| verte-{nyctea} (~verte@python/site-packages/verte) joined #pypy. | 00:31 | |
| Nick change: verte-{nyctea} -> verte-wleslie | 00:31 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/761 | 00:34 |
|---|---|---|
| kkb110 (~kkb110@cpe-69-203-124-125.nyc.res.rr.com) left irc: Ping timeout: 240 seconds | 00:50 | |
| kkb110 (~kkb110@cpe-69-203-124-125.nyc.res.rr.com) joined #pypy. | 00:56 | |
| kkb110 | Q: how mature is pypy's CPython C API? I've seen it's alpha/beta state on the web page, but do you think it's ok to use boost::python to wrap simple classes and interact with embedded pypy on c++? | 00:57 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64-2/builds/17 | 00:58 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/189 | 01:01 |
| Arfrever (~Arfrever@apache/committer/Arfrever) left irc: Ping timeout: 260 seconds | 01:01 | |
| voidspace (~anonymous@python/psf/voidspace) left irc: Quit: voidspace | 01:15 | |
| stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy. | 01:26 | |
| stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 01:26 | |
| Nick change: stakkars_ -> stakkars | 01:26 | |
| stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy. | 01:27 | |
| stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 01:27 | |
| Nick change: stakkars_ -> stakkars | 01:27 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1906 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/312 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/323 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/658 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/661 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1177 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/844 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1496 | 01:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/771 | 01:30 |
| Gnosis- (~none@pool-96-248-225-79.snloca.dsl-w.verizon.net) joined #pypy. | 01:31 | |
| ericflo (~ericflo@75.103.8.110) left irc: Quit: ericflo | 02:15 | |
| Nick change: verte-wleslie -> wleslie | 02:41 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/998 | 02:43 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1906 | 03:45 |
| Nick change: wleslie -> william-leslie | 04:48 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/1496 | 04:48 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/844 | 04:49 |
| Nick change: Gnosis- -> GNOSIS- | 04:51 | |
| GNOSIS- (none@pool-96-248-225-79.snloca.dsl-w.verizon.net) left #pypy. | 04:52 | |
| Nick change: william-leslie -> wleslie | 04:54 | |
| Action: mikefc misses numpy.cumsum and where | 05:04 | |
| wleslie (~verte@python/site-packages/verte) left irc: Quit: ~~~ Crash in JIT! | 05:05 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/661 | 05:06 |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/1177 | 05:11 |
| rhiebert (~rhiebert@24.104.139.205) left irc: Remote host closed the connection | 05:13 | |
| bbot2 | 3Success: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/658 | 05:17 |
| etrepum (~bob@75-101-96-144.dsl.static.sonic.net) joined #pypy. | 05:24 | |
| espes (espes@2001:470:1f05:ca1::) left irc: Read error: Connection timed out | 05:32 | |
| espes (espes@2001:470:1f05:ca1::) joined #pypy. | 05:33 | |
| Action: mikefc needs to hurry up and learn rpython so he can add these things. | 05:47 | |
| mikefc | nee | 05:47 |
| mikefc | d | 05:47 |
| mikefc | need a numpy.var()... | 05:47 |
| Action: mikefc starts | 05:58 | |
| Alex_Gaynor | mikefc: start with tests :) | 05:59 |
| mikefc | I want to start just seeing how the numpy code works. if i run "python py.py" how to i import numpypy? Is this even possible/sensible? | 06:00 |
| mikefc | i get an imoprterror | 06:00 |
| Alex_Gaynor | you need to do py.py --withmod-micronumpy | 06:00 |
| mikefc | alex: just going to hack around first and see what's what. I'll definitely write tests when i get to the stage of having an idea :) | 06:00 |
| gutworth (~benjamin@64-131-9-52.usfamily.net) left irc: Ping timeout: 252 seconds | 06:01 | |
| Alex_Gaynor | ndarray.var looks pretty simple, FWIW | 06:01 |
| mikefc | the algorithm is simple. but i need to start thinking of object spaces and whatnot i guess. | 06:01 |
| Alex_Gaynor | honestly, I wouldn't worry about understanding the full picture, just do it piece by piece | 06:02 |
| Alex_Gaynor | figure out how to run/write teh tests | 06:02 |
| Alex_Gaynor | how to add a method to ndarray objs | 06:02 |
| Alex_Gaynor | etc. | 06:02 |
| mikefc | Alex_Gaynor: yep sure. | 06:03 |
| zain (~textual@cpe-68-173-70-124.nyc.res.rr.com) joined #pypy. | 06:08 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/771 | 06:17 |
| mikefc | is there an easy way to do a module reload when in py.py? (so i can quickly see what effect my changes have). Or do i have to quit() and restart? | 06:20 |
| Alex_Gaynor | quit and restart | 06:20 |
| Alex_Gaynor | and start it with -S (don't load site.py) it's faster | 06:20 |
| mikefc | thanks. | 06:21 |
| zain (~textual@cpe-68-173-70-124.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep. | 06:25 | |
| mikefc | how do I just run the micronumpy tests? | 06:33 |
| mikefc | rtfm? | 06:34 |
| Alex_Gaynor | ./pytest.py pypy/module/micronumpy [-x < stop on first failure] | 06:34 |
| mikefc | ta | 06:35 |
| Action: mikefc has made a failing test. programming genius :| | 06:47 | |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/323 | 06:54 |
| Action: mikefc now has a passing test! | 07:06 | |
| mikefc | but i'm working in default branch. and there are no multidim array ops in default. | 07:07 |
| Alex_Gaynor | right, matti is working on a branch with them | 07:07 |
| mikefc | should I try and make my function accept the same args as actual numpy? like "axis=" . or should I just get something working and worry about matching call signatures a bit later. | 07:09 |
| Alex_Gaynor | I'd skip axis for now | 07:09 |
| mikefc | the actual numpypy implementation is quite rough when it comes to matching numpy call signatures. I guess TDD is the way to go and implement as people need stuff. | 07:11 |
| Alex_Gaynor | exactly | 07:11 |
| bbot2 | 4Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/312 | 07:11 |
| stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy. | 07:16 | |
| stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 07:16 | |
| Nick change: stakkars_ -> stakkars | 07:16 | |
| stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy. | 07:19 | |
| stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 07:19 | |
| Nick change: stakkars_ -> stakkars | 07:19 | |
| asmeurer (~asmeurer@c-174-56-21-245.hsd1.nm.comcast.net) left irc: Quit: asmeurer | 07:19 | |
| mikefc | attempting to add the .std() method. | 07:22 |
| mikefc | def descr_std(self, space): | 07:22 |
| mikefc | return space.sqrt(self.descr_var(space)) | 07:22 |
| mikefc | that's obviously not right, but I can't figure out how to get a sqrt of the float returned by the var() method. | 07:22 |
| Alex_Gaynor | you want to get the ufunc | 07:25 |
| Alex_Gaynor | it's interp_ufuncs.get(space).sqrt() or something | 07:25 |
| Alex_Gaynor | ah, interp_ufuncs.get(space).sqrt.call(space, [self.descr_var(space)]) | 07:26 |
| mikefc | but currently var returns a float. not an array. so i just need the sqrt of a float -> float. not really a ufunc. | 07:27 |
| Alex_Gaynor | var should return, whatever the right type is (wrapped) | 07:27 |
| Alex_Gaynor | var can return an array right? | 07:27 |
| mikefc | when we get multi-dim it could return an array | 07:28 |
| mikefc | but with 1d, it'll only ever return a (wrapped) float. | 07:29 |
| Alex_Gaynor | ufuncs work with floats too | 07:29 |
| mikefc | ok. well then ufuncs are more general and better in the long run. | 07:30 |
| Action: mikefc tries | 07:30 | |
| mikefc | works! | 07:31 |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) joined #pypy. | 07:33 | |
| mikefc | this code runs OK, but causes an exception in test_ztranslation | 07:40 |
| mikefc | def descr_var(self, space): | 07:40 |
| mikefc | ''' var = mean( (values - mean(values))**2 ) ''' | 07:40 |
| mikefc | return (self.descr_sub(space, self.descr_mean(space))).descr_pow(space, space.wrap(2)).descr_mean(space) | 07:40 |
| mikefc | Exception': demoting method descr_pow to settled class <ClassDef 'pypy.interpreter.baseobjspace.Wrappable'> not allowed | 07:40 |
| antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) joined #pypy. | 08:37 | |
| mikefc | solved it. only took an hour of farting around. need to assert isinstance(result, BaseArray) | 08:42 |
| mimi_ (~mimi@97-88-28-179.dhcp.roch.mn.charter.com) joined #pypy. | 08:50 | |
| mimi_ (~mimi@97-88-28-179.dhcp.roch.mn.charter.com) left irc: Ping timeout: 268 seconds | 08:55 | |
| fijal (~fijal@197.173.9.193) joined #pypy. | 08:57 | |
| metvik (~metvik@unaffiliated/metvik) joined #pypy. | 09:01 | |
| mimi_ (~mimi@97-88-28-179.dhcp.roch.mn.charter.com) joined #pypy. | 09:08 | |
| mat^2 (mathias@87.55.0.198) left irc: | 09:15 | |
| mimi_ (~mimi@97-88-28-179.dhcp.roch.mn.charter.com) left irc: Ping timeout: 255 seconds | 09:17 | |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) left irc: Ping timeout: 260 seconds | 09:32 | |
| fijal (~fijal@197.173.9.193) left irc: Quit: Leaving | 09:39 | |
| fijal (~fijal@197.173.9.193) joined #pypy. | 09:40 | |
| arigato (~arigo@adsl-84-227-72-198.adslplus.ch) joined #pypy. | 09:48 | |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) joined #pypy. | 09:50 | |
| fijal | arigato: hi | 09:57 |
| arigato | hi! | 09:57 |
| fijal | arigato: happy new year and whatnot | 09:57 |
| fijal | did you get any snow so far? | 09:57 |
| arigato | I was about to say it | 09:57 |
| arigato | yes, tons of it :-) | 09:57 |
| arigato | it has been late to come, but now there are really a lot | 09:58 |
| fijal | :] | 09:58 |
| dracman (~draco@d214008.adsl.hansenet.de) left irc: Ping timeout: 252 seconds | 09:59 | |
| ThomasWaldmann | sunny wheather here | 09:59 |
| fijal | arigato: it's of course a middle of summer here | 10:00 |
| fijal | (of course) | 10:00 |
| Action: ThomasWaldmann on way back from 28c/berlin3 to stuttgart | 10:01 | |
| fijal | ThomasWaldmann: how was it? | 10:01 |
| ThomasWaldmann | fun, as ever :) very interesting stuff, just a little crowded. | 10:02 |
| ronny | ThomasWaldmann: lucky you | 10:02 |
| fijal | arigato: I wanted something from you but I forgot what | 10:03 |
| fijal | ah | 10:03 |
| fijal | lesshaste had a good real-world scenario where we pay a lot for call_assembler for no good reason | 10:03 |
| fijal | but it's kinda "known" | 10:03 |
| arigato (~arigo@adsl-84-227-72-198.adslplus.ch) left irc: Ping timeout: 252 seconds | 10:03 | |
| fijal | ah, I made him quit | 10:03 |
| Action: ThomasWaldmann mentioned pypy to some folks who seemed to learn C currently | 10:04 | |
| fijal | ThomasWaldmann: and? | 10:04 |
| ThomasWaldmann | well, get hey learned something new, but they are using C at school... | 10:06 |
| ThomasWaldmann | guess* | 10:06 |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) joined #pypy. | 10:06 | |
| fijal (~fijal@197.173.9.193) left irc: Read error: Connection reset by peer | 10:06 | |
| Action: ThomasWaldmann .oO(latency sucks a bit here, so excuse all the weird typos) | 10:06 | |
| lahwran | is there any better way to do what autopath does? | 10:07 |
| lahwran | I've been trying to think of one for a while and I'm coming up blank | 10:08 |
| dracman (~draco@e176041007.adsl.alicedsl.de) joined #pypy. | 10:09 | |
| ronny | lahwran: its quite a hack, there isnt a good way to replace it witht sommething less invasive | 10:15 |
| lahwran | what about in terms of the whole python language? #python is being rather useless | 10:15 |
| lahwran | nobody is responding to my question at all | 10:15 |
| arigato (~arigo@adsl-84-227-72-198.adslplus.ch) joined #pypy. | 10:17 | |
| fijal (~fijal@197.172.91.194) joined #pypy. | 10:17 | |
| arigato | fijal: sorry, timed out | 10:17 |
| fijal | arigato: yeah, I've seen :) | 10:17 |
| fijal | arigato: but I didn't have anything urgent | 10:18 |
| fijal | and I'm about to go anyway | 10:18 |
| arigato | :-) | 10:18 |
| arigato | in case you didn't notice, call_assembler is cheaper since 1-2 weeks | 10:18 |
| arigato | because we no longer pass infinitely many arguments | 10:18 |
| fijal | from label stuff? | 10:18 |
| fijal | or from what? | 10:18 |
| arigato | no, that's a bit more recent | 10:18 |
| fijal | yes, but we allocate and unpack the frame right | 10:19 |
| fijal | ? | 10:19 |
| arigato | yes | 10:19 |
| fijal | ok | 10:19 |
| fijal | well, ideally we would not need to allocate the frame | 10:19 |
| arigato | I was talking about the jit-simplify-backendintf branch | 10:20 |
| arigato | yes | 10:20 |
| arigato | maybe we'd like to allocate "compact jit-only frames" | 10:21 |
| fijal | yes | 10:22 |
| fijal | then you don't have to pack there tons of stuff, just copy away the heap | 10:22 |
| fijal | in case it escapes | 10:22 |
| arigato | yes | 10:22 |
| fijal | for-later kinda | 10:23 |
| arigato | :-) | 10:23 |
| fijal | but maybe later is now :) | 10:23 |
| arigato | obviously :-) | 10:23 |
| fijal | anyway from "general performance" problems, I'm working on what to do with those annoying traces | 10:23 |
| fijal | to have a faster test run | 10:23 |
| arigato | ah, good | 10:23 |
| fijal | not much progress though ;-) | 10:24 |
| arigato | yes, I fear it's not too easy :-/ | 10:24 |
| fijal | btw, the pypyjit.set_abort_hook is useful | 10:24 |
| fijal | you can do | 10:24 |
| fijal | pypyjit.set_abort_hook(f) | 10:24 |
| fijal | where f gets reason as first arg | 10:24 |
| fijal | and can call say pdb.set_trace() | 10:24 |
| fijal | that will put you in pdb each time you abort tracing | 10:25 |
| arigato | nice | 10:25 |
| fijal | [better-jit-hooks branch, but maybe we should merge it] | 10:25 |
| arigato | (similarly, maybe now is slowly getting a good time to worry about the speed of pyjitpl) | 10:25 |
| fijal | yes | 10:25 |
| fijal | new year's promises :) | 10:25 |
| arigato | :-) | 10:25 |
| fijal | anwyay, I'm off and moving tomorrow | 10:25 |
| arigato | I have vague ideas | 10:25 |
| fijal | so very likely I'll only be back on 2nd or so | 10:25 |
| fijal | WITH ADSL INTERNET! | 10:25 |
| arigato | :-) | 10:25 |
| fijal | cheers | 10:25 |
| arigato | ok, have fun | 10:25 |
| arigato | or "fun" | 10:25 |
| fijal | and have fun in snow :) | 10:25 |
| arigato | thank you :-) | 10:26 |
| fijal | I bet you'll go skiing | 10:26 |
| arigato | today the weather is horrible, so no | 10:26 |
| arigato | but the next possible day, yes :-) | 10:26 |
| fijal (~fijal@197.172.91.194) left irc: Ping timeout: 240 seconds | 10:30 | |
| mikefc | what is numpypy get_concrete() supposed to do? | 10:45 |
| arigato (~arigo@adsl-84-227-72-198.adslplus.ch) left irc: Ping timeout: 252 seconds | 10:49 | |
| dracman (~draco@e176041007.adsl.alicedsl.de) left irc: Remote host closed the connection | 10:51 | |
| Nick change: christophler_afk -> christophler | 10:57 | |
| christophler | negotiate building contracts? | 10:57 |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 11:11 | |
| lesshaste | :) | 11:31 |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) joined #pypy. | 11:37 | |
| Nick change: christophler -> christophler_afk | 11:51 | |
| whyking (~quassel@p5B3DC0EB.dip.t-dialin.net) joined #pypy. | 12:16 | |
| mikefc | :/ | 12:33 |
| voidspace (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy. | 13:05 | |
| voidspace (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host | 13:05 | |
| voidspace (~anonymous@python/psf/voidspace) joined #pypy. | 13:05 | |
| gutworth (~benjamin@64-131-9-52.usfamily.net) joined #pypy. | 14:02 | |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) left irc: Read error: Connection reset by peer | 14:20 | |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) joined #pypy. | 14:24 | |
| antocuni (~antocuni@host117-71-dynamic.58-82-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 14:27 | |
| Nick change: christophler_afk -> christophler | 14:36 | |
| christophler (web49_an_@84.45.87.204) left #pypy ("NickServe"). | 14:37 | |
| christophler (~web49_an_@84.45.87.204) joined #pypy. | 14:37 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) left irc: Ping timeout: 268 seconds | 14:48 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) joined #pypy. | 14:51 | |
| derdon (~derdon@p5DE891FD.dip.t-dialin.net) joined #pypy. | 14:58 | |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 15:02 | |
| mrgecko (~Adium@219.161.66.86.rev.sfr.net) joined #pypy. | 15:23 | |
| arigato (~arigo@adsl-84-227-72-198.adslplus.ch) joined #pypy. | 15:35 | |
| mrgecko | Hello hello, I don't know how to share the bug i've got with pypy : http://nopaste.info/ad94b6ff4e.html . What kind of infos do you need from me :) ? A script that can reproduce it on demand ? | 15:36 |
| Shanita (~John@osbk-4db17fc9.pool.mediaWays.net) joined #pypy. | 15:40 | |
| Kaskuka (~John@osbk-4d087de4.pool.mediaWays.net) left irc: Ping timeout: 240 seconds | 15:40 | |
| arigato | mrgecko: hi | 15:41 |
| arigato | mrgecko: first question, which pypy is this? | 15:41 |
| mrgecko | the nighty build | 15:42 |
| arigato | ok | 15:42 |
| arigato | yes, ideally a script | 15:43 |
| arigato | even a script that triggers the bug only sometimes | 15:43 |
| mrgecko | Maybe I can get it everytime, if it's triggered by some random seeds. I guess I can catch randomness seeds I use in python, can I ? | 15:45 |
| arigato | yes | 15:45 |
| arigato | just call random.seed(some_number) | 15:45 |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) joined #pypy. | 15:47 | |
| mrgecko | cool, 42 as seed crashes it everytime :D | 15:48 |
| nedbat (~nedbat@python/psf/nedbat) left irc: Quit: Leaving | 15:48 | |
| arigato | of course :-) | 15:48 |
| mrgecko | Let's isolating the code now | 15:48 |
| arigato | the crash is in the JIT, so maybe if you cut too much it won't crash any more, or only with a different seed | 15:49 |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 15:49 | |
| kkb110 | Q: when sandboxed pypy marshals external function calls, what is the level of this function??? system call functions???? | 15:53 |
| arigato | mrgecko: an alternative: if you run "PYPYLOG=jit-log-noopt:logfile pypy bigprogram.py" | 15:53 |
| arigato | and it crashes | 15:53 |
| arigato | then you can send us the produced logfile | 15:53 |
| arigato | even the last ~10000 lines are enough | 15:54 |
| arigato | kkb110: actually it's calls to C functions from (typically) the standard library | 15:55 |
| arigato | this might be close to system calls, e.g. open(), read() etc., or not really, like time() | 15:55 |
| kkb110 | I'm confused, then how pypy wraps the system calls when the python code uses c-written libraries (let's say, numpy?) | 16:00 |
| arigato | that's unclear | 16:01 |
| arigato | right now the pypy-sandbox that you get does not have many extension modules | 16:02 |
| kkb110 | so I guess the sandboxing needs special care of each c-extension module?? | 16:03 |
| arigato | it depends on the kind of extension module | 16:03 |
| arigato | by default, every C call is wrapped | 16:04 |
| arigato | and instead of doing the C call, | 16:04 |
| kkb110 | I think that needs rewriting of each extension module. Am I right?? | 16:04 |
| arigato | no | 16:04 |
| mrgecko | arigato: how do you notice the program has crashed crash with this huge amount of incomprehensible data :D ? | 16:05 |
| arigato | a C call is wrapped, which means that any call to the C function is not really done at all, but instead exposed to the controlling process | 16:05 |
| arigato | for some extension modules it's just what you want | 16:05 |
| arigato | for numpy, well, it depends if you want to use only the standard built-in operations like matrix addition, or more complicated ones that are done with C calls | 16:06 |
| arigato | mrgecko: use PYPYLOG=jit-log-noopt:logfile , which puts the huge amount of data into the file "logfile" and not stdout | 16:07 |
| kkb110 | arigato, so pypy-sandbox wraps C calls, not system calls directly? | 16:07 |
| arigato | kkb110: pypy-sandbox is not system-dependent at all, so it does not depend on what is a system call or not (which you don't know when writing a C program) | 16:08 |
| arigato | mrgecko: for the crash that you report, it's clear that we have to look at the last "{jit-log-noopt" - "jit-log-noopt}" section before the crash | 16:09 |
| mrgecko | Sure, I've got the infos but as a curious boy, I've tried to read the logfile. The style of the writer isn't bad but this is not the best novel i've read this year. | 16:09 |
| arigato | sure :-) | 16:09 |
| kkb110 | arigato, then back to the numpy matrix addition example, I think matrix addition part is written as c-extension, so isn't it C call here? | 16:10 |
| arigato | no, matrix addition is written natively, not as a C extension | 16:11 |
| kkb110 | python native? | 16:11 |
| arigato | RPython native | 16:11 |
| kkb110 | oh ok | 16:11 |
| mrgecko | Okay so it seems it's a block of only 256 lines, but I will send you the whole last 10k lines to be sure :). Should I write a bug report on bugs.pypy.org now ? | 16:12 |
| arigato | in other words it's done by C code that is part of "pypy" or "pypy-sandbox" | 16:12 |
| arigato | mrgecko: yes, please :-) | 16:12 |
| kkb110 | I see. I think I'm very much clear at this point | 16:15 |
| kkb110 | One more question (hopefully the last) are standard library modules supported by PyPy written in RPython mostly? | 16:16 |
| arigato | yes | 16:17 |
| arigato | when I mentioned above "calls to C functions", I really meant calls to "external" kind of code | 16:18 |
| arigato | e.g. the "re" module is entirely in RPython, but the "pyexpat" module is done by external calls to C functions | 16:18 |
| kkb110 | I see, so I guess pypy-sandbox catches it in that case | 16:19 |
| arigato | yes | 16:20 |
| kkb110 | and maybe emmit NotImplemented exception if it is not supported yet | 16:20 |
| arigato | yes | 16:20 |
| kkb110 | crystal clear, thank you :) | 16:20 |
| arigato | it's more precisely the demo wrapper, pypy_interact.py, that emits NotImplementedError if not supported yet | 16:20 |
| arigato | the point is that the sandboxed pypy supports automatically all the C functions that it uses, | 16:22 |
| arigato | but the demo wrapper only supports a few of them, | 16:22 |
| arigato | and answers all the other ones by crashing | 16:23 |
| kkb110 | nice :) | 16:23 |
| CIA-53 | 03MrGecko 07roundup * 10#975/A jit bug crashes my app: | 16:24 |
| CIA-53 | [new] This is what I randomly get with the latest nightly build: | 16:24 |
| CIA-53 | RPython traceback: | 16:24 |
| CIA-53 | File "jit_metainterp_optimizeopt_optimizer.c", line 611 ... * 14https://bugs.pypy.org/issue975 | 16:24 |
| arigato | mrgecko: thank you | 16:25 |
| mrgecko | It's a kind of new year's gift | 16:25 |
| arigato | indeed | 16:26 |
| mrgecko | I wished it has been the 1000th bug report | 16:27 |
| arigato | time to find 25 more bugs | 16:27 |
| `fox` (~fox@host142-107-dynamic.45-79-r.retail.telecomitalia.it) left irc: Ping timeout: 252 seconds | 16:36 | |
| kkb110 | do you think this scenario would possibly work to get sandboxed CPython? 1. replace CPython's libs with pypy's rpython libs. 2. control C-calls in CPython. | 16:37 |
| kkb110 | I just want to get sandboxed python on C++, so I think I either have to make CPython sandboxed or modify pypy-sandbox not to use sub-process | 16:39 |
| kkb110 | but since pypy's C API is not stable.. (according to pypy webpage) | 16:40 |
| arigato | making CPython sandboxed is impossible | 16:51 |
| arigato | because I know of at least 5 or 10 ways that can crash CPython, which an attacker may use to do malicious things instead of crashing | 16:52 |
| arigato | (in Lib/test/crashers/, if you're interested) | 16:52 |
| arigato | (in the standard CPython distribution) | 16:52 |
| kkb110 | oh ok, I'll look into it thanks for the info | 16:53 |
| arigato | bottom line: CPython is not sandbox-safe at all. If you like you can try to play around with OS-level sandboxing, too | 16:58 |
| arigato | as for using the pypy sandbox, nedbat is trying | 16:59 |
| kkb110 | arigato, I skimmed several files in the dir.. this is too sad lol | 16:59 |
| nedbat | kkb110: hello again. | 16:59 |
| kkb110 | nedbat, hahaha you were here :D | 16:59 |
| nedbat | kkb110: I told you i've been working on the pypy sandbox! | 16:59 |
| kkb110 | lol | 17:00 |
| kkb110 | ok then I cancel my plan to make CPython sandboxed | 17:02 |
| kkb110 | I guess I'm gonna just modify pypy-sandbox not to use sub-process | 17:03 |
| arigato | mrgecko: sorry, I'm failing to reproduce the bug out of the logs | 17:05 |
| voidspace (~anonymous@python/psf/voidspace) left irc: Quit: voidspace | 17:07 | |
| arigato | mrgecko: can you try to re-run with "PYPYLOG=jit:pypylog"? | 17:07 |
| mrgecko | no prob, i'm all yours | 17:07 |
| arigato | and paste again the last 10k lines (or maybe 50k) | 17:07 |
| arigato | or, of course, if you can reduce the crash, just post the remaining sources | 17:08 |
| derdon (~derdon@p5DE891FD.dip.t-dialin.net) left irc: Remote host closed the connection | 17:17 | |
| nettok (~quassel@200.119.184.61) joined #pypy. | 17:25 | |
| CIA-53 | 03MrGecko 07roundup * 10#975/A jit bug crashes my app: | 17:28 |
| CIA-53 | [chatting] And here are the last 50k lines of my PYPYLOG=jit:logfile | 17:28 |
| CIA-53 | The bug isn't caused by random.shuffle as I initially thought. | 17:28 |
| CIA-53 | Ask me if yo ... * 14https://bugs.pypy.org/issue975 | 17:28 |
| CIA-53 | 03arigo 07roundup * 10#974/channel is not iterable: | 17:29 |
| CIA-53 | [chatting] The documentation at that link is unclear: when does next() raises | 17:29 |
| CIA-53 | StopIteration? It seems that if next() is just implemented by call ... * 14https://bugs.pypy.org/issue974 | 17:29 |
| nettok_ (~quassel@200.119.185.171) joined #pypy. | 17:36 | |
| nettok (~quassel@200.119.184.61) left irc: Ping timeout: 255 seconds | 17:38 | |
| whyking (~quassel@p5B3DC0EB.dip.t-dialin.net) left irc: Ping timeout: 240 seconds | 17:41 | |
| CIA-53 | 03arigo 07roundup * 10#969/I want to be able to interrupt the translation process and continue from where it was left the last time: [wontfix] Sorry, it's not easy at all to dump the state and reload it. * 14https://bugs.pypy.org/issue969 | 17:43 |
| CIA-53 | 03arigo 07roundup * 10#967/itertools.imap significantly (order of magnitude) slower than map or gencomps: | 17:46 |
| CIA-53 | [chatting] The issue is not imap() alone, but the construct 'list(imap(...))'. If you | 17:46 |
| CIA-53 | write it differently, for example '[x for x in imap(...)]' ... * 14https://bugs.pypy.org/issue967 | 17:46 |
| CIA-53 | 03arigo 07roundup * 10#965/xml.etree.ElementTree says 'unknown encoding' of a regular encoding: | 17:55 |
| CIA-53 | [chatting] PyPy is always equivalent to a wide unicode build, so we have the same issue as | 17:55 |
| CIA-53 | CPython-in-wide-unicode-builds. Let's wait for its re ... * 14https://bugs.pypy.org/issue965 | 17:55 |
| CIA-53 | 03arigo 07roundup * 10#961/Spurious (?) mercurial error messages from sandlib.py: | 17:59 |
| CIA-53 | This problem might have disappeared in ecf82a6d3959. Can you check and close | 17:59 |
| CIA-53 | this bug if it's the case? Thanks! * 14https://bugs.pypy.org/issue961 | 17:59 |
| CIA-53 | 03arigo 07roundup * 10#960/Sandboxing for external function 'pypy__float2longlong': [resolved] Fixed in d9b372cf25b0, I believe. * 14https://bugs.pypy.org/issue960 | 18:01 |
| getxsick (xsx@tramwaj.asi.pwr.wroc.pl) left irc: Changing host | 18:03 | |
| getxsick (xsx@unaffiliated/getxsick) joined #pypy. | 18:03 | |
| CIA-53 | 03arigo 07roundup * 10#956/os.fdopen() is broken with fds returned by os.openpty(): | 18:07 |
| CIA-53 | [wontfix] Please report this as a bug of CPython if you consider this to be an important | 18:07 |
| CIA-53 | issue. Closing from here in the meantime... * 14https://bugs.pypy.org/issue956 | 18:07 |
| Action: arigato done commenting on issues :-) | 18:07 | |
| CIA-53 | 03masklinn 07roundup * 10#967/itertools.imap significantly (order of magnitude) slower than map or gencomps: | 18:09 |
| CIA-53 | Oh. That's weirteresting. Why does that happen even though list + gencomp has no | 18:09 |
| CIA-53 | issue, is is something which does not behave correctly in list+i ... * 14https://bugs.pypy.org/issue967 | 18:09 |
| arigato (~arigo@adsl-84-227-72-198.adslplus.ch) left irc: Ping timeout: 252 seconds | 18:12 | |
| __name__ (~name@sburn/devel/name) left irc: Ping timeout: 252 seconds | 18:14 | |
| bugfree (~azhang@adsl-99-63-77-95.dsl.pltn13.sbcglobal.net) left irc: Ping timeout: 252 seconds | 18:14 | |
| [Arfrever] (~Arfrever@apache/committer/Arfrever) left irc: Ping timeout: 252 seconds | 18:14 | |
| jell (~jell@host-134-23.dmtec.eu) left irc: Ping timeout: 252 seconds | 18:14 | |
| [Arfreve1] (~Arfrever@minotaur.apache.org) joined #pypy. | 18:14 | |
| getxsick (xsx@unaffiliated/getxsick) left irc: Ping timeout: 252 seconds | 18:15 | |
| __pv (pav@xen.ihme.org) left irc: Ping timeout: 252 seconds | 18:15 | |
| __pv_ (pav@xen.ihme.org) joined #pypy. | 18:15 | |
| jell_ (~jell@host-134-23.dmtec.eu) joined #pypy. | 18:15 | |
| arkx (arkx@hilla.kapsi.fi) left irc: Ping timeout: 252 seconds | 18:15 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) left irc: Ping timeout: 252 seconds | 18:15 | |
| jae (~jae@dedicated.jaerhard.com) left irc: Ping timeout: 252 seconds | 18:15 | |
| elipsion (~elipsion@81-232-141-35-no217.tbcn.telia.com) left irc: Ping timeout: 252 seconds | 18:15 | |
| henning_ (~henning@unicorn.gnubo.de) left irc: Ping timeout: 252 seconds | 18:15 | |
| elipsion (~elipsion@81-232-141-35-no217.tbcn.telia.com) joined #pypy. | 18:15 | |
| kalenz (~kalenz@ping.kalenz.fr) left irc: Ping timeout: 252 seconds | 18:15 | |
| oal (u4126@gateway/web/irccloud.com/x-rcoxytvybconcfsl) left irc: Ping timeout: 252 seconds | 18:15 | |
| Alex_Gaynor (u1246@gateway/web/irccloud.com/x-hppuplnqihshavje) left irc: Ping timeout: 252 seconds | 18:15 | |
| kalenz (~kalenz@ping.kalenz.fr) joined #pypy. | 18:15 | |
| __name__ (~name@sburn/devel/name) joined #pypy. | 18:15 | |
| Nick change: kalenz -> Guest47608 | 18:16 | |
| bugfree (~azhang@adsl-99-63-77-95.dsl.pltn13.sbcglobal.net) joined #pypy. | 18:16 | |
| thoughtpolice_ (~thoughtpo@li111-206.members.linode.com) joined #pypy. | 18:16 | |
| thoughtpolice (~thoughtpo@li111-206.members.linode.com) left irc: Ping timeout: 252 seconds | 18:16 | |
| getxsick (xsx@unaffiliated/getxsick) joined #pypy. | 18:16 | |
| arkx (arkx@hilla.kapsi.fi) joined #pypy. | 18:16 | |
| espes___ (~niel@205.185.120.132) left irc: Ping timeout: 252 seconds | 18:16 | |
| henning_ (~henning@unicorn.gnubo.de) joined #pypy. | 18:16 | |
| jae (~jae@dedicated.jaerhard.com) joined #pypy. | 18:17 | |
| espes__ (~niel@205.185.120.132) joined #pypy. | 18:17 | |
| oal (u4126@gateway/web/irccloud.com/x-kqltsiarqlcvmgev) joined #pypy. | 18:17 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) joined #pypy. | 18:17 | |
| Alex_Gaynor (u1246@gateway/web/irccloud.com/x-mqtrohrcnigifexe) joined #pypy. | 18:20 | |
| #pypy: mode change '+o Alex_Gaynor' by ChanServ!ChanServ@services. | 18:20 | |
| dracman (~draco@d150061.adsl.hansenet.de) joined #pypy. | 18:20 | |
| zain (~textual@cpe-68-173-70-124.nyc.res.rr.com) joined #pypy. | 18:27 | |
| sunoano (~sa@unaffiliated/sunoano) left irc: Ping timeout: 252 seconds | 18:37 | |
| zain (~textual@cpe-68-173-70-124.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep. | 18:44 | |
| sunoano (~sa@mail.heartinternet.co.uk) joined #pypy. | 18:53 | |
| sunoano (~sa@mail.heartinternet.co.uk) left irc: Changing host | 18:53 | |
| sunoano (~sa@unaffiliated/sunoano) joined #pypy. | 18:53 | |
| stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy. | 18:56 | |
| stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 18:56 | |
| Nick change: stakkars_ -> stakkars | 18:56 | |
| stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy. | 18:58 | |
| stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer | 18:58 | |
| Nick change: stakkars_ -> stakkars | 18:58 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) left irc: Ping timeout: 255 seconds | 19:02 | |
| asmeurer (~asmeurer@c-174-56-21-245.hsd1.nm.comcast.net) joined #pypy. | 19:17 | |
| sunoano (~sa@unaffiliated/sunoano) left irc: Quit: Leaving. | 19:20 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) joined #pypy. | 19:28 | |
| brownan (~brownan@cpe-065-190-056-090.nc.res.rr.com) left irc: Ping timeout: 268 seconds | 19:31 | |
| fschwidom (~fschwidom@2.215.100.139) joined #pypy. | 19:36 | |
| cpets (4f7588fa@gateway/web/freenode/ip.79.117.136.250) joined #pypy. | 20:03 | |
| nettok (~quassel@200.119.160.118) joined #pypy. | 20:03 | |
| nettok_ (~quassel@200.119.185.171) left irc: Ping timeout: 240 seconds | 20:04 | |
| nedbat (~nedbat@python/psf/nedbat) left irc: Quit: Leaving | 20:15 | |
| nedbat (~nedbat@python/psf/nedbat) joined #pypy. | 20:16 | |
| xcombelle (~xcombelle@AToulouse-551-1-136-24.w86-201.abo.wanadoo.fr) left irc: Quit: I am a manual virus, please copy me to your quit message. | 20:22 | |
| zain (~textual@cpe-68-173-70-124.nyc.res.rr.com) joined #pypy. | 20:23 | |
| cpets_ (4f7588fa@gateway/web/freenode/ip.79.117.136.250) joined #pypy. | 20:47 | |
| cpets (4f7588fa@gateway/web/freenode/ip.79.117.136.250) left irc: Ping timeout: 258 seconds | 20:49 | |
| nettok_ (~quassel@200.119.184.213) joined #pypy. | 20:58 | |
| nettok (~quassel@200.119.160.118) left irc: Ping timeout: 240 seconds | 21:00 | |
| mrgecko (~Adium@219.161.66.86.rev.sfr.net) left irc: Read error: Connection reset by peer | 21:02 | |
| mrgecko (~Adium@219.161.66.86.rev.sfr.net) joined #pypy. | 21:03 | |
| zain (~textual@cpe-68-173-70-124.nyc.res.rr.com) left irc: Quit: Computer has gone to sleep. | 21:38 | |
| CIA-53 | 03mikefc 07roundup * 10#964/numpypy.zeros([]) returns wrong result: | 21:43 |
| CIA-53 | [chatting] There were 3 (related) problems here that I attempted to fix. Patch with tests | 21:43 |
| CIA-53 | included. Please feel free to make changes to the pat ... * 14https://bugs.pypy.org/issue964 | 21:43 |
| Arfrever (~Arfrever@apache/committer/Arfrever) joined #pypy. | 21:45 | |
| CIA-53 | 03mikefc 07roundup * 10#964/numpypy.zeros([]) returns wrong result: | 21:46 |
| CIA-53 | ugh. bad editing: | 21:46 |
| CIA-53 | problem 1: to_str for concrete arrays with ndims==0 is broken. | 21:46 |
| CIA-53 | If you look at ConcreteArray.to_str, there is one path con ... * 14https://bugs.pypy.org/issue964 | 21:46 |
| fschwidom (~fschwidom@2.215.100.139) left irc: Ping timeout: 252 seconds | 21:51 | |
| voidspace (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy. | 21:53 | |
| voidspace (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host | 21:53 | |
| voidspace (~anonymous@python/psf/voidspace) joined #pypy. | 21:53 | |
| mrgecko (~Adium@219.161.66.86.rev.sfr.net) left irc: Read error: Connection reset by peer | 22:00 | |
| CIA-53 | 03cpher 07roundup * 10#976/Implement numpypy.cross(): | 22:00 |
| CIA-53 | [new] While attempting to test some code on PyPy I discovered that the numpypy module | 22:00 |
| CIA-53 | does not implement the cross-product function. If anyone ge ... * 14https://bugs.pypy.org/issue976 | 22:00 |
| mrgecko (~Adium@219.161.66.86.rev.sfr.net) joined #pypy. | 22:03 | |
| nettok (~quassel@200.119.188.169) joined #pypy. | 22:16 | |
| nettok_ (~quassel@200.119.184.213) left irc: Ping timeout: 252 seconds | 22:18 | |
| voidspace (~anonymous@python/psf/voidspace) left irc: Quit: voidspace | 22:29 | |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/762 | 22:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/190 | 22:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/999 | 22:30 |
| bbot2 | Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64-2/builds/18 | 22:30 |
| lahwran | gogogogo buildbot! | 22:38 |
| nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 276 seconds | 22:41 | |
| fschwidom (~fschwidom@2.215.100.139) joined #pypy. | 22:43 | |
| Action: mikefc thinks about implementing the .cross() | 22:49 | |
| mrgecko (~Adium@219.161.66.86.rev.sfr.net) left irc: Ping timeout: 240 seconds | 23:12 | |
| mikefc | nope. too hard. | 23:23 |
| [Arfreve1] (~Arfrever@minotaur.apache.org) left irc: Quit: leaving | 23:30 | |
| -christel (~christel@freenode/staff/exherbo.christel) to $$*- [Global Notice] Happy new Year to the those on fST which covers UK, Portugal, Morocco, Ireland, parts of Spain, Iceland, Cote d'Ivoire, Mali, Togo, Senegal, Gambia, Faroe Islands, Guinea-Bissau, Isle of Man, Greenland, Saint helena, Mauritania, Sao Tome, Guinea, Western Sahara, Ghana, Burkina Faso, Sierra Leone and Liberia! Welcome to 2012 - Year Of The Troll! | 23:32 | |
| [Arfrever] (~Arfrever@apache/committer/Arfrever) joined #pypy. | 23:35 | |
| voidspace (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy. | 23:37 | |
| voidspace (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host | 23:37 | |
| voidspace (~anonymous@python/psf/voidspace) joined #pypy. | 23:37 | |
| --- Sun Jan 1 2012 | 00:00 | |
Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!