well, you're looking through a narrow window; I've been here for months, and there's a *lot* of useful chatter. | 19:08 |
| MostAwesomeDude | Maybe people are just doing other things in life at this particular moment. :3 | 19:08 |
| elipsion | i kind of guessed ;) | 19:09 |
| elipsion | isn't there supposed to be a sprint going on as well? | 19:10 |
| MostAwesomeDude | Maybe. I don't track sprints since I'm geographically removed. | 19:11 |
| Alex_Gaynor | There is one. | 19:11 |
| MostAwesomeDude | I thought there was one this weekend, or maybe last weekend? | 19:11 |
| Rhy0lite | Alex_Gaynor: pong | 19:22 |
| Alex_Gaynor | Rhy0lite: I can't seem to get gcc to compile, any idea what "fatal error: bits/predefs.h: No such file or directory" indicates? google isn't helpful on the matter | 19:23 |
| Rhy0lite | gcc trunk is in sort of a bad state at the moment because a directory reorg yesterday | 19:24 |
| Alex_Gaynor | ah, perfect | 19:24 |
| exarkun | l2branch nubs | 19:24 |
| Alex_Gaynor | huh, svn up seems to have resolved it | 19:25 |
| k_bx (~k_bx@94.244.19.62) joined #pypy. | 19:25 | |
| Rhy0lite | ok | 19:25 |
| Alex_Gaynor | nevermind, still fails :) | 19:25 |
| Rhy0lite | Richi's proposed patch is fairly easy, I think | 19:26 |
| Rhy0lite | from his comment 18 in the bugzilla | 19:26 |
| Alex_Gaynor | I suspect so as well, but I'd like to get gcc compiling first :) | 19:26 |
| Rhy0lite | yes, but we can talk theoretically | 19:26 |
| Rhy0lite | basically change if (!DECL_ARTIFICIAL (lab)) | 19:27 |
| Alex_Gaynor | err, I gotta run to calss now,but I can do so in 20 minutes :) | 19:27 |
| daniloaf (~daniloaf@177.42.188.153) left irc: Quit: Fui embora | 19:27 | |
| Rhy0lite | to also test for the new flag | 19:27 |
| Rhy0lite | we can chat later | 19:27 |
| Rhy0lite | Alex's proposed change is more involved | 19:27 |
| kenaan | 12fijal numpy-multidim 11b6ce14bbf83d 15/pypy/: make parser not rpython (we'll think about it later) and use the same trick as we used in test_newgs. Runnin... | 19:29 |
| fijal | good | 19:33 |
| kenaan | 12fijal default 1119bc61988c39 15/pypy/: Merge rgc-mem-pressure. This branch adds memory pressure in some crucial points where C allocates a lot, but struct... | 19:35 |
| kenaan | 12fijal rgc-mem-pressure 110ede8b92968e 15/: closed merged branch | 19:35 |
| kenaan | 12fijal default 118ecb5f0cd990 15/pypy/: merge default | 19:35 |
| apanda (~apanda@dhcp-45-222.EECS.Berkeley.EDU) left irc: Read error: Connection reset by peer | 19:37 | |
| aurojit (~apanda@dhcp-45-222.EECS.Berkeley.EDU) joined #pypy. | 19:37 | |
| fijal | "good" | 19:37 |
| Action: fijal has a reasonable test_zjit for multidim arrays | 19:37 | |
| kushal (~kdas@114.143.164.178) joined #pypy. | 19:38 | |
| kushal (~kdas@114.143.164.178) left irc: Changing host | 19:38 | |
| kushal (~kdas@fedora/kushal) joined #pypy. | 19:38 | |
| Rhy0lite (dje@nat/ibm/x-loatbeckkpypsdko) left irc: Quit: Leaving | 19:39 | |
| Alex_Gaynor | fijal: how long does it take? | 19:43 |
| kenaan | 12mattip numpy-multidim 117dbb08ac308d 15/pypy/module/micronumpy/interp_numarray.py: pep-8, use StringBuilder | 19:46 |
| kenaan | 12mattip numpy-multidim 118e0658ce330e 15/pypy/: Merge str and repr cleanup | 19:46 |
| mattip (~chatzilla@109.64.113.99) joined #pypy. | 19:47 | |
| rguillebert (~rguillebe@46.7.250.249) left irc: Ping timeout: 256 seconds | 19:47 | |
| mattip | fijal: Take a look? | 19:47 |
| kushal (~kdas@fedora/kushal) left irc: Quit: This computer has gone to sleep | 19:48 | |
| fijal | Alex_Gaynor: how long does what take? | 19:57 |
| fijal | mattip: ok | 19:57 |
| Alex_Gaynor | fijal: running zjit tests | 19:57 |
| fijal | Alex_Gaynor: not any longer | 19:58 |
| Alex_Gaynor | fijal: so that's what, 30 seconds or something? | 19:58 |
| fijal | 19s one | 19:59 |
| fijal | 25 all | 19:59 |
| Alex_Gaynor | cool | 19:59 |
| fijal | Alex_Gaynor: max/min is broken | 20:00 |
| fijal | it creates slow bridges | 20:00 |
| Alex_Gaynor | on default? | 20:00 |
| kenaan | 12arigo stm 11e67329e4d516 15/pypy/: setarrayitem. | 20:00 |
| kenaan | 12arigo stm 11104a651020e1 15/pypy/: getarrayitem. Disabled for now because it's missing the C impl. | 20:00 |
| fijal | well, yes | 20:00 |
| kenaan | 12arigo stm 11d5f6a1b6e66c 15/pypy/translator/stm/test/test_funcgen.py: Start a test for the complicated logic in funcgen.py, even though it mirrors closely the (tested) logic in rstm.py. | 20:00 |
| Alex_Gaynor | how do you make it not create bridges? | 20:00 |
| fijal | but tests on default are not updated | 20:00 |
| fijal | I dunno | 20:00 |
| fijal | or make bridges better? | 20:00 |
| Alex_Gaynor | are bridges that bad? | 20:00 |
| fijal | well | 20:00 |
| fijal | they fail the test | 20:00 |
| fijal | run it with --viewloops | 20:00 |
| fijal | on numpy-multidim | 20:01 |
| kenaan | 12hakanardo default 117202b0d9cb70 15/pypy/module/__builtin__/: fix an overlow bug | 20:03 |
| kenaan | 12hakanardo default 117cd8e99541db 15/pypy/: hg merge | 20:03 |
| aboudreault (~alanb@osgeo/member/aboudreault) left irc: Ping timeout: 258 seconds | 20:03 | |
| arigato (~arigo@c-c4c4e055.1321-1-64736c11.cust.bredbandsbolaget.se) left irc: Ping timeout: 244 seconds | 20:04 | |
| Action: fijal ponders if there is a way that avoids division | 20:17 | |
| Alex_Gaynor | fijal: a way o what? | 20:18 |
| fijal | way of handling multidimensional arrays | 20:18 |
| fijal | I'll read about strides | 20:18 |
| ericflo (~ericflo@75.103.8.110) left irc: Ping timeout: 252 seconds | 20:20 | |
| ericflo (~ericflo@75.103.8.110) joined #pypy. | 20:20 | |
| rekamso (~textual@216.190.27.14) left irc: Ping timeout: 245 seconds | 20:21 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy. | 20:21 | |
| jblaze_ipad (~jblazeipa@mobile-166-147-114-174.mycingular.net) left irc: Quit: Colloquy for iPad - http://colloquy.mobi | 20:24 | |
| Alex_Gaynor | fijal: someone explained something about them yesterday, apparently numpy uses the same storage for slices as arrays | 20:24 |
| dalke (~Adium@c-2ec29d45-74736162.cust.telenor.se) left irc: Ping timeout: 252 seconds | 20:25 | |
| fijal | well | 20:25 |
| mwhudson (mwh@linaro/mwhudson) joined #pypy. | 20:25 | |
| fijal | yes | 20:25 |
| fijal | if you modify a slice, you modify the original array | 20:25 |
| Alex_Gaynor | fijal: right but I'm saying they use the same C struct to represent them both | 20:26 |
| fijal | ah | 20:26 |
| fijal | ok | 20:26 |
| Alex_Gaynor | so perhaps we should look into copying their representation, as I bet it works well with fortran and other libs | 20:26 |
| fijal | the representation (as in memory layout) is well known | 20:27 |
| fijal | we use the same | 20:27 |
| Alex_Gaynor | even for slices? | 20:27 |
| fijal | one of the two anyway | 20:27 |
| fijal | slices don't introduce a new memory layout | 20:27 |
| fijal | and I don't think fortran code cares about any sort of meta-info | 20:28 |
| fijal | well, I *do* know | 20:28 |
| Alex_Gaynor | :) | 20:28 |
| dalke (~Adium@c-2ec2b474-74736162.cust.telenor.se) joined #pypy. | 20:32 | |
| fijal | can you always express a view of a view as a single view? | 20:32 |
| fijal | I guess yes | 20:32 |
| Alex_Gaynor | I hope so, but I don't swear to it | 20:33 |
| mattip | I am pretty sure it works, that's why my to_str recursively calls itself with a narrower view of the original ndarray | 20:34 |
| fijal | I know it works | 20:36 |
| fijal | I'm just thinking maybe we can make it easier | 20:36 |
| fijal | mattip: pep8 please as well :) | 20:37 |
| mattip | The 'numpy NDimArray branch implemented indexing wit a vector in setitem getitem rather than a single index, | 20:37 |
| mattip | (oops) | 20:37 |
| mattip | so there is no division. | 20:38 |
| fijal | mattip: what editor do you use? | 20:38 |
| mattip | vim, irc in ChatZilla | 20:38 |
| fijal | you can install yourself some sort of vim extension that will color stuff if it's not pep8 | 20:39 |
| fijal | it uses pep8 and pyflakes, based on flymake | 20:39 |
| fijal | super-useful | 20:39 |
| Action: fijal will look into strides tomorrow | 20:39 | |
| mattip | I did pep8 the interp_numarray.py, all the code I wrote passed. | 20:40 |
| mattip | I didn't want to touch stuff written by others, since the web site says | 20:41 |
| mattip | something about "almost pep-8, use the current code as an idea of what is accepted" | 20:41 |
| fijal | descr_repr misses spaces after commas | 20:43 |
| fijal | I can add them | 20:43 |
| fijal | mattip: otherwise looks good so far | 20:44 |
| kenaan | 12fijal numpy-multidim 11a5e0435c51ef 15/pypy/module/micronumpy/interp_numarray.py: pep8 | 20:44 |
| fijal | mattip: can't isinstance be substituted with a method? | 20:44 |
| fijal | also you don't have to check if to_str == '[]' | 20:45 |
| fijal | meh :/ | 20:45 |
| Action: fijal nicely crashed py.test | 20:45 | |
| fijal | mattip: I'll call it a day, it works nicely | 20:45 |
| aurojit (~apanda@dhcp-45-222.EECS.Berkeley.EDU) left irc: Remote host closed the connection | 20:46 | |
| mattip | :) | 20:46 |
| fijal | it definitely can be simplified though :) | 20:46 |
| fijal | but it's definitely beyond my current abilities | 20:47 |
| mattip | pleasant dreams | 20:47 |
| mattip (chatzilla@109.64.113.99) left #pypy. | 20:48 | |
| mat^2 (~mathias@212.130.113.35) left irc: | 20:50 | |
| kenaan | 12fijal pypy.org[extradoc] 11a86881b1306a 15/: update numbers | 20:51 |
| lizardo (~lizardo@189.2.128.130) left irc: Quit: Leaving | 20:53 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) left irc: Ping timeout: 240 seconds | 20:57 | |
| Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy. | 20:59 | |
| Insectoid (~Q@c-98-230-33-116.hsd1.fl.comcast.net) joined #pypy. | 21:02 | |
| Insectoid | I'm probably going to say something stupid -- I ask forgiveness ahead of time. | 21:02 |
| fijal | Insectoid: that's ok :) | 21:02 |
| timonator | yeah, i did that a few times myself and nobody was mad at me | 21:03 |
| fprimex (~fprimex@brent-macbook.sc.fsu.edu) left irc: Quit: alsjdfh | 21:04 | |
| Insectoid | So, I've been trying to build wxPython with PyPy. I've had some interesting luck. The first issue I ran across is that the PyPy headers define ob_refcnt as a long. An int was expected -- I think this may be a case of not using Py_ssize_t, which seems to be an int on windows and a long elsewhere, but I honestly don't know much about C so I could be wrong. | 21:05 |
| fijal | I don't actually know :) | 21:07 |
| fijal | Insectoid: is it actually a problem? | 21:07 |
| fijal | or just a warning? | 21:07 |
| Insectoid | Next, PyClassObject, the struct from pypy_decl, is missing several fields (or is if it's supposed to take the place of the one from classobject.h. wxPython is specifically complaining about cl_bases -- I presume the list of base classes, which doesn't appear to be there. | 21:08 |
| Insectoid | It is a problem, yes, I get build errors. | 21:08 |
| fijal | ah | 21:08 |
| fijal | hm | 21:08 |
| Insectoid | Should I pastebin a build? | 21:09 |
| fijal | http://morepypy.blogspot.com/2010/05/running-wxpython-on-top-of-pypy.html | 21:10 |
| fijal | have you seen this? | 21:10 |
| fijal | Insectoid: I doubt I can help | 21:10 |
| fijal | maybe amaury_ | 21:10 |
| fijal | ? | 21:10 |
| Insectoid | Yeah, that's what gave me the idea to even try this. | 21:10 |
| fijal | ok | 21:10 |
| fijal | did you apply the patch by chance? | 21:10 |
| exarkun | Insectoid: Open tickets in the issue tracker. | 21:10 |
| Insectoid | ... Foollishly, I must not have read enough. I remember reading that like a year ago, I'll go check out a patch now. Told you I'd say something stupid ;-) | 21:11 |
| amaury_ | Insectoid: on which platform are you compiling? | 21:12 |
| amaury_ | Insectoid: for PyClassObject, I added #ifdef in wxPython code | 21:13 |
| amaury_ | pypy is similar to Python3 in this area: no PyClassObject! | 21:13 |
| Insectoid | amaury_: Windows, i'm patching now. | 21:14 |
| Insectoid | amaury_: I this #ifdef in the patch, or should I do it manually? | 21:14 |
| amaury_ | it should be in the patch | 21:15 |
| Insectoid | Fantastic. If this works I'll go bug them in #wxPython to please integrate this patch if you don't mind. | 21:15 |
| amaury_ | wxPython is migrating to another system | 21:15 |
| amaury_ | not SWIG-based | 21:16 |
| Insectoid | The Phoenix thing? | 21:16 |
| amaury_ | yes | 21:16 |
| amaury_ | gives me headache | 21:16 |
| Insectoid | I didn't think that had actually gotten anywhere | 21:16 |
| amaury_ | They say it works | 21:16 |
| amaury_ | somewhat | 21:16 |
| amaury_ | but Phoenix uses metaclasses with a tp_alloc slot | 21:16 |
| amaury_ | and I'm fighting with cpyext to make this work | 21:17 |
| apanda (~apanda@dhcp-45-222.EECS.Berkeley.EDU) joined #pypy. | 21:19 | |
| fijal | amaury_: metaclasses in RPython are a mess :/ | 21:19 |
| amaury_ | in this case, it's not RPython | 21:19 |
| amaury_ | because types are discovered at runtime | 21:19 |
| amaury_ | IIRC it's not possible to define metaclasses in RPython | 21:23 |
| amaury_ | I tried a long time ago | 21:23 |
| jonanin (~jonanin@24-183-50-140.dhcp.mdsn.wi.charter.com) left irc: Remote host closed the connection | 21:23 | |
| derdon (~derdon@pD9E1DFDC.dip.t-dialin.net) joined #pypy. | 21:25 | |
| voidspace (anonymous@python/psf/voidspace) left irc: Quit: voidspace | 21:31 | |
| rekamso (~textual@63-228-127-70.dia.static.qwest.net) joined #pypy. | 21:33 | |
| amaury_ | Insectoid: did you manage to compile wxPython? | 21:34 |
| fijal | amaury_: well, we should fix that one day | 21:38 |
| amaury_ | only if it's useful | 21:39 |
| amaury_ | So far, we worked this around with applevel code | 21:39 |
| fijal | pfff | 21:40 |
| fijal | well, the trend is to move more to applevel anyway | 21:40 |
| aisipos (~aisipos@vpn.lax.truecarcorp.com) joined #pypy. | 21:41 | |
| dalke (~Adium@c-2ec2b474-74736162.cust.telenor.se) left irc: Ping timeout: 255 seconds | 21:44 | |
| dalke (~Adium@c-2ec22819-74736162.cust.telenor.se) joined #pypy. | 21:44 | |
| Insectoid | amaury_: Patch was a bit stale, finally got it sorted (not very experienced in C, sorry it took so long.) Now all seems to work until we get to linking, at which point it can't find python27.lib -- how do I tell it to link against libpypy instead of libpython? | 21:48 |
| fijal | Insectoid: you compile it using pypy right? | 21:48 |
| Insectoid | fijal: Yes. | 21:48 |
| amaury_ | Insectoid: isn't there a .lib file in pypy-win32.zip? | 21:49 |
| Insectoid | where would I find this? In the pypy snapshot I downloaded, in the root? There is none. | 21:50 |
| Insectoid | according to cygwin's find, there are no .lib files in pypy-c-jit-48653-82489bdede61-win32> | 21:51 |
| amaury_ | hum | 21:52 |
| fijal | there should be | 21:52 |
| kkris (~kris@80-123-45-230.adsl.highway.telekom.at) left irc: Remote host closed the connection | 21:52 | |
| daniloaf (~daniloaf@187.114.150.249) joined #pypy. | 21:52 | |
| amaury_ | no, it's not automatic | 21:52 |
| amaury_ | there is a libpypy-c.lib, but it stays in /tmp/usession/testing-1 | 21:53 |
| jokey | oi, approaching the 4,7 average again :) man this site is better than coffee in the morning :D | 21:54 |
| Insectoid | Should I be building with a different pypy? | 21:55 |
| fijal | amaury_: it should be bundled by package.py no? | 22:00 |
| amaury_ | we should, yes | 22:01 |
| amaury_ | but package.py does not look at /tmp/usession IIRC | 22:01 |
| amaury_ | and the driver is already too complicated for me | 22:02 |
| amaury_ | (the translation driver) | 22:02 |
| fijal | I can fix the driver | 22:02 |
| fijal | not today though | 22:02 |
| amaury_ | the funny part is that any version of libpypy-c.lib would work | 22:03 |
| dalke (~Adium@c-2ec22819-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 22:03 | |
| dalke (~Adium@c-2ec22819-74736162.cust.telenor.se) joined #pypy. | 22:03 | |
| dalke (~Adium@c-2ec22819-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 22:03 | |
| dalke1 (~Adium@c-2ec22819-74736162.cust.telenor.se) joined #pypy. | 22:03 | |
| amaury_ | it's even possible to build one quickly by hand | 22:03 |
| amaury_ | err, with the help of the compiler of course | 22:04 |
| Insectoid | What, if anything, should I do now? Evidently the snapshot builds do not require what I need. Do the release builds? Can I build it myself? Do I need to get PyPy from bitbucket and translate it? | 22:06 |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer | 22:06 | |
| dalke1 (~Adium@c-2ec22819-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 22:06 | |
| dalke (~Adium@c-2ec22819-74736162.cust.telenor.se) joined #pypy. | 22:06 | |
| gutworth (~benjamin@dhcp-128-36-89-186.central.yale.edu) left irc: Ping timeout: 255 seconds | 22:08 | |
| mwhudson (mwh@linaro/mwhudson) left irc: Read error: Operation timed out | 22:15 | |
| amaury_ | Insectoid: compiling pypy yourself is a good exercise anyway | 22:17 |
| amaury_ | and it's probably necessary to use the same compiler for pypy and wxPython | 22:18 |
| dalke1 (~Adium@c-2ec22819-74736162.cust.telenor.se) joined #pypy. | 22:21 | |
| dalke (~Adium@c-2ec22819-74736162.cust.telenor.se) left irc: Read error: Connection reset by peer | 22:21 | |
| Insectoid | amaury_: I'm concerned as I am on Windows x64 and only have 4G of ram. | 22:21 |
| tlynn (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) joined #pypy. | 22:22 | |
| amaury_ | you'll need python 32bit anyway | 22:22 |
| amaury_ | 4G is enough | 22:22 |
| Insectoid | Righto, so clone from bitbucket, translate and package it? Do I have to do anything special to keep this lib I need? | 22:24 |
| amaury_ | yes :-| | 22:24 |
| amaury_ | in %TMP%\usession-*\testing_1 | 22:25 |
| amaury_ | where all files are built | 22:25 |
| amaury_ | there will be a libpypy-c.lib | 22:25 |
| dalke1 (~Adium@c-2ec22819-74736162.cust.telenor.se) left irc: Ping timeout: 240 seconds | 22:25 | |
| amaury_ | your should copy it somewhere before this temp dir is deleted | 22:25 |
| Insectoid | Should I rename it to python27.lib? the linker command line specifically referenced that. | 22:26 |
| amaury_ | yes, probably | 22:26 |
| amaury_ | I'm sorry I used to know that very well | 22:26 |
| Insectoid | Thank you for your help. Clone in progress, we'll see how this goes. | 22:26 |
| amaury_ | but I don't have the win32 machine I had at the time | 22:27 |
| amaury_ | Insectoid: you'll probably want to install third party libraries as well | 22:27 |
| amaury_ | and compile them | 22:27 |
| amaury_ | zlib, bz2, expat | 22:27 |
| amaury_ | Please read this http://codespeak.net/pypy/dist/pypy/doc/windows.html | 22:28 |
| Action: amaury_ -> bed | 22:30 | |
| antocuni__ (~antocuni@host44-121-dynamic.10-79-r.retail.telecomitalia.it) left irc: Ping timeout: 260 seconds | 22:31 | |
| whitelynx (~whitelynx@li117-47.members.linode.com) left irc: Quit: Ex-Chat | 22:34 | |
| costy_petrisor (~JDI@79.117.215.111) left irc: Quit: Leaving. | 22:42 | |
| amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 260 seconds | 22:45 | |
| antocuni__ (~antocuni@host44-121-dynamic.10-79-r.retail.telecomitalia.it) joined #pypy. | 22:45 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/701 | 22:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/126 | 22:47 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/930 | 22:47 |
| dmsk (~dmsk@62.61.145.89.generic-hostname.arrownet.dk) joined #pypy. | 22:49 | |
| gutworth (~benjamin@128.36.89.186) joined #pypy. | 22:57 | |
| elliottcable (~me@ell.io) left irc: K-Lined | 23:03 | |
| Trundle (~andy@python/site-packages/trundle) left irc: Ping timeout: 248 seconds | 23:07 | |
| k_bx (~k_bx@94.244.19.62) left irc: Ping timeout: 260 seconds | 23:11 | |
| gtaylor (~gtaylor@108-196-160-120.lightspeed.tukrga.sbcglobal.net) left irc: Remote host closed the connection | 23:16 | |
| rguillebert (~rguillebe@46.7.250.249) joined #pypy. | 23:20 | |
| aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) joined #pypy. | 23:20 | |
| [Arfrever] (~Arfrever@apache/committer/Arfrever) left irc: Read error: Connection reset by peer | 23:23 | |
| [Arfrever] (~Arfrever@apache/committer/Arfrever) joined #pypy. | 23:24 | |
| tilgovi (~randall@c-67-180-48-70.hsd1.ca.comcast.net) joined #pypy. | 23:27 | |
| tilgovi (~randall@c-67-180-48-70.hsd1.ca.comcast.net) left irc: Changing host | 23:27 | |
| tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy. | 23:27 | |
| antocuni__ (~antocuni@host44-121-dynamic.10-79-r.retail.telecomitalia.it) left irc: Remote host closed the connection | 23:27 | |
| whitelynx (~whitelynx@75.110.189.230) joined #pypy. | 23:29 | |
| tlynn_ (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) joined #pypy. | 23:31 | |
| tlynn (~tlynn@cpc6-cmbg14-2-0-cust121.5-4.cable.virginmedia.com) left irc: Ping timeout: 240 seconds | 23:33 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep | 23:41 | |
| Vorpal (~AnMaster@unaffiliated/anmaster) left irc: Ping timeout: 276 seconds | 23:42 | |
| aat (~aat@cpe-72-225-174-173.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep. | 23:45 | |
| mitchellh (~mitchellh@c-69-181-107-107.hsd1.ca.comcast.net) joined #pypy. | 23:45 | |
| tilgovi (~randall@couchdb/developer/tilgovi) left irc: Ping timeout: 256 seconds | 23:54 | |
| --- Fri Nov 4 2011 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!