#pypy IRC log for Tuesday, 2012-01-17

whitelynx|work (~whitelynx@63.241.75.144) left irc: Quit: Ex-Chat00:01
voidspace (~voidspace@87-194-212-65.bethere.co.uk) joined #pypy.00:15
voidspace (~voidspace@87-194-212-65.bethere.co.uk) left irc: Changing host00:15
voidspace (~voidspace@python/psf/voidspace) joined #pypy.00:15
mfoord (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy.00:17
mfoord (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host00:17
mfoord (~anonymous@python/psf/voidspace) joined #pypy.00:17
amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Ping timeout: 248 seconds00:22
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.00:23
amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy.00:26
EnCuKou (~encukou@ip-94-113-220-25.net.upcbroadband.cz) left irc: Ping timeout: 240 seconds00:36
mutux (~muratux@139.179.102.153) left irc: Read error: Connection reset by peer01:03
squiddy (~squiddy@g229037172.adsl.alicedsl.de) left irc: Quit: Leaving01:17
nickik (~nick@217.150.234.222) left irc: Quit: Verlassend01:19
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Remote host closed the connection01:20
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/20601:30
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64-2/builds/3401:30
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 276 seconds01:42
bbot24Failure: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/77801:49
icrazyhack (~horieyui@183.60.100.166) left irc: Quit: http://www.cnblogs.com/crazyhack01:54
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/192702:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/34002:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/68002:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/33002:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/67802:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/119702:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/86102:00
bbot2Started: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/151402:00
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/79102:00
nedbat (~nedbat@python/psf/nedbat) joined #pypy.02:06
mfoord (~anonymous@python/psf/voidspace) left irc: Quit: mfoord02:07
nettok (~quassel@190.149.184.238) joined #pypy.02:24
Shanita (~John@osbk-4db152de.pool.mediaWays.net) joined #pypy.03:03
Kaskuka (~John@osbk-4db163b1.pool.mediaWays.net) left irc: Ping timeout: 240 seconds03:04
nshepperd (~asdfg@ppp118-209-23-185.lns20.mel4.internode.on.net) left irc: Ping timeout: 240 seconds03:05
bbot23Success: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/101503:10
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 248 seconds03:15
dracman (~draco@212.255.29.213) left irc: Ping timeout: 252 seconds03:18
papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) joined #pypy.03:19
whitelynx (~whitelynx@75.111.197.204) left irc: Ping timeout: 255 seconds03:22
bgola (~bgola@c951826e.virtua.com.br) left irc: Remote host closed the connection03:22
bgola (~bgola@c951826e.virtua.com.br) joined #pypy.03:23
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-win-x86-32/builds/33003:24
dracman (~draco@212.255.43.181) joined #pypy.03:25
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.03:26
bgola (~bgola@c951826e.virtua.com.br) left irc: Ping timeout: 248 seconds03:28
dukebilgewater (~dukebilge@cpe-74-78-69-183.maine.res.rr.com) left irc: Quit: dukebilgewater03:31
oal (u4126@gateway/web/irccloud.com/x-lhdzgqgbidfplosv) left irc: Remote host closed the connection03:37
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-dlgnnjmsxhmyelkc) left irc: Remote host closed the connection03:38
bfirsh (u1308@gateway/web/irccloud.com/x-xcpfqjeimrkfzbxe) left irc: Remote host closed the connection03:38
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.03:39
oal (u4126@gateway/web/irccloud.com/x-hpsfptwsglulpoot) joined #pypy.03:52
bgola (~bgola@c951826e.virtua.com.br) joined #pypy.03:52
Justin__ (~Justin@114-36-59-117.dynamic.hinet.net) joined #pypy.03:55
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/03:57
nedbat (~nedbat@python/psf/nedbat) joined #pypy.03:57
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-wszeoynejeyzyaqp) joined #pypy.03:57
#pypy: mode change '+o Alex_Gaynor' by ChanServ!ChanServ@services.03:57
Justin_ (~Justin@114-34-52-12.HINET-IP.hinet.net) left irc: Ping timeout: 240 seconds03:57
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.03:59
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Client Quit04:00
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.04:03
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/192704:12
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/04:17
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 245 seconds04:19
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #pypy.04:26
justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) joined #pypy.04:28
kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/04:51
anish_ (anish@nat/redhat/x-mpffyrleibflqbsd) joined #pypy.04:57
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) left irc: Quit: DasIch05:02
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) left irc: Quit: Leaving05:09
bbot23Success: 15http://buildbot.pypy.org/builders/pypy-c-Ojit-no-jit-linux-x86-32/builds/86105:13
nettok (~quassel@190.149.184.238) left irc: Ping timeout: 252 seconds05:14
bbot23Success: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-32/builds/151405:14
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.05:17
nirbheek (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Ping timeout: 255 seconds05:19
bbot23Success: 15http://buildbot.pypy.org/builders/pypy-c-app-level-linux-x86-64/builds/67805:20
bbot23Success: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-32/builds/119705:33
k_bx (~k_bx@unstuck-evidence.volia.net) joined #pypy.05:39
bbot23Success: 15http://buildbot.pypy.org/builders/pypy-c-jit-linux-x86-64/builds/68005:48
AlekSi (~AlekSi@85.235.191.82) joined #pypy.05:53
justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) left irc: Ping timeout: 240 seconds05:57
mattip (c1a9466d@gateway/web/freenode/ip.193.169.70.109) joined #pypy.05:59
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Read error: Connection timed out05:59
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.06:02
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) joined #pypy.06:03
mattipa numpy epiphany: reduce and matrix multiplication are just fancy forms of broadcast,06:05
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) left irc: Client Quit06:05
mattipreduce needs an identity to get the first value correctly06:06
mattipbut dot doesn't06:06
mattip (c1a9466d@gateway/web/freenode/ip.193.169.70.109) left irc: Quit: Page closed06:15
mikefc_a koen from mattip06:21
timotimois that anything like a koan?06:26
mikefc_nothing at all the same.06:34
mikefc_:)06:34
mikefc_mine's the jewish version?06:35
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-64/builds/79106:48
amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) left irc: Read error: Operation timed out06:49
bbot24Failure: 15http://buildbot.pypy.org/builders/pypy-c-jit-macosx-x86-64/builds/34006:52
therve (therve@2a01:e0b:1:116:ca0a:a9ff:fe03:3be7) left #pypy ("Ex-Chat").06:58
asmeurer_ (~asmeurer@129.138.218.98) left irc: Quit: asmeurer_07:10
fijalmorning07:16
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Read error: Connection timed out07:17
squiddy (~squiddy@g229037172.adsl.alicedsl.de) joined #pypy.07:18
k_bx (~k_bx@unstuck-evidence.volia.net) left irc: Ping timeout: 252 seconds07:20
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.07:20
JaRoel (~jaroel|4d@sink.jaroel.nl) left irc: Remote host closed the connection07:25
k_bx (~k_bx@unstuck-evidence.volia.net) joined #pypy.07:43
k_bx (~k_bx@unstuck-evidence.volia.net) left irc: Ping timeout: 248 seconds07:57
papercrane (~papercran@c-76-103-172-115.hsd1.ca.comcast.net) left irc: Quit: Computer has gone to sleep.07:59
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Read error: Connection timed out08:01
k_bx (~k_bx@195.20.130.1) joined #pypy.08:04
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.08:04
nanonyme (nanonyme@unaffiliated/nanonyme) left irc: Ping timeout: 260 seconds08:04
nanonyme (nanonyme@unaffiliated/nanonyme) joined #pypy.08:04
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer08:09
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.08:09
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy.08:17
oal (u4126@gateway/web/irccloud.com/x-hpsfptwsglulpoot) left irc: Remote host closed the connection08:25
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-wszeoynejeyzyaqp) left irc: Read error: Connection reset by peer08:25
oal (u4126@gateway/web/irccloud.com/x-yptucbmimjcxgnyg) joined #pypy.08:26
mattip (c1a9466d@gateway/web/freenode/ip.193.169.70.109) joined #pypy.08:36
mattipthis may be the wrong time of day to ask but,08:36
mattipdoes anyone have a 15 minute presentation on pypy?08:37
mattip(pypy the python interpreter, less about pypy as a translation platform)08:38
fijalyeah08:38
fijalmattip: what do you want to talk about?08:38
fijalall our talks are online08:38
mher (~Adium@tethras5.ucd.ie) joined #pypy.08:38
mattipextradocs?08:38
fijalhttps://bitbucket.org/pypy/extradoc/src/extradoc/talk08:39
fijalyes08:39
fijalif you say something about the topic, I can try to find a precise one08:39
fijalalso, pypy tends to outgrow it's presentations quickly :/08:39
mattipI want to show off how fast it is and why someone should care about pushing it forward08:39
fijalhttps://bitbucket.org/pypy/extradoc/src/extradoc/talk/ep2011/talk08:40
fijalthis maybe?08:40
mattipwhere that someone currently uses matlab alot08:40
fijalwe didn't talk about numpy yet a lot :/08:40
fijalmattip: for when you need it?08:40
fijalI have to prepare a numpy-related presentation anyway08:40
mattipthe presentation is 29/2/1208:41
oal (u4126@gateway/web/irccloud.com/x-yptucbmimjcxgnyg) left irc: Remote host closed the connection08:41
fijalah ok08:41
mattipbut I need to show a draft soonish.08:41
fijalmattip: mine is some time 22/23rd of Feb08:42
fijalok08:42
fijalI can definitely help08:42
fijal15 min you say?08:42
mattiplikewise?08:42
mattipyes, really short08:42
fijalok08:44
mattipthe ep2011 talk look good for a start, I will use that to get the draft approved and then work on it a bit.08:45
fijalwell, it does not talk about numpy08:47
fijalso what do you want to talk about?08:47
mattipsomething like "hey this is not just an academic project, it can help you right now"08:47
oal (u4126@gateway/web/irccloud.com/x-xrgmfvojqecpvhsv) joined #pypy.08:48
mattipyour laplace benchmark should be there08:48
mattipI also want to generate more developers, or get allocated time from work to contribute.08:49
fijalvideo processing maybe?08:49
fijalcool08:49
fijalthat would be really awesome08:49
fijalI would show how quickly we progress08:49
mattipYes, forgot about the cool video demo.08:49
fijaland that "running web apps" is not really our only goal08:50
fijalyou might want to read pieces by travis oliphant and peter wang to know exactly what they accuse us of08:50
mattipwhere?08:50
mattipI would like to graphically show progress with numpy, like the speed benchmark shows constant improvement.08:51
fijalhttp://technicaldiscovery.blogspot.com/2011/10/thoughts-on-porting-numpy-to-pypy.html08:51
mattipyou should say: "how would you show that?"08:52
Action: mattip reading08:52
fijalhttp://blog.streamitive.com/2011/10/17/numpy-isnt-about-fast-arrays/08:53
fijalhttp://blog.streamitive.com/2011/10/19/more-thoughts-on-arrays-in-pypy/08:53
fijalmattip: there is a pretty good slide on europython talk I think which is from speed08:53
fijalwe don't have such detailed data yet though08:53
mattipI'm trying to envision the graph. It is a sloped line. That's as far as I got.08:54
mattipsomething else, time is the xaxis.08:54
mattipnumber of classes in numpypy? number of functions exposed in the module?08:55
fijalI dunno08:56
fijalsize of codebase?08:56
fijalbut it's not telling the entire story08:56
kenaan12fijal numpy-indexing-by-arrays-2 11bec89443f75e 15/pypy/module/micronumpy/: indexing by bool arrays of the same shape, step 108:56
fijalsince codebase is extremely dense08:56
fijalI would just plot by hand features on a timeline08:56
fijallike dtypes/multidim arrays etc08:57
antocuni (~antocuni@89-40.5-85.cust.bluewin.ch) joined #pypy.08:57
fijalantocuni: hi08:57
antocunihi08:57
mattiptimeline is good.08:57
fijalmattip: would be cool if we made a blogpost with that, so I can help08:58
fijalantocuni: I'm thinking about adding sphinx & generating cpython docs to nightly bench08:58
antocuniit sounds like a real world sensible benchmark, yes08:58
antocunihow do we compare?08:59
stakkars_ (~tismer@i59F7726A.versanet.de) joined #pypy.08:59
Action: fijal is still running08:59
fijalI think badly08:59
Action: mattip back to the real world09:00
mattipfijal: thanks.09:00
fijalantocuni: 5:14 vs 3:509:01
fijalantocuni: we loose09:01
fijal3:0509:01
antocuni"good"09:01
antocunia world in which we win for every benchmark is boring :-)09:01
mattip (c1a9466d@gateway/web/freenode/ip.193.169.70.109) left irc: Quit: Page closed09:02
fijalantocuni: a world in which we compete only against cpython is boring :/09:04
rguillebert (~rguillebe@89-40.5-85.cust.bluewin.ch) joined #pypy.09:04
antocunitrue enough09:04
fijalpom pom pom09:07
kenaan12benol jvm-improvements 111afe14470fa3 15/.hgignore: Add files generated by PyCharm to .hgignore09:10
kenaan12benol jvm-improvements 1190cb8d420991 15/pypy/: Fix userspace builders in ootype  Implement the getlength() method of StringBuilders in ootype.09:10
kenaan12benol jvm-improvements 11143a2edf9601 15/pypy/: Fix compute_unique_id to support built-ins in ootype.  Otherwise the translation fails because it doesn't ...09:10
kenaan12benol jvm-improvements 116b650a500d68 15/pypy/rlib/: Declare oo_primitives that should implement some rffi operations.  For now the actual implementations are ...09:10
kenaan12benol jvm-improvements 11c59ec9806ac6 15/pypy/translator/: Add a missing cast from Unsigned to UnsignedLongLong in the JVM backend.09:10
kenaan12benol jvm-improvements 11e31e85a8d333 15/pypy/translator/jvm/opcodes.py: Handle the 'jit_is_virtual' opcode by always returning False09:10
kenaan12benol jvm-improvements 115ba62496112e 15/pypy/: Implemented float2longlong and longlong2float for the JVM.  Also removed the oo_primitive for pypy__rotate...09:10
kenaan12benol jvm-improvements 1179b211ffff3c 15/pypy/translator/jvm/src/pypy/PyPy.java: Simpler implementations of float2longlong and longlong2float.09:10
kenaan12benol jvm-improvements 11259f2ab80ae4 15/pypy/rpython/: Fix the implementation of compute_unique_id for _builtin_type.09:10
antocunigood, I merged the pull request09:11
antocunilet's run the tests now09:11
fijalantocuni: we should do a post about stats09:12
fijaland 201109:12
fijallike09:12
fijal"in 2011 pypy grew from X to Y"09:12
fijal"in 2011 pypy got into fedora, debian, arch & gentoo"09:12
bbot2Started: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1928 [12antocuni, jvm-improvements]09:13
antocunifijal: good idea09:13
fijalantocuni: feel like doing something ni that area ;-)09:15
antocuniyes, but I'm not sure I'll manage to do it during the sprint09:16
antocuniI want to use this time to code :-)09:16
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-wsfyymxqtfzowcvo) joined #pypy.09:16
#pypy: mode change '+o Alex_Gaynor' by ChanServ!ChanServ@services.09:16
amaury_ (~amaury_@74.125.57.34) joined #pypy.09:17
fijalAlex_Gaynor: hi09:20
k_bx (~k_bx@195.20.130.1) left irc: Remote host closed the connection09:20
nanonymefijal, in 2012 it ate the world which promptly ended09:22
lesshastehi fijal 09:29
lesshasteI think I missed a message last night09:29
kenaan12fijal numpy-indexing-by-arrays-2 1135596eb33a6f 15/pypy/module/micronumpy/: add & and |09:29
fijallesshaste: which one?09:31
lesshastethe last thing I said was about your persuading me not to reply to Mark09:31
lesshasteanything after that09:31
lesshasteback in 20 mins09:35
lesshaste (~lesshaste@87-194-206-189.bethere.co.uk) left irc: Quit: Leaving09:35
lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) joined #pypy.09:38
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds09:39
arigato (~arigo@89-40.5-85.cust.bluewin.ch) joined #pypy.09:41
G2P (~G2P@fw-asn1.ornis.com) joined #pypy.09:43
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.09:52
kenaan12hager ppc-jit-backend 11909126039943 15/pypy/jit/backend/ppc/ppcgen/regalloc.py: check var.type in make_sure_var_in_reg09:53
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 252 seconds10:02
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy.10:05
JaRoel (~jaroel|4d@office.fourdigits.nl) joined #pypy.10:07
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds10:09
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 252 seconds10:10
kenaan12fijal numpy-indexing-by-arrays-2 1123d3ca4e7881 15/pypy/module/micronumpy/test/test_numarray.py: good, fix the test10:10
kenaan12arigo extradoc 11eb50533e45ea 15/sprintinfo/leysin-winter-2012/planning.txt: (all) planning for today10:11
kenaan12fijal numpy-indexing-by-arrays-2 11fed67b9508a6 15/pypy/module/micronumpy/interp_numarray.py: jit merge points10:16
mikefcugh. just found another case where code gets jitted into oblivion.  now trying to make a testcase10:17
JaRoel (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 252 seconds10:17
gorakhargosh (~user@14.96.13.208) joined #pypy.10:17
bbot24Failure: 15http://buildbot.pypy.org/builders/own-linux-x86-32/builds/1928 [12antocuni, jvm-improvements]10:18
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy.10:22
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.10:26
kenaan12fijal numpy-indexing-by-arrays-2 1192eaf5297003 15/pypy/module/micronumpy/: fix test_zjit. Skip the test that fails every time someone touches the world.10:27
JaRoel (~jaroel|4d@office.fourdigits.nl) joined #pypy.10:27
Justin__ (Justin@114-36-59-117.dynamic.hinet.net) left #pypy.10:29
kenaan12antocuni jvm-improvements 1188e6de55b13e 15/: close about-to-be-merged branch10:29
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 244 seconds10:29
kenaan12antocuni default 11c1b66ebfb441 15/: (benol) merge the jvm-improvements branch, which fixes (again :-)) the translation for the JVM backend10:29
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Read error: Connection timed out10:31
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.10:34
kvda (~kvda@124-169-134-221.dyn.iinet.net.au) joined #pypy.10:35
lizardo (~lizardo@189.2.128.130) joined #pypy.10:56
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Read error: Operation timed out10:56
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) joined #pypy.10:59
anish_ (anish@nat/redhat/x-mpffyrleibflqbsd) left irc: Ping timeout: 244 seconds11:00
nirbheek_ (~nirbheek@gentoo/developer/flyingspaghettimonster/nirbheek) left irc: Quit: Leaving11:05
witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy.11:06
witulski (stupsi@fwstups.cs.uni-duesseldorf.de) left #pypy.11:11
voidspace (~voidspace@python/psf/voidspace) left irc: Quit: Uhm... gotta go11:15
kenaan12fijal default 1189436d06f0ae 15/pypy/objspace/std/unicodeobject.py: a minor improvement11:16
kenaan12fijal numpy-indexing-by-arrays-2 11fe9bca1da2b6 15/pypy/module/micronumpy/: progress on arr[arr_of_bools] = arr11:16
kenaan12fijal default 1181aaf2060a57 15/pypy/objspace/std/unicodeobject.py: merge11:16
anish_ (anish@nat/redhat/x-joqpfjfoetvxawix) joined #pypy.11:16
fijalok, I'll fight with getting sphinx to run nightly11:17
AlekSi (~AlekSi@85.235.191.82) left irc: Remote host closed the connection11:22
ciphergothHow come when I run the sandbox, it invokes gcc?11:22
ciphergothand how come it wants to stat its own executable?11:23
timotimothe first part is the platform probing to see what a couple of features of your platform are (but i forgot which)11:24
timotimoprobably stuff like calling conventions? different sizes of datatypes maybe?11:25
ciphergothplatcheck_7.c11:25
kenaan12rguillebert py3k 1116202ef48d39 15/pypy/module/sys/interp_encoding.py: On linux, set ascii as the default locale if none can be found11:25
kenaan12rguillebert py3k 117948fef3ed5a 15/: Merge heads11:25
ciphergothand then module_0.c to make externmod.so11:26
mfoord (~anonymous@87-194-212-65.bethere.co.uk) joined #pypy.11:27
mfoord (~anonymous@87-194-212-65.bethere.co.uk) left irc: Changing host11:27
mfoord (~anonymous@python/psf/voidspace) joined #pypy.11:27
ciphergothhurm - so the overhead of executing a new sandbox is pretty high!11:28
timotimoi suspect with a bit of work this platform check could be constant folded away11:29
timotimoi just have no idea how it works :)11:29
ciphergothDoes PyPy support running more than one instance of the Python interpreter in a single process?11:33
timotimonot yet, but it's probably not very hard (what with the object spces already being neatly separated and all)11:33
lukegb (~lukegb@2a01:4f8:101:2403::2:1) left irc: Read error: Operation timed out11:34
ciphergothwould be very handy for lots of things11:34
lukegb (~lukegb@2a01:4f8:101:2403::2:1) joined #pypy.11:34
ciphergothJust to say this explicitly: PyPy is *incredibly* cool in so many ways.  Amazing work and congratulations to all involved.11:35
ciphergothI'm wondering about how one might provide something like Codecademy for writing Python, and a sandboxed Python interpreter is an essential component11:35
timotimoyes, it didn't fall short of its goal to encourage experimentation and what at first seemed like a silly/dumb idea at the beginning turned out much, much better than most people expected11:36
xorAxAx (~alexander@moinmoin/coreteam/alexander) left irc: Ping timeout: 248 seconds11:36
mcdonc_ (~mcdonc@ip72-209-213-54.dc.dc.cox.net) joined #pypy.11:37
timotimohave you looked at the python interpreter that's translated to javascript by llvm?11:37
timotimoi think it's got "emscripten" in its name somewhere11:37
ciphergothWhose idea was rpython?  That seems like the real breakthrough11:37
apollo13 (~apollo13@unaffiliated/apollo13) left irc: Ping timeout: 248 seconds11:37
mcdonc (~mcdonc@ip72-209-213-54.dc.dc.cox.net) left irc: Ping timeout: 248 seconds11:37
xorAxAx (~alexander@moinmoin/coreteam/alexander) joined #pypy.11:37
timotimodunno, i'm relatively new to the project11:37
mitsuhiko (~mitsuhiko@ubuntu/member/mitsuhiko) left irc: Ping timeout: 248 seconds11:37
JaredW (~JaredW@118.142.17.238) left irc: Remote host closed the connection11:38
JaredW (~JaredW@118.142.17.238) joined #pypy.11:38
lizardo_ (~lizardo@189.2.128.130) joined #pypy.11:38
timotimohttp://pypy.org/people.html � this lists people who worked on the rpython translation toolchain, but not who came up with the idea in the first place11:40
TheNoodle (~noodle@sarcasticsupport.com) left irc: Ping timeout: 248 seconds11:40
AlekSi (~AlekSi@85.235.191.82) joined #pypy.11:40
amaury_ (~amaury_@74.125.57.34) left irc: Ping timeout: 252 seconds11:40
apollo13 (~apollo13@unaffiliated/apollo13) joined #pypy.11:40
raydeo (~raydeo@merickel.org) left irc: Ping timeout: 248 seconds11:40
pybot (~pybot@stackless.com) left irc: Ping timeout: 248 seconds11:40
-ERROR from server- Closing Link: stackless.com (Ping timeout: 248 seconds)11:40
pybot joined #pypy.11:40
-ChanServ (ChanServ@services.)- [#pypy] Welcome!11:40
-ChanServ (ChanServ@services.)- [#stackless] Welcome to #stackless. This channel is for the discussion of matters relating to Stackless Python. Remember that others in this channel may not be in the same timezone as you, so stick around for an answer (even if only in your channel history) if you ask a question. The mailing list is also recommended as being a faster place to get help (http://www.stackless.com/mailman/listinfo/stackless).11:41
tos9 (~tos9@unaffiliated/tos9) left irc: Ping timeout: 248 seconds11:41
tos9 (~tos9@unaffiliated/tos9) joined #pypy.11:41
lizardo (~lizardo@189.2.128.130) left irc: Ping timeout: 248 seconds11:42
Nick change: lizardo_ -> lizardo11:44
fijaltimotimo: what do you mean?11:46
mitsuhiko (~mitsuhiko@ubuntu/member/mitsuhiko) joined #pypy.11:46
fijalciphergoth: if you get a nightly it should not do any of those any more11:46
fijalciphergoth: it was fixed and we're planning doing a new release soonish11:47
timotimowhat do i mean with what?11:47
fijal<timotimo> http://pypy.org/people.html � this lists people who worked on the rpython translation toolchain, but not who came up with the idea in the first place11:48
fijallast thing you said :)11:48
timotimoright. well, it says armin, antonio and alex worked on it11:49
squiddy (~squiddy@g229037172.adsl.alicedsl.de) left irc: Quit: Leaving11:49
fijaleveryone on the list worked on it11:49
timotimobut it does not say when it started and who came up with the idea to build rpython11:49
fijalis it important?11:50
fijalpypy is not a one-man project11:50
hpktimotimo: rpython was basically there from the very start around 2003 and continously worked on by many people11:51
timotimohm, that's true, such an information really wouldn't be meaningful11:51
jacob22The seed of making a language that would both be interpretable by CPython and translatable to C came from discussions between hpk, armin and Christian Tismer, but there were many people involved in hashing out what this would mean. Essentially everybody at the first sprint in Hildesheim in 2003.11:57
jacob22And then the concept evolved over a couple of years, with many more contributors.11:58
nedbat (~nedbat@python/psf/nedbat) joined #pypy.12:04
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds12:08
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.12:10
amaury_ (~amaury_@74.125.57.34) joined #pypy.12:10
apexo (~apexo@2a01:238:436b:8301:5054:ff:fe87:82fb) left irc: Quit: ZNC - http://znc.sourceforge.net12:12
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds12:17
aboudreault (~alanb@osgeo/member/aboudreault) joined #pypy.12:17
kenaan12fijal extradoc 11ebe7ed17d1d5 15/planning/micronumpy.txt: add a section12:21
kenaan12fijal numpy-indexing-by-arrays-2 116fe5770303bb 15/pypy/module/micronumpy/: fix setitem with bool index12:28
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.12:29
kenaan12fijal numpy-indexing-by-arrays-2 11b48aeb33bb41 15/pypy/module/micronumpy/interp_numarray.py: introduce a jit driver, although it's a bit pointless right now12:32
kenaan12fijal extradoc 11daaef6427a03 15/planning/micronumpy.txt: add two more items12:33
fijal_ (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.12:36
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds12:39
Nick change: fijal_ -> fijal12:45
icrazyhack (~horieyui@183.93.198.23) joined #pypy.12:47
amaury (amaury_@nat/google/x-svrdhglkcccoxwsf) joined #pypy.12:48
kenaan12fijal default 11865005356c2f 15/pypy/objspace/std/: (fijal, cfbolz) improve error reporting12:48
amaury_ (~amaury_@74.125.57.34) left irc: Ping timeout: 240 seconds12:50
anish_ (anish@nat/redhat/x-joqpfjfoetvxawix) left irc: Ping timeout: 260 seconds12:53
sven_hager (~sven@fwstups.cs.uni-duesseldorf.de) joined #pypy.12:54
nedbat (~nedbat@python/psf/nedbat) left irc: Ping timeout: 248 seconds13:00
gorakhargosh (~user@14.96.13.208) left irc: Quit: ERC Version 5.3 (IRC client for Emacs)13:00
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Remote host closed the connection13:00
aboudreault (~alanb@osgeo/member/aboudreault) left irc: Ping timeout: 240 seconds13:01
int3__ (~int3__@bb116-14-91-196.singnet.com.sg) joined #pypy.13:03
rguilleberthi amaury, it crashed because the locale didn't exist on the system so I put ascii by default but I'm not sure it's good enough13:04
berkerpeksag (~berker@212.156.129.218) joined #pypy.13:06
berdario (~Tiibiidii@host158-89-dynamic.0-87-r.retail.telecomitalia.it) joined #pypy.13:09
nedbat (~nedbat@python/psf/nedbat) joined #pypy.13:14
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) joined #pypy.13:18
jnoller (~jnoller@vpn.nasuni.com) joined #pypy.13:28
fijaljnoller: hi13:29
jnollerhowdy13:29
berkerpeksag (~berker@212.156.129.218) left irc: Read error: Connection reset by peer13:30
berkerpeksag (~berker@212.156.129.218) joined #pypy.13:31
aboudreault (~alanb@osgeo/member/aboudreault) joined #pypy.13:31
amauryrguillebert: it's good enough for now13:32
rguillebertok13:33
amauryCPython also return 'ascii' it seems13:33
rguillebertyes but I think it gets it from an external function13:34
amauryfor LANG=foo it also returns 'ascii'13:41
kenaan12arigo stm 1129e9345db0e6 15/pypy/translator/stm/: (antocuni, arigo) Test and fix: handle the case of stm_{get,set}field() called after descriptor_init() but outside a tr...13:52
kenaan12arigo stm 114ef670f3a925 15/pypy/translator/: (antocuni, arigo)  Refactored targetdemo.py to use the new interface.  Allow debug_start/debug_print/debug_stop to work...13:52
kenaan12arigo default 113d6e00235d87 15/pypy/config/pypyoption.py: Use default=False, and enable it only in -O2/O3/Ojit, like the other optimizations.  Fixes an issue if weakrefs are...13:52
kenaan12arigo default 11f34f0c11299f 15/: merge heads13:52
bfirsh (u1308@gateway/web/irccloud.com/x-lzfctubzlskvzarf) joined #pypy.14:00
justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) joined #pypy.14:01
fprimex (~fprimex@brent-macbook.sc.fsu.edu) joined #pypy.14:03
mfoord (~anonymous@python/psf/voidspace) left irc: Quit: mfoord14:10
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) left irc: Quit: DasIch14:11
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) joined #pypy.14:11
ciphergothfijal: thanks for your help in Twitter!14:22
ciphergothI'd figured out that I don't have to restart the controlling process14:22
ciphergothbut I *do* have to have one sandbox process per sandbox, and kill and restart it to get a fresh one14:23
ciphergothlooks like that could be changed but isn't on the roadmap14:23
fijalwell, but invocation of the sandbox process does not run gcc14:24
fijalor anything like that14:24
fijalit does stat the executable14:24
fijalciphergoth: nedbat is the one handling sandboxing issues14:24
khs (~khs@2001:700:300:2120:725a:b6ff:fee5:a44) joined #pypy.14:24
ciphergothfijal: oh! it's pypy_interact that's running gcc?14:25
fijalyes14:25
fijalnot even any more14:25
gtaylor (~gtaylor@108-196-160-156.lightspeed.tukrga.sbcglobal.net) joined #pypy.14:25
ciphergothah, phew14:25
fijaland tons of import and tons of crap14:25
ciphergothaha14:25
fijalyes, it's crap, but not very important crap :)14:25
fijalit's also better on trunk14:25
Action: ciphergoth does hg up default14:26
fijalciphergoth: you have to build sandbox yourself btw :/14:26
fijalthere is no nightly build14:26
ciphergothno problem14:26
fijalbesides, depending how secure you want it to be, you might want to compile it with -O2 and not with -Ojit14:26
ciphergothOK rerunning translate now...14:27
ciphergothand am using --opt=214:27
fijal--sandbox14:27
fijalright?14:27
icrazyhack (~horieyui@183.93.198.23) left irc: Ping timeout: 252 seconds14:27
ciphergoththough I plan on running this on a VM that won't do anything else, so I don't need sooper duper security14:28
ciphergoth python translate.py --sandbox --opt=2  targetpypystandalone.py14:28
fijalyeah, then you can do -Ojit14:28
fijalthat's relatively secure14:28
fijalbtu for example there are usually more ways to segfault the JIT than non-jit version14:28
ciphergothOK will do that next time14:28
fijalfor obvious reasons :)14:28
fijalwhat are you doing?14:29
ciphergothfijal: I want to write a codecademy equivalent for Python14:29
fijalcool14:29
nedbatciphergoth: that's funny, that's what I'm doing...14:29
ciphergothI wrote this: http://dev.lshift.net/paul/rurple/14:29
ciphergothnedbat: oooh!14:29
fijaluh14:30
ciphergothnedbat: I have not even started14:30
ciphergothjust investigating the tools14:30
fijalmaybe you guys should cooperate instead :)14:30
fijalbut who am I to tell you14:30
nedbatciphergoth: yeah, watch out, that's how I became the maintainer of the sandbox! :)14:30
ciphergoththink that's a good idea!14:30
ciphergothI don't even understand the sandbox14:30
fijalnedbat: admitedly it does not require much work :)14:30
nedbatciphergoth: me either! (not really)14:31
ciphergothit is quite a lot of work!14:31
nedbatfijal: no, but it could use more.14:31
ciphergoththe whole chain from web page to sandboxed process is pretty complex!14:31
ciphergothnedbat: I'm a bit worried about performance. How expensive is it to start a sandboxed process14:31
nedbatciphergoth: it's not cheap.14:32
Da_Blitz (~Da_Blitz@203.56.250.63) left irc: Ping timeout: 248 seconds14:32
nedbatciphergoth: lahwran and I have been kicking around some ideas for speeding it up.14:33
lahwranit still has plenty of room for speedup14:33
lahwranunfortunately I've been busier than I'd have liked, and I haven't had time to work on it for a few weeks14:33
lahwranciphergoth: at the moment it can start up and tear down a process in a half second on my p4ht 3ghz (not a good processor)14:34
Da_Blitz (~Da_Blitz@203.56.250.63) joined #pypy.14:34
ciphergothwould trying to have multiple sandboxes in one process make sense?14:34
lahwrannot really, no14:34
lahwranunless you mean multiple sandbox slaves in one pypy_interact, in which case, sure14:35
lahwranyou should be aware that the sandbox is a two-process system when you run pypy_interact14:35
lahwranI've been testing it with a four (!) process system, a "client" which sends a request to a "server", which then fires up the "interact" process, which finally starts the "pypy-c-sandbox" process14:36
nedbati've been using a django process which runs the interact part and starts the sandbox process14:36
lahwranhorribly unnecessary, but I was attempting to produce something where I could distribute the client to some friends and say "try me"14:36
lahwranreally, most of the "hard" work of detaching it from the pypy code is done, and it's just a matter of run-of-the-mill optimization14:38
fijalnedbat: but don't you run pypy_interact per sandbox process?14:38
fijalnedbat: the running of sandbox itself should be like 0.2s14:38
fijalI would think14:38
nedbatfijal: I don't run pypy_interact as its own process.  I use that code in my django process.14:39
fijalnedbat: so django process is alive and running and tearing down one sandboxed process takes how much?14:39
nedbatfijal: I don't have real measurements.  At the browser, running a chunk of code takes 1s or so?14:40
justinpeel (~justinpee@c-76-23-44-72.hsd1.ut.comcast.net) left irc: Read error: Operation timed out14:40
ciphergothnedbat: so what do you have working at the moment? Sounds like at the very least you have14:41
fijalnedbat: that's way too long14:41
ciphergotha way of launching sandboxes from inside your browser14:41
fijalciphergoth: in principle it should be possible to run that in-process14:41
ciphergothfrom inside your webs erver rather14:41
ciphergothfijal: *confused*14:41
fijalciphergoth: as in the two-process model is not something inherently built-in14:42
nedbatfijal: run *what* in process?14:42
fijalnedbat: sandbox inside pypy-interact14:42
fijalnedbat: but I think 1s comes from somewhere ele14:42
fijalelse14:42
ciphergothfijal: you're talking about multiple instances of the Python interpreter in a single process?14:42
nedbatfijal: yeah, i don't know where the 1s is from.14:42
ciphergothso the process contains both the controlling interpreter and the sandboxed interpreter?14:43
fijalciphergoth: yes14:43
ciphergothbut a single garbage colllector for both14:43
fijalciphergoth: probably more than one14:44
fijalciphergoth: you can't control the amount of memory very good with a single GC14:44
ciphergothmakes sense14:44
fijalnedbat: let me put it that way. if your website becomes popular enough the time of start is actually a problem, I'm willing to help you trim it :)14:44
ciphergothso what is shared besides the PID?  IIRC process startup itself is very cheap under Linux14:44
nedbatfijal: thanks!14:45
fijalas there are no pieces of design that prevents it from happening14:45
nedbatciphergoth: http://choosepython.com/gym/2/lists14:45
fijalit's just that we never cared14:45
ciphergothfijal: we'll hold you to that :-)14:45
fijal:]14:45
whitelynx (~whitelynx@63.241.75.144) joined #pypy.14:45
fijalciphergoth: communication is cheaper14:46
ciphergothnedbat: WOW!14:46
fijalso say instead of passing in strings you can just pass memory references14:46
fijalall and all it's trading security for performance in a way14:46
fijalbut you can do tricks14:46
fijallike setup python interpreter (this is taking some time) and then fork for each request14:46
ciphergothnedbat: amazing work! Is the source available anywhere?14:46
ciphergothnedbat: getting to where you're up to isn't easy!14:47
ciphergothnedbat: but I want to do something even harder14:47
ciphergothI want it to be like Rurple-NG, with an integrated debugger14:47
fijalciphergoth: ah, you know you can do crazy stuff in pypy with debugger?14:48
ciphergothso you can single-step through your program14:48
ciphergothfijal: no? I just used settrace14:48
fijalyou can fake tracebacks14:48
fijalfor example14:48
fijalusing transparent proxies14:48
nedbatciphergoth: thanks, you should also see this: http://people.csail.mit.edu/pgbovine/python/14:48
ciphergothfijal: don't udnerstand sorry!14:48
fijalciphergoth: in cpython you can't create new tracebacks14:49
ciphergothnedbat: is that also pypy-based?14:49
nedbatciphergoth: no, I don;t think so14:49
ciphergothfijal: where should I go to larn more abotu pypy and the debugger?14:50
fijalciphergoth: it's not debugger specific, but you can read about transparent proxies14:50
fijalhttp://doc.pypy.org/en/latest/objspace-proxies.html#transparent-proxies14:51
ciphergothhmm, wonder how that other link sandboxes the Python code?14:51
Da_Blitzwasent there a seccomp branch in the pypy repo?14:51
exnihilo (~nathan@descartes.its.unc.edu) left irc: Remote host closed the connection14:51
nedbatciphergoth: he has a link to his repo.14:51
ciphergothnetbat: is the source for http://choosepython.com/ online?14:51
ciphergothI see that - looking now14:52
nedbatciphergoth: it is: https://github.com/nedbat/choosy14:52
exnihilo (~nathan@descartes.its.unc.edu) joined #pypy.14:52
ciphergothO M G14:53
ciphergotham very pleased to find so much of the work I thought was ahead of me already done14:53
nedbatciphergoth: :) it makes me wonder who else is out there working on stuff like this.14:54
mat^2 (~mathias@212.130.113.35) joined #pypy.14:55
fijalciphergoth: in general (sandbox is a sad exception) pypy release is usually centuries old in terms of features/performance14:58
AlekSi_ (~AlekSi@85.235.191.82) joined #pypy.14:58
fijaleven though we try to release it every 2 months or so14:58
nedbatciphergoth: drop me an email, and we'll talk about what to do together.14:58
bivab (~david@89-40.5-85.cust.bluewin.ch) joined #pypy.14:59
ciphergothnedbat: I shall!15:00
fijalnedbat: btw, coverage is awesome15:00
nedbatfijal: :) thanks15:00
ciphergothwow, translate is absolutely bringing my machine to a standstill15:01
AlekSi (~AlekSi@85.235.191.82) left irc: Ping timeout: 240 seconds15:01
ciphergothnedbat: paul@ciphergoth.org15:01
ciphergoththanks15:01
mat^2 (~mathias@212.130.113.35) left irc: 15:02
kenaan12fijal release-1.8.x 112b32f9c1533d 15/pypy/module/sys/version.py: Create a release branch15:03
fijalnedbat: speaking of which15:04
ciphergothwhat's coverage?15:04
kenaan12fijal default 11a4261375b359 15/pypy/module/cpyext/include/patchlevel.h: mark patchlevel as well15:04
kenaan12fijal default 1196334e1f84f3 15/pypy/module/: update it here as well15:04
nedbatuh-oh, feature request in-coming!15:04
nedbatciphergoth: http://nedbatchelder.com/code/coverage/15:05
antocuni"I don't know what is the question, but pdb(++) is the answer"15:06
arigatoin the series "did you know it": UnicodeError is a subclass of ValueError15:08
mat^2 (~mathias@212.130.113.35) joined #pypy.15:08
fijalnice15:08
ciphergothwow - how is it possible to know which lines of a Python program are executable when they're not executed?15:08
fijalciphergoth: what?15:09
ciphergothfijal: what coverage does seems impossible to me15:09
fijalciphergoth: dunno, ask ned15:09
nedbatciphergoth: luckily the compiler leaves some information in the code objects.15:10
ciphergothnedbat: ah!15:10
nedbatciphergoth: coverage.py used to re-parse the code, etc, but it was complex and sometimes wrong.  Now it just uses the line number table in the code object.15:10
ciphergothso you scan all the code objects looking for lines that might be executable - coolness!15:10
nedbatciphergoth: but for branch coverage, it tries to understand the bytecode, and is complex and sometimes wrong. :(15:10
ciphergothit's very cool15:12
exnihiloExcuse my ignorance, if I wanted to suspend a frame (to be resumed later), what would be the best way to go about implementing that?  Also, would it be possible to make a deepcopy of said paused frame object?15:13
exnihiloI've looked at the code in eval and a few other places but I still don't have much context15:13
exarkuna deep copy of a frame object is likely to explode and kill everyone nearby15:14
exnihiloI recognize that it is a non trivial issue :D15:15
stakkars_ (~tismer@i59F7726A.versanet.de) left irc: Quit: schnarch15:15
Da_Blitzexnihilo: have you looked at stackless?15:16
voidspace (~anonymous@python/psf/voidspace) joined #pypy.15:16
exnihilono, pypy was my first stop.  My project is symbolic objects that short circuit their execution context to return generative symbolic expressions15:16
fijalexnihilo: pypy and stackless can pickle frames15:17
fijalpickled frame is suspended15:17
fijalalthough I would suggest suspending using stackless features (like greenlets)15:17
fijalor generators15:17
exnihiloI wanted to copy the generator interface as much as possible.  the issue is that I don't want symbolic expressions to be one time use things, or I'm just making thunk over15:18
fijalexnihilo: you're much better with other languages though15:18
fijalI would not suggest deepcopying the frame15:18
exnihilowell. the way I was looking at it was to escape out of the inner loop with an exception and a copy of the frame, roughly15:19
exnihilobut I'm open to suggestions15:19
timotimonedbat: did i just start a process that does nothing but loop on your server? :\15:19
Action: nedbat knew it was a bad idea to post that url...15:20
exnihiloI just need to be able to hope out of something like str(Symbol()), rather than have the frame complete 15:20
nedbattimotimo: probably, were you trying to break it?  What did you do?15:20
kenaan12rguillebert py3k 11f3ca10db7f5e 15/pypy/interpreter/pyparser/test/test_pyparse.py: (antocuni,romain) Adapt the test to py3k15:22
kenaan12rguillebert py3k 115dd563b2360f 15/pypy/interpreter/: (antocuni, romain) started to parse the new tuple unpacking (only the case with no parenthesis is implemented so...15:22
CIA-1703cfbolz 07roundup * 10#1001/Improve RPython documentation: 15:22
CIA-17[new] We should try to make the RPython documentation better. Here's a list of some15:22
CIA-17things that Laurie had a hard time figuring out:15:22
CIA-17- "The di ... * 14https://bugs.pypy.org/issue100115:22
CIA-1703cfbolz 07roundup * 10#1002/Improve JIT documentation: 15:22
CIA-17[new] Applying the JIT to an interpreter should have some better documentation. All15:22
CIA-17the information is basically there (in blog posts, papers, do ... * 14https://bugs.pypy.org/issue100215:22
nedbattimotimo: no answer?15:24
timotimonedbat: while True: pass15:27
timotimojust to see if i get a time-out or something15:28
witulski (~stupsi@fwstups.cs.uni-duesseldorf.de) joined #pypy.15:28
nedbattimotimo: i might not have timeouts in place yet... :(15:28
timotimothat tutor thing with the visualisation is HOT15:29
nedbattimotimo: yes, it's very cool.15:30
timotimoi'm just looking at a naive implementation of quicksort in it15:31
timotimoit's really really sweet15:31
timotimodo you know who made it and how?15:31
nedbattimotimo: there's a link to the repo on it somewhere15:31
ciphergothyou can go back as well as forward!15:33
nedbatciphergoth: single-steps the program on the server, then sends the entire run history to the browser as json.15:35
ciphergothah!15:35
nedbatciphergoth: *he single-steps.15:35
bgola (~bgola@c951826e.virtua.com.br) left irc: Ping timeout: 252 seconds15:35
ciphergothdownsides: no interaction, and you can't do anything with a program that doesn't halt.15:37
ciphergothUpsides: you can go backwards!15:37
nedbatciphergoth: downside: he doesn't allow any imports.15:37
ciphergothhmm think that's OK for this job15:37
nedbatciphergoth: ah, that's where you and I may differ: I want to be able to do exercises like, "see what textwrap.dedent does"15:38
ciphergothOK15:38
ciphergothI want to allow users to drive a robot round a maze15:40
ciphergothunfortunately there's no easy way to add to the list of functions which are deferred to the sandbox15:41
fijalciphergoth: you have to embed it in some sort of API15:41
fijalno easy way15:42
ciphergothhow hard is that?15:42
fijalyou can for example serialize commands and send it over a socket15:42
Da_Blitzthat reminds me of a BP website where you could drive model dump trucks around a model town via flash15:42
fijalso say a file descriptor 1001 would always mean something15:42
ciphergoththat gets pretty complex compared to directly deferring15:43
nedbatciphergoth: or even a known filepath:  /dev/robot is the socket to write to.15:43
ciphergothyes15:43
nedbatciphergoth: write an api on the sandbox side that simply pickles and writes to the "socket".15:44
AlekSi_ (~AlekSi@85.235.191.82) left irc: Remote host closed the connection15:45
ciphergothas I say that gets moderately complicated compared to simulating an API call15:46
ciphergothbut may be the only way forward15:46
fijalciphergoth: well15:46
fijalciphergoth: there is no really secure way to calling python API15:46
fijalciphergoth: feel free to come with a proposal15:46
fijalyou can even use something already existing like xmlrpc or amp or something15:47
fijalor execnet15:47
ciphergothsure and that's simple from the interpreter point of view15:48
nedbatciphergoth: remember, this isn't a real socket.  One .py file (running in the sandbox) will say,   f.write(pickle.dumps(obj))    and in the interact code will be:  def write(d):  obj = pickle.loads(d)15:48
ciphergothbut the controlling process has a much more interesting time of it15:48
ciphergothyou can guarantee that it'll always write the whole thing in one syscall?15:50
ciphergothis unpickling safe?15:50
raydeo (~raydeo@merickel.org) joined #pypy.15:51
lucianciphergoth: afaik, not really. you could construct an object that executes random things when unpickled15:51
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Read error: Connection reset by peer15:52
Action: Da_Blitz dosent know if its helpful but i do have a lib to isolate a python process fromt he rest of the running system15:52
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.15:53
squiddy (~squiddy@g229037172.adsl.alicedsl.de) joined #pypy.15:53
mwhudson (~mwh@linaro/mwhudson) left irc: Ping timeout: 240 seconds15:54
nedbatciphergoth: marshal then.15:55
nedbatciphergoth: or limit the data that can go in.15:55
nedbatciphergoth: use json if you really want to be sure.15:55
nedbatDa_Blitz: how does your lib work?15:56
ciphergothmarshal is fine15:56
ciphergothmarshal is what the sandbox itself usees15:56
timotimoevery now and then interest flames up in my head to restart the programming game engine that would use python for its entities' scripts15:56
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Remote host closed the connection15:56
ciphergoththough the manual page explicitly disclaims such use15:56
ciphergothWarning The marshal module is not intended to be secure against erroneous or maliciously constructed data. Never unmarshal data received from an untrusted or unauthenticated source.15:57
Da_Blitznedbat: linux namespaces, bind mounts and some other LXC inspired stuff15:57
nedbatciphergoth: in this case, it's your api code creating the marshal, it isn't untrusted.15:58
nedbatciphergoth: that warning is about arbitrarily-constructed bytestreams being unmarshalled.15:58
ciphergothwhat's to stop the code in the sandbox opening /dev/robot and sending the stuff itself?15:58
ciphergothor using reflection to alter what my code sends?15:58
nedbatciphergoth: ah, that is a good point.15:58
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.15:58
nedbatciphergoth: json it is.15:58
ciphergothaargh not json15:59
nedbatciphergoth: why not?15:59
ciphergothwhat about AMP or PB?15:59
ciphergothslooooooow15:59
timotimoare you seriously worried about performance at this level?15:59
nedbatciphergoth: what bandwidth are you expecting across this boundary?15:59
felipejcruz (~felipecru@mvx-187-16-96-106.mundivox.com) joined #pypy.15:59
fijalthe cape town heat wave is killing my computer :/16:00
ciphergothI suppose it's acceptable really, just a reflex16:00
nedbatciphergoth: i'll know to stand back the next time I tap your knee...16:03
witulski (stupsi@fwstups.cs.uni-duesseldorf.de) left #pypy.16:03
dmalcolm (~david@c-24-61-12-82.hsd1.ma.comcast.net) joined #pypy.16:05
durin42 (~durin@173-15-14-174-Illinois.hfc.comcastbusiness.net) joined #pypy.16:05
durin42 (~durin@173-15-14-174-Illinois.hfc.comcastbusiness.net) left irc: Changing host16:05
durin42 (~durin@adium/durin42) joined #pypy.16:05
felipejcruz (~felipecru@mvx-187-16-96-106.mundivox.com) left irc: Ping timeout: 276 seconds16:06
mwhudson (~mwh@linaro/mwhudson) joined #pypy.16:07
tumbleweedfijal: I was also hearing horrible noises out of a hard drive, last night :/16:10
fijaltumbleweed: are you coming for Tue dinner with jeremy btw?16:10
tumbleweedyup16:10
fijaldo you know where/what time?16:10
tumbleweedas usual, we meet at UCT at 18:30. So I'll know then16:11
fijaloh ok16:11
fijalsms me or make jeremy do that16:11
tumbleweeddoes jerith have your number? I don't16:11
asmeurer__ (~asmeurer@129.138.218.98) joined #pypy.16:12
fijalhe does16:12
fijalarigato: do you feel like replying to the guy on ML abotu memory problems?16:15
fijalor should I try to do it?16:15
exnihilofijal: so, from reading the docs, if I wrap code to be evuated in continulets, I will be able to achieve the short circuit behavior I want by playing games with f_back?16:16
fijalexnihilo: I don't know upfront16:16
fijalexnihilo: I hope not16:16
fijalarigato: I'll reply16:16
exnihiloheh :| I think I need to just get something working with a test case so someone who knows interpreters better can tell me how to do things in a less insane way16:17
kenaan12antocuni py3k 11dc00842db627 15/pypy/interpreter/pyparser/test/test_pyparse.py: (antocuni, romain): most of the time the parser does not need a space, thus we instantiate it only for the tests (o...16:20
kenaan12fijal benchmarks 1170c28be43161 15/: add cpython documentation generation by sphinx16:21
Da_Blitzexnihilo: you really dont want to play with continulets, greenlets or stackless is a better idea as they exisit for cpython as well16:21
kenaan12fijal benchmarks 1177aa6fd8e4ec 15/: merge16:21
fijalpushed16:22
exnihiloda_blitz: I appreciate your input.  I am pretty bad at describing exactly what I want, I think some code will made what I mean clearer :)16:24
fijalexnihilo: I think our point is "if you want to play with frames for your production code, it's probably a very bad idea"16:24
Da_Blitznp, feel free to send me a link when you have somthing, i have played with stackles and greenlets before16:25
exnihilofijal: this isn't for production code, this is some proof of concept stuff with symbolic programming in python16:26
exnihiloI know frames are generally bad mojo for production code :)16:26
fijalok16:27
kenaan12antocuni py3k 1146e3b0574cc8 15/pypy/interpreter/pyparser/test/test_pyparse.py: bah, if we inherit we run all the test twice16:28
mitsuhiko (~mitsuhiko@ubuntu/member/mitsuhiko) left irc: Excess Flood16:29
kenaan12rguillebert py3k 11e110ee00c88b 15/pypy/interpreter/pyparser/data/Grammar3.2: (antocuni, romain) port the py3k grammar for the new tuple unpacking (with parenthesis)16:30
bgola (~bgola@c951826e.virtua.com.br) joined #pypy.16:32
kenaan12fijal benchmarks 11c01ab59c2a28 15/: fix the benchmark16:37
fijalok, good16:39
fijalantocuni: we have a benchmark we're slower on :)16:39
timotimoawesome!16:45
berkerpeksag (~berker@212.156.129.218) left irc: Read error: Connection reset by peer16:46
berkerpeksag (~berker@212.156.129.218) joined #pypy.16:47
berkerpeksag (~berker@212.156.129.218) left irc: Quit: Instantbird 1.116:56
bivab_ (~david@89-40.5-85.cust.bluewin.ch) joined #pypy.16:59
Alex_Gaynorfijal: do you have a link to numpypy readiness?17:01
bivab (~david@89-40.5-85.cust.bluewin.ch) left irc: Ping timeout: 240 seconds17:03
Nick change: bivab_ -> bivab17:03
khs (~khs@2001:700:300:2120:725a:b6ff:fee5:a44) left irc: Quit: Leaving17:04
oxnard (~oxnard@ws29-ilnk.elkhart.net) joined #pypy.17:10
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) joined #pypy.17:16
JaRoel (~jaroel|4d@office.fourdigits.nl) left irc: Ping timeout: 240 seconds17:17
mitsuhiko (~mitsuhiko@ubuntu/member/mitsuhiko) joined #pypy.17:24
amaury (amaury_@nat/google/x-svrdhglkcccoxwsf) left irc: Ping timeout: 260 seconds17:26
mitsuhiko (~mitsuhiko@ubuntu/member/mitsuhiko) left irc: Excess Flood17:27
sven_hager (~sven@fwstups.cs.uni-duesseldorf.de) left irc: Quit: Leaving17:30
mitsuhiko (~mitsuhiko@ubuntu/member/mitsuhiko) joined #pypy.17:30
lucian (~lucian@93-97-174-114.zone5.bethere.co.uk) left irc: Ping timeout: 240 seconds17:32
G2P (~G2P@fw-asn1.ornis.com) left irc: Quit: Leaving.17:44
timotimoin a descr_foo method, how can i set a default value? foo=space.wrap("Bar")?17:44
timotimow_foo, that is17:44
timotimono, that makes no sense17:44
Alex_Gaynortimotimo: w_foo="Bar", it'll automatically be wrapped17:45
timotimono, that makes no sense17:47
timotimosorry, wrong console :)17:47
timotimoif i want to check for the content ,will i call space.unwrap on it?17:48
mattip (4fb1711d@gateway/web/freenode/ip.79.177.113.29) joined #pypy.17:51
kenaan12rguillebert py3k 11fbf25d09b601 15/pypy/interpreter/: (antocuni,romain) Add support for the new unpacking at the ast level fixed list comprehension17:51
Alex_Gaynortimotimo: No, you can't use space.unwrap in RPython code, to unwrap a str you use space.str_w17:53
timotimoturns out my code was totally okay, but i incorrectly read the fail/traceback of my test case17:55
cocoatomo (~cocoatomo@p2118-ipbf901souka.saitama.ocn.ne.jp) joined #pypy.17:55
timotimonow i've "implemented" ndarray.ravel and i'll implement numpypy.ravel, too17:56
cocoatomo (~cocoatomo@p2118-ipbf901souka.saitama.ocn.ne.jp) left irc: Remote host closed the connection17:56
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) left irc: Quit: DasIch18:01
timotimowhat pastebin is cool nowadays?18:02
timotimohttps://gist.github.com/1627867 � review plox18:03
Alex_Gaynortimotimo: paste.pocoo18:04
Alex_Gaynoryou can't use space.unwrap in RPython code, you should use space.str_w18:04
Alex_Gaynorbut really you should use @unwrap_spec(order=str)18:04
Alex_Gaynoryour tests are missing asserts for some of the line18:05
mher (~Adium@tethras5.ucd.ie) left irc: Quit: Leaving.18:06
Alex_Gaynortimotimo: numpypy.ravel should really already be in the fromnumeric file, shouldn't it?18:06
timotimooh?18:07
timotimowell, it wasn't in the nightly build18:08
Alex_Gaynorwell, if it's not in there already,that's certainly where it belonds18:08
Alex_Gaynorperhaps mikefc or fijal could explain that file18:08
timotimowhere do i even find it?18:09
Alex_Gaynorlib_pypy/numpypy18:09
timotimooh, duh, there it is18:09
Alex_Gaynorit would be good if this file had a docstring explaining how to update it18:10
timotimoright, i didn't look for numpypy.ravel, so i didn't see the not implemented thingie18:10
arigato (~arigo@89-40.5-85.cust.bluewin.ch) left irc: Quit: See you18:10
timotimoah, i suppose it's generated then. yes, a bit of documentation would be great18:10
bivab (~david@89-40.5-85.cust.bluewin.ch) left irc: Quit: bivab18:10
timotimoAlex_Gaynor: with "some of the line", were you referring to the case where a is not an array yet?18:12
timotimowhere do i put tests for the ravel function from the "fromnumeric" file? it seems like i can't just import ravel from "_numpypy" in my test18:15
fijalhello18:18
timotimohello fijal :)18:18
fijalAlex_Gaynor: like baroquesoftware.com:500018:18
fijal?18:19
timotimoif you explain fromnumeric to me, i could write that bit of docstring18:19
Alex_Gaynorfijal: Firefox can't establish a connection to the server at baroquesoftware.com:5000.18:19
jaffachief (~textual@host86-170-15-228.range86-170.btcentralplus.com) joined #pypy.18:19
fijaloh yes my machine crashed18:19
AlekSi (~AlekSi@176.57.74.133) joined #pypy.18:19
fijalhow about now?18:20
Alex_Gaynorfijal: yes, cool18:20
Alex_Gaynorfijal: Can I make a pony request?18:20
timotimooooh, that's colorful18:21
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) joined #pypy.18:21
fijalAlex_Gaynor: go ahead18:23
kenaan12alex_gaynor default 118dea12b8117a 15/: move some stuff that's in fromnumeric out of the appnumpy file.18:23
fijalAlex_Gaynor: I want to work on this but I have tons of shit on my list18:23
fijalfeel like taking over something?18:23
fijaltimotimo: do you need some explaining?18:23
timotimoalex claimed i do18:24
Alex_Gaynorfijal: yeah, I want to make it so that it shows types and if you click one it inspects the methods, and for functions it shows which parameters we have18:24
Alex_Gaynorfijal: I can do it though, I need lunch and tracebin hacking first though18:25
kenaan12alex_gaynor default 11593bcb6b2e6f 15/lib_pypy/numpypy/test/test_fromnumeric.py: oops, fix tests18:25
fijalAlex_Gaynor: yeah, that sounds cool18:26
fijal__doc__ maybe?18:26
fprimex (~fprimex@brent-macbook.sc.fsu.edu) left irc: Quit: meeting18:26
Alex_Gaynorfijal: you mean check that we have a matching __doc__?18:26
fijalwe don't18:27
fijalbut it might link to doc18:27
fijalI'll add array.stuff18:27
fijaland the legend18:27
fijalright now18:27
Alex_Gaynordo it for all types :)18:27
fijaloh ok18:27
fijalcool18:27
Alex_Gaynorit's not like it's harder :)18:27
fijalbesides, I want to generate this as static html nightly18:27
fijalhow about that?18:27
Alex_Gaynorgetting food now18:27
fijalok, you deal with that18:27
Alex_Gaynorwith what, food?  ok :)18:27
kenaan12alex_gaynor default 11f602adef8b38 15/lib_pypy/numpypy/test/test_fromnumeric.py: uncomment a test18:28
timotimois there a way to run the doctests? or are we on a strict anti-doctest policy?18:29
fijaltimotimo: unless they come with some other software, yes that's the policy18:29
papercrane (~papercran@75.101.111.82) joined #pypy.18:31
voidspace (~anonymous@python/psf/voidspace) left irc: Quit: voidspace18:32
dzen~/win 618:33
Alex_Gaynorfijal: ok, black with food :)18:36
nopf (~froschif@147.172.224.4) joined #pypy.18:36
fijalgood18:37
AlekSi (~AlekSi@176.57.74.133) left irc: Remote host closed the connection18:37
timotimowhy do you raise NotImplemented?18:38
timotimoinsetad of NotImplementedError?18:38
fijalfor what?18:38
timotimoi believe NotImplemented is for comparison oparators only18:38
fijaldo we?18:38
fijalwe should noty18:38
fijalit's a bug18:38
Alex_Gaynorit is, raising NotImplemented won't work18:38
timotimoyup :)18:38
Alex_Gaynorfijal: btw, David's talk on annotation/flow space was the best explanation I've ever seen for normal people18:39
fijalcool18:39
fijalkernel change, reboot18:40
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Read error: Connection reset by peer18:40
Alex_Gaynoro_O18:40
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.18:43
kenaan12alex_gaynor default 1133bd443587ba 15/pypy/module/_codecs/interp_codecs.py: split _codecs.lookup_codec into parts with and without a loop, so the JIT can inline most of lookup_codec, ev...18:51
Alex_Gaynorfijal: ping18:52
fijalpong18:52
Alex_Gaynorfijal: feel like doing the thing so that debug_merge_point for the python jitdrivers get turned into a subclass with code and opcode idx attrs?18:53
fijalyeah18:53
Alex_Gaynorfijal: sweet, thanks18:53
Alex_GaynorI'm doing the trace splitting for tracebin ATM18:53
Alex_Gaynorfijal: and putting random optimizations into _codecs :)18:54
fijalI have a pending change for parsing18:54
fijalare you using jitlogparser module?18:54
Alex_Gaynorno, all in memory18:54
fijalstill18:54
timotimowhoops, i've been working on my strbuf_by_default branch18:54
fijalpart of it operates on loops18:54
fijalno?18:54
fijallike there is a part loading and the other part18:54
fijalI'm fine with doing the other part again18:55
Alex_Gaynor?18:55
fijalAlex_Gaynor: you load it and parse it but then you process it quite a bit18:55
fijalbefore showing it in jitviewer18:55
Alex_Gaynorno, I never parse18:55
fijalyes18:55
fijalthat's what you do for jitviewer18:55
Alex_GaynorI'm talking about for tracebin18:55
fijalyes, I know18:55
Alex_Gaynortracebin never parses18:55
fijalI KNOW18:55
Alex_Gaynorok :)18:55
Alex_Gaynorsorry, I think I'm confused18:56
fijalI KNOW I KNOW I KNOW18:56
fijalok18:56
fijalnow the first part parser, the second part reconnects them makes inlines etc.18:56
fijalthere is no point in redoing all the work18:56
fijaleither reuse it or replace it if you feel like rewriting18:56
fijalit's in jitlogparser, look at tests maybe18:57
Alex_Gaynork, thanks18:57
Alex_Gaynorfirst I need the DebugMergePoint metadata though :)18:57
fijalI would not mind someone rewriting it again18:57
fijalit's just that would be cool if he learns ;-)18:57
Alex_GaynorI may rewrite it, we'll see18:58
fijalAlex_Gaynor: anyway, I don't want to end up with two implementations of this18:59
Alex_Gaynorfijal: sure19:00
fijalboth half-working at least19:01
Alex_Gaynorof course :)19:02
timotimonow i don't really know how to continue with my ravel implementation ...?19:03
timotimodo i just commit the code for ndarray.ravel and wait for others to do the numpypy.ravel thing?19:03
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) left irc: Ping timeout: 240 seconds19:05
fijal (~fijal@196-210-150-33.dynamic.isadsl.co.za) joined #pypy.19:12
fijal<fijal> I broke my packaging system :/19:12
fijal<fijal> Alex_Gaynor: well, most of all figuring out all the possible combinations19:12
fijal<fijal> like:19:12
fijal<fijal> no debug_merge_points in chunks19:12
fijal<fijal> a no info loop that has inline something with info etc. etc.19:12
fijal<fijal> it would be cool if you can come up with design that handles it19:12
fijal<fijal> but right nwo a lot of those cases had been hacked in19:12
fijal<fijal> Alex_Gaynor: generally the parsed and unparsed list should behave roughly similary19:12
fijal<fijal> as far as the interface is concerned19:12
fijalAlex_Gaynor: ^^^19:13
Alex_Gaynorfijal: I'm not sure I follow19:13
Alex_Gaynorfijal: the first step is just exposing the metadata on the ResOp instance for the hook19:14
fijalthe thing is you have enough right now to be able to just run jitlogparser19:14
fijalno?19:14
Alex_Gaynoronly if I turned the resops back into a str19:14
Alex_Gaynor(+/- having input args and such)19:14
Alex_GaynorI don't want to be serializing and deserializing like that, if feels backwards19:15
fijalno19:15
fijalyou don't parse it19:15
fijalyou just use split_trace or so19:15
Alex_Gaynorfrom where?19:16
fijalyeah, it's a bit tangled now19:16
Alex_GaynorI mean, yes in theory this is fine, but I need the metadata on the objs from the hook19:17
fijalAlex_Gaynor: you should be able to call Function.from_operations19:17
amaury_ (~amaury_@46-127-23-192.dynamic.hispeed.ch) joined #pypy.19:17
fijalAlex_Gaynor: what I'm saying is the code is there19:18
fijalyou just have to untangle it from other parts19:18
Alex_Gaynorfijal: yes, but I don't have the metadata to use the code19:18
fijalwhich part?19:20
Alex_Gaynorfijal: I need the code object and the instruction index for each debug_merge_point resop that is a python one19:20
amaury_rguillebert: Hi19:20
fijalAlex_Gaynor: but not to run this code :)19:20
amaury_do you know that ast.py is generated?19:20
Alex_Gaynorfijal: because this parses it from the string?19:21
fijalAlex_Gaynor: yes19:21
Alex_Gaynorfijal: I don't want to do that though, it breaks :)19:21
fijalAlex_Gaynor: this *is* backwards but let's start there19:21
Alex_Gaynorfijal: which function parses the debug_merge_point arg?19:22
fijalTraceForOpcode19:22
fijalparse_code_data19:22
Alex_Gaynorok, cool19:23
jaffachief (~textual@host86-170-15-228.range86-170.btcentralplus.com) left irc: Quit: Computer has gone to sleep.19:25
kenaan12amauryfa py3k 11d682184291a6 15/lib_pypy/: Remove more L suffixes and long() calls19:27
kenaan12amauryfa py3k 1103901e509c25 15/pypy/module/termios/interp_termios.py: Another fix in termios19:27
kenaan12amauryfa py3k 112c6715a45957 15/lib_pypy/_pypy_interact.py: Add a reminder about the broken pyrepl.19:27
kenaan12amauryfa py3k 115b195e604794 15/pypy/interpreter/astcompiler/: ast.py is generated. Add the Starred in Python.asdl, and regenerate with "python interpreter/astcompiler/tools/asdl...19:27
Alex_Gaynorfijal: this way sucks :/ you have to re-read and re-exec the file to get the code obj19:32
stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy.19:36
AlekSi (~AlekSi@176.57.74.133) joined #pypy.19:36
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer19:36
Nick change: stakkars_ -> stakkars19:36
mattip (4fb1711d@gateway/web/freenode/ip.79.177.113.29) left irc: Quit: Page closed19:37
fijalAlex_Gaynor: ya, this sucks19:39
Alex_Gaynorfijal: let's do the metadata first, it makes so much more sense19:39
fijalAlex_Gaynor: can you rework this code to be nicer and I'll add metadata?19:39
Alex_Gaynorfijal: I'm going to rewrite it TBH, then I'll port it back to jitlogparser19:40
fijaldon't underestimate it btw19:41
fijalwriting it from scratch is not a very good idea19:41
Alex_Gaynorwell, I'm reading from this code19:41
Alex_Gaynorso it's not like I"m ignorant of it :)19:41
fijalyou'll soon be :)19:41
Alex_Gaynor:D19:41
fijalAlex_Gaynor: I'm actually serious, work from this code, don't start it from scratch19:43
fijalbecause it seems you need like 5 lines no?19:43
Alex_Gaynorfijal: yes :)19:44
fijal5 lines of diff from what it's there19:44
fijalif you want to like connect bridges and shit, you need more than 5 lines19:44
Alex_Gaynorwell, I'm doing this TDD, one bit at a time, so I don't have bridge yet :019:44
Alex_Gaynorbut yes, I agree with you19:45
fijalI just *really* don't want in the situation where jitviewer works for some loops and tracebin for some others :/19:46
Alex_Gaynorfijal: anyway, I now have: Exception AttributeError: "'ResOperation' object has no attribute 'py_code'" in jit hook <bound method Recorder.on_compile of <tracebin.recorder.Recorder object at 0x00007f72e75558a0>> ignored19:46
Alex_Gaynorwant me to work on the numpy status thing now?19:46
kenaan12fijal default 1180570f9631d7 15/pypy/tool/jitlogparser/: a test and a fix19:46
kenaan12fijal default 1131b1dbaa7024 15/: merge19:46
fijalyes please19:46
Alex_Gaynork19:46
fijalI'll do it in a second19:46
fijalonce I get my system in a stable state19:46
Alex_Gaynorawesome19:46
JaRoel|4d (~jaroel|4d@office.fourdigits.nl) left irc: Remote host closed the connection19:50
timotimohttps://gist.github.com/1627867 � may i commit this to default?19:51
AlekSi (~AlekSi@176.57.74.133) left irc: Remote host closed the connection19:51
Alex_Gaynorno, this doesn't translate for one19:54
fijaltimotimo: does it pass tests?19:55
timotimothe one i wrote19:55
fijaltimotimo: run all19:55
fijalspecifically it'll fail on test_ztranslation19:55
timotimoi should re-run all tests just to be super-paranoid19:55
timotimovery good point19:55
fijalwhy do you want ravel that does nto support anything?19:56
fijalAlex_Gaynor: how do I specify the parent typedef?19:57
Alex_Gaynorfijal: TypeDef("name", (parent,))19:58
Alex_GaynorI think parent can also be a single item and not a tuple19:58
timotimobecause the rest of numpypy doesn't support order arguments yet19:58
Alex_Gaynorfijal: I think I'm writing too much crazy pypy code, I'm forgetting how to write a simple script19:58
fijaltimotimo: fix it first :)19:58
timotimoohgod.19:58
timotimothis way i'll never be able to contribute anything worthwhile :\20:00
rguilleberthi amaury_ 20:02
fenrrir (~fenrrir@187.40.155.188) joined #pypy.20:02
rguilleberterf !20:02
fijaltimotimo: hey, you can add more ufuncs20:02
amaury_rguillebert: ?20:02
fijaltimotimo: add .where20:02
timotimookay, i'll have to read what those do first, though20:02
rguillebertI knew I had to modify Python.asdl eventually20:03
rguillebertbut I didn't know when20:03
rguillebertand basically when antonio and I worked on it, we didn't think about it at all20:04
timotimofijal: do you mean the "where" optional keyword argument?20:04
rguilleberthe told me he had never worked on the parser either20:04
bivab (~david@89-40.5-85.cust.bluewin.ch) joined #pypy.20:05
JaRoel|4d (~jaroel|4d@2001:980:7133:1:5841:e335:a7dc:a982) joined #pypy.20:10
fijaltimotimo: numpy.where20:10
JaRoel|4d (~jaroel|4d@2001:980:7133:1:5841:e335:a7dc:a982) left irc: Remote host closed the connection20:10
JaRoel|4d (~jaroel|4d@sink.jaroel.nl) joined #pypy.20:10
fijalAlex_Gaynor: for example dmp will know only bytecode no and pycode20:12
Alex_Gaynorfijal: why not?20:12
timotimofijal: should i implement that at applevel? also it seems like i should implement numpy.nonzero, too?20:13
ericflo (~ericflo@75.103.8.110) joined #pypy.20:16
fijalAlex_Gaynor: I mean it wont have the precise lineno20:18
fijaltimotimo: copy maybe?20:18
Alex_Gaynorfijal: oh, lineno is easy you can get it by running disassembler20:18
fijalAlex_Gaynor: "easy"20:18
fijalok :)20:18
Alex_Gaynorfijal: if you have the code obj it's easy20:18
fijalok20:21
Alex_Gaynorfijal: is your server down again?20:21
mikefcAlex_Gaynor: you may to close my https://bugs.pypy.org/issue998 if you've fixed fromnumeric in your own way.20:21
mikefcmy patch is probably now obsolete20:21
fijalAlex_Gaynor: might be20:22
fijalwant me to put it up again?20:23
Alex_Gaynorfijal: eh, whatever I found it in a different browser tab :)20:24
Alex_Gaynorso long as I don't reload I'm ok20:24
fijaldone20:24
fijalanyway20:24
Alex_Gaynorok, adding method support now20:25
Alex_Gaynorwell, let me upload this so you can see it now20:25
mikefctimotimo: i can talk about what i did to from numeric if you like.20:34
DasIch_ (~DasIch@p4FFDF136.dip.t-dialin.net) joined #pypy.20:34
DasIch (~DasIch@p4FFDED18.dip.t-dialin.net) left irc: Ping timeout: 245 seconds20:35
Nick change: DasIch_ -> DasIch20:35
mikefcnot that it was very interesting.  20:36
mikefcthere seem to be about 8 places in numpy where functions are described, but all they do is call the ndarray method.20:36
mikefc8==exaggeration20:36
mikefcbut it's definitely not 120:36
rekamso (~textual@216.190.27.14) joined #pypy.20:40
jnoller (~jnoller@vpn.nasuni.com) left irc: Quit: Leaving...20:45
fijalmikefc: I appreciate what you did :)\20:45
fijalmikefc: would be cool if you do more of it20:45
fijalAlex_Gaynor: ok, feel like reviewing/translating/merging?20:46
kenaan12fijal better-jit-hooks-2 110513eb237097 15/pypy/: dmp creation20:46
Alex_Gaynorfijal: gimme 2 minutes, finishing up the next part of numpypystatus20:46
kenaan12fijal better-jit-hooks-2 111ef43f689d9b 15/pypy/module/pypyjit/: implement jithooks20:46
fijalAlex_Gaynor: I'll run translation in the screen on tannit20:48
fijalannotating those hooks is a little nightmare btw :/20:49
mikefcfijal: i realise that anything i "import" from numpy now is going to be redone later down the road, but I feel it's necessary to do so in order to bootstrap some parts of numpy.  I wish I had more time to hack :(20:50
mikefcbut it's fun trying to wrap my head around interp level20:51
fijalmikefc: :)20:51
fenrrir (~fenrrir@187.40.155.188) left irc: Quit: Leaving20:51
fijalmikefc: anyway, I think we're doing pretty good20:53
Alex_Gaynorfijal: CalledProcessError: Command '['/usr/bin/python', '-c', '\nimport types\nimport numpy as numpy\n\nobj = getattr(numpy, "ComplexWarning")\nfor name in dir(obj):\n    if name.startswith("_"):\n        continue\n    obj = getattr(obj, name)\n    kind = "U"\n    if isinstance(obj, types.TypeType):\n        kind = "T"\n    print kind, ":", name\n']' returned non-zero exit status 120:53
Alex_Gaynorworst error message ever20:53
Alex_Gaynordo you see the problem?20:53
Alex_Gaynoreh, I do20:54
fijalno?20:54
fijalI guess you can't pass such shit to -c, can you?20:54
fijalor \\n or so20:54
mwhudsonthe line after the if isn't indented enough?20:54
fijal-c should not be after \n I believe20:55
fijalbut yeah, \\n is my favorite candidate20:55
fijalAlex_Gaynor: anyway, the previous version was IMO better20:55
fijalyou're doing an abomination20:55
fijalat least do python -20:55
fijaland pipe it there20:55
fijalor echo | python or so20:55
Alex_Gaynorfijal: wait until you see the code20:55
fijaldon't-want-to20:56
Alex_Gaynorfijal: http://paste.pocoo.org/show/536390/ teh results are kind of nice, even if the code isn't20:58
fijalAlex_Gaynor: are you from the single-file-to-rule-them-all school by chance?20:58
fijalwhy did you just make php out of my 5 files?20:58
Alex_Gaynorfijal: I'm not :)20:58
Alex_Gaynorfijal: I just... started coding and suddently... abomination20:59
Topic changed on #pypy by mwhudson!mwh@linaro/mwhudson: PyPy, a fast and compliant python interpreter | IRC logs: http://tismerysoft.de/pypy/irc-logs/pypy/ | <Alex_Gaynor> fijal: I just... started coding and suddently... abomination20:59
Alex_Gaynorfijal: I'm going to get a snack, then I'll review your branch.  Feel like cleaning up my disaster :D21:00
fijalno :)21:00
fijalI feel like taking shower it was >30 today21:00
Alex_Gaynorfijal: I'll reivew your branc anyways :)21:02
fijalmaybe later21:02
fijalbut maybe I'll go to bed21:02
lizardo (~lizardo@189.2.128.130) left irc: Quit: Leaving21:04
kenaan12alex_gaynor better-jit-hooks-2 11b906e1141138 15/pypy/module/pypyjit/interp_resop.py: Some naming changes.21:17
Alex_Gaynorfijal: my only feedback before merge is: Should the attribute on the DMP be descr, or perhaps greenkey?21:17
fijalgreenkey_repr?21:19
fijalwell, dunno21:19
fijalgreenkey is better21:19
fijalAlex_Gaynor: feel free to change?21:19
fijalit does translate FYI21:19
Alex_Gaynoryes, I'll change21:19
Alex_Gaynorafter I change I'll merge I guess?21:19
fijalyeah21:19
timotimowe don't have masked arrays yet, do we?21:22
[revolting_blob!sunshine@107.6.5.104] --abuse LeDucky21:23
fijalnope21:23
fijalis numpy really using BLAS for dot?21:24
fijalanyway, sleep21:25
timotimohuh, why can numpy do stuff like "if x is masked: ... if y is masked: ..." ?!21:25
timotimonot sure if it's awesome or really dumb21:26
Alex_Gaynorfijal: can you review that push and if it looks good merge?21:27
kenaan12alex_gaynor better-jit-hooks-2 116354270bb576 15/pypy/module/pypyjit/: change DMP.descr to DMP.greenkey21:27
fijalAlex_Gaynor: just merge it21:29
Alex_Gaynorfijal: great21:29
Alex_Gaynorfijal: now I get to translate again21:30
kenaan12alex_gaynor better-jit-hooks-2 11886f6fbce8ef 15/: Close branch for merge21:30
kenaan12alex_gaynor default 11c6ea8ef23064 15/pypy/: (fijal mostly) Merged better-jit-hooks-2.  This exposes metadata about DebugMergePoints in jit hooks, specifi...21:30
fijalAlex_Gaynor: you have a translated version in tmp/usession-better-jit-hooks-2-0/testing_121:31
fijalon tannit21:31
fijalbefore your changes though21:31
Alex_Gaynorfijal: exactly :)21:31
fijalis it a big deal?21:31
fijalno :)21:31
Alex_Gaynornot sure, but I'm stepping away from my computer for a bit anyways, so it's fine to translate again21:31
fijalok21:32
fijalAlex_Gaynor: does interp_attrproperty_w prevent you from overwriting it btw?21:32
Alex_Gaynorfijal: yes, it's fget only21:32
fijalk21:32
Alex_Gaynorfijal: FYI there's a python-ideas thread about a pure python interface for creating buffers (i.e. __buffer__)(21:33
fijalnice21:34
Alex_Gaynorsomeday we should fix our buffers :/21:34
fijalit's hard/pointless21:35
fijalAlex_Gaynor: I'll try finish indexing by arrays tomorrow21:38
fijalany idea what next?21:38
Alex_Gaynorfijal: dunno, I have a giant list of functions/types/methods that aren't implemented :)21:38
Alex_Gaynorfijal: 2 good ones: copy=False to array constructor and subclass support.  On a more meta level: ability to run numpy's tests21:39
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.21:45
aboudreault (~alanb@osgeo/member/aboudreault) left irc: Ping timeout: 255 seconds21:47
bivab (~david@89-40.5-85.cust.bluewin.ch) left irc: Quit: bivab21:54
stakkars_ (~tismer@p5DCE1AF1.dip.t-dialin.net) joined #pypy.22:04
jaffachief (~textual@host86-170-15-228.range86-170.btcentralplus.com) joined #pypy.22:13
fprimex (~fprimex@fl-67-233-144-239.dhcp.embarqhsd.net) joined #pypy.22:39
PiotrSikora (~none@nginx/adept/piotrsikora) left irc: Excess Flood22:41
PiotrSikora (~none@nginx/adept/piotrsikora) joined #pypy.22:41
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer22:43
Tobu (~Tobu@unaffiliated/tobu) joined #pypy.22:47
antocuni (~antocuni@89-40.5-85.cust.bluewin.ch) left irc: Ping timeout: 252 seconds22:47
tilgovi_ (~randall@couchdb/developer/tilgovi) joined #pypy.22:50
squiddy (~squiddy@g229037172.adsl.alicedsl.de) left irc: Quit: Leaving22:50
fijalAlex_Gaynor: pick random one and implement it you suggest?22:50
Alex_Gaynorfijal: Did you see my 3 concrete reccomendations?22:50
fijalya22:50
fijalthe tes thingie not gonna happen22:51
fijalfor now22:51
Alex_Gaynorhow about copy=False so we can use numpy.asarray directly?22:51
fijalyeah22:51
mikefc_fijal: got time to approve a patch or two?  (simple numpypy layout reorg)22:52
fijalmikefc_: sure22:52
mikefc_http://paste.pocoo.org/show/536428/ 22:52
mikefc_^ move the numpypy tests to module/test_lib_pypy/numpypy22:52
mikefc_it's just a file move22:53
jaffachief (~textual@host86-170-15-228.range86-170.btcentralplus.com) left irc: Quit: Computer has gone to sleep.22:53
fijalmikefc_: sounds good, commit it22:53
Action: mikefc_ tries22:53
fijalAlex_Gaynor: I frankly wonder where "numpy is efficient" meme comes from22:54
Alex_Gaynorno idea22:56
kenaan12cool...@gmail.com default 11759d96c835e6 15/: move apptests for numpypy to module/test_lib_pypy22:57
kvda (~kvda@124-169-134-221.dyn.iinet.net.au) left irc: Quit: x___x22:58
mikefc_woo.22:59
fijalmikefc_: can you put a full name on your profile btw?22:59
fijalAlex_Gaynor: only dot is using blas as far as I can see22:59
mikefc_i thought i did.22:59
fijalmaybe we should do the same22:59
Alex_Gaynorfijal: well, there's specific submodules which expose tons of blas stuff22:59
Alex_GaynorAFAIK22:59
fijalyes yes of course22:59
fijalbut it's not transparent22:59
bbot2Started: 15http://buildbot.pypy.org/builders/own-macosx-x86-32/builds/77923:00
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/20723:00
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/101623:00
bbot2Started: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64-2/builds/3523:00
fijalAlex_Gaynor: have another slower module btw23:00
berdario (~Tiibiidii@host158-89-dynamic.0-87-r.retail.telecomitalia.it) left irc: Quit: Ex-Chat23:00
Alex_Gaynorfijal: ?23:00
Tobu (~Tobu@unaffiliated/tobu) left irc: Ping timeout: 252 seconds23:01
fijalgenshi23:01
Alex_Gaynorfijal: sphinx is slow too btw, I saw you added the benchmark23:01
Alex_Gaynorfijal: the benchmark does a make clean before rerunning right?23:02
fijalhttp://paste.pocoo.org/show/536429/23:02
fijalAlex_Gaynor: yeah23:02
fijalAlex_Gaynor: via hodgestar23:02
fijalwho made genshi run on pypy at least23:02
Alex_Gaynorgotta go do stuff with boiling water, BRB23:02
mikefc_fijal: http://paste.pocoo.org/show/536431/    this moves fromnumeric just into the core/ subdir in numpypy. to follow the layout of numpy.23:08
mat^2 (~mathias@212.130.113.35) left irc: Ping timeout: 252 seconds23:08
fijalmikefc_: that is fine as well, without asking :)23:08
mikefc_ok. 23:09
fenrrir (~rodrigo@189.70.243.190) joined #pypy.23:09
mikefc_trying to be careful not to break anything....23:11
kenaan12mikefc default 11e4314d54ea7f 15/lib_pypy/numpypy/: Move fromnumeric into core/ subdirectory to follow numpy's layout23:11
fijalyeah23:11
fijalbut there are layers23:11
fijalif you commit stuff to lib_pypy/numpypy you can only break numpypy and it23:11
fijal's tests23:11
fijalwhich is fine23:11
fijalif you however commit code to module/micronumpy, you might break translation23:11
fijaland armin will complain23:11
fijalas long as you can only break numpypy and it's tests it's generally fine23:12
mikefc_cool. that takes the pressure off :)23:12
fijalso if you want to commit stuff into module, it's good to run *all* tests and ask for a review23:12
fijalespecially that it usually takes little time23:12
fijalbut lib_pypy is fine in general23:12
Tobu (~Tobu@unaffiliated/tobu) joined #pypy.23:13
int3__ (~int3__@bb116-14-91-196.singnet.com.sg) left irc: Ping timeout: 255 seconds23:17
int3__ (~int3__@bb121-7-238-234.singnet.com.sg) joined #pypy.23:17
stakkars_ (~tismer@p5DCE1AF1.dip.t-dialin.net) left irc: Quit: schnarch23:18
mikefc_fijal: 2 more little patches?   the transpose patch i sent you before: http://paste.pocoo.org/show/536436/ 23:19
fijalif not hasattr(a, 'T')23:20
fijalis this really how numpy does it?23:20
mikefc_no.23:20
mikefc_numpy does something much fancier. let me paste it.23:20
int3__ (~int3__@bb121-7-238-234.singnet.com.sg) left irc: Remote host closed the connection23:21
int3__ (~int3__@bb121-7-238-234.singnet.com.sg) joined #pypy.23:21
mikefc_http://paste.pocoo.org/show/536438/23:22
fijaloh23:23
fijalmikefc_: can we do something like that?23:23
mikefc_relying on __array_wrap__ possibly existing.  this is what i mean when i say i know all the numpy stuff will probably have to be reimported.23:23
fijal__array_wrap__ might be None for now23:24
fijaland this should work23:24
fijalno?23:24
fijalhm23:24
fijalmikefc_: no, I think we're fine for now with your stuff :)23:24
mikefc_i think we'll get there eventually, but "my stuff" was just copied from the old appnumpy code. i.e. i'm lazy23:25
bfirsh (u1308@gateway/web/irccloud.com/x-lzfctubzlskvzarf) left irc: Remote host closed the connection23:25
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-wsfyymxqtfzowcvo) left irc: Remote host closed the connection23:25
oal (u4126@gateway/web/irccloud.com/x-xrgmfvojqecpvhsv) left irc: Remote host closed the connection23:25
bfirsh (u1308@gateway/web/irccloud.com/x-spqsfteezyengfsz) joined #pypy.23:25
mikefc_i don't think we have an asarray() yet.23:25
mikefc_i may write a _wrapit() work-a-like. but there's so much else (more interesting!) to do23:26
mikefc_fijal: for now is that transpose patch ok?23:27
fijalmikefc_: yes23:27
fijalyeah, I think we're fine23:27
fijaldon't we have .transpose()?23:28
mikefc_we have the .T method. but now the numpy.transpose() function.23:28
mikefc_s/now/not23:28
mikefc_until now :)23:28
Nick change: tilgovi_ -> tilgovi23:29
kenaan12mikefc default 11bc30fb8eed3e 15/: Add numpypy.transpose() function to call the array method23:29
gtaylor (~gtaylor@108-196-160-156.lightspeed.tukrga.sbcglobal.net) left irc: Quit: Konversation terminated!23:30
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-ouchyoetporourmg) joined #pypy.23:31
#pypy: mode change '+o Alex_Gaynor' by ChanServ!ChanServ@services.23:31
whitelynx (~whitelynx@63.241.75.144) left irc: Quit: Ex-Chat23:31
kenaan12mikefc default 1107bd76d923fd 15/lib_pypy/numpypy/core/fromnumeric.py: Use the proper exception for NotImplementedError23:32
mikefc_fijal: thanks for the assistance. I think I should probably work in branches from now on.23:33
fijalcool23:33
rguillebert (~rguillebe@89-40.5-85.cust.bluewin.ch) left irc: Quit: Quitte23:37
kenaan12fijal extradoc 117288adc71948 15/blog/draft/pypy-2011.rst: start drafting blog post23:38
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer23:40
stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy.23:40
mikefc_fijal: is this you? http://baroquesoftware.com:5000/23:40
stakkars_ (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer23:40
fijalAlex_Gaynor: want to come up with some interesting numbers?23:41
fijalmikefc_: yes, why?23:41
kenaan12fijal extradoc 115ddb005ca8e2 15/blog/draft/pypy-2011.rst: some made up claims23:41
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy.23:41
Alex_Gaynorfijal: 423:41
fijalAlex_Gaynor: look at extradoc23:41
mikefc_fijal: looks good :)23:41
fijalmikefc_: alex managed to make an abomination from it though23:42
Alex_Gaynorfijal: technically I built my abominiation from scratch23:43
fijalAlex_Gaynor: how that makes you less guilty?23:44
Alex_Gaynorfijal: well, your pressure sourcecode isn't defiled this way :P23:44
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer23:44
fijalanyway, I lost link23:44
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy.23:44
fijalbit.ly/wANBxu23:45
fijalAlex_Gaynor: ^^^23:45
bbot24Failure: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64-2/builds/3523:45
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) left irc: Read error: Connection reset by peer23:45
stakkars (~tismer@p5DDB7A14.dip.t-dialin.net) joined #pypy.23:45
Alex_Gaynornot bad23:45
fijalno23:45
fijalthat's a year of dev23:45
Alex_GaynorI know, we're awesome :)23:46
fijalAlex_Gaynor: anyway, I made up some stats23:46
fijalfeel free to add more23:46
fijaland I'll try sleeping23:46
fijalAlex_Gaynor: and put a legend in the abomination23:47
Alex_Gaynorfijal: there's no legend needed in my version, it's very obvious23:47
fijalok23:47
fijalAlex_Gaynor: so how about you hook it up with buildbot?23:47
fijalso it runs it every night and puts it somewhere?23:48
mattip (4fb1711d@gateway/web/freenode/ip.79.177.113.29) joined #pypy.23:48
mattiphi23:48
tilgovi (~randall@couchdb/developer/tilgovi) left irc: Read error: Connection reset by peer23:49
Alex_Gaynorfijal: that would require understanding the buildbot :D23:49
Alex_Gaynorfijal: I need to improve the UI a bit before I make this public23:49
Alex_Gaynorfijal: I just pushed some updates to tracebin-client FWIW23:49
mattipI'm still plugging away at dot().23:50
fijalmattip: just use blas23:50
mattipFor small arrays it may be better than calling blas.23:50
mattipBut yes, blas should be next.23:51
fijalwe don't have a good story for small arrays anyway23:51
fijalbut yeah, I'm mostly joking23:51
fijalmattip: can you at least make it raise in case the case is unsupported?23:51
fijalinstead of returning bogus results23:51
bbot24Failure: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-64/builds/20723:51
Alex_Gaynorfijal: you broke benchmarks it appears23:52
fijaldid I?23:52
mattipuhh, the default branch doesn't know about multidimensional arrays in descr_dot.23:52
mattipI think.23:52
fijaloh fuck you23:52
bbot24Failure: 15http://buildbot.pypy.org/builders/jit-benchmark-linux-x86-32/builds/101623:53
Action: mattip ducking23:53
fijalmattip: not you, python's os module23:53
mattipwhew!23:53
kenaan12fijal benchmarks 11b6376749f6e5 15/benchmarks.py: unbreak benchmarks23:53
fijalAlex_Gaynor: why does python have so much worse interface than bash?23:53
fijallike, rm -rf23:54
fijalhow hard is that?23:54
Alex_Gaynorshutil.rmtree23:54
fijalyeah23:54
fijalAlex_Gaynor: but not good enough23:54
fijalit still raises OSError if nothing's there23:54
Alex_GaynorWhatever, try to write an if statement in bash, then complain23:54
fijal:]23:54
fijalok23:54
fijalI'll try to sleep instead23:55
fijalcya23:55
mattipI can't wait to hear fijal's response when I commit the first version of this dot() mess.23:55
Alex_Gaynorfijal: bye23:55
fijalAlex_Gaynor: commit abomination next to buildbot's stuff, I'll integrate it23:55
Alex_Gaynor (u1246@gateway/web/irccloud.com/x-ouchyoetporourmg) left irc: Remote host closed the connection23:55
fijalmattip: I'm judgemental, what can ISay :)23:55
mattips'ok. I'll just take my meds.23:55
fijalhehe23:56
fijalmattip: cool, would be good to have a working dot btw23:56
fijalmattip: I suppose it's fine to force everything for now btw23:56
ousado (~ousado@unaffiliated/ousado) left irc: Read error: Connection reset by peer23:56
mattipIm not even dreaming about lazy yet.23:57
mattipit's just really ugly right now.23:57
tilgovi (~randall@couchdb/developer/tilgovi) joined #pypy.23:58
--- Wed Jan 18 201200:00

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