#pypy IRC log for Monday, 2010-12-13

Ademan (~dan@adsl-71-141-236-146.dsl.snfc21.pacbell.net) joined #pypy.00:00
kenaan03afa 10r8001614 pypy/branch/fast-forward 05A(test_unpack.py) 07M(objspace.py): PyPy provides better error messages than CPython ...00:01
amaury_ (5645c561@gateway/web/freenode/ip. left irc: Quit: Page closed00:02
Vorpal (~AnMaster@unaffiliated/anmaster) joined #pypy.00:02
pedronis (~pedronis@253-109.104-92.cust.bluewin.ch) left irc: Ping timeout: 272 seconds00:04
Trundle (~andy@python/site-packages/trundle) left irc: Remote host closed the connection00:11
pedronis (~pedronis@253-109.104-92.cust.bluewin.ch) joined #pypy.00:16
CIA-903strombrg 07roundup * 10#608/Gibberish from ctypes: Yes, you certainly may ask. I'll try to find time for this soon, but I'm starting a new contract tomorrow morning, and I have a bunch of paperwor ... * 14https://codespeak.net/issue/pypy-dev/issue60800:28
mat^2 (~mathias@ left irc: Ping timeout: 240 seconds00:31
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) joined #pypy.01:22
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) joined #pypy.01:38
idnar (~quassel@unaffiliated/idnar) left irc: Ping timeout: 276 seconds01:47
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Read error: Connection reset by peer02:05
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.02:06
Action: Varriount is away (Shower)02:20
exarkungood to know02:23
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Read error: Connection reset by peer02:29
mat^2 (~mathias@ joined #pypy.02:29
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.02:30
ousado (~ousado@unaffiliated/ousado) left irc: Ping timeout: 241 seconds02:33
ousado_ (~ousado@p5B3C163F.dip0.t-ipconnect.de) joined #pypy.02:33
dcolish (~dcolish@li95-246.members.linode.com) left irc: Ping timeout: 245 seconds02:39
skulltone (~elbowtone@hudsondoyle.geek.nz) left irc: Ping timeout: 255 seconds02:39
VarriountAnyone here know anything about procedural land generation?02:46
xorAxAxask __doc__02:46
exarkunThere have been hundreds of thousands of man hours of development on terrain generation software, with thousands of papers written on the subject02:47
Varriountyes, google searching proves that02:48
jellVarriount: look at fractal land generation algorithms02:48
exarkunSo ask a question that isn't best answered by "google it" :p  and for bonus points, make it relevant to PyPy.02:48
VarriountHow fast would pypy be using a procedural generator which manipulates a 16 by 16 chunk of data?02:49
Varriount^ Relevant to pypy02:49
exarkunheh heh heh.02:49
exarkuntotally unanswerable, though. :)02:50
MostAwesomeDudeVarriount: $ PYTHONPATH=path/to/beta pypy path/to/beta/tools/simplexbench.py02:50
MostAwesomeDudeIt's not like I haven't been looking into it. :302:50
jell"realy fast" - it's good answer ;)02:50
VarriountMostAwesomeDude: Stop following me around02:50
jellMostAwesomeDude: isn't that expectable?02:50
MostAwesomeDudeVarriount: I was here a long time ago. :302:51
VarriountYour still stalking me02:51
VarriountJust.. Er.. Preceptively02:51
MostAwesomeDudejell: Yeah, I saw good 2-4x speedups with PyPy. Not as dramatic as I'd hoped, but my initial conclusion was that CPython already did a good job on the code in question.02:51
Action: MostAwesomeDude should probably disassemble and read the raw assembly02:52
MostAwesomeDudeHey, I didn't take that x86 assembly course for nothing.02:52
jell"python is for prototyping" is really connected with cpython02:52
jellso that "cpython already did a good job" - is mostly academical02:53
jellMostAwesomeDude: btw. for what you did that course?02:53
MostAwesomeDudeThat's the theory, which is why I haven't already __slot__'d my classes and inlined all my locals. But, still.02:53
MostAwesomeDudejell: Standard CS course required at university.02:54
MostAwesomeDudeI wish they taught it in ARM assembly instead, but CLV.02:54
jelli had 8051 on mine02:54
jelland that was not so realy bad for most students02:55
MostAwesomeDudeOh, my grades? A.02:55
jellbut mostly boring for that realy interested :P02:56
MostAwesomeDudeFirst A in a non-music course for me. I was so happy! I'm still bad at school though.02:56
jelli'm worse...02:56
jellless than week ago google rejected my candidacy because of my terrible spoken english...02:57
MostAwesomeDudeUf. Keep practicing.02:57
jellso my school was probably worse ;P02:57
jellmy school, or my school skills02:58
jellMostAwesomeDude: i will...02:58
jelli have to speak more than i write :P02:59
dcolish (~dcolish@li95-246.members.linode.com) joined #pypy.03:01
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) left irc: Ping timeout: 240 seconds03:02
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) joined #pypy.03:04
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) left irc: Client Quit03:06
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) joined #pypy.03:07
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) left irc: Client Quit03:07
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) joined #pypy.03:08
dcolish (~dcolish@li95-246.members.linode.com) left irc: Ping timeout: 245 seconds03:11
CIA-903strombrg 07roundup * 10#608/Gibberish from ctypes: I've attached a copy of the files fed to backshift that lead to the gibberish from ctypes (gdbm). I tried to check them into my SVN, but SVN is c ... * 14https://codespeak.net/issue/pypy-dev/issue60803:13
dcolish (~dcolish@li95-246.members.linode.com) joined #pypy.03:18
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) left irc: Ping timeout: 245 seconds03:21
elbowtone (~elbowtone@hudsondoyle.geek.nz) joined #pypy.03:25
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) joined #pypy.03:27
jcea_BT (jcea@jabber.hst.ru) left #pypy.03:43
si14 (~si14@ joined #pypy.03:57
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) left irc: Remote host closed the connection03:58
si14 (~si14@ left irc: Remote host closed the connection03:59
Varriount (~Varriount@pool-72-83-125-170.washdc.fios.verizon.net) left irc: Quit: Is this really cold, or heat gone awry?04:05
gutworth (~benjamin@64-131-13-109.usfamily.net) left irc: Ping timeout: 272 seconds04:18
panni_ (hannes@ip-178-203-85-85.unitymediagroup.de) left irc: Quit: ( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )04:23
vborja (~quassel@ joined #pypy.04:24
nettok (~quassel@ joined #pypy.04:33
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Quit: schnarch04:34
DasIch (~DasIch@p5DC5FEDD.dip.t-dialin.net) left irc: Ping timeout: 265 seconds04:57
DasIch (~DasIch@p5DC5FC28.dip.t-dialin.net) joined #pypy.05:09
kenaan03agaynor 10r8001714 pypy/branch/fast-forward/pypy/module/_bisect 07M(interp_bisect.py test_bisect.py): Fixed an error in the bisect tests, verify that lo is non-negative. ...05:44
dju (dju@fsf/member/dju) left irc: Quit: Quitte05:51
kenaan03agaynor 10r8001814 pypy/branch/fast-forward/pypy 07M(function.py test_descroperation.py): Delay verifying that classmethod is provided with a non-callable. ...05:57
kenaan03agaynor 10r8001914 pypy/branch/fast-forward/pypy/objspace/std 07M(test_typeobject.py typeobject.py): A unicode string for __slots__ isn't considered as an iterable. ...06:06
dju (dju@fsf/member/dju) joined #pypy.06:06
stakkars (~tismer@i59F7FA29.versanet.de) joined #pypy.06:06
dju (dju@fsf/member/dju) left irc: Max SendQ exceeded06:10
dju (dju@fsf/member/dju) joined #pypy.06:10
Alex_GaynorFor amaury/arigato (logs): the test_itertools from the fast-forawrd build I triggered last night has an RPython level MemoryError from inside the GC06:12
fijal (~fijal@41-134-41-226.dsl.mweb.co.za) joined #pypy.06:16
stakkars_ (~tismer@i59F7FA29.versanet.de) joined #pypy.06:18
stakkars__ (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.06:19
stakkars (~tismer@i59F7FA29.versanet.de) left irc: Ping timeout: 272 seconds06:22
Nick change: stakkars__ -> stakkars06:22
stakkars_ (~tismer@i59F7FA29.versanet.de) left irc: Ping timeout: 272 seconds06:22
nettok_ (~quassel@ joined #pypy.06:24
nettok (~quassel@ left irc: Ping timeout: 250 seconds06:27
kenaan03agaynor 10r8002014 pypy/branch/fast-forward/pypy/objspace/std 07M(stringobject.py test_stringobject.py): Make "\t\t".expandtabs(sys.maxint) raise an OverflowError with a useful error message. ...06:32
fijalAlex_Gaynor: shouldn't you be sleeping?06:33
Alex_Gaynorfijal: it's only 1:3006:34
fijalI suppose so, yes06:34
fijalwell, I forget I'm 1hr away from EU06:34
Alex_Gaynorbesides I'm being sort of productive, can't stop now06:34
stakkars_ (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.06:37
fijalgood I guess06:40
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Ping timeout: 264 seconds06:42
Nick change: stakkars_ -> stakkars06:42
Alex_Gaynorfijal: http://paste.pocoo.org/show/304184/ is legal rpython right?06:44
fijalI don't think so06:44
fijallemme check06:44
fijalI've seen somewhere assert step == 106:45
fijalI'm not sure any more06:47
Alex_Gaynorfijal: I thought itwas that rpython doesn't support step for slicing, but I can't remember exactly06:47
fijalit's easy enough to check in a test though :)06:48
Alex_Gaynorit's also easy to commit and then hit the buildbot :D06:48
fijalwell, reply rate is lower06:49
Alex_Gaynorif we write tests for all translation, tests will take even longer than they do now06:49
fijal>           raise Exception,"unknown op: %r" % op06:50
fijalE           Exception: unknown op: v0 = newslice((None), (None), (2))06:50
fijalso the answer is no06:50
Alex_Gaynorsigh :(06:50
fijaltook me 5s to write a test \06:50
fijalI love comments about NINJA-IDE06:53
fijal"look it's pythonic, it's written in Python!"06:53
fijal"it's not very innovative"06:53
Alex_GaynorI tried it yesterday, it was ok, not very special though.06:53
fijal"there is not much to innovate..."06:53
fijalwhich I completely disagree with :)06:55
ErKa (keryell@keryell.pck.nerim.net) joined #pypy.07:01
Alex_Gaynorfijal: why isn't unicode in ObjSpace.MethodTable07:04
antocuni (~antocuni@host244-124-dynamic.20-79-r.retail.telecomitalia.it) joined #pypy.07:12
lmoura (~lauromour@ left irc: Ping timeout: 265 seconds07:12
Alex_Gaynormorning antocuni 07:18
antocunihi alex07:18
stakkars_ (~tismer@i59F7FA29.versanet.de) joined #pypy.07:19
nettok_ (~quassel@ left irc: Ping timeout: 265 seconds07:20
tav (~tav@ left irc: Quit: tav07:22
lmoura (~lauromour@ joined #pypy.07:24
kenaan03afa 10r8002114 07M(pypy/branch/fast-forward/lib-python/modified-2.7.0/test/test_optparse.py): Really apply decorator to the function :-( ...07:25
cwillu_at_work (~cwillu@cwillu.com) left irc: Ping timeout: 240 seconds07:26
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Read error: Connection reset by peer07:26
Nick change: stakkars_ -> stakkars07:26
stakkars_ (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.07:27
stakkars__ (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.07:28
stakkars (~tismer@i59F7FA29.versanet.de) left irc: Ping timeout: 250 seconds07:31
Nick change: stakkars__ -> stakkars07:31
Nick change: stakkars -> 20QACW42207:31
Nick change: stakkars_ -> 84XABEB1K07:31
20QACW422 (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Client Quit07:32
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.07:35
stakkars_ (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.07:36
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Read error: No route to host07:36
Nick change: stakkars_ -> stakkars07:36
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Client Quit07:36
kenaan03afa 10r8002214 pypy/branch/fast-forward/pypy/module/itertools 07M(interp_itertools.py test_itertools.py): Fix an infinite loop with itertools.product(x, repeat=0) ...07:50
qbproger (~qbproger@ left irc: Remote host closed the connection07:50
mvt (~mvantelli@ joined #pypy.08:08
idnar (~quassel@unaffiliated/idnar) joined #pypy.08:08
idnar (~quassel@unaffiliated/idnar) left irc: Remote host closed the connection08:08
idnar (~quassel@unaffiliated/idnar) joined #pypy.08:08
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.08:09
ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 264 seconds08:18
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Ping timeout: 264 seconds08:21
84XABEB1K (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Ping timeout: 240 seconds08:24
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.08:25
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Remote host closed the connection08:28
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.08:28
amaury_ (5bc3480e@gateway/web/freenode/ip. joined #pypy.08:34
amaury_Alex_Gaynor: Hi08:35
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Ping timeout: 276 seconds08:47
hruske (~Gasper@ joined #pypy.08:47
verte (~verte@python/site-packages/verte) joined #pypy.08:52
Alex_Gaynoramaury_: hi08:52
amaury_I probably fixed the memory error in test_itertools08:52
amaury_but there are indeed issues in low-memory conditions08:53
Alex_Gaynoramaury_: I guess deep down there is still a bug that an RPython MemoryError escaped08:53
amaury_yes, like: gc.collect() needs a little memory to work08:53
kenaan03antocuni 10r8002314 07M(pypy/branch/jitypes2/lib_pypy/_ctypes/function.py): actually pass the *value* of the pointer buffer (i.e., the address of the param which has been wrapped with byref). test_voidresult passes ...08:54
Alex_Gaynorwell its 4am08:56
Alex_Gaynorbed time :)08:56
ErKa (keryell@m3.wifi.enstb.org) joined #pypy.08:56
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.08:58
stakkars_ (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.08:58
Ademanhow should one parse binary data from rpython?09:03
antocuniAdeman: rlib.rstruct09:03
Ademanantocuni: I saw that, it didn't look often used, and the comment in runpack.py wasn't encouraging heh09:04
Ademanbut that's what I should use?09:04
antocuniwell, this is what you have :-)09:04
antocuniif it's not enough, you could probably improve it09:04
Ademanhehe, it's probably fine, I was mostly reacting to the scary comment haha09:05
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Quit: stakkars09:18
Nick change: stakkars_ -> stakkars09:18
ericvrp_ (~ericvrp_@ericvrp.demon.nl) joined #pypy.09:30
idnar (~quassel@unaffiliated/idnar) left irc: Remote host closed the connection09:38
kenaan03cfbolz 10r8002414 07M(pypy/trunk/pypy/interpreter/argument.py): fix test of r79907. apparently, no existing test needed this condition. ...09:39
Arfrever (~Arfrever@gentoo/developer/Arfrever) joined #pypy.09:44
kenaan03cfbolz 10r8002514 pypy/trunk/pypy/objspace/std 07M(dictmultiobject.py test_dictmultiobject.py): implement an EmptyDictImplementation, that most dictionaries start out as. This ...09:45
voidspace (~voidspace@python/psf/voidspace) joined #pypy.09:48
kenaan03antocuni 10r8002614 07M(pypy/branch/jitypes2/lib_pypy/_ctypes/function.py): support string arguments and results. Makes test_stringresult passing ...10:01
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) joined #pypy.10:02
Action: fijal has tickets for pypy sprint10:19
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) left irc: Remote host closed the connection10:25
verte (~verte@python/site-packages/verte) left irc: Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org10:26
verte (~verte@python/site-packages/verte) joined #pypy.10:27
kenaan03fijal 10r8002714 07M(pypy/extradoc/sprintinfo/leysin-winter-2011/people.txt): I'm coming, hooray! ...10:28
antocunifijal: cool :-)10:29
jcea_BT (~jcea@jabber.hst.ru) joined #pypy.10:31
idnaria (~quassel@unaffiliated/idnar) joined #pypy.10:38
Nick change: idnaria -> idnar10:39
lizardo (~lizardo@ joined #pypy.10:40
cwillu_at_work (~cwillu@cwillu.com) joined #pypy.10:42
ambv (~ambv@host- joined #pypy.10:43
mat^2 (~mathias@ left irc: 10:44
arigato (~arigo@fwstups.cs.uni-duesseldorf.de) joined #pypy.10:46
kenaan03david 10r8002814 pypy/branch/arm-backend/pypy/jit/backend/arm 07M(opassembler.py regalloc.py): Fix calls with parameters on the stack and guard_exception ...10:55
arigato (~arigo@fwstups.cs.uni-duesseldorf.de) left irc: Ping timeout: 276 seconds11:02
iPeter (~iPeter@80-218-231-105.dclient.hispeed.ch) joined #pypy.11:14
iPeter (iPeter@80-218-231-105.dclient.hispeed.ch) left #pypy.11:17
romao (~romao@unaffiliated/errado) joined #pypy.11:17
antocuni (~antocuni@host244-124-dynamic.20-79-r.retail.telecomitalia.it) left irc: Ping timeout: 276 seconds11:21
ErKa (keryell@m3.wifi.enstb.org) left irc: Read error: Operation timed out11:21
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Quit: stakkars11:36
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.11:39
arigato (~arigo@fwstups.cs.uni-duesseldorf.de) joined #pypy.11:42
arigatofijal: cool indeed :-)11:50
amaury_arigato: pypy-c translate.py does not crash anymore11:58
amaury_but it's slow: http://paste.pocoo.org/show/304248/11:58
Kami_ (~kami@unaffiliated/kami-/x-9078513) left irc: Ping timeout: 276 seconds11:59
Kami_ (~kami@unaffiliated/kami-/x-9078513) joined #pypy.12:00
amaury_http://speed.pypy.org/changes/ layout is not as nice as before12:00
arigatowell it's missing colors, apparently?12:01
amaury_there was also a cache issue12:01
amaury_yes, colors are missing12:01
arigatoany clue what caused the change?12:01
arigatoI don't see anything bad, btw12:03
amaury_no, but Miquel was active on pypy-dev recently12:03
arigatoah, did you just mean to point out the missing colors, and not a regression? :-)12:04
amaury_there was a regression, until I forced to reload the css12:04
amaury_now colors are still missing12:05
danchrrelative bars seem far too small to be useful: http://speed.pypy.org/comparison/?exe=1%2B41%2C1%2B172%2C1%2BL&ben=1%2C2%2C25%2C3%2C4%2C5%2C22%2C6%2C7%2C8%2C23%2C24%2C9%2C10%2C11%2C12%2C13%2C14%2C15%2C16%2C17%2C18%2C19%2C20%2C26&env=1&hor=true&bas=2%2B35&chart=relative+bars12:07
danchrseems quite odd that the chart goes from -110x to 20x despite the fact that none of the entries exceed 10x&12:08
danchralso, nice brief URL ;)12:15
antocuni (~antocuni@host237-120-dynamic.11-79-r.retail.telecomitalia.it) joined #pypy.12:17
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Quit: stakkars12:31
ErKa (keryell@m3.wifi.enstb.org) joined #pypy.12:32
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.12:35
verte (~verte@python/site-packages/verte) left irc: Quit: ~~~ Crash in JIT!12:40
kenaan03arigo 10r8002914 pypy/branch/jit-stackcheck/pypy 07M(llmodel.py model.py assembler.py regalloc.py rx86.py test_ztranslation.py pyjitpl.py rstack.py): Implement stack checks at the beginning of the entry code for ...12:40
CIA-903Ago 07roundup * 10#592/trackgcroot.py fails with AssertionError on Gentoo x64: [resolved] Recompiling gcc and glibc with -hardened solved the issue. * 14https://codespeak.net/issue/pypy-dev/issue59212:43
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Read error: Connection reset by peer12:43
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.12:43
antocunironny: ping12:43
dan_waterworth (~dan@ joined #pypy.12:47
kenaan03arigo 10r8003014 pypy/trunk/pypy 07M(__init__.py interp_posix.py test_posix2.py rposix.py ll_os.py test_extfunc.py): Merge branch/more-posix, adding some more functions to the 'os' module. ...12:47
kenaan03arigo 10r8003114 06D(pypy/branch/more-posix/): Merged. ...12:48
kenaan03arigo 10r8003214 05A(pypy/branch/getopt-appmain/): A branch in which to refactor the option parsing done by appmain.py, ...12:49
arigatofirst, I will try to merge fast-forward into app_main.py, because it already changed some number of details12:50
antocuniarigato: while you are at it, what about having the possibility to override app_main with a custom file after translation?12:51
antocuniwould be very useful for developing/testing12:51
arigatousing what?  an option on the command-line..?12:51
antocunior an env variable, yes12:52
arigatonot sure I see the point -- we need app_main.py to parse the command line12:52
arigatoand if it's an env var, then it opens the question of "but it's not ignored if I pass -E"12:53
ronnyantocuni: pong12:53
antocuniwhatever, it's just a feature that I missed sometimes, but I suppose we can live without :-)12:54
arigatoah, you mean mostly as a way to test changes you just made to app_main.py12:54
arigatoyou know that you can also just run app_main.py?12:55
antocunironny: so, is there any other blocker before we can upload the "official" repo to bitbucket?12:55
ronnyantocuni: im not aware of any problems12:55
antocuniarigato: yes, but sometimes it does something slightly different12:55
ronnyantocuni: the question is should i keep doing it on my laptop, or do we want to move it to one of the servers12:56
antocunironny: as you prefer, but I think your laptop is fine12:56
antocunias soon as the repo is on bb and buildbot pulls from it, we can send an email to pypy-dev saying that at time X we will "officially" switch, and svn repo will be made readonly12:57
ronnyantocuni: btw, i just noticed, we didnt deal with some minor side things like languuages and extradoc12:59
antocunironny: yes, I know12:59
antocunibut we can do it later12:59
ronnyi suppose we can just forbid writing to trunk, branches, release12:59
antocuniyes. I don't know how to actually do it, though :-)13:00
antocunironny: but before doing it I want to be sure that buildbot works fine13:00
ronnyantocuni: can you create a pypy/pypy repo on bb and give me admin?13:03
antocunironny: I'm doing it :-)13:04
ronnyantocuni: the upload?13:05
antocunironny: no, the creation of the pypy repo. I just did it13:06
ronnyok, pushing now13:07
ronnywill take a while, only dsl at home13:08
antocunimaybe we will finally manage to do the switch :-)13:08
arigatoI should come up with yet another obscure way to do branches on svn...13:08
Topic changed on #pypy by antocuni!antocuni@host237-120-dynamic.11-79-r.retail.telecomitalia.it: PyPy, the flexible snake (IRC logs: http://www.tismer.com/pypy/irc-logs/pypy/?C=M;O=D ) | finally, mercurial migration is happening!13:08
antocuniarigato: :-)13:08
ronnyarigato: i fear that the mercurial api will be to your liking and you will actually implement subtree merges13:10
kenaan03arigo 10r8003314 pypy/branch/jit-stackcheck/pypy/jit/backend/x86 07M(assembler.py rx86.py): Fixes for 64-bit. ...13:11
danchrso, when will the buildbots use Mercurial checkouts? :)13:12
mvt (~mvantelli@ left irc: Ping timeout: 240 seconds13:12
antocunidanchr: very soon, we already have an updated buildbot config to do that: https://bitbucket.org/pypy/buildbot13:13
danchrneat :)13:13
danchrI should finish my sys._mercurial patch13:13
antocunidanchr: are you doing it for pypy? cool, where is it?13:14
amaury_danchr: is it similar to CPython?13:14
danchrthe output is based on the equivalent patch to CPython13:14
danchror rather, the result is13:14
danchrI just need to fix sys.subversion handle to allow the revision to be a string13:16
danchrantocuni: this is it  http://bitbucket.org/danchr/pypy-patches/src/tip/sys-mercurial.diff13:18
mvt (~mvantelli@ joined #pypy.13:18
danchralso, I have a patch to add a .hgignore file13:19
antocunidanchr: uhm, what happens if I don't have hg in my path?13:19
danchrgood point13:19
danchrwell, if you have PyPy from a Mercurial checkout but no hg executable, it'll currently blow up13:20
antocunialso, probably it'd be a good idea to revert HGPLAIN to whatever value it had before13:20
antocunidanchr: I think that the current logic for svn tries hard to work even if you have an svn checkout but no executable13:21
danchrthat's hard for Mercurial, as the dirstate is a binary file and _definitely_ an internal implementation detail13:21
danchralso, it doesn't store the tags&13:21
cwillu_at_work (~cwillu@cwillu.com) left irc: Ping timeout: 265 seconds13:22
arigatoamaury_: do you know the point, in fast-forward/../app_main.py, of having two names for the -i option?  "inspect" and "interactive"13:22
danchrantocuni: is there a preferred way to check for an executable in PyPy?13:23
antocuniI think that the pylib has something for that13:24
Action: antocuni looks13:24
antocuniyes, that one13:24
pedronis (~pedronis@253-109.104-92.cust.bluewin.ch) left irc: Ping timeout: 265 seconds13:27
antocunironny: I'm going afk, be back in ~1hr13:30
ronnyantocuni: its up13:30
antocunigood, so I can start to clone it and have it finished for when I'm back :-)13:31
mat^2 (~mathias@ joined #pypy.13:33
amaury_arigato: the two names also appear in the sys.flags structure13:33
arigatoyes, I wonder why13:33
arigatoyou mean that CPython has a sys.flags starting from 2.6, ok, I didn't know13:33
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) left irc: Quit: stakkars13:35
arigato"PYTHONINSPECT=1 python" shows a python where inspect but not interactive is set, indeed13:35
arigatowhat the exact rules is on CPython, I have no way to know but reading the source I suppose13:36
amaury_IIUC, "inspect" causes the interpreter to show the >>> prompt13:36
amaury_"interactive" is related to tty or buffering options13:37
arigatoso in PyPy it's different (i.e. a bug)13:37
arigato"inspect" is never read13:37
arigatoalso, we most probably don't care, but "python -33 -3" shows a sys.flags where py3k_warning is 313:39
amaury_I heard someone here saying that we should not implement this -3 flag13:39
arigatoyes, that's another issue13:40
pedronis (~pedronis@253-109.104-92.cust.bluewin.ch) joined #pypy.13:40
arigatoI meant to say the same e.g. with "python -dd -d"13:40
arigatoif you run a script that sets PYTHONINSPECT to '1', then you end up in a prompt where sys.flags shows both inspect=0 and interactive=013:41
arigatoI've no real clue if there is a strong logic at all13:42
amaury_Py_GETENV("PYTHONINSPECT") is called twice in the same function...13:46
arigatoyes, it's really a huge custom grown-up mess, disguised into some nice superficial tools like sys.flags13:48
cwillu_at_work (~cwillu@cwillu.com) joined #pypy.13:50
amaury_danchr: about sys.mercurial: did you look at module/sys/version.py?13:51
danchramaury_: that's what I'm patching :)13:51
amaury_"""we hack the number directly out of the .svn directory to avoid to depend on an external 'svn' executable in the path."""13:51
danchrthat's going to be frightfully hard for Mercurial, and downright impossible for tags&13:52
danchrwhen an archive is made, Mercurial puts a '.hg_archival.txt' file in there, but I haven't added any logic to parse it13:53
stakkars (~tismer@77-21-85-35-dynip.superkabel.de) joined #pypy.13:54
danchrit's worth noting that the current hack might not work for Subversion 1.7 (if it's ever released)13:55
kenaan03arigo 10r8003414 pypy/branch/fast-forward/pypy/translator/goal 07M(app_main.py test_app_main.py): With this particular (custom) logic, then we mimic more closely CPython's ...13:58
Action: danchr clones PyPy14:08
tav (~tav@ joined #pypy.14:12
bsod1 (~osa1@ joined #pypy.14:23
danchrokay, here's a tested version of the sys._mercurial patch: http://paste.pocoo.org/show/304302/14:23
danchrshall I push it to Subversion?14:24
amaury_will it work on Windows?14:24
danchrI have absolutely no idea, but I think so :)14:25
danchror rather, I can't think of a reason why it shouldn't14:25
exarkundanchr: you didn't test on windows?  but you have "python.exe" in your example output :)14:25
danchrexarkun: that's what CPython calls its build results :P14:26
exarkunon Windows yes...14:26
danchrand on Mac OS X too14:26
Action: exarkun skepticals14:26
danchruntil it's installed, then it's called something else14:26
exarkunthis is all translation-time stuff, I guess?14:26
danchrthe patch? yes14:27
exarkunSo that os.putenv isn't going to screw over anyone who tries to run hg on PyPy14:27
exarkunI guess it's a little broken for funny path names14:27
danchroh, yeah it might be14:27
exarkunYou should be able to use subprocess.call() instead of os.popen() pretty easily, I think.14:27
exarkunwell actually I have no idea.  Let's say "I hope" :)14:28
amaury_on Windows, a sensible place to look for hg is os.path.join(sys.prefix, 'Scripts', 'hg')14:28
danchrerm, I don't think subprocess.call would work, as I need standard output14:29
danchramaury_: I think most Windows users are going to have 'hg' on their path, and probably obtain it by installing TortoiseHg14:30
danchrI think TortoiseHg installs a frozen CPython executable14:30
amaury_it's not my case, at least14:30
danchrin that case, it'll just fail to get any version info, which I'd consider acceptable14:31
danchrI don't know how common having 'hg' installed, but not on your path is, but I suspect not very&14:32
amaury_I don't have any "python" in my path...14:35
bobbyz_ (~bobbyz@ joined #pypy.14:39
exarkundanchr: subprocess.Popen then14:41
danchrexarkun: yeah, I'm making it work right now :)14:41
danchran updated version: http://bitbucket.org/danchr/pypy-patches/src/tip/sys-mercurial.diff14:46
Action: exarkun doesn't see anything wrong14:51
antocunidanchr: you can checkin directly on the mercurial repo if you want, I just added you as a writer :-)14:57
danchrthanks :)14:57
ronnyi think i'll have to figure a incrementive replay updater later on14:57
antocunironny: I thought you already had it14:57
ronnyit wonr wrong somewhere14:58
ronnyand i completely killed it for a very robust recreate from piped svn dump14:58
ronnyi should investigate in --startrev14:58
Action: danchr uses it for hacking on PyPy14:58
antocunironny: if it's easier for you, I'm fine for saying that at some point we will just stop using svn and start using mercurial14:59
antocuniwithout incremental update14:59
ronnyantocuni: the actual convert is still incremental14:59
ronnyjust the replay metadata14:59
ronnyis created by piping a complete svn dump into the tool14:59
antocuniah, ok14:59
ronny(takes about 8 minutes)14:59
antocuniwell, but then it should not be a problem15:00
antocuniI don't think you will have to run it many times15:00
ronnycurrently a few times every day should suffice15:00
danchran updated version of the patch that I'm currently testing: http://paste.pocoo.org/show/304323/15:01
danchrronny: would you perhaps take a look?15:01
antocuni_ (~antocuni@host96-125-dynamic.11-79-r.retail.telecomitalia.it) joined #pypy.15:03
Action: antocuni_ hates his network :-(15:03
antocuni_<antocuni> well, but then it should not be a problem15:04
antocuni_<antocuni> I don't think you will have to run it many times15:04
amaury_arigato: another crash on windows15:04
amaury_I recompile with lldebug15:04
antocuni (~antocuni@host237-120-dynamic.11-79-r.retail.telecomitalia.it) left irc: Ping timeout: 276 seconds15:05
amaury_arigato: approximative call stack: http://paste.pocoo.org/show/30432715:05
danchrso, the conversion didn't create a .hgignore file15:07
danchrshall I commit one to Subversion?15:07
ronnydanchr: please do so15:14
ronnydanchr: also the patch seems fine15:15
kenaan03danchr 10r8003514 05A(pypy/trunk/.hgignore): Add .hgignore file. ...15:15
danchrronny: I'm testing a new one which does two things15:16
danchr1) disable user configuration by setting HGRCPATH to os.devnull15:16
danchr2) disregard 'tip' and use the branch15:16
danchrthis is it: http://paste.pocoo.org/show/304335/15:17
danchr(I'll update the commit message with improved output once the build finishes)15:18
Action: danchr pushes it15:31
kenaan03danchr 10r8003614 pypy/trunk/pypy/module/sys 07M(__init__.py version.py): Add sys._mercurial attribute. ...15:32
kenaan03david 10r8003714 pypy/branch/arm-backend/pypy/jit/backend/arm 07M(assembler.py codebuilder.py opassembler.py regalloc.py): Make backend translate again ...15:34
amaury_arigato: I have the error "in pypy_g_OptVirtualize_optimize_GETARRAYITEM_GC: fixed getitem out of bounds"15:39
amaury_index=0, length=0...15:39
amaury_should I try to dump the current loop?15:40
panni_ (hannes@ip-178-203-85-85.unitymediagroup.de) joined #pypy.15:42
amaury_and how?15:48
antocuni__ (~antocuni@host37-57-dynamic.56-82-r.retail.telecomitalia.it) joined #pypy.15:50
antocuni_ (~antocuni@host96-125-dynamic.11-79-r.retail.telecomitalia.it) left irc: Ping timeout: 276 seconds15:52
arigatoin your case, so that we see also the unoptimized loops15:56
mcdonc_ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) joined #pypy.15:56
mcdonc_ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) left irc: Read error: No route to host15:57
idnar (~quassel@unaffiliated/idnar) left irc: Read error: Operation timed out16:03
ErKa (keryell@m3.wifi.enstb.org) left irc: Ping timeout: 240 seconds16:03
mcdonc__ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) joined #pypy.16:05
idnar (~quassel@unaffiliated/idnar) joined #pypy.16:06
mcdonc__ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) left irc: Read error: No route to host16:07
antocuni__ronny, danchr: do you have any clue why I get this? http://paste.pocoo.org/show/304356/16:08
Nick change: antocuni__ -> antocuni16:08
danchrbreakage in bitbucket?16:09
danchrtry using SSH :)16:09
kenaan03arigo 10r8003814 pypy/branch/getopt-appmain/pypy/translator/goal 07M(app_main.py test_app_main.py): Whack at app_main until I get a result that I'm kind of happy with. ...16:10
antocuniindeed, ssh worked16:10
antocuniuhm, I tried to setup bitbucket and CIA.vc to write commits on this channel, but it seems it didn't work16:11
dcolish (~dcolish@li95-246.members.linode.com) got netsplit.16:12
JStoker (jstoker@unaffiliated/jstoker) got netsplit.16:12
ivan (~ivan@unaffiliated/ivan/x-000001) got netsplit.16:12
tote (me@c-edcfe655.018-24-73746f48.cust.bredbandsbolaget.se) got netsplit.16:12
antocuniok, it seems that CIA saw the commits16:12
antocuninow, I just need to understand how to convince it to put the messages on the channel16:13
amaury_arigato: I'm surprised to see RPythonAnnotator.flowin() in the jitcode!16:14
amaury_ah no sorry, I'm running translate.py16:14
amaury_of course16:14
dcolish (~dcolish@li95-246.members.linode.com) returned to #pypy.16:16
JStoker (jstoker@unaffiliated/jstoker) returned to #pypy.16:16
ivan (~ivan@unaffiliated/ivan/x-000001) returned to #pypy.16:16
tote (me@c-edcfe655.018-24-73746f48.cust.bredbandsbolaget.se) returned to #pypy.16:16
xorAxAxantocuni: there is a filter active16:17
xorAxAxantocuni: birkenfeld and I set it up, then it was switched to kenaan. why CIA again?16:17
antocunixorAxAx: I was trying to get bitbucket commits16:17
antocuniand for kenaan I think we have to write something custom16:17
xorAxAxok, makes sense16:18
antocuniI'm fine to use keenan on the long run, but I'd like to have cia in the meantime16:18
xorAxAxi have another project running on bitbucket which stopped working with CIA as well16:18
antocunixorAxAx: well, looking here http://cia.vc/stats/project/pypy it seems that the commits got to cia16:19
jimbaker (~jimbaker@c-24-8-39-78.hsd1.co.comcast.net) joined #pypy.16:19
antocuniI think we just need to instruct the bot to print them16:19
antocunixorAxAx: do you know how to enable it again?16:19
xorAxAxyes, of course16:19
xorAxAxwe need to modify the filter16:20
xorAxAxantocuni: which branch do you send?16:20
xorAxAxyes, we need to somehow identify the commits coming from bitbucket16:21
bgola (~Bruno@ joined #pypy.16:21
xorAxAxcurrently CIA only shows the roundup branch16:21
antocuniah, is it maybe the "module" field in the bitbucket setting?16:21
xorAxAxwe can also filter on module, yes16:21
antocuniso far, I just configured project==pypy and module==""16:22
xorAxAxthen set the module to bitbucket16:22
xorAxAxand push another changeset16:22
antocunidid you already modify the filter?16:23
Action: antocuni tries16:23
xorAxAxah, it didnt save it :-(16:23
danchrronny: could you perhaps try running in incremental update?16:24
xorAxAxantocuni: now try again16:24
mcdonc__ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) joined #pypy.16:24
Action: antocuni pushes16:25
CIA-903antocuni * 6f6dc96c294b r39482 10bitbucket/: import svn externals in this branch too * 14http://bitbucket.org/pypy/pypy/changeset/6f6dc96c294b/16:25
antocunixorAxAx: thanks :-)16:25
xorAxAxno problem :)16:26
Nick change: murmelpumpe -> timonator16:26
mcdonc__ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) left irc: Read error: Connection reset by peer16:26
Nick change: timonator -> timos16:27
danchrantocuni: you should be able to use subrepos for those with Mercurial 1.7.1 and later16:27
antocunidanchr: yes, but later :-)16:28
antocuniI think that the big step is to migrate the whole pypy16:28
danchralso, you might want to configure an email service for mailing the list :)16:28
antocunithen, we can migrate the rest one by one16:28
antocuniI did16:28
danchrd'oh :)16:29
antocunihowever, I don't like the bitbucket default one because it uses "Bitbucket" in the From field16:29
amaury_arigato: so, now I have a large log file (280Mb), is the last "jit-log-noopt-loop" relevant?16:29
danchrI think you'll have to host hgweb yourself to get around that16:29
danchr(or hgweb & SSH)16:30
arigatoamaury_: yes16:30
nettok (~quassel@ joined #pypy.16:30
arigatoamaury_: keep in mind that as long as you are in the debugger, the log may not have been flushed16:30
arigatoyou can try to call  fflush(pypy_debug_file)16:30
arigatopypy_debug_file is a normal FILE* variable in C16:31
amaury_no, the program exited already16:32
amaury_is it possible to cut part of the log and feed it to some test?16:33
arigatonot directly16:33
arigatoyou need to rewrite some parts, like the descrs, into the same format as test_optimizeopt.py16:33
amaury_I'll try16:35
amaury_the loop has 340 operations16:35
bsod1 (~osa1@ left irc: Ping timeout: 260 seconds16:40
mvt (~mvantelli@ left irc: Quit: Leaving16:44
ronnydanchr: started one, should be up in a minute16:46
ronnydanchr: someone has to merge afterwards i think16:46
danchrdealing with that on the Mercurial side is easy :)16:47
antocunipff, it also seems that bb emails does not show the branch16:47
mcdonc__ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) joined #pypy.16:47
zooko (~user@173-164-32-246-colorado.hfc.comcastbusiness.net) joined #pypy.16:48
mcdonc__ (~mcdonc@168-92-165-82.ipv4.firstcomm.com) left irc: Read error: Connection reset by peer16:49
dmalcolm (~david@nat/redhat/session) joined #pypy.16:50
dmalcolm (~david@nat/redhat/session) left irc: Changing host16:50
dmalcolm (~david@nat/redhat/x-gdzhbqlmqizllibo) joined #pypy.16:50
antocuni_ (~antocuni@host65-124-dynamic.7-79-r.retail.telecomitalia.it) joined #pypy.16:53
woulf (~quassel@p5796F5C3.dip.t-dialin.net) joined #pypy.16:54
antocuni (~antocuni@host37-57-dynamic.56-82-r.retail.telecomitalia.it) left irc: Ping timeout: 276 seconds16:55
CIA-903bivab * c1dae116dd3e r39483 10bitbucket/pypy/jit/backend/arm/ (4 files): Make backend translate again * 14http://bitbucket.org/pypy/pypy/changeset/c1dae116dd3e/16:57
CIA-903danchr * 805815216ca7 r39484 10bitbucket/.hgignore: Add .hgignore file. * 14http://bitbucket.org/pypy/pypy/changeset/805815216ca7/16:57
CIA-903danchr * 1afa61e2bcb8 r39485 10bitbucket/pypy/module/sys/ (__init__.py version.py): (log message trimmed)16:57
CIA-9Add sys._mercurial attribute.16:57
CIA-9Based on the equivalent patch to CPython 2.7:16:57
CIA-9Example output: * 14http://bitbucket.org/pypy/pypy/changeset/1afa61e2bcb8/16:57
CIA-903arigo * 5e8eb0237e7f r39486 10bitbucket/pypy/translator/goal/ (app_main.py test2/test_app_main.py): 16:57
CIA-9Whack at app_main until I get a result that I'm kind of happy with.16:57
CIA-9Add tests for parse_command_line(). * 14http://bitbucket.org/pypy/pypy/changeset/5e8eb0237e7f/16:57
antocuni_personally, I like keenan formatting more than CIA's one16:58
antocuni_anyway, I don't care too much right now :-)16:58
fijalcan't we just move to bitbucket while we're at it?16:58
antocuni_fijal: let me setup buildbot first16:58
fijalinstead of running replay script couple times a day?16:58
danchrhosting Mercurial yourself isn't all that hard FWIW16:58
Action: antocuni_ stops the current buildbot16:58
fijaldanchr: yeah, but I guess we want more than just mercurial16:58
fijalwe want wiki for sure16:59
danchrmakes sense16:59
fijalprobably moving away from roundup would not hurt either16:59
danchrthe bitbucket tracker is a bit quirky, but it works16:59
danchrI think Atlassian plans to integrate JIRA or something16:59
danchror rather, I suspect they do16:59
fijalroundup is not really great16:59
fijalso my expectations are low16:59
danchrthe two bug trackers I use most are Mercurial (which I think is roundap) and hgsubversion (which is bitbucket)17:00
antocuni_uhm, is it expected that buildbot.pypy.org still gives me the index page even if buildbot is stopped?17:00
danchrI'd say they're about equal, but the main problem with Bitbucket is that their auto-closing is buggy17:00
kenaan03arigo 10r8003914 pypy/trunk/pypy 07M(llmodel.py model.py assembler.py regalloc.py rx86.py test_ztranslation.py pyjitpl.py rstack.py lloperation.py stack.h): Merge branch/jit-stackcheck. ...17:00
fijalantocuni_: Shift+R17:00
kenaan03arigo 10r8004014 06D(pypy/branch/jit-stackcheck/): Branch merged. ...17:01
fijalarigato: that was quick :)17:01
danchrarigato: now you'll have to ask ronny to convert that too :)17:01
antocuni_uhm, I already tried to refresh the page. Anyway, now it's offline, good17:01
arigatodanchr: as long as I don't see an announcement on pypy-dev, I'll continue to make and merge branches :-)17:02
arigatoantocuni_: btw, I can help you making parts of the codespeak svn repo read-only17:02
arigatowhen you need it17:02
antocuni_arigato: thanks17:03
antocuni_I'll tell you when to do that17:03
Action: fijal would like to have discussion "what we do with profiling" on the sprint17:03
fijalthis is the first thing I see that is python-3-only17:05
zookoFWIW I use and adminstrate trac for http://tahoe-lafs.org and am satisfied with it.17:05
fijalzooko: I suppose exarkun would happily disagree17:06
antocuni_uhm, cloning the pypy repo from bitbucket on codespeak is super-slow17:07
antocuni_no clue why17:07
exarkunI only spent an hour this weekend on Twisted trac maintenance17:07
antocuni_uhm, actually everything is slow on codespeak17:09
antocuni_there is dot taking 99% of cpu17:09
antocuni_I can't17:09
antocuni_it's apache17:09
antocuni_maybe we should kill the "use codespeak is you don't have dot" feature17:10
antocuni_ok, now it finished17:10
arigatoI killed it17:10
antocuni_ah :-)17:10
zookoexarkun: I know there are often problems with some sort of consistency thing on twisted's trac -- possibly multithreaded access to the svn repo causing some sort of other pile-up.17:10
exarkuntrac uses 100% cpu pretty much all the time17:10
arigatoor we could at least add a limit on cpu time, yes17:10
zookoI haven't experienced precisely the same problems on the tahoe-lafs trac, but it has different (and mostly lesser) revision control load.17:11
zookoMine doesn't use a lot of CPU.17:11
fijalarigato: you killed "dot on codespeak" feature?17:11
zookotahoe's probably has much lower user load too.17:11
zookoThan twistedmatrix.com17:11
arigatofijal: no17:11
fijalarigato: maybe we should :)17:12
fijalI guess that's what anto suggested17:12
ronnydanchr: just takes 10 minutes17:12
zookoYes, mine's idling at 0% CPU according to top.17:12
exarkunokay I guess it must not be 100% usage all the time, since it has only used 583 CPU minutes since it restarted last night17:12
ronnyarigato: although it'll be a bit od a pain, the more you add, the more pain we have17:12
exarkunThat's closer to 50% I guess17:12
ronny(not that its much)17:12
exarkunAlthough postgres is also doing lots of stuff17:12
zookoYeah mine has 30 minutes CPU and it was restarted... 17:12
arigatoronny: I'm fine with freezes or something, as long as you announce it first17:13
arigatojust talking about it here is not enough17:13
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) joined #pypy.17:14
kenaan03fijal 10r8004114 pypy/branch/out-of-line-guards/pypy 07M(runner.py jtransform.py history.py pyjitpl.py rclass.py): Progress on out-of-line guards. Use weakref for storing affected codes ...17:14
exarkunmy latest complain about trac, though, is that the standard user registration plugin (no registration built in) lets users corrupt your password file17:15
zookoEh, at least a day or two ago, can't find it.17:15
zookoWe use sqlite instead of postgres and we have the latest stable version of trac -- 0.12 -- where twistedmatrix.com had an older version last time I looked.17:16
zookoBut if I had to guess, I would guess that we would probably have similar problems to you if we had similar load of users and revisions.17:16
zookoHm, how does it let users corrupt your password file?17:16
zookoWe use some registration thing that stores auth information in a file that isn't your unix accounts file. An "htpasswd".17:17
antocuni_uhm, we have "almost" a problem on codespeak17:17
antocuni_there are only 264M left on the home partition17:17
antocuni_which is where buildbot lives17:17
fijaldanchr: pypy/module/_stackless/test/17:18
fijalwhy is this in .hgignore?17:18
ronnyarigato: let me kill a big file17:19
ronnyeh antocuni_ 17:19
antocuni_arigato: could you create me a directory on /dev/sda1 where to put all the old buildbot stuff?17:19
danchrthe .hgignore was generated by hgsubversion, but I removed or collapsed some of it17:19
kenaan03arigo 10r8004214 07M(pypy/branch/getopt-appmain/pypy/translator/goal/app_main.py): Fix. ...17:20
CIA-903arigo * 74bbe61137d9 r39488 10bitbucket/pypy/ (10 files in 8 dirs): 17:20
CIA-9Merge branch/jit-stackcheck.17:20
CIA-9Reintroduces (cheap) stack checks in JITted code. * 14http://bitbucket.org/pypy/pypy/changeset/74bbe61137d9/17:20
CIA-903convert-repo * 7bacafd654eb r39487 10bitbucket/: closed branch jit-stackcheck * 14http://bitbucket.org/pypy/pypy/changeset/7bacafd654eb/17:20
ivan (~ivan@unaffiliated/ivan/x-000001) left irc: Ping timeout: 240 seconds17:20
ambv (~ambv@host- left irc: Quit: enthusiastic non-action17:21
antocuni_ronny: thanks for having removed the file, but I fear is still not enough for holding a backup of buildbot logs17:21
antocuni_I suppose I could move them to /tmp, but I'm a bit scaried :-)17:22
exarkunYou might just throw away sufficiently old buildbot logs17:22
exarkunThey're not really good for anything17:22
antocuni_yes, it's just that I didn't want to think about what to keep and what not :-)17:22
arigatocheck them in the pypy repo...  hum17:23
Nick change: antocuni_ -> antocuni17:23
exarkunAlso, sufficiently recent versions of buildbot will compress newly written logs and let you configure a max number to keep (it will delete extras beyond that automatically)17:23
arigatowhere are the logs so far?17:23
antocuniin ~buildmaster/pypy17:23
antocuniI want to copy the whole dir to the new buildbot, so we keep the history of e.g. the tests run in the past days17:24
arigato400 MB "only", so it should fit on ~antocuni --- also bzip2'ed it should be small17:25
antocuniarigato: 400MB? It's 1.717:25
fijalI'm unable to ssh to codespeak17:25
arigatoantocuni: are you counting subdirs too?17:25
antocunidu -h17:25
arigatoah, ok, I see17:26
arigatothe subdirs are not readable by ~arigo, so I missed them --- also they are compressed17:26
antocuniyes, and the test results are actually in the subdirs, so we really want them :-)17:26
arigatowhere should I create the dir?17:27
antocunianywhere outside /home, I think17:27
arigato'/' is already a bit of a mess, if you ask me17:27
antocuni /backup?17:27
arigatono clue what it's used for17:28
antocuni /backup/old-svn-based-buildbot17:28
antocuniwell, if everything goes well we can kill it soon17:28
antocuniI just want to play safe17:28
ivan (~ivan@unaffiliated/ivan/x-000001) joined #pypy.17:29
Moku (~John@ left irc: Ping timeout: 245 seconds17:30
Moku (~John@osbk-4d087b2b.pool.mediaWays.net) joined #pypy.17:30
fijalcodespeak is unusable17:34
fijalI can't ssh, can't svn up17:34
arigatoworks from here17:35
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) left irc: Read error: Connection timed out17:35
antocuniworks from here too17:35
arigatomight be a bit slow, if anto is busy copying files from one partition to another17:35
antocuniyes, but you should definitely be able to ssh in17:35
fijalnot in like 5 minutes17:36
fijalbetter now17:36
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) joined #pypy.17:36
fijalnot by much though17:36
fijalI could ssh, still svn up is taking forever17:36
antocunithe big copy is finished17:39
fijalI've cleaned up some stuff17:42
arigatoantocuni: sorry if I seem to insist, but if you're planning to have us switch very soon, then you should announce it to pypy-dev17:44
antocuniarigato: yes, but as I told I'd like to be sure that at least buildbot works :-)17:44
antocuniI'm doing it right now17:44
Action: arigato -> dinner17:46
arigatosee you17:46
antocunisee you17:46
arigato (~arigo@fwstups.cs.uni-duesseldorf.de) left irc: Quit: See you17:46
fijalI have to go and fetch something17:47
antocuniok, the new buildbot seems to be up and running17:51
antocuniand it correctly show the old test results17:52
antocunialthough in a wrong order it seems, but I think it's because of a change I did to play well with hg (and it seems to play not-so-well with svn, do we care?)17:52
Action: antocuni starts a test build17:52
ronnyantocuni: might be related to revision numbers17:53
antocuniI remember I changed how they are sorted, because of course we cannot sort alphabetically17:53
CIA-903fijal * e407bdf735c5 r39489 10bitbucket/pypy/ (5 files in 4 dirs): 17:53
CIA-9Progress on out-of-line guards. Use weakref for storing affected codes17:53
CIA-9and cheat (horribly) to pretend that LoopToken is a low-level pointer * 14http://bitbucket.org/pypy/pypy/changeset/e407bdf735c5/17:53
CIA-903arigo * e73c414c7186 r39490 10bitbucket/pypy/translator/goal/app_main.py: Fix. * 14http://bitbucket.org/pypy/pypy/changeset/e73c414c7186/17:53
bsod1 (~osa1@ joined #pypy.17:56
antocunironny: please stop to push changes to the repo17:56
antocunielse we keep creating new heads17:56
antocuniwe need to pick a head which has the svn externals merged, else it buildbot cannot work17:57
antocuniah cool, it seems that the latest two checkins did not create heads in the default branch :-)17:58
CIA-903antocuni * e5b73981fc8d r39491 10bitbucket/: merge heads * 14http://bitbucket.org/pypy/pypy/changeset/e5b73981fc8d/18:00
antocuniok, the build is in-progress18:01
__name__ (~name@sburn/devel/name) joined #pypy.18:01
antocuniI'll check is later this evening18:02
antocuniand if everything is ok, I'll send an announcement to pypy-dev18:02
ronnyantocuni: i noticed to late that i had it running, i made myself a new command for later, tell me when i can mess up things again18:04
antocunironny: not before tomorrow morning, I'd say18:05
antocuniso that if everything goes well, we can have nightly builds this night18:05
ErKa (~keryell@keryell.pck.nerim.net) joined #pypy.18:16
romao (~romao@unaffiliated/errado) left irc: Ping timeout: 240 seconds18:17
antocuni (~antocuni@host65-124-dynamic.7-79-r.retail.telecomitalia.it) left irc: Ping timeout: 276 seconds18:18
danchrquick conversions of the subrepos to Mercurial  http://hg.villiom.dk/python/pypy/18:22
jimbaker (~jimbaker@c-24-8-39-78.hsd1.co.comcast.net) left irc: Quit: jimbaker18:26
ErKa (~keryell@keryell.pck.nerim.net) left irc: Ping timeout: 264 seconds18:26
jimbaker (~jimbaker@c-24-8-39-78.hsd1.co.comcast.net) joined #pypy.18:26
zooko (~user@173-164-32-246-colorado.hfc.comcastbusiness.net) left irc: Quit: ERC Version 5.3 (IRC client for Emacs)18:38
fijal (~fijal@41-134-41-226.dsl.mweb.co.za) left irc: Ping timeout: 240 seconds18:47
ErKa (~keryell@keryell.pck.nerim.net) joined #pypy.18:48
santagada (~lsantagad@ joined #pypy.18:54
santagadawhat is happening with mercurial? is something mirroring commits?18:56
kenaan03david 10r8004314 pypy/branch/arm-backend/pypy/jit/backend/arm 07M(regalloc.py regalloc.py): Next register allocation fix ...19:01
zooko (~user@173-164-32-246-colorado.hfc.comcastbusiness.net) joined #pypy.19:06
dan_waterworth (~dan@ left irc: Quit: Leaving19:07
dmalcolmRandom newbie question: I'm attempting to patch code generation to supply python-level source code in C-level comments (in translator/c/funcgen.py) .  I was going to use the "offset" of the objspace.flow.model.SpaceOperation instances, but they're almost all the default (-1).  Am I going to need to go through the code fixing this first, or am I missing something19:08
Alex_Gaynordmalcolm: I fear you'll need to wait for {arigato,cfbolz,fijal} who know this part of the code.19:08
dmalcolmah, thanks19:09
Action: dmalcolm hopes he's correctly understanding the various levels of the system19:09
Alex_Gaynorflow is the very first part of the translation tool chain19:09
dmalcolm(fwiw, am looking at making the autogenerated .c code easier to read)19:14
amaury_nice goal19:14
Alex_Gaynoramaury_: DO you know if there's any reason buildbot shouldn't be triggered ATM?19:16
amaury_no, why?19:18
amaury_maybe they were already updated to track the hg repo19:18
jcea_BT (jcea@jabber.hst.ru) left #pypy.19:25
danchrI think they were19:27
amaury_ (5bc3480e@gateway/web/freenode/ip. left irc: Quit: Page closed19:28
ErKa (~keryell@keryell.pck.nerim.net) left irc: Read error: Operation timed out19:29
jcea_BT (~jcea@jabber.hst.ru) joined #pypy.19:30
kenaan03hakanardo 10r8004414 pypy/branch/jit-unroll-loops/pypy 07M(compile.py test_memmgr.py test_optimizebasic.py test_pypy_c.py): fixed tests ...19:52
cwillu_at_work (~cwillu@cwillu.com) left irc: Ping timeout: 272 seconds19:59
santagada (~lsantagad@ left irc: Quit: Ex-Chat20:01
Alex_Gaynortranslating with cpython is insane :(20:15
idnar (~quassel@unaffiliated/idnar) left irc: Ping timeout: 260 seconds20:17
Moku (~John@osbk-4d087b2b.pool.mediaWays.net) left irc: Read error: Connection reset by peer20:20
mat^2 (~mathias@ left irc: Ping timeout: 260 seconds20:22
Moku (~John@osbk-4d087b2b.pool.mediaWays.net) joined #pypy.20:23
exarkunHey twisted_tcp is getting slower what are you doing, http://speed.pypy.org/timeline/?exe=1%2C3&base=2%2B35&ben=twisted_tcp&env=tannit&revs=5020:36
mat^2 (~mathias@ joined #pypy.20:36
lizardo (~lizardo@ left irc: Quit: Leaving20:42
antocuni (~antocuni@host65-124-dynamic.7-79-r.retail.telecomitalia.it) joined #pypy.20:51
bsod1 (~osa1@ left irc: Remote host closed the connection20:55
antocunicool, t20:56
antocunithe hg based buildbot "worked20:56
antocunithere are tons of failing tests, but probably they are legitimate20:57
Alex_Gaynorantocuni: What do we need to enter to do branches?20:58
ronnyantocuni: did they fail before?20:58
ronnyAlex_Gaynor: hg branch thename20:58
ronnythen the next commit will be on that branch20:58
Alex_Gaynorronny: no, I meant to trigger a build of that branch in the buildbot20:59
Alex_Gaynorbefore we entered `branches/fast-forward` for example20:59
antocunironny: I don't know, as there are some checkins since the yesterday20:59
ronnyi see20:59
antocunimost of the failures are related to the jit, I think it's because of armin's stackcheck branch20:59
antocuniAlex_Gaynor: just the name of the mercurial branch21:00
antocuniAlex_Gaynor: however, keep in mind that branches don't work out of the box21:00
antocunibecause they miss externals21:00
antocuniI'll write an howto to deal with active branches21:01
Alex_Gaynorantocuni: so if I just triggered a fast-forward build, is it going to break because of that?21:01
antocunie.g., it misses the testrunner directory21:01
lucian (~lucian@78-86-217-168.zone2.bethere.co.uk) left irc: Remote host closed the connection21:01
ronnyantocuni: is testrunner a subrepo or a copy?21:02
antocuniAlex_Gaynor: look at 85b74c8de1c921:02
antocunironny: it's an svn rubrepo21:02
antocuniwe will probably want to make it an hg subrepo at some point21:03
ronnyoh, i see21:03
ronnyi gess i'll pester the bitbucket guys a bit21:04
antocunifor what?21:04
ronnywrt something akin to organisations21:05
ronny(the github thing to have a group of people manage the repos of a project021:05
ronnyi know they are working on something, but i have no idea how close to completion21:05
antocuniah, so that if we create a new repo under pypy a selected group of people automatically become admin?21:05
ronnyantocuni: no, a group of people will be allowed to create repos under pypy and be admin for them21:06
antocuniwhile you are at it, tell them that I'd like 1) the name of the committer as the sender in the email and 2) the branch name also in the email21:07
ronnyyou might want to open a bug for that21:08
antocuniI think I'll do21:09
antocuniwhere is the official issue tracker?21:09
Arfrever (~Arfrever@gentoo/developer/Arfrever) left irc: Quit: Ex+re (KVIrc 4)21:16
Alex_Gaynorantocuni: http://bitbucket.org/site/master/issues/new21:20
idnar (~quassel@unaffiliated/idnar) joined #pypy.21:22
stakkars_ (~tismer@brln-4db94d1a.pool.mediaWays.net) joined #pypy.21:29
Trundle (~andy@python/site-packages/trundle) joined #pypy.21:31
woulf (~quassel@p5796F5C3.dip.t-dialin.net) left irc: Remote host closed the connection21:31
Ademanwait, did I read wrong or are we close to having hg support?21:55
stakkars_ (~tismer@brln-4db94d1a.pool.mediaWays.net) left irc: Quit: hasta la vista21:59
dcolish (~dcolish@li95-246.members.linode.com) left irc: Quit: Coyote finally caught me22:05
dcolish (~dcolish@ joined #pypy.22:08
danchrantocuni: I converted some of the subrepos to Mercurial  they're at http://hg.villiom.dk/python/pypy/22:10
antocunidanchr: thanks22:10
danchryou might have to hack the PyPy build infrastructure a bit to use them as-is22:11
antocunithe question is more whether we want to migrate them or not22:12
antocunifor most, I'd say yes22:12
antocunibug e.g. pyrepl is really an external project22:13
danchrcan't help you with that one, but I figured the conversions might be useful if you do :)22:13
antocunisure :-)22:13
antocunidid you use the authormap?22:13
danchralso a tagpath and I renamed the one branch in pyrepl so it wasn't called ../branch/pyrepl22:14
antocuniah, I never realized that there is a branch in pyrepl22:15
danchrI contemplated stripping it, but didn't see the point22:16
danchryou can also close it if you like :)22:16
antocunieh, one thing at a time22:16
antocunifirst, I want to officially switch22:16
mwhudsonyeah, if pyrepl is going to be maintained anywhere else, it's going to be at launchpad.net/pyrepl22:16
antocunithen, we'll take care of the rest :-)22:16
danchrworking with Bazaar is a pain22:17
antocunimwhudson: then let's keep it on svn, else it will be a nightmare for us22:17
mwhudsonantocuni: oh, to include it in pypy still?22:17
antocuniyes, so far pyrepl is included by default22:17
antocuniwe need it for the prompt22:17
mwhudsonmaybe i should use it as an excuse to learn mercurial then :)22:17
antocuniwell, we don't "need" it, but we use it for readline22:18
mwhudsondanchr: the pyrepl import should be one directory deeper22:20
danchrmwhudson: I'm probably biased, but my experience is that Bazaar sucks compared to Mercurial :)22:20
mwhudsondanchr: funnily enough...22:20
danchrmwhudson: yeah, that's why I suggested that using them as-is would require a bit of hackery to PyPy22:21
antocunielbowtone: vcs22:21
elbowtonecodeville is clearly the best VCS ever22:21
antocuniok, let's migrate to that22:21
mwhudsondanchr: well, not even to be useful to pypy, just to be a sensible import22:21
mwhudson(i realise this is because the layout in svn is whack)22:22
antocuniah yes, indeed22:22
danchrI can do that, with a second pass conversion with 'hg convert'22:22
antocuniin pypy we import just one dir of pyrepl, not the whole repo22:22
antocunithe same for greenlet, iirc22:22
antocunithis make it also more difficult to migrate pyrepl to mercurial, because we would need some custom logic to put it on sys.path22:23
antocuni(or use a symlink)22:23
mwhudsonjust do the conversion right :-)22:23
danchrif that's what you'd like, sure :)22:24
antocuniwell, my point is that you cannot import just a dir inside a subrepo, you need to import the whole subrepo22:24
danchrah, I see the toplevel in pyrepl is empty22:25
mwhudsonantocuni: repo in svn means, if anything, "the whole damn thing"22:26
antocuniI know22:26
ronnysvn's hierachial structure makes some things more easy22:26
mwhudsonone of the design decisions with svn that didn't play out so well on balance i think..22:26
antocunimwhudson: but currently in pypy we import trunk/pyrepl/pyrepl22:26
mwhudsonantocuni: right22:26
ronnyi suppose pypy could get a contrib area where one can put reasonably well structured subrepos of stuff like py, py.test, pyrepl22:27
mwhudsonprobably because i or hpk did something dumb when we converted from cvs i guess22:27
ronnyalternatively it could be solved like in pida (where we just have a script grab the externals)22:27
danchrthe problem with Subversion is that important concepts like tags and branches are conventions, and thus not enforced22:27
ronnyi think i'll tweet about the official hg repo now22:29
antocunironny: wait22:30
antocuniI'd like to send a mail to pypy-dev first22:31
antocuniI'd say that as long as the svn repo is writable, the hg one is not "official"22:31
ronnyi didnt tell its the primary one now, just that its up22:31
antocuniok ok :-)22:31
antocuninot really a problem22:31
amaury_ (5645c561@gateway/web/freenode/ip. joined #pypy.22:34
danchrhmmm, seems like I should implement filemap renaming in hgsubversion to get pyrepl right&22:34
amaury_are the builbots already tied to hg?22:35
antocuniamaury_: yes22:35
amaury_but they don't work yet?22:36
antocunino, they should work22:36
antocunihowever, before you can run it on a branch you need to import the svn externals on that branch22:36
antocuniI'm writing an howto right now22:36
amaury_so they don't work yet.22:37
antocuniwell, they do for trunk :-)22:37
amaury_will we continue to work with branches the same way as we did with svn?22:38
amaury_for example, it's difficult to follow the work done in a branch22:39
amaury_pypy-svn entries don't show the branch22:39
amaury_and bitbucket interface cannot show a branch's log22:40
antocuniyes, I realized that :-(22:40
antocuniI think we can ask the bb guys to implement it, or at worst write a small cgi hook that sends the email as we want22:40
ArneBab (~quassel@HSI-KBW-109-192-087-031.hsi6.kabel-badenwuerttemberg.de) left irc: Read error: Connection reset by peer22:40
antocuniI'd say that everything related to this is open to discussion22:41
ArneBab (~quassel@HSI-KBW-109-192-087-031.hsi6.kabel-badenwuerttemberg.de) joined #pypy.22:41
amaury_when should we use branches vs. forks?22:42
zooko (user@173-164-32-246-colorado.hfc.comcastbusiness.net) left #pypy ("#tahoe-lafs").22:42
antocunipersonally, I'd prefer to always use branches, so that the history is kept in one place22:43
antocunibut of course we cannot stop people to do forks22:43
antocuninote that if you do a fork you loose all the niceties like commit emails, irc messages, and buildbot22:43
amaury_note that if you do a branch you loose all the niceties like commit emails, irc messages, and buildbot22:44
amaury_because everything is messed22:44
danchrI'd say use clones or MQ as a staging area; use them until you have something worth sharing?22:45
antocuniuh? how would it be different than now?22:45
danchrnot much perhaps22:46
amaury_today branches are clearly named in emails and IRC22:46
antocuniI agree22:48
antocuniwe will have to fix it22:48
antocunibut first, let's complete the migration :-022:48
kenaan03antocuni 10r8004514 pypy/extradoc/planning/hg-migration 05A(active-branches-howto.txt) 06D(merge-howto.txt): rename the file, and add a paragraph for how to import svn externals ...22:53
amaury_do buildbots really need 19 minutes to "hg update"?22:54
antocuniamaury_: the first time they need to clone the whole repo22:56
antocunithen, they will just pull the new commits22:56
amaury_yes, but the task starts with "rm -rf"22:57
antocuniI'm quite sure than we I ported buildbot to hg I checked that they didn't clone the whole repo all the times22:58
antocuniso, I don't know what's happening :-(22:58
antocunianyway, I'll investigate tomorrow because now it's getting late22:58
Action: amaury_ -> sleep as well22:59
amaury_ (5645c561@gateway/web/freenode/ip. left irc: Quit: Page closed22:59
kenaan03afa 10r8004614 pypy/branch/fast-forward 05A(.hgignore) 07M(argument.py asmmemmgr.py llmodel.py test_asmmemmgr.py model.py assembler.py codebuf.py regalloc.py rx86.py test_ztranslation.py viewcode.py pyjitpl.py __init__.py interp_ ...23:03
antocuni (~antocuni@host65-124-dynamic.7-79-r.retail.telecomitalia.it) left irc: Ping timeout: 276 seconds23:07
Nick change: ousado_ -> ousado23:13
ousado (~ousado@p5B3C163F.dip0.t-ipconnect.de) left irc: Changing host23:13
ousado (~ousado@unaffiliated/ousado) joined #pypy.23:13
bobbyz_ (~bobbyz@ left irc: Read error: Operation timed out23:24
__name__ (~name@sburn/devel/name) left irc: Remote host closed the connection23:32
bgola (~Bruno@ left irc: Ping timeout: 240 seconds23:32
cwillu_at_work (~cwillu@cwillu.com) joined #pypy.23:35
cwillu_at_work (~cwillu@cwillu.com) left irc: Ping timeout: 240 seconds23:46
cwillu_at_work (~cwillu@cwillu.com) joined #pypy.23:49
--- Tue Dec 14 201000:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!