[00:43] aleale (~redorlik@cpe.atm0-0-0-129140.0x3ef2fa3a.bynxx3.customer.tele.dk) left irc: "http://www.blackened.com/blackened/";
[00:49] yuuh (ore@dsl-62-220-10-98.berlikomm.net) left irc: "utz utz utz"
----- silence for 38 minutes ----- [01:27] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "[x]chat"
----- silence for 20 minutes ----- [01:47] senra (~rodrigo@200-158-47-74.dsl.telesp.net.br) joined #pypy.
----- silence for 47 minutes ----- [02:34] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) left irc: "http://fredrikj.net";
----- silence for 37 minutes ----- [03:11] senra (~rodrigo@200-158-47-74.dsl.telesp.net.br) left irc: "Leaving"
----- silence for 2 hr and 21 minutes ----- [05:32] hpk (~hpk@merlinux.de) left irc: Read error: 110 (Connection timed out)
[05:43] lac (~jacob@lac.silver.supporter.pdpc) got netsplit.
[05:49] lac (~jacob@lac.silver.supporter.pdpc) got lost in the net-split.
[05:52] Rhamphoryncus (~rhamph@rhamphoryncus.user) left irc: ""What do you mean, how many hit points do I have?" || Support ISO 8601! http://www.cl.cam.ac.uk/~mgk25/iso-time.html";
----- silence for 30 minutes ----- [06:22] lac (~jacob@c-51c6e055.1321-1-64736c11.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 1 hr and 52 minutes ----- [08:14] adim (~adim@logilab.net2.nerim.net) joined #pypy.
----- silence for 23 minutes ----- [08:37] arre (ac@ratthing-b3fa.strakt.com) joined #pypy.
----- silence for 51 minutes ----- [09:28] aleale (~redorlik@cpe.atm0-0-0-129140.0x3ef2fa3a.bynxx3.customer.tele.dk) joined #pypy.
[09:41] <arre> Hi Anders.
[09:42] <arre> Are you still having problems with test_intobject?
----- silence for 41 minutes ----- [10:23] <aleale> I havent had time to retry. I am trying to get the codecs to work with your unicode implementation
----- silence for 30 minutes ----- [10:53] <aleale> arre: from a fresh checkout everything works. I must have screwed up something yesterday, sorry
[10:55] <arre> aleale: Don't be. Me and Samuele fix some snags with the marged unicode branch. You might have gotten the bad version.
[10:57] yuuh (ore@dsl-62-220-10-98.berlikomm.net) joined #pypy.
[11:04] <aleale> could be.
----- silence for 1 hr ----- [12:04] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
[12:18] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) joined #pypy.
[12:18] <cfbolz> hi!
[12:19] <arigo> hi Carl!
[12:19] <arigo> your exams?
[12:19] <stakkars> hi!
[12:20] <cfbolz> 1.3
[12:20] <cfbolz> yay!
[12:20] <stakkars> :)
[12:21] Action: arigo is unfamiliar with the scale
[12:21] <stakkars> 1.0 is best
[12:21] <cfbolz> 4.0 is worst
[12:22] <arigo> good :-)
[12:22] <arigo> great :-)
[12:22] <cfbolz> :-)
[12:23] <cfbolz> I got very nervous and made some blatant mistakes in the middle. Thus 1.3 is quite a surprise, I thought it would be much worse
[12:24] <arigo> that was a spoken exam, I guess
[12:24] <arigo> beginnings and endings are most important :-)
[12:25] <cfbolz> yes. Otherwise I wouldn't know the scale yet
[12:25] <cfbolz> grade I mean
[12:25] <arigo> sure
[12:25] <cfbolz> arigo: Yes. After the last stupid exam I asked to change topic, so that the end wouldn't be totally screwed up.
[12:26] <cfbolz> s/exam/error
[12:29] <aleale> Hi, congrats to you Carl. Do we still have pickle problems? testresults show 86% again
[12:30] <aleale> I thougth it would be 89%
[12:31] <cfbolz> aleale: thanks -- more time for PyPy now :-)
----- silence for 55 minutes ----- [13:26] faassen (~faassen@a80-127-80-154.adsl.xs4all.nl) joined #pypy.
[13:34] arre (ac@ratthing-b3fa.strakt.com) left #pypy.
[13:43] yuuh (ore@dsl-62-220-10-98.berlikomm.net) left irc: Read error: 104 (Connection reset by peer)
[13:43] stakkars (gdyggr@dsl-62-220-10-98.berlikomm.net) left irc: Read error: 54 (Connection reset by peer)
----- silence for 18 minutes ----- [14:01] stakkars (~tismer@i528C1C2E.versanet.de) joined #pypy.
[14:02] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[14:11] <cfbolz> hi eric!
[14:20] <cfbolz> arigo: I
[14:20] <cfbolz> 'm trying to read up a bit on llypes
[14:21] <cfbolz> It would be possible to implement reference counting in lltypes only, right?
[14:21] <cfbolz> withouth much help from the backend
[14:21] <cfbolz> I mean
[14:27] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) left irc: "Leaving"
----- silence for 37 minutes ----- [15:04] jacob22 (jacob@enzo.strakt.com) joined #pypy.
[15:09] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 16 minutes ----- [15:25] <ericvrp> I have llvm mostly working on the codespeak server (x86-64) . Had to patch translator/llvm a little also but now 64 tests pass and only 7 fail.
[15:26] <arigo> I wonder if Holger didn't succeed in installing it on codespeak too
[15:27] <ericvrp> Maybe, we discussed this yesterday and we agreed I look into it. Maybe he tried a previous version?
[15:28] <arigo> ah, possibly.
[15:28] <arigo> that was definitely before the new llvm release.
[15:28] <arigo> good, then :-)
[15:37] <stakkars> arigo: hmm, I should have checked in,earlier. You new hashable base class conflicts with my constant optimization.
[15:38] <arigo> shouldn't be to hard to resolve, is it?
[15:39] <arigo> s/to/too
[15:39] <stakkars> well, maybe a bit.
[15:39] <stakkars> my constant does not have a key. I didn't see it yet,but I guess your baseclass has akey?
[15:40] <arigo> yes, the base class is almost exactly like Constant
[15:40] <stakkars> I reverted the logic. The expensive key tuples are no longer needed.
[15:40] <arigo> I guess you can move the optimization there.
[15:40] <stakkars> key is a computed attribute, and I create a new subtype for every const type is seen.
[15:41] <arigo> what's the purpose of the subtyping?
[15:41] <stakkars> no I cannot I think. Oh well, I can, if I can teach hashable to use its subclass to do the trick.
[15:42] <stakkars> save quite some megabytes for not creating almost equally many tuples as constants.
[15:42] <arigo> why do you need subclassing instead of just computing the key attribute as was previously done now in the constructor?
[15:42] <arigo> s/now//
[15:43] <stakkars> wait -- previously key was not computed,or I missed it.
[15:43] <arigo> key was computed by __init__ and stored in the attribute.
[15:44] yuuh (ore@i528C1C2E.versanet.de) joined #pypy.
[15:46] <stakkars> yes. you don't need to store key.
[15:46] <arigo> right
[15:46] <arigo> the obvious way to avoid storing it is as a property, which recomputes it the same way. no subclassing needed?
[15:46] <stakkars> yes, but computationally quite expensive.
[15:47] <stakkars> the other thing is to get rid of key completely, even the computed field,
[15:47] <stakkars> becuase the way the COnstant is working, it is perfect as a dict member, nokeys needed.
[15:47] <arigo> yes, but the .key is used to compute equality and hash of the Constant itself.
[15:48] <arigo> if you can compute the same result in a better way, fine
[15:48] <stakkars> I can
[15:48] <arigo> but for now it's rather a premature optimization if it gets in the way of refactoring
[15:48] <stakkars> I agree.
[15:49] <stakkars> The effect was that I could swap logic:
[15:49] <stakkars> we have 100000 constants, with less than 9000 mutables. So the immutables get a computed key,
[15:49] <stakkars> the mutables are a subclass with an explicit keyfield. :-)
[15:50] <stakkars> I didn'texpect refactoring here.
[15:50] <arigo> I didn't either, two days ago :-)
[15:51] <stakkars> I would suggest to add pickling support, which I have, to the Constant (better not the hashable)
[15:51] <arigo> makes sense
[15:52] <stakkars> it is just handy, because by slottifying, you remove the builtin ability.
[15:52] <arigo> ah, I see
[15:52] <stakkars> so in a sense it is a matter of keeping the interface compatible ;-)
[15:52] <arigo> this lowers the risk of having parts of the user program "sneak" in the pickle
[15:53] <stakkars> this is true. Well, it cannot happen at all, because I changed concept.
[15:54] <stakkars> I just recognized that the wholepickling process is nothing else but a very reduced
[15:54] <stakkars> version of genc, but producing a different product.
[15:55] <stakkars> I will get exactly what we want, because I use the traversal methods which we want to use. simple
[15:55] <stakkars> makes the pickler just another (very tiny) backend.
[15:56] <arigo> :-)
[15:56] <stakkars> makes sense?
[15:56] <arigo> sure, I said the same thing yesterday ;-)
[15:56] <stakkars> I didn't recognize it. Now I understand that one.
[15:56] <stakkars> yes.
[15:57] <stakkars> basically, tis could even be an interface for other backends, in the sense
[15:57] <stakkars> that pickling does really nothing but the little object collecting.
[15:57] <stakkars> the code generation process is an extension. (but maybe har to formulate)
[15:58] <stakkars> hard
[15:58] <stakkars> a nice side effect is that we are not so restricted. genc handles it with and without annotator.
[15:58] <stakkars> so the same applies to genpickle. It is the same, just no C code.
[16:01] Action: stakkars going to work in the sun
[16:01] <arigo> :-)
[16:02] <stakkars> hey, basically the whole intospective process could be thought as a parser,
[16:03] <stakkars> and codegenerators define their callbacks. (XLSish, not soo nice)
[16:03] <stakkars> s/XLS/XML/
[16:04] <stakkars> should be written in stackless style.
[16:05] <stakkars> the parser walks all the objects and produces a stream,
[16:05] <stakkars> which is then dispatched to several stateful tasklets which handle their stuff in a loop.
[16:05] <stakkars> much later maybe :-]
[16:05] <arigo> :-))
[16:06] <stakkars> works great, the whole hugeproject that I were in is done this way.
[16:07] <stakkars> (ip) away
----- silence for 17 minutes ----- [16:24] Rhamphoryncus (~rhamph@rhamphoryncus.user) joined #pypy.
[16:26] mayall (~mayall@212-70-201-231.ath.dialup.tee.gr) joined #pypy.
[16:32] stakkars (~tismer@i528C1C2E.versanet.de) left irc: Read error: 60 (Operation timed out)
[16:47] mayall (~mayall@212-70-201-231.ath.dialup.tee.gr) left irc: "ChatZilla 0.9.52B [Mozilla rv:1.6/20040113]"
[16:59] adim (adim@logilab.net2.nerim.net) left #pypy.
[17:14] Rhamphoryncus (~rhamph@rhamphoryncus.user) left irc: Read error: 104 (Connection reset by peer)
[17:17] Rhamphoryncus (~rhamph@rhamphoryncus.user) joined #pypy.
[17:21] cfbolz (~cfbolz@hdlb-d9b95eae.pool.mediaWays.net) joined #pypy.
[17:25] cfbolz (~cfbolz@hdlb-d9b95eae.pool.mediaWays.net) left irc: Client Quit
----- silence for 38 minutes ----- [18:03] faassen (~faassen@a80-127-80-154.adsl.xs4all.nl) left irc: "Download Gaim: http://gaim.sourceforge.net/";
----- silence for 1 hr and 20 minutes ----- [19:23] Nick change: pedronis -> pedronis_afk
[19:23] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Remote closed the connection
----- silence for 44 minutes ----- [20:07] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: "Chatzilla 0.9.68.3 [Firefox 1.0.4/20050516]"
[20:08] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[20:18] mayall (~mayall@212-70-201-138.ath.dialup.tee.gr) joined #pypy.
[20:18] <mayall> pypeople?
[20:23] mayall (~mayall@212-70-201-138.ath.dialup.tee.gr) left irc: "ChatZilla 0.9.52B [Mozilla rv:1.6/20040113]"
[20:24] mayall (~mayall@212-70-201-138.ath.dialup.tee.gr) joined #pypy.
[20:34] mayall (~mayall@212-70-201-138.ath.dialup.tee.gr) left irc: "ChatZilla 0.9.52B [Mozilla rv:1.6/20040113]"
[20:35] mayall (~mayall@212-70-201-138.ath.dialup.tee.gr) joined #pypy.
[20:44] mayall (~mayall@212-70-201-138.ath.dialup.tee.gr) left irc: "ChatZilla 0.9.52B [Mozilla rv:1.6/20040113]"
----- silence for 40 minutes ----- [21:24] tic (~vision@c-ba6e73d5.019-35-67626717.cust.bredbandsbolaget.se) left irc: Read error: 60 (Operation timed out)
[21:37] Nick change: pedronis_afk -> pedronis
[21:38] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 25 minutes ----- [22:03] yuuh (ore@i528C1C2E.versanet.de) left irc: "utz utz utz"
----- silence for 51 minutes ----- [22:54] yuuh (~yuuhu@i528C1C2E.versanet.de) joined #pypy.
[23:02] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) joined #pypy.
----- silence for 17 minutes ----- [23:19] suman (~suman1@221.135.200.203) joined #pypy.
[23:22] <suman> hi all, i am new to pypy and want to contribute to it... i have read the architecture document and the gettibng started document.. after readig them i felt that the right way to do it would be fix some bugs or do some tasks... but i found no todo's ..also is this the right way to start?
[23:27] <suman> is there a todo list or something like that?
[23:28] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) left irc: "Leaving"
[23:39] <arigo> hi suman
[23:40] <arigo> what parts of PyPy are you more interested in?
[23:40] <arigo> after reading the architecture document?
[23:43] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) joined #pypy.
[23:45] <arigo> basically, we are collecting these days issues about things left to do
[23:46] <arigo> the list is unsorted, and not classified in any way as "would be a nice starting task" or not
[23:48] <cfbolz> we need to become more friendly to beginners
[23:48] <arigo> indeed
[23:49] <arigo> suman: at the moment, I would suggest that you write a mail to pypy-dev
[23:49] <arigo> suman: if you have some preferences, tell us
[23:50] <arigo> suman: in any case we'll reply and organize some list of suggestion
[23:58] arigo (~arigo@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: Remote closed the connection
[23:59] <aleale> uit
[23:59] aleale (~redorlik@cpe.atm0-0-0-129140.0x3ef2fa3a.bynxx3.customer.tele.dk) left irc: "Foolproof operation: All parameters are hard coded."
[00:00] --- Wed May 25 2005