[00:11] <ericvrp> sabi: I think I fixed it in the current svn version.
[00:12] <sabi> ok, cool.
[00:15] <ericvrp> bye
[00:15] <ericvrp> oh, and thanks again!
[00:21] <sabi> :)
[00:33] seb-- (n=seb-@adsl-66-159-194-51.dslextreme.com) left #pypy ("Leaving").
----- silence for 51 minutes ----- [01:24] cfbolz (n=carlson@hdlb-d9b950cd.pool.mediaWays.net) joined #pypy.
[01:39] cfbolz (n=carlson@hdlb-d9b950cd.pool.mediaWays.net) left irc: "Verlassend"
----- silence for 51 minutes ----- [02:30] <sabi> Python 2.4.1 (pypy 0.7.0 build) on linux2
[02:30] <sabi> yay! llvm version translated
[02:30] Action: sabi can do work now :)
[02:45] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) got netsplit.
[02:46] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) returned to #pypy.
----- silence for 1 hr and 35 minutes ----- [04:21] Rhamphoryncus (n=rhamph@unaffiliated/rhamphoryncus) joined #pypy.
----- silence for 1 hr and 4 minutes ----- [05:25] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) left irc: Read error: 60 (Operation timed out)
[05:28] yuuh (i=ore@i577B4BEA.versanet.de) left irc: "utz utz utz"
[05:30] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) joined #pypy.
[05:35] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) got netsplit.
[05:41] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) got lost in the net-split.
[05:50] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) joined #pypy.
[05:54] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) got netsplit.
[05:55] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Good morning, all. In just under 5 minutes, we'll be restarting most of the network. It will be very noisy and we hope it goes smoothly. Wish us luck and see you on the other side!
[05:56] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) returned to #pypy.
[05:59] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) got netsplit.
[05:59] sabi (i=njriley@shrug.csl.uiuc.edu) got netsplit.
[05:59] thingie24 (n=asdf@valhalla.ccp.cc) got netsplit.
[05:59] mwh (n=user@82-33-185-193.cable.ubr01.azte.blueyonder.co.uk) got netsplit.
[05:59] dialtone (n=dialtone@host126-9.pool80116.interbusiness.it) got netsplit.
[05:59] xorAxAx (n=xorAxax@moinmoin/coreteam/alexander) got netsplit.
[05:59] Rhamphoryncus (n=rhamph@unaffiliated/rhamphoryncus) got netsplit.
[05:59] lypanov (n=lypanov@arcturus.theinternetco.net) got netsplit.
[05:59] tic (n=vision@c-a46e73d5.019-35-67626717.cust.bredbandsbolaget.se) got netsplit.
[05:59] zookofamilytime (n=user@blk-215-95-202.eastlink.ca) got netsplit.
[05:59] ericvrp (n=chatzill@ericvrp.demon.nl) got netsplit.
[05:59] ironfroggy (n=calvin@pcp01874264pcs.chmbrs01.pa.comcast.net) got netsplit.
[05:59] idnar (i=mithrand@unaffiliated/idnar) got netsplit.
[05:59] mwh`` (n=user@pc150.maths.bris.ac.uk) got netsplit.
[06:00] xorAxAx (n=xorAxax@83.137.100.36) joined #pypy.
[06:00] seb- (n=seb-@adsl-66-159-194-51.dslextreme.com) returned to #pypy.
[06:00] mwh (n=user@82.33.185.193) joined #pypy.
[06:00] idnar (i=mithrand@165.146.97.119) joined #pypy.
[06:00] thingie24 (n=asdf@213.167.138.65) joined #pypy.
[06:00] tic (n=vision@c-a46e73d5.019-35-67626717.cust.bredbandsbolaget.se) returned to #pypy.
[06:00] ericvrp (n=chatzill@82.161.1.42) joined #pypy.
[06:00] mwh`` (n=user@pc150.maths.bris.ac.uk) returned to #pypy.
[06:00] Rhamphoryncus (n=rhamph@unaffiliated/rhamphoryncus) returned to #pypy.
[06:00] lypanov (n=lypanov@arcturus.theinternetco.net) returned to #pypy.
[06:01] tic (n=vision@c-a46e73d5.019-35-67626717.cust.bredbandsbolaget.se) got netsplit.
[06:01] lypanov (n=lypanov@arcturus.theinternetco.net) got netsplit.
[06:01] Rhamphoryncus (n=rhamph@unaffiliated/rhamphoryncus) got netsplit.
[06:01] dialtone (n=dialtone@host126-9.pool80116.interbusiness.it) returned to #pypy.
[06:01] ironfroggy (n=calvin@pcp01874264pcs.chmbrs01.pa.comcast.net) returned to #pypy.
[06:02] pyb0t joined #pypy.
[06:03] drewp (n=nnnnndre@a.mx.bigasterisk.com) joined #pypy.
[06:04] xorAxAx (n=xorAxax@thinkmo.de) joined #pypy.
[06:05] sabi (i=njriley@shrug.csl.uiuc.edu) joined #pypy.
[06:05] idnar (i=mithrand@unaffiliated/idnar) joined #pypy.
[06:05] lypanov (n=lypanov@arcturus.theinternetco.net) joined #pypy.
----- silence for 2 hr and 11 minutes ----- [08:16] adim (n=adim@logilab.net2.nerim.net) joined #pypy.
----- silence for 1 hr and 59 minutes ----- [10:15] idnar (i=mithrand@unaffiliated/idnar) left irc: Nick collision from services.
[10:15] idnar_ (i=mithrand@unaffiliated/idnar) joined #pypy.
----- silence for 1 hr and 16 minutes ----- [11:31] aleale (n=andersle@clogs.dfki.uni-sb.de) joined #pypy.
[11:35] pedronis (n=Samuele_@c-398b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
[11:35] pedronis (n=Samuele_@c-398b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Client Quit
[11:50] cfbolz (n=a@b0bar.physi.uni-heidelberg.de) joined #pypy.
[11:50] <cfbolz> hi all!
[11:52] <lypanov> moin cfbolz :)
[12:04] <mwh> pyb0t: are you dead?
[12:04] <cfbolz> no he is not
[12:04] <cfbolz> (or is this a command he can understand?)
[12:04] <mwh> ben's mail on pypy-de
[12:04] <mwh> v
[12:05] pedronis (n=Samuele_@c-398b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
[12:06] <cfbolz> christian told me that the hoster of nimrod.terra-link has exceeded some sort of volume and so the logs will be away untill the 1st
[12:06] <cfbolz> Hi samuele!
[12:06] <mwh> ah, ok
[12:06] <cfbolz> mwh: we are still logged, though, we just can't access it
[12:06] <pedronis> hi
[12:09] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Hi all. Just a reminder: there's information on the current state of the network on http://freenode.net/news.shtml .... we're currently blocking private messages from unregistered users. One of the effects of this change is to reduce private message spam. We're working on providing a NickServ mode by which registered users can turn off the filtering.
[12:10] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] This feature is not yet a permanent thing. You can help by providing feedback. Please email policy@freenode.net and indicate if you (1) would prefer this blocking feature be permanent, (2) would prefer to see it temporary, just during spambot attacks, or (3) would prefer not to see private message blocking from unregistered users. Thank you!
[12:12] <mwh> cfbolz: ah
[12:12] <mwh> do you wnat to answer the mail, then?
[12:14] <cfbolz> ok
[12:14] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] I should mention that when unregistered users try to send messages and are blocked, they're referred to information on registration. Also, staff will be able to read messages from unregistered users so that they can report problems.
[12:14] <cfbolz> at least if I find out how to do that without a mouse (mine has just died)
[12:15] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Again, please provide feedback as outlined above, to policy@freenode.net . Thank you for your patience, and have a great morning!
[12:17] <mwh> heh
[12:25] dialtone (n=dialtone@host126-9.pool80116.interbusiness.it) left irc: "Leaving"
[12:25] dialtone (n=dialtone@host126-9.pool80116.interbusiness.it) joined #pypy.
[12:32] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Hi all. Just to clarify based on a comment from a user. Staff will be able to read private messages from unregistered users directed to the *staffer*, not to other users. Thanks.
[12:39] <cfbolz> whew, test_gc now takes nearly ten minutes
[12:39] <xorAxAx> better than 2 days :)
[12:40] <mwh> cfbolz: congratulations on the mouseless email :)
[12:40] <cfbolz> thanks
[12:41] <cfbolz> thunderbird is surprisingly annoying without mouse, but firefox is even worse
[12:41] <mwh> heh
[12:41] Nick change: idnar_ -> idnar
[12:41] <mwh> i have a mac and i hardly ever touch the damn thing :)
[12:42] <cfbolz> "thing" being the mac or the mouse?
[12:42] <mwh> mouse
[12:42] <mwh> or, well, trackpad
[12:43] <cfbolz> :-)
[12:43] <cfbolz> I basically don't use it except for surfing
[12:43] <xorAxAx> try opera without a mouse :-)
[12:43] <cfbolz> those link things are really a bad idea
[12:44] <mwh> ah, yes, i use it for them, i have to admit
[12:44] <mwh> you can drive mose we browsers without a mouse, it's just very, very annoying
[12:45] <cfbolz> indeed
[12:47] Action: cfbolz -> food
[12:51] <lypanov> xorAxAx: opera has keyshortcuts for almost everything :)
[12:52] <xorAxAx> lypanov: yep :)
[12:52] <xorAxAx> you can even use it if you are blind
[12:56] thingie24 (n=asdf@valhalla.ccp.cc) left irc:
[13:03] Action: cfbolz back
----- silence for 27 minutes ----- [13:30] <ericvrp> hi carl
[13:30] <cfbolz> hi eric!
[13:32] <ericvrp> is your SOC deadline today?
[13:32] <cfbolz> yes
[13:32] <cfbolz> I am mostly finished. still working on documentation
[13:33] <ericvrp> happy with the result?
[13:33] <cfbolz> kind of
[13:34] <cfbolz> there is nothing I am particularly unhappy about
[13:34] <cfbolz> but it feels not entirely perfect yet
[13:35] <ericvrp> nice. Something else: I found out that llvm has an experimental escape analysis being worked on. Will give that a try.
[13:38] <ericvrp> 75 euro at the nearest online shop (can't currently send private messageson IRC)
[13:40] <xorAxAx> yeah, lilo is censoring
[13:40] <xorAxAx> hint - you can invite him to your own channel
[13:40] <xorAxAx> /join #ericvrp etc.
[13:42] <cfbolz> ericvrp: do you want me to get a copy for you?
[13:42] <ericvrp> yes please
[13:43] <ericvrp> do you have a paypal account?
[13:45] <cfbolz> no
[13:45] <cfbolz> just give me cash in paris or whenever
[13:45] <ericvrp> ok, thank you. I will mail you my address
[13:45] <cfbolz> good
[13:46] <ericvrp> thank man, have to go again. till later
[13:46] <cfbolz> bye
[13:47] <lypanov> cfbolz: your soc was the pluggable gc?
[13:47] <cfbolz> plugability was not one of the main goals
[13:48] <cfbolz> rather the fact that you can write it in python
[13:48] Action: lypanov nods
[13:48] <lypanov> seen the jalapeno framework for this?
[13:48] Action: lypanov is most interested in implementing ulterior reference counting at some point in the future for his own project
[13:51] <cfbolz> jalapeno? link?
[13:54] nik (n=chatzill@44.74.62.81.cust.bluewin.ch) joined #pypy.
[13:54] <cfbolz> hi nik
[13:54] <nik> hi all
[13:57] <pedronis> cfbolz: jalapeno was the old name of jikes rvm
[13:57] <cfbolz> ah i see
[13:57] <cfbolz> so the answer is yes
[13:57] __alex (n=alex@141.22.64.4) joined #pypy.
[13:58] <lypanov> pedronis: ah :)
[13:59] <pedronis> when it was just an IBM research project and still not open source
[14:03] nik (n=chatzill@44.74.62.81.cust.bluewin.ch) left #pypy.
[14:07] nikh (n=chatzill@44.74.62.81.cust.bluewin.ch) joined #pypy.
[14:15] <cfbolz> pedronis: is there anything you specifically want me to do? otherwise I will just concentrate on documentation
[14:16] <pedronis> the gcs run themself under interpreter now?
[14:16] <cfbolz> yes
[14:16] <cfbolz> that's why it's so slow
[14:17] <pedronis> you also did a new gc?
[14:17] <cfbolz> yes, a _very_ simple deferred refcounting
[14:17] <cfbolz> works on llinterpreter as well
[14:17] <cfbolz> it's kind of naive
[14:17] <pedronis> no, I think for now is ok
[14:18] <pedronis> it will probably need further refactoring to be really used with the backends but that's a different issues
[14:18] <cfbolz> of course
[14:18] <cfbolz> I will work on that if the issue arises
[14:19] <pedronis> yes, anyway the backend themself will need refactoring
[14:19] <cfbolz> indeed
[14:19] <pedronis> like the gc policy interface
[14:19] <cfbolz> yes
[14:20] <cfbolz> I guess looking at the GcWrapper thing will help finding out what the policy will have to do -- the responsibility is kind of the same
[14:21] <pedronis> nikh: I'm reading them, but it seeems I cannot answer until I register my nick
[14:22] <cfbolz> just use a channel
[14:23] <xorAxAx> dont we all love lilo?
[14:23] <nikh> pedronis: i see
[14:23] <lypanov> pedronis: /msn NickServ register blah
[14:23] <lypanov> pedronis: /msn NickServ identify blah
[14:23] <cfbolz> don't we all just hate spammers?
[14:23] <lypanov> s/msn/msg/
[14:23] <nikh> pedronis: join #gothenburg-planning for a moment?
[14:24] <lypanov> cfbolz: yup
[14:24] <xorAxAx> cfbolz: i havent seen any spam on freenode yet
[14:24] <xorAxAx> cfbolz: and i am on a lot of channels here
[14:24] <xorAxAx> (for months=
[14:24] <cfbolz> guess you are lucky
[14:25] <xorAxAx> yeah
[14:25] <cfbolz> I haven't received much eMail spam either, but I've been using the web for years
[14:25] <xorAxAx> then you havent been active on the usenet :-)
[14:28] arigo (n=arigo@pdpc/supporter/sustaining/arigo) joined #pypy.
[14:28] <cfbolz> hi armin
[14:29] <arigo> hi
[14:29] <pedronis> hi
[14:29] Action: arigo tries half-heartedly to get pygame running on a mac
[14:30] <pedronis> download python2.4 and pygame from bob
[14:30] <arigo> ok
[14:30] <pedronis> it seems that the speed enhancement was indeed useful
[14:31] <cfbolz> arigo: you have a mac?
[14:32] <arigo> right now, here at work, yes
[14:32] <cfbolz> ah, I see
[14:32] <arigo> but it should be magically turned into a PC soon anyway
[14:32] <pedronis> arigo: translated pypy is around only 100-120 times slower than CPython (with thread disabled)
[14:33] <pedronis> but enabling thread can add another 50% overhead
[14:33] <arigo> good
[14:33] <arigo> ah, another low-hanging fruit :-)
[14:33] <arigo> using sys.setcheckinterval()
[14:33] <pedronis> wich is not unexpected given that we release the GIL on each bytecode
[14:35] arigo (n=arigo@pdpc/supporter/sustaining/arigo) left irc: "Leaving"
[14:36] arigo (n=arigo@pdpc/supporter/sustaining/arigo) joined #pypy.
[14:37] Action: arigo switched to his beloved laptop :-)
[14:37] <arigo> btw, any clue why this mac I was on takes to translate_pypy?
[14:38] <arigo> it's supposed to have 1.25GB of RAM, an 1.8GHz machine
[14:38] <pedronis> never tried to translate on a Mac
[14:38] <pedronis> you should ask holger
[14:39] <arigo> ok
[14:39] <arigo> do you want to do the checkinterval thing?
[14:40] <pedronis> I'm checking in a change to richards to use time.time
[14:41] <pedronis> time.clock overflows
[14:43] <pedronis> I'm running all the tests
[14:43] <pedronis> I had to do fix in annotation
[14:44] <pedronis> because the results of type could make the contains assert unhappy
[14:44] <pedronis> because is_type_of was attached to cached PBC
[14:44] <pedronis> strange that it was not triggered before
[14:45] <arigo> argh
[14:45] <arigo> indeed
[14:45] <pedronis> I'm just constructing the PBC directly
[14:46] <pedronis> bookkeeper.getpbc is a bit odd now
[14:46] <pedronis> because is not creating classdefs anymore
[14:46] <pedronis> so the comment doesn't apply
[14:46] <pedronis> and the caching is just a speed memory thing
[14:47] <arigo> ah, right
[14:49] <pedronis> indeed we have some places that already construct PBCs directly
[14:50] <pedronis> but I don't know what would happend speed/memory wise if we removed the caching/getpbc completely
[14:52] <pedronis> arigo: I'm getting timeout problems from the thread tests
[14:52] <arigo> ah
[14:53] <pedronis> related to the speed increase?
[14:54] arre (n=ac@1-1-5-33a.gfa.gbg.bostream.se) joined #pypy.
[14:54] <pedronis> cfbolz: are failures in test_gc expected?
[14:55] <cfbolz> no
[14:55] <arigo> pedronis: more likely related to the fact that the test are now always enabled (before, you needed py.test --usemodules=thread)
[14:55] <cfbolz> pedronis: what architecture, which tests?
[14:55] <pedronis> arigo: umph, I see, that's bad
[14:56] <arigo> I'm not too surprized
[14:56] <pedronis> well, but it means the test are not that useful
[14:56] yuuh (n=yuuhu@i577B4C89.versanet.de) joined #pypy.
[14:56] <pedronis> if they are so prone to fail
[14:57] zooko (n=user@blk-215-95-202.eastlink.ca) joined #pypy.
[14:57] <arigo> I just added the checkinterval code, and I get timeouts -- expectedly, as threads run even less often than before
[14:58] zooko (n=user@blk-215-95-202.eastlink.ca) left irc: Remote closed the connection
[14:58] <arigo> time to think about how system calls could release the GIL?
[14:58] <arigo> maybe even in py.py ?
[14:59] <pedronis> we should start the refactoring too, otherwise adding new feature will always look shinier
[15:00] <arigo> which of the many refactoring do you have in mind?
[15:00] <pedronis> translate_pypy and translator
[15:00] <arigo> ah
[15:01] <pedronis> cfbolz: test_llinterp_lists
[15:01] <cfbolz> :-)
[15:02] <cfbolz> which class?
[15:02] <pedronis> tuples
[15:02] <cfbolz> there are something like six tests for every name
[15:02] <pedronis> string_concatanation
[15:03] <pedronis> cfbolz: finding stuff among the tons of log output is not completely trivial
[15:04] <pedronis> cfbolz: SemiSpace it seems
[15:05] <cfbolz> could you pipe the stuff into a file and put it somwhere?
[15:05] <pedronis> I can try
[15:11] <pedronis> cfbolz: they seem index out of bound problems in fixed_size
[15:11] <pedronis> it takes a bit to run them
[15:13] <cfbolz> I know
[15:13] <cfbolz> index out of bound in fixed_size can mean anythin
[15:13] <cfbolz> because the arg to fixed_size is a typeid read from somewhere out of memory
[15:14] <pedronis> well it seems that that typeid is wrong
[15:14] <cfbolz> yes, or it's reading from some random memory location
[15:19] <pedronis> cfbolz: there's a file gc_failure in my codespeak home dir
[15:19] <cfbolz> thanks
[15:19] <cfbolz> what sort of machine is it>
[15:19] <cfbolz> ?
[15:20] mwh`` (n=user@pc150.maths.bris.ac.uk) left irc: "biab"
[15:21] <pedronis> an amd 64 but running at 32 bits
[15:24] <pedronis> cfbolz: L l P are all 4
[15:28] <lypanov> arigo: i keep on wondering why noone ever bothers to get those frameworks working on the mac
[15:28] <arigo> pygame?
[15:28] <lypanov> aye
[15:28] <lypanov> and all the varioous ruby ones
[15:28] <lypanov> nothing works :P
[15:29] <arigo> yes, Bob seems to be a bit alone in caring
[15:29] <arigo> but his OSX 10.3 / Python 2.3 version of Pygame just crashed spectacularly for me
[15:29] <lypanov> ouch oops
[15:29] <pedronis> arigo: you want 2.4
[15:30] <pedronis> 2.3+pygame was very crash prone for me too
[15:30] <arigo> I'll try to recompile Python -- not right now though
[15:30] <pedronis> compiling pygame is not too bad on the Mac, the hassle are the dependecy to have sound and video working
[15:35] mwh` (n=user@pc150.maths.bris.ac.uk) joined #pypy.
[15:35] Action: lypanov could never get any of it working as all of the builds assumed linux
[15:35] <lypanov> (this was the various ruby things though)
[15:36] <lypanov> (which surprises me, as most of the non japanese ruby guys run on macs)
[15:36] <lypanov> (but evidently don't play games :P)
[15:36] <lypanov> arigo: btw. what are you using pygame for?
[15:36] <lypanov> i need something to do dumping of my execution structure
[15:36] <xorAxAx> flowchart visualisation?
[15:36] Action: lypanov nods
[15:37] <lypanov> basically the same thing then
[15:37] <lypanov> its impossible to track anything anymore :/
[15:38] Nick change: pedronis -> pedronis_away
[15:50] <arigo> yes, we're using pygame to display control flow graphs
[15:50] <arigo> in combination with the Graphviz 'dot', it's quite useful
[15:55] Action: lypanov nods
[15:55] <lypanov> i was using dot but it can't show extra information on hover
[15:56] <lypanov> so was going to make a html version
[15:56] <arigo> sounds familiar :-)
[15:56] <arigo> we eventually made our own pygame viewer and definitely don't regret it
[15:56] <arigo> messing with html is painful and limited
[15:57] Action: sabi did some stuff with dot-generated svg, but it's nowhere as flexible as pygame
[16:00] <lypanov> arigo: got screenshots of the pygame view?
[16:01] <sabi> i never tried to get the flowgraphs working in pygame on mac, but arigo's pypy presentation worked ok :)
[16:01] <lypanov> :)
[16:01] <arigo> let me take a few shots...
[16:02] <lypanov> thx :)
[16:08] <arigo> http://codespeak.net/svn/pypy/dist/pypy/doc/image/
[16:12] <lypanov> very cool :)
[16:15] <arigo> clicking around moves and zooms...
[16:16] <arigo> it should be easy to extract an independent tool that reads a .dot file, but of course the real interest is the extra information when the mouse is over a name in red (which also work like hyperlinks)
[16:17] pedronis_away (n=Samuele_@c-398b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Read error: 110 (Connection timed out)
[16:17] Action: lypanov nods
[16:17] <lypanov> i currently generate a dot file
[16:17] <lypanov> but i can't fit enough information in it
[16:18] <arigo> yes, here we generate an compile a dot file dynamically and insert more info back when we read the result (based on words)
[16:26] arre (n=ac@1-1-5-33a.gfa.gbg.bostream.se) left irc: Remote closed the connection
[16:35] pedronis (i=pedronis@ratthing-b246.strakt.com) joined #pypy.
----- silence for 30 minutes ----- [17:05] nikh (n=chatzill@44.74.62.81.cust.bluewin.ch) left #pypy.
[17:06] <cfbolz> pedronis: the failures are quite mysterious, I tried to reproduce them but no luck
[17:06] <cfbolz> the problem is that the semispace gc works (on your machine) if you run it directly
[17:07] <cfbolz> but not if you run it on the llinterpreter
[17:13] <pedronis> what can be different in those cases
[17:14] <cfbolz> the setup is a bit different
[17:14] <cfbolz> but still
[17:14] <cfbolz> on codespeak I get failures too
[17:14] <cfbolz> but different ones
[17:15] <pedronis> mmh
[17:15] <pedronis> I run them again on 2.4 and they passed
[17:15] <pedronis> the previous run was 2.3
[17:16] <cfbolz> ouch. now it gets really obscure
[17:18] <cfbolz> ok, some of the codespeak failures are expected
[17:19] <cfbolz> they test growth of the heap in bytes
[17:19] <pedronis> it's confirmed they fail on 2.3 and pass on 2.4
[17:20] <cfbolz> any ideas?
[17:21] <cfbolz> did array or struct change?
[17:21] <pedronis> yes, I'm wondering whether there were changes or bug fixes
[17:27] <mwh`> arigo: ping
[17:27] <arigo> pong
[17:28] <mwh`> when we discussed manmonths for UoB can you remember what we decided?
[17:28] <arigo> yes
[17:28] <arigo> (bea has made a copy of that too, I think)
[17:28] <mwh`> oh, hang on, i've found my notes too
[17:28] <arigo> :-)
[17:29] <mwh`> you've read bea's mail on pypy-funding?
[17:29] <arigo> not yet
[17:31] <mwh`> ok
[17:31] <mwh`> i think we need to come up with some blurb about how wonderful the UoB and/or me are
[17:31] <arigo> I guess we need to involve pedronis on this too
[17:32] <pedronis> cfbolz: do they work for you on 2.3?
[17:33] <cfbolz> no
[17:33] <pedronis> so you are seeing the same problem
[17:34] <cfbolz> on codespeak I see different ones
[17:35] <cfbolz> I still try to find a local computer with 2.3 installed
[17:35] <pedronis> either is a bug in 2.3 or something related to the changes in arithmetic behavior
[17:36] <cfbolz> ah, that's an idea
[17:37] <cfbolz> but it is mainly shifts that changed, isn't it?
[17:39] <mwh`> my it's suddenly got very dark outside
[17:39] <pedronis> cfbolz: yes, it is shifts
[17:39] <cfbolz> I don't think I use any
[17:50] aleale (n=andersle@clogs.dfki.uni-sb.de) left irc: "BitchX-1.1-final -- just do it."
[17:51] <pedronis> given that it works on 2.4 and the tests are deterministic
[17:51] <pedronis> it seems 2.3 reading garbage
[17:53] <cfbolz> it's all quite mysterious. I'm at the moment trying to do a diff between the log output of the two versions
[17:53] tirkal (n=user@unaffiliated/tirkal) joined #pypy.
[17:53] tirkal (n=user@unaffiliated/tirkal) left #pypy.
----- silence for 16 minutes ----- [18:09] lypanov (n=lypanov@arcturus.theinternetco.net) left irc: "leaving"
[18:16] <cfbolz> ok, I'm not going to look into it now. otherwise the docs won't be finished until tomorrow
[18:17] <cfbolz> pedronis: is that ok?
[18:19] <pedronis> yes, they pass on 2.4
[18:19] <pedronis> it seems we are very good at triggering all kind of strange behavior
[18:20] Action: hpk has cosmic ray bugs on translation on dist/ currently
[18:21] <cfbolz> pedronis: indeed
[18:22] <pedronis> hpk: what kind of failure are you getting
[18:23] <hpk> scrolled out of screen now, i'll record it next time around
[18:25] Action: pedronis I'm starting a translation
[18:32] Nick change: hpk -> hpk_out
[18:41] __alex (n=alex@141.22.64.4) left irc: Remote closed the connection
----- silence for 17 minutes ----- [18:58] Rhamphoryncus (n=rhamph@unaffiliated/rhamphoryncus) left irc: "Support ISO 8601! http://www.cl.cam.ac.uk/~mgk25/iso-time.html";
[19:03] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Hi all. As you may be aware, we're currently testing server changes to block private messages from unregistered users. We've working on a companion feature to NickServ, SET UNFILTERED ON, which allows registered users to turn back on messages from unregistered ones.
[19:04] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] We're currently testing SET UNFILTERED ON on the freenode testnet (server irc.freenode.net port *9001*).... if you need this feature, please go test it and let us know on #hyperion on the production network, if there are any problems. For more info, stop by #hyperion .... thank you!
[19:07] <pedronis> arigo: it seems that the large overhead when threads are enabled is still there even if releasing less often
[19:07] adim (n=adim@logilab.net2.nerim.net) left irc: Remote closed the connection
[19:09] <arigo> maybe we call getexecutioncontext() quite often?
[19:10] <pedronis> mmh, that's a possibility
[19:18] <pedronis> arigo: we get it on function entry and on list/dict comparison
[19:18] <arigo> doesn't sound large-overhead-ish
[19:23] <pedronis> I get 456 pystone without threads and 286 with
----- silence for 25 minutes ----- [19:48] Nick change: hpk_out -> hpk
[19:58] <cfbolz> I am going home now
[19:58] <cfbolz> see you tomorrow
[19:58] <arigo> see you!
[19:58] <cfbolz> pedronis: if you still come up with anything, drop me an email
[19:59] <pedronis> I don't think there's anything, the bulk of it is there
[19:59] <pedronis> see you
[19:59] cfbolz (n=a@b0bar.physi.uni-heidelberg.de) left irc: "Leaving"
[20:01] <pedronis> 'm about to leave too
[20:01] <pedronis> see you
[20:01] Action: pedronis leaving
[20:01] <arigo> bye
[20:02] <mwh> arigo: where are you currently?
[20:02] pedronis (i=pedronis@ratthing-b246.strakt.com) left irc: "Chatzilla 0.9.67 [Firefox 1.0.2/20050325]"
[20:02] <arigo> at Duesseldorf
[20:02] <mwh> pdpc/supporter/sustaining/arigo isn't very enlightening :)
[20:02] <mwh> ah
[20:03] <arigo> hum, yes I tried to shake off this cloak but they didn't answer my mail
[20:03] <sabi> yay freenode!
[20:03] <mwh> heh
[20:16] <xorAxAx> Thanks for using Freenode!11111one
[20:16] <xorAxAx> arigo: /stats p
[20:16] <xorAxAx> arigo: then query one of the guys that you see in the list
[20:23] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Hi all. In just a moment, we'll be upgrading services to provide registered users with the ability to talk to unregistered. For more information, "/msg nickserv help set unfiltered" once the new version is up
[20:30] -lilo (i=levin@freenode/staff/pdpc.levin) to $*- [Global Notice] Hi all. As of this moment, if you are a registered user and need to receive messages from unregistered users, please just "/msg nickserv set unfiltered on". Let us know if there are any problems!
----- silence for 23 minutes ----- [20:53] pedronis_away (n=Samuele_@c-398b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
[20:53] Nick change: pedronis_away -> pedronis
[20:57] arigo (n=arigo@pdpc/supporter/sustaining/arigo) left irc: "[x]chat"
[21:05] ericvrp2 (n=eric@ericvrp.demon.nl) joined #pypy.
----- silence for 32 minutes ----- [21:37] ericvrp2 (n=eric@ericvrp.demon.nl) left irc: Read error: 113 (No route to host)
[21:40] ericvrp2 (n=eric@ericvrp.demon.nl) joined #pypy.
----- silence for 43 minutes ----- [22:23] ericvrp2 (n=eric@ericvrp.demon.nl) left irc: "Lost terminal"
----- silence for 1 hr and 32 minutes ----- [23:55] dialtone (n=dialtone@host126-9.pool80116.interbusiness.it) left irc: "Leaving"
[23:56] dialtone (n=dialtone@host126-9.pool80116.interbusiness.it) joined #pypy.
[00:00] --- Thu Sep 1 2005