[02:01] yuuh (~yuuhu@i528C1F8D.versanet.de) got netsplit.
[02:07] yuuh (~yuuhu@i528C1F8D.versanet.de) got lost in the net-split.
[02:07] yuuh (fomizwul@i528C1F8D.versanet.de) joined #pypy.
[02:13] stakkars_ (eklkrgby@i528C1F8D.versanet.de) joined #pypy.
[02:18] stakkars (~tismer@rosine27.inf.fu-berlin.de) left irc: Read error: 113 (No route to host)
----- silence for 1 hr and 32 minutes ----- [03:50] yuuh (fomizwul@i528C1F8D.versanet.de) left irc: "utz utz utz"
[03:57] llun\ved\ (~triest@rtay-164-107-198-196.resnet.ohio-state.edu) joined #pypy.
----- silence for 35 minutes ----- [04:32] llun\ved\ (triest@rtay-164-107-198-196.resnet.ohio-state.edu) left #pypy.
----- silence for 1 hr and 35 minutes ----- [06:07] tea (~tea@cap049-090.kcn.ne.jp) joined #pypy.
----- silence for 20 minutes ----- [06:27] stakkars_ (eklkrgby@i528C1F8D.versanet.de) left irc: Read error: 113 (No route to host)
----- silence for 1 hr and 44 minutes ----- [08:11] adim (~adim@logilab.net2.nerim.net) joined #pypy.
----- silence for 50 minutes ----- [09:01] aleale (~redorlik@cpe.atm0-0-0-129140.0x3ef2fa3a.bynxx3.customer.tele.dk) joined #pypy.
[09:11] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: "Chatzilla 0.9.68.3 [Firefox 1.0.4/20050516]"
[09:11] Nick change: hpk_away -> hpk_
[09:12] hpk_ (~hpk@merlinux.de) left irc: Read error: 104 (Connection reset by peer)
[09:12] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[09:12] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: Client Quit
[09:15] arre (ac@ratthing-b3fa.strakt.com) joined #pypy.
[09:15] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[09:29] hpk_ (~hpk@merlinux.de) joined #pypy.
----- silence for 25 minutes ----- [09:54] Gromit (~bear@B74cc.b.pppool.de) joined #pypy.
----- silence for 22 minutes ----- [10:16] <hpk_> aleale: hi anders, ayt?
[10:17] Nick change: hpk_ -> hpk
----- silence for 26 minutes ----- [10:43] <aleale> hi
[10:48] <arre> Good morning everyone.
[10:57] <hpk> good morning
[10:58] <arre> How do you feel about getting the new unicode in before the release?
[10:58] <hpk> not too enthusiastic but it depends
[10:58] <hpk> did you try to merge and see what happens?
[10:59] <hpk> we shifted things around quite a lot around
[10:59] <arre> The only thing I'm missing right now is unicodedata.normalize()
[10:59] <arre> I han't tried to merge yet, but i'm ready to.
[10:59] <hpk> give it a try and see what happens
[10:59] <hpk> but i tend to think we should do this for the next release
[11:00] Action: hpk is favouring a "0.7" at EuroPython, based on CPython 2.4 and with various fixes and improvements
[11:01] <hpk> and maybe even including a tool to translate small python programs
----- silence for 22 minutes ----- [11:23] <hpk> aleale: i answered your issue (twice in fact :-)
[11:34] <aleale> arre: I think we should wait (I have tried to merge, and had a lot of trouble)
[11:34] <aleale> but maybe it is only my lack of svn knowledge
[11:36] <aleale> hpk: thanks
----- silence for 21 minutes ----- [11:57] Gerald (~bear@B74ae.b.pppool.de) joined #pypy.
[11:58] Nick change: Gerald -> Gromit_
[12:01] Gromit (~bear@B74cc.b.pppool.de) left irc: Nick collision from services.
[12:01] Nick change: Gromit_ -> Gromit
----- silence for 27 minutes ----- [12:28] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) joined #pypy.
----- silence for 29 minutes ----- [12:57] Gerald (~bear@B7482.b.pppool.de) joined #pypy.
[12:58] <arre> aleale: I'll start working on the merging but i won't commit any until after the release.
[12:58] Nick change: Gerald -> Gromit_
[12:59] arigo (~arigo@ratthing-b409.strakt.com) joined #pypy.
[12:59] Gromit (~bear@B74ae.b.pppool.de) left irc: Nick collision from services.
[12:59] Nick change: Gromit_ -> Gromit
[13:07] Gromit (bear@B7482.b.pppool.de) left #pypy ("Client Exiting").
----- silence for 20 minutes ----- [13:27] yuuh (urcahtdh@i528C1F8D.versanet.de) joined #pypy.
[13:28] <pedronis> arigo: notice that app_inspect was marked not marked as NOT_RPYTHON before the comment was added, did you try to see what happens if you NOT_RPYTHON after the change?
[13:28] <arigo> pedronis: argh. The *comment* influences the ApplevelClass.
[13:29] <arigo> uh? no, you're not talking about callable(), right
[13:29] <pedronis> I'm talking about callable, but maybe confused
[13:29] <pedronis> let me check
[13:30] <hpk> arigo, pedronis: hi
[13:30] <arigo> btw, I just re-disabled use_geninterp from translate_pypy (was re-enabled by mistake in gateway.py)
[13:30] <arigo> hpk: hi
[13:31] <pedronis> arigo: the module app_inspect did not needed the NOT_RPYTHON comment before the enable print changes
[13:32] <arigo> ah, ok
[13:32] <pedronis> both NOT_RPYTHON and the comment above the isinstance line were added
[13:32] <pedronis> in the enabling print commit
[13:32] <pedronis> wondering whether your change make it "RPYTHON" again or not
[13:33] <arigo> pedronis: does it make sense now in annrpython to "continue anyway" if there are blocked blocks?
[13:33] <arigo> we shouldn't have any of these now, maybe we should again raise an exception?
[13:33] <pedronis> yes
[13:34] <pedronis> I thing we removed all sources of spurious blocked blocks
[13:34] <pedronis> if there are some is a real problem typically
[13:34] Action: hpk i wondering how everybody relates to the release ... (feeling a bit lonely lately)
[13:34] Action: hpk -> lunch
[13:34] <pedronis> hpk: we are trying to ship something that can still annotate all of PyPy properly
[13:34] <arigo> hpk: we are hunting bad annotator behavior probably due to print statements
[13:39] <arigo> pedronis: the interface of the flow space is really broken; that's why Christian needs the hasattr in gateway.py.
[13:39] <arigo> pedronis: the flow space should be able to analyse a PyCode object instead of a real CPython function.
[13:39] <arigo> pedronis: but I guess it's not something to fix now...
[13:40] <pedronis> I see
[13:41] <pedronis> using a hasattr instead of adding a method to the space is still somewhat orthogonal
[13:41] <arigo> no, I'm thinking about something different
[13:41] <pedronis> to the fact that there's a need to reconstruct the function
[13:42] <arigo> gateway uses call_args(w_func), so all that is really needed is to look at this Function object, write a simple_call to it, and analyse its func_code
[13:42] <arigo> but of course this requires changing the whole meaning of simple_call and the annotator and translator
[13:42] yuuh (urcahtdh@i528C1F8D.versanet.de) left irc: Read error: 145 (Connection timed out)
[13:43] <arigo> i.e. we should try to turn CPython functions into Functions (or otherwise some container) for the flow graphs, and not the other way around.
[13:47] <arigo> pedronis: I don't see any blocked blocks at all
[13:47] <pedronis> arigo: I said I suspect the last one
[13:47] <pedronis> was releated to geninterp being enabled
[13:47] <arigo> oh, right
[13:48] <arigo> good, then
[13:48] <pedronis> still strange, but won't show if it is disabled
[13:53] <aleale> I have test_file_extra and test_import failing
[13:53] <arigo> did you clean up your .pyc files?
[13:54] <aleale> I'll try that, sorry
[14:01] <aleale> it still fails
[14:02] rxe (~rxe@host-212-158-203-122.bulldogdsl.com) joined #pypy.
[14:02] <rxe> hi
[14:03] <arigo> hi!
[14:03] <pedronis> arigo: yes the NOT_RPYTHON is necessary now in app_inspect, unless there's a way to lazily access _instance because geninterp chokes on it
[14:04] <rxe> arigo: i am still in time for M.5?
[14:04] <arigo> rxe: I guess so
[14:04] <arigo> depends on how much you want to change the world :-)
[14:04] <rxe> arigo: :-) I am just checking docs changes & minor stuff to traceobjspace
[14:05] <arigo> go ahead :-)
[14:06] <arigo> ah, we've been moving the doc about traceobjspace around a bit already
[14:07] <rxe> yup conflicts galore i see...
[14:08] <arigo> it's now described in objspace.txt
[14:08] <arigo> and there is an example in getting_started.txt.
[14:08] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) joined #pypy.
[14:09] <rxe> arigo: cool! but wish i had an internet connection at home - ok resolving :-)
[14:17] rxe (rxe@host-212-158-203-122.bulldogdsl.com) left #pypy.
[14:20] Action: pedronis going to the office
[14:20] pedronis (~Samuele_P@c-3a8b70d5.022-54-67626719.cust.bredbandsbolaget.se) left irc: "Chatzilla 0.9.68a [Firefox 1.0.2/20050317]"
[14:21] Action: hpk is back
[14:22] <hpk> arigo: it's just that half-the-people are not communicating and the other half is mostly working on translation which is not exactly a M0.5 target
[14:22] <arigo> yes, I understand your concerns
[14:23] <arigo> I wouldn't say Samuele and me are mostly working on translation; it took half an hour this "morning" to see what's wrong with translate_pypy.py
[14:24] <arigo> and I definitely understand the point of view that translate_pypy.py should be shipped in a fully working state.
[14:24] <hpk> well, if you really want that then we should have made that an issue or talked about it before
[14:24] <hpk> and it also enforces my "freeze dist" suggestion
[14:25] <hpk> becasue by tomorrow translate_pypy will be broken again
[14:25] <hpk> and we do advertise 'svn co ...dist'
[14:26] <arigo> it's all about what the release really is for.
[14:27] <hpk> sure
[14:27] <hpk> but it's also about what we decided
[14:27] <arigo> we decided not to put more work in translation, but that's a bit different
[14:28] <arigo> translate_pypy.py works already, and it's one of the first real new things we can offer, so we can spend the tiny amount of efforts it requires just to make sure it indeed works precisely in the release tgz and the svn tag.
[14:29] <aleale> so what do we need more ?
[14:29] <aleale> I think the documentation is ok ?
[14:29] <arigo> hpk: it should be a py.tested feature that we want to pass, if you like.
[14:30] <hpk> there just is/was a lot of ground work to do
[14:30] <hpk> i basically worked 14-16 hours the last days
[14:30] <hpk> and missed being able to discuss things
[14:32] <hpk> IMO targetting a 0.7 at EuroPython with python2.4/translate programs/lots of other pending stuff makes sense
[14:33] <aleale> I agree
[14:33] <arigo> definitely
[14:34] <aleale> Is it only me who have failing tests ?
[14:35] <hpk> aleale: last time i checked all passed or skipped
[14:35] <aleale> lib/test2/test_file_extra.py fails
[14:36] <arigo> aleale: traceback?
[14:36] <aleale> and \module\__builtin__\test\test_import.py
[14:36] <aleale> do you want the traceback here ?
[14:36] <arigo> hum
[14:36] <arigo> e-mail them to me, for example (arigo@tunes.org)
[14:38] <aleale> done
[14:39] <hpk> arigo: could you fix pypy-talk-pycon2005/pycon.txt:53: (ERROR/3) Unexpected indentation ?
[14:39] <hpk> the restifier is a big agressive and looks everywhere :-)
[14:39] <hpk> (we are linking html-generated stuff in extradoc from other places)
[14:40] <arigo> i see :-)
[14:40] rxe (~rxe@host-212-158-203-122.bulldogdsl.com) joined #pypy.
[14:40] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) joined #pypy.
[14:40] <cfbolz> hi!
[14:41] <hpk> hi carl!
[14:41] <arigo> hi!
[14:41] <hpk> so the only really remaining issues are geninterp documentation and the mission statement
[14:42] <arigo> aleale: didn't get your mail :-(
[14:42] <hpk> apart from actually packaging the stuff
[14:43] <hpk> arigo: would you like to write the mission statement in screen? or do you have something ready or in mind?
[14:44] <arigo> ok for screen, but let's wait for Samuele who I expect to show up any minute now
[14:45] <hpk> we are only talking about three sentences that need to be reviewed afterwards anyway but ok (i don't believe in the efficiency doc-screen sesssions with >2 people basically unless SHPY is there :-)
[14:46] <rxe> hi
[14:47] <hpk> hi richard
[14:47] <rxe> hpk: hi
[14:47] <rxe> sorry i've been absent
[14:47] <rxe> no internet connection
[14:47] <rxe> i wrote some docs on trace object space, but someone has beat me to it
[14:47] <rxe> what to do? maybe some kind of intelligent merge?
[14:48] <hpk> yes, just do as you see fit
[14:49] <cfbolz> rxe: do you mean the stuff in getting_started?
[14:52] <aleale> arigo: I tried to sent the tb from another account - hope it works this time
[14:53] <hpk> arigo: the main thing i was missing was feedback for the various changes i did but at some point i decided to just go ahead ...
[14:53] <arigo> aleale: ooooups. bad Windows expectations from me.
[14:53] <rxe> cfbolz: no - that was nice :-) I mean there were conflicts in objspace.txt
[14:54] <cfbolz> ok
[14:54] <rxe> cfbolz: nice to meet you btw. :-)
[14:54] <cfbolz> rxe: hi! :-)
[14:54] <hpk> rxe: you are coming to EP + the sprints, aren't you?
[14:54] <arigo> hpk: the current documentation situation is very good, if I can say so without feeling pushed into doing so
[14:54] <aleale> arigo: ok - I expected something like that
[14:55] <arigo> aleale: I'm not sure how universal newlines should work on Windows
[14:55] <arigo> aleale: e.g. file('abc', 'rU').read() should accept any kind of end-of-lines, but return them how?
[14:55] <cfbolz> hpk: I read all your changes and found them quite excellent -- sorry, couldn't do more due to time constraints
[14:55] <aleale> arigo: what can I do to help
[14:55] <aleale> ?
[14:55] <rxe> hpk: need to delay that answer... most likely yes.
[14:56] <hpk> arigo, cfbolz: thanks, actually i wasn't quite fishing for compliments but simply to uncertainty on my side
[14:56] <arigo> aleale: take the test_rw_universal function,
[14:56] <arigo> aleale: and try the same steps in plain Python with a plain file object
[14:57] <arigo> aleale: then see if the final f.read() returns \n or \r\n
[14:57] pedronis (pedronis@ratthing-b246.strakt.com) joined #pypy.
[14:57] <hpk> rxe: i would much enjoy it
[14:58] <aleale> arigo: I have to go get the kids now, will tonight (20.00) be early enough?
[14:58] <arigo> aleale: I guess not, but if needed I can reboot this machine on Windows and try myself
[15:01] <aleale> ok i can give it 10 min
[15:01] <arigo> aleale: wait, we figured out that the test is most probably broken
[15:02] <rxe> py.test rocks as a frontend to reST! :-)
[15:02] <rxe> hpk: thanks - me too. I would very much like to go.
[15:04] <arigo> pedronis: ok, mission statement time
[15:08] <arigo> screen -x arigo/arigo ?
[15:08] <aleale> arigo: in CPython it works as you expected
[15:08] <aleale> arigo: gotta go - seeya
[15:08] <arigo> see you
[15:08] rxe (~rxe@host-212-158-203-122.bulldogdsl.com) left irc: Read error: 104 (Connection reset by peer)
[15:09] rxe (~rxe@host-212-158-203-122.bulldogdsl.com) joined #pypy.
[15:18] <arigo> hpk: #pypy-miss?
[15:18] <hpk> arigo: ok
[15:26] rxe (rxe@host-212-158-203-122.bulldogdsl.com) left #pypy.
[15:26] rxe (~rxe@host-212-158-203-122.bulldogdsl.com) joined #pypy.
[15:34] yuuh (qdzoouk@dsl-62-220-10-128.berlikomm.net) joined #pypy.
[15:38] rxe (rxe@host-212-158-203-122.bulldogdsl.com) left #pypy.
[15:44] <cfbolz> hpk: I reformated the translation section of getting_started. Is it ok to check in (or too late to proofread)?
[15:45] <hpk> cfbolz: just go ahead
[15:45] <hpk> we need to do a last round of reading the docs anyway
[15:45] <cfbolz> ok
[15:50] yuuh (qdzoouk@dsl-62-220-10-128.berlikomm.net) left irc: "utz utz utz"
[15:52] stakkars (kxzlimqk@dsl-62-220-10-128.berlikomm.net) joined #pypy.
[15:52] <cfbolz> hi!
[15:52] <stakkars> hi
[15:53] <arigo> hi!
[16:04] <hpk> hi
[16:05] <stakkars> after classobjinterp and exceptionsinterp are evaporated,
[16:06] <stakkars> do you still want me to write a chapter about geninterp,
[16:06] <stakkars> or is it better to not talk about it at all?
[16:06] <hpk> yes, you can't escape it :-)
[16:06] <stakkars> how much timeleft?
[16:06] <hpk> the documentation is really also for people wanting to understand PyPy's code base
[16:07] <hpk> not really much, do you have something already?
[16:07] <stakkars> hey, then do we have a detailedexplanation of gateway.py and the horrendous tricks how it invents types and functions?
[16:08] <hpk> not yet
[16:08] <hpk> but documentation is clearly an iterative process
[16:08] <stakkars> I just wanted to say that *that* is really hard to understand (tried to pickle it tonite)
[16:09] <stakkars> so I add a chapter to translation, at the end maybe, because it is nothing that people will need urgently?
[16:10] <hpk> it's just something where people want to know how it works and relates to the rest of the code base
[16:10] <hpk> (including me :-)
[16:10] <stakkars> ok that will be short since it is meanwhile almost exactly what genc does, without annotation etc.
[16:12] <stakkars> I see that sys2 is gone. how was it solved?
[16:12] <hpk> genc is not documented much, either (but that's also because it is not yet in shape and using the RTyper properly)
[16:13] <stakkars> ouch, that means my stuff explains exactly that part, for genc too!
[16:13] <hpk> yes, at best
[16:13] <hpk> sys2 could go because there was no risk of a .py file doing "import sys" in pypy/module anymore
[16:14] <stakkars> fine
[16:14] <stakkars> do I extend the already quite lengthy translation.txt or start with a new one?
[16:15] <hpk> put it into translation.txt
[16:15] <hpk> it is also a good idea to use a link to geninter.py
[16:16] <stakkars> do we have an exact release time?
[16:17] <hpk> see https://codespeak.net/issue/pypy-dev/issue53 if you missed reading it
[16:17] <hpk> no, we don't
[16:17] <stakkars> I missed everything, was kind of sick. Too much hacking the last days
[16:17] <hpk> but it'd like to stop around 6pm and then do strictly consolidation/last fixes and the packaging
[16:18] <stakkars> can't get a good docu together before 8, we have guests here and I need to escape to the FU
[16:19] <stakkars> tracker sorting by id does not sort numerically, by the way
[16:21] <hpk> that's a known problem of the particular backend
[16:22] <hpk> then do it until 8
[16:27] <stakkars> 2 tests don't work here
[16:27] <stakkars> test_import (doesn't look windows specific)
[16:28] <stakkars> test_py (does, uses cmdexec)
[16:28] <stakkars> let's stick with this.
[16:28] <hpk> can you make an issue out of it with the tracebacks?
[16:28] <stakkars> yes
[16:29] <stakkars> current orany smilestone?
[16:29] <hpk> M0.5
[16:29] <stakkars> I don't believe we can solve it today
[16:29] <hpk> you can also post it to the "make sure the tests work" M0.5 issue (or however it's called)
[16:30] <cfbolz> arigo and aleale discuessed test_import
[16:30] <stakkars> ough, I see they are both py.process related.
[16:31] <stakkars> and I have to write docu, so who will fix this today
[16:31] <cfbolz> IIRC the conclusion was that the test is broken
[16:31] <stakkars> great, please someone disable it --- i'm escaping from here
----- silence for 44 minutes ----- [17:15] <cfbolz> did richard say whether his trace changes are finished?
[17:29] <arigo> I can try the windows tests myself, but I don't promise anything
[17:34] stakkars (kxzlimqk@dsl-62-220-10-128.berlikomm.net) left irc: Read error: 113 (No route to host)
[17:35] <cfbolz> BTW: I just watched the PyCon "slides" for the first time -- really cool
[17:38] <arigo> :-)
[17:38] <hpk> yeah, documentation even makes sense for ourselves :-)
[17:38] <cfbolz> I especially love the part where the fruit is shuffled by the interpreter
[17:41] Action: hpk is away a bit
[17:41] <cfbolz> Ok, I'm going home. See you later.
[17:41] <hpk> see you
[17:42] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) left irc: "Leaving"
[17:46] <arre> Time to head home. I'll here from you later...
[17:47] <arre> p.s. On the unicodefront: I'v merged my branch and all tests looks like passing, but i'll hold commiting anything untill after release.
[17:47] arre (ac@ratthing-b3fa.strakt.com) left irc: "using sirc version 2.211+KSIRC/1.3.11"
[17:48] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) joined #pypy.
[17:49] <cfbolz> did we agree on how much slower PyPy is?
[17:49] <cfbolz> the numbers still disagree
[17:59] Action: arigo and samuele heaving for an early dinner before the last rush
[17:59] Nick change: arigo -> arigo|dinner
[18:01] Gromit (~bear@B74d4.b.pppool.de) joined #pypy.
[18:14] <hpk> arigo|dinner: side-note: you didn't say who "we" is in your commit.
[18:16] <cfbolz> I changed the speed to 2000 times slower, ok?
[18:18] lac (~jacob@lac.silver.supporter.pdpc) joined #pypy.
[18:18] <cfbolz> hi!
[18:18] <lac> hi all
[18:20] <hpk> hi laura
[18:25] <cfbolz> hpk: I'm going to proofread some docs, should I start anywhere particular?
[18:26] <hpk> good question, hum
[18:27] <hpk> i wouldn't know
[18:27] <hpk> you could have a look if we don't repeat our "app<->interplevel" descriptions all the time
[18:28] <hpk> the architecture section about it is probably the best thing to reference to, should there be a duplications
[18:30] <cfbolz> ok. should I put a link from the section about the interpreter-level console in getting_started to the section in architecture?
[18:31] <hpk> yes, i'd add a sentence to the end of 2.2.1
[18:31] <cfbolz> ok
[18:31] <hpk> is it really called "affix" btw?
[18:31] <hpk> never heard that word before
[18:32] <lac> affix is a good english word.
[18:32] <cfbolz> yes, but what does it mean?
[18:33] <hpk> postfix probably :-)
[18:33] <idnar> affix is the generic term for prefixes and suffixes
[18:33] <hpk> prefix i meant
[18:33] <hpk> so you have prefix, suffix, afix, and postfix?
[18:33] <cfbolz> :-)
[18:33] <hpk> affix
[18:33] <idnar> postfix means the same as suffix, and affix could mean either :)
[18:33] <cfbolz> we need prefix then?
[18:34] <hpk> hehe, seems so
[18:34] <idnar> hang on
[18:34] <hpk> prefix i a better computer-english term i dare to guess
[18:34] <idnar> dictionary says I'm wrong, affix is an addition to the end specifically
[18:34] <lac> affix is the end only
[18:34] <idnar> so affix, suffix, and postfix are the same
[18:34] <cfbolz> ouch. we need prefix then even more
[18:34] <hpk> indeed.
[18:34] <cfbolz> I'll do it
[18:36] <hpk> your previous changes to getting_started look good, btw
[18:36] <cfbolz> ok. The translation section looked really stupid before
[18:36] Action: hpk considers changing getting_started to getting-started once and for all
[18:36] <cfbolz> good idea
[18:37] <hpk> i'll do it when you are finished
[18:37] <cfbolz> ok
[18:37] <hpk> because i need to tweak the web code as well
[18:38] <hpk> in my personal ranking the coding-guide needs the most efforts in the future
[18:38] <cfbolz> future == post 0.5?
[18:39] <hpk> yip
[18:39] <cfbolz> agreed
[18:39] <cfbolz> but not today
[18:39] <hpk> surely not
[18:39] <hpk> i am really exhausted
[18:39] <cfbolz> yes, me too. And I have an exam coming on Tuesday :-)
[18:39] <hpk> woa
[18:40] <hpk> i'll schedule a post-m0.5 meeting then right on wednesday to keep you active :-)
[18:40] <cfbolz> :-)
[18:41] <hpk> we should try all this without :-) to see how long it takes until we pull the guns
[18:42] <cfbolz> hehe. Would be an interesting experiment on how much humour depends on anything facial expressions
[18:42] <cfbolz> s/anything/
[18:42] <hpk> it's damn easy to misunderstand each other digitally
[18:42] <cfbolz> indeed
[18:43] <cfbolz> ouch: <Ctrl-D> won't work under Windows to exit interpreter-level console, right?
[18:43] <hpk> right
[18:43] <cfbolz> do you know what it is there?
[18:43] <cfbolz> I don't remember
[18:43] <hpk> wasn't it <control-Z> or something?
[18:43] <hpk> control-z return or something funny
[18:43] <cfbolz> right, right.
[18:46] <hpk> is this really correct english?:
[18:46] <hpk> A compiler/interpreter is close to as complex as software ever gets.
[18:47] <cfbolz> sounds wrong to me
[18:48] <hpk> me, too.
[18:49] <hpk> what about:
[18:49] <hpk> Writing Compilers and Interpreters are among the most
[18:49] <hpk> complex endeavours in software development.
[18:49] <hpk> in exchange?
[18:49] <cfbolz> nice
[18:50] Action: hpk should visit an english course some day ...
[18:50] <hpk> is it "is" or "are" here?
[18:50] <cfbolz> i'd say is, but I'm not sure
[18:52] <cfbolz> did richard change the output of the trace object space?
[18:52] <hpk> no idea
[18:52] <hpk> but cross-checking richard's changes is actually a good idea
[18:55] <cfbolz> ouch, the output looks different
[18:56] Nick change: arigo|dinner -> arigo
[18:56] <cfbolz> hpk: or did someone beautify this?
[18:56] <cfbolz> I mean the example in getting_started
[18:57] <hpk> hopefully not
[18:57] <hpk> but i don't know who inserted it
[18:57] <cfbolz> I did
[18:57] <cfbolz> anyway, I'll insert an up-to-date version
[18:57] <hpk> ok
[18:58] Action: hpk is meanwhile fiddling with architecture.txt and wistling innocently to not raise arigo's attention
[19:00] <arigo> mmmmh?
[19:00] <hpk> just wait for my monster diff
[19:01] <arigo> should I go off to Windows-land and look at the py.test failures?
[19:01] <hpk> if you dare to, yes
[19:01] <hpk> and check that control-d, control-z or whatever works with the console switching
[19:01] <arigo> ok
[19:06] <cfbolz> I'm finally going to eat something, too. See you.
[19:06] cfbolz (~cfbolz@edv01.physi.uni-heidelberg.de) left irc: "Leaving"
[19:16] Action: hpk ducks to escape arigo's "architecture.txt is already perfect" wrath
[19:17] arigo (~arigo@ratthing-b409.strakt.com) left irc: Remote closed the connection
[19:26] -lilo (lilo@levin-pdpc.staff.freenode) to $*- [Global Notice] Good evening, all. News up on the freenode website: (1) Major downtime on Sunday. (2) Status of fundraiser. Please take a look at http://freenode.net/news.shtml ! Thank you for your support for freenode, and have a great evening!
[19:32] adim (~adim@logilab.net2.nerim.net) left irc: Remote closed the connection
[19:40] <aleale> hpk: in the interactive pypy interpreter control-c get you to the interpeter level console
[19:40] <hpk> on win32
[19:40] <aleale> and control-d gets you back
[19:41] <hpk> sounds good
[19:41] <aleale> another control-d shuts pypy down. Just like it says in the documentation
[19:41] <aleale> on win32 yes
[19:48] arigo (~arigo@ratthing-b409.strakt.com) joined #pypy.
[19:53] stakkars (~tismer@rosine27.inf.fu-berlin.de) joined #pypy.
[19:55] <stakkars> did a little bit, but want to do quite much more.
[19:56] <arigo> stakkars: I just checked in fixes for Windows ME in pypy and in the py lib
[19:57] <stakkars> test_all ?
[19:57] <arigo> stakkars: just a minute...
[19:57] <arigo> stakkars: svn up again
[19:58] <arigo> stakkars: can you see if it solves your Windows problems too?
[19:58] <arigo> fixed interpreter/test/test_py.py, module/__builtin__/test/test_import.py
[19:58] <stakkars> aye aye
[19:58] <stakkars> testing now...
[19:58] <arigo> plus some Windows-ME-specific issues that I got.
[19:59] <arigo> thanks
[20:00] <stakkars> what issues? I sawjust disabling of something for win32
[20:00] <arigo> there is no SYSTEMROOT env var, for example
[20:01] <arigo> (that's a py lib fix, so you might not see it in pypy-svn)
[20:01] <stakkars> arghh, winnows
[20:01] <arigo> also COMMAND.COM and CMD.EXE use different quotings, the details of which I ignore...
[20:02] <stakkars> yes, that's a lot of fun.
[20:02] <stakkars> test_import still fails
[20:03] <arigo> which test and how?
[20:03] <stakkars> test_PYTHONPATH_takes_precedence
[20:03] <stakkars> cmdexec, it simply doesn't give returncodes.Similar to something we already disabled
[20:03] <arigo> uh? should be skipped now
[20:04] <arigo> aaaargh.
[20:04] <arigo> sorry, delete the last two lines
[20:04] <arigo> test_PYTHONPATH_takes_precedence should work now, becasuse it's not about exit codes
[20:05] <arigo> you svn-up'ed the py lib too?
[20:05] <stakkars> the whole dist
[20:05] <stakkars> output = py.process.cmdexec('''"%s" "%s" -c "import urllib"'''
[20:05] <stakkars> (sys.executable, pypypath) )
[20:05] <stakkars> assert output.strip() == '42'
[20:05] <stakkars> assert '' == '42'
[20:05] <stakkars> + where '' = ''.strip()
[20:06] <stakkars> line 165
[20:06] <arigo> what if you go to
[20:06] <arigo> py\process\cmdexec.py
[20:06] <arigo> lines 89-99
[20:06] <arigo> the whole block about cmd_quoting
[20:06] <arigo> and trash it
[20:06] <hpk> stakkars: the actual docstring of getinterplevel.py should be copied to the motivation section in the documentation
[20:07] <hpk> it's a lot cleaer than talking about "translating python to python"
[20:07] <hpk> clearer
[20:07] <stakkars> if you like it, I'll do it.
[20:08] <hpk> yes, but of course, i don't mean copying over the whole docstring verbatim, just the first bits
[20:09] <hpk> stakkars: notet hat samuele just fixed some typos -> svn up before you change
[20:10] <stakkars> arigo: if I trash it, I crash with a syntaxerror from cmd
[20:10] <pedronis> they were not really typos, there were some strange chars in the file for " '
[20:10] <hpk> whatever
[20:10] <stakkars> yeah. Because some of you are using veeeerylong lines and I had no tool, I started editing with Word which supports this,
[20:11] <stakkars> but it added some funny chars,so I reverted to ultraedit.
[20:11] <hpk> actually, i converted some stuff this morning to use hard CRs
[20:11] <arigo> stakkars: any space in your path to dist?
[20:12] <arigo> (not that it is forbidden, just trying to figure out what's going on)
[20:13] <stakkars> no
[20:14] <stakkars> I remember,that cmd quoting is from me, it is needed for winXP
[20:15] <stakkars> hpk: what exactlydo I use from the doctring, thesecond para? And where to put it, do you want me to drop something for it?
[20:15] <hpk> i can do it and you can cross-check/remodify? it's easier than explaining it here
[20:16] <stakkars> go ahead, but please not much
[20:17] <arigo> stakkars: can you (for fun) mail me the output of 'HELP CMD' ?
[20:17] <hpk> sidenote: you are using some terms that are not used throughout the rest of the documentation
[20:17] <hpk> stakkars: like "app-space"
[20:19] cfbolz (~cfbolz@hdlb-d9b946ac.pool.mediaWays.net) joined #pypy.
[20:20] <cfbolz> back (although I will be shifting in and out again).
[20:20] <cfbolz> Is the number of lines in architecture still correct?
[20:21] Gromit (~bear@B74d4.b.pppool.de) left irc: Read error: 110 (Connection timed out)
[20:23] <arigo> cfbolz: I think they are still a rough approximation
[20:23] <cfbolz> ok
[20:24] <stakkars> hpk: this wordis not in the file.
[20:24] <hpk> stakkars: second sentence in geninterplevel.py in the docstring
[20:25] <stakkars> gack! what should it be?
[20:26] <stakkars> application level?
[20:27] <hpk> yes, i am committing solely changes to motivation (and a link to the appropriate architecture section)
[20:28] <stakkars> arigo:mailsent
[20:30] <hpk> stakkars: commited the motivation refactor, hope you like it
[20:30] <cfbolz> stakkars: I think it's very good that all sections in translation.txt use the same example
[20:31] Nick change: lac -> lac-out4dinner
[20:32] <lac-out4dinner> take care all, back later. I can finally proofread something if you like, but looks like you are doing just great without me.
[20:33] <cfbolz> BTW: Is there something *I* should proofread? I'm feeling kind of useless right now.
[20:33] <stakkars> hpk: yes it is quite nice,although I don't believe in what I said in the
[20:33] <stakkars> docstring longer time ago:
[20:34] <stakkars> """RPython to C translation hopefully can produce more efficient code"""
[20:34] <stakkars> is nonsense.
[20:35] <stakkars> this could be done on-the-fly by flowing stuff and then interpret theflowgraph,forinstance.
[20:35] <stakkars> In a sense, I believe that this backend will not live very long.
[20:36] <hpk> i think it would be nice to rename the chapter to "The Python interplevel backend", btw
[20:36] <hpk> because we usually have the name than one can find in the source in the section header
[20:36] <stakkars> exactly that did I not do, intentionally.
[20:36] <hpk> s/than/that/
[20:36] <hpk> huh?
[20:37] <stakkars> the name is ugly
[20:37] <hpk> well, this is about consistency and making it easy for people for finding the docs :-)
[20:37] <stakkars> The C backend is a wrong name then, too
[20:38] <hpk> no, because no section mentions the 'Gen' prefix
[20:38] <stakkars> aha
[20:38] <stakkars> I'd reallylike tospend more time on missing content :-)
[20:39] <lac-out4dinner> is what we call 'backend' what IBM calls 'Virtual Machine'?
[20:39] <stakkars> no
[20:39] <lac-out4dinner> didn't think so. thanks.
[20:40] <stakkars> it can be a virtual machine. Since I create Python, my output is for a virtualmachine.
[20:40] <stakkars> The C output is for a realmachine. I think it's just not really the point.
[20:40] <stakkars> It is in a sense the atrget machine, whatever that is.
[20:40] <stakkars> target
[20:41] <hpk> cfbolz: i think there is not much to proofread anymore
[20:41] <cfbolz> anything else to do?
[20:41] <stakkars> hpk: is it hard to colorize the python code, and more important, how can I set links
[20:41] <stakkars> to the mebedded code fragment?
[20:42] <hpk> colorization: i don't know
[20:42] <hpk> links: what exactly do you mean?
[20:42] <hpk> you mean like pointing to somewhere inside the code?
[20:43] <hpk> cfbolz: do you feel like writing a news item for the front page?
[20:43] <stakkars> well, yes, linking to something, or some markup,,,
[20:43] <arigo> stakkars: does interpreter\test\test_py.py work for you?
[20:43] <stakkars> the problem is that I want to comment on code, and it would be nicest if the code was in a box
[20:44] <hpk> stakkars: i don't think you can easily link to inside a "verbatim"-block
[20:44] <stakkars> and I can put side notes. Maybe too much for REsT
[20:44] <hpk> stakkars: look at the ReST docu, it's quite extensive
[20:44] <cfbolz> hpk: not really. Well, I could start.
[20:44] <hpk> and tell me what you found :-)
[20:44] <stakkars> commenting on something without interspersing it is not nice for the reader and the writer
[20:44] <aleale> arigo: test_py works for me
[20:44] <stakkars> hey, a table might be possible!
[20:45] <stakkars> but tables cannot contain verbatim, can they?
[20:45] <arigo> aleale: and not module\__builtin__\test_import.test_PYTHONPATH_takes_precedence ?
[20:45] <hpk> you can always do what i did for the thunk example:
[20:45] <hpk> http://codespeak.net/pypy/index.cgi?getting-started#lazily-computed-objects
[20:45] <hpk> stakkars: they probably can
[20:46] <hpk> stakkars: but let's please discuss formatting problems some other time
[20:46] <stakkars> what is special about this? Little chunklets with lines between?
[20:47] <aleale> arigo:no not test_import.<test_PYTH...>
[20:47] Action: stakkars feels like "here is the answer, but don't comment on it"
[20:48] cfbolz (~cfbolz@hdlb-d9b946ac.pool.mediaWays.net) left irc: "Verlassend"
[20:49] <hpk> stakkars: it's just that i am trying hard since yesterday to move towards getting the release out and i am exhausted
[20:49] <hpk> and i don't know exactly and would have to test putting verbatim stuff in a table
[20:49] <hpk> i have an idea but would have to look it up in the ReST doku
[20:49] <arigo> aleale: can you go to test_import.py line 163
[20:49] <arigo> and add a space near the end of the line, between the final " and the ''' ?
[20:50] <stakkars> hpk: no no, it is just fine. Will merge my explainments into the output code.
[20:51] <aleale> should I leave cmdexec alone?
[20:51] <arigo> yes, just a single extra space character
[20:51] <arigo> output = py.process.cmdexec('''"%s" "%s" -c "import urllib" ''' %
[20:52] <hpk> arigo: btw, i guess for win32 cmdexec() could just reuse commands.getstatusoutput and raise on status!=0 ?
[20:53] <arigo> "commands" is Unix-only
[20:53] <hpk> ah great fun
[20:53] <arigo> sure
[20:53] <aleale> it stilll fails - no output recorded
[20:53] <hpk> sounds familiar actually, i probably thought of this before
[20:55] <aleale> os.popen3 is blocking, right? not like subprocess.Popen ?
[20:55] <aleale> I used subprocess to make lib-python/conftest.py work on win2k
[20:56] <arigo> aleale: we can't figure out why test_py works and not test_import, because they use exactly the same cmdexec()
[20:56] <arigo> aleale: subprocess is a Python 2.4-only module, btw
[20:56] <stakkars> arigo:what about the arguments?
[20:57] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: "Chatzilla 0.9.68.3 [Firefox 1.0.4/20050516]"
[20:57] <stakkars> we can use the timemachine and use this module, put it somewhere :-)
[20:57] <aleale> arigo: I know but works for 2.3 as well - you just have to install it yourself
[20:58] <aleale> its interpreter/test/test_py, right ?
[20:58] <arigo> yes
[20:58] <aleale> trying it again
[20:59] <aleale> There is a tiny s - I have overlooked before - it skips something
[21:01] <hpk> arigo: i am starting to think about releasing, you still hold that freezing dist at this point is not a good idea, i guess
[21:01] <arigo> I'll just skip this test on Windows too
[21:01] <hpk> arigo: (or not a good idea at all, i don't remember the details :-)
[21:02] <arigo> it's really testing something that we know works; the problem is really on how to check that on Windows.
[21:02] <aleale> arigo: reason: Skipped: cannot detect process exit code for now
[21:02] <arigo> hpk: sorry, misunderstood you
[21:02] <arigo> aleale: ok, but the rest of test_py.py passes if I understood correctly
[21:02] <aleale> yes
[21:02] <arigo> hpk: yes I don't see much the point of moving away from dist/ development for now
[21:03] <aleale> arigo: interpreter\test\test_py.py[5] ....s
[21:05] <arigo> skipping the test on windows.
[21:05] <aleale> ok
[21:05] <arigo> skipping the test_import I mean.
[21:08] <hpk> pedronis: have you looked at the LICENSE stuff i wrote?
[21:08] cfbolz (~cfbolz@hdlb-d9b95ec8.pool.mediaWays.net) joined #pypy.
[21:08] <hpk> to be consistent, i am afraid we need a LICENSE file in lib-python
[21:09] Action: hpk somehow guesses that hardly anyone really examined the LICENSE file hard, right?
[21:09] <hpk> no offense meant, just asking
[21:09] <pedronis> hpk: I looked at it, no strong opinions
[21:10] <hpk> hpk: you did actually put a note about lib-python/2.3.4 into the top level LICENSE file
[21:10] <pedronis> I saw that
[21:10] <cfbolz> :-)
[21:10] <pedronis> anyway IANAL
[21:11] <hpk> me neither
[21:11] <pedronis> you want one if you really want hard-examitation of it
[21:11] <hpk> "want"? well
[21:11] <hpk> i have the impression to a certain level it's a kind of neccessity for a project like this
[21:11] <pedronis> well if
[21:12] <pedronis> hpk: yes
[21:12] <hpk> basically i think it should at least look consistent and complete to us
[21:12] <cfbolz> guenter jantzen and ben young are still in the list -- they sent a mail to pypy-dev that they don't want this
[21:12] <hpk> and i know that doesn't imply that lawyers don't neccesarily share that
[21:13] <hpk> cfbolz: yes, i saw that
[21:13] <pedronis> well, I have looked at it, it seems ok, no strong opinions
[21:13] <hpk> basically i think we wouldn't it wrong if we just killed a lot of names from that list
[21:13] <cfbolz> (that was just the thing I found when I examined the LICENSE file)
[21:14] <hpk> yes, i am not really happy with this list of copyright holders
[21:14] <hpk> but i couldn't come up with an easy systematic approach to cut out names
[21:14] <hpk> although i basically think the list could be reduced to 7-8 names
[21:15] <pedronis> well, the problem is at some point we may need contributions agreements from all them
[21:15] <hpk> and the rest be listed as contributors
[21:15] <hpk> pedronis: i guess so, yes
[21:16] <cfbolz> not for 0.6, anyway
[21:18] <hpk> right
[21:18] <hpk> should '0.6' appear anywhere in the sys. attributes?
[21:20] <cfbolz> maybe in the startup message?
[21:21] <hpk> woa, the download is going to be some 4 MBs
[21:21] <cfbolz> that's tiny. Compare this to checking out
[21:21] <aleale> you have been busy for two years :)
[21:22] <cfbolz> do you think this is much or not?
[21:22] <hpk> it's quite a bit
[21:22] <hpk> i didn't think about it so far
[21:22] <hpk> but i didn't think in MBs
[21:22] <hpk> but of course that is lib-python
[21:23] <arigo> hpk: I'm fine with no 0.6 in any .py file at this point
[21:23] <aleale> it includes the std lib and py lib
[21:23] <hpk> arigo: ok
[21:23] <hpk> aleale: sure, but the py lib is 300k or something
[21:23] <hpk> aleale: it's really lib-python
[21:24] <aleale> hpk: true
[21:26] <cfbolz> what's the correct way to link to the release announcement?
[21:26] <cfbolz> http://codespeak.net/pypy/index.cgi?doc/release-0.6.html?
[21:27] <hpk> cfbolz: that's safe, yes
[21:27] <cfbolz> ok. I started the news item.
[21:30] <cfbolz> ah. Should I check in? It will be on the website immediately, right? (and it's not finished)
[21:32] <hpk> hum, put it in a self-contained file in documentation? we can delete it when merging it
[21:33] <aleale> hpk: lib-python is 9 Mb uncompressed on my machine (only py files)
[21:36] <stakkars> how do I enable doctest on a REsT file,using test_all?
[21:37] <cfbolz> hpk: ok, checked in as documentation/news-0.6.txt -- not very good, I fear.
[21:37] <hpk> stakkars: that's basically disabled
[21:37] <hpk> see https://codespeak.net/issue/pypy-dev/issue4
[21:38] <arigo> can I check in a typo in translation.txt?
[21:38] <hpk> sure
[21:38] <cfbolz> stakkars: I can check your examples by hand, if you want.
[21:39] <stakkars> cfbolz: thanks, that's not the point. I wanted to ensure that the system complains if we change interface but not the doc.
[21:39] <cfbolz> I'm SO stupid! I changed the web page, too.
[21:39] <hpk> even with a ReST error :-)
[21:39] <cfbolz> sorry, I'm getting really too tired.
[21:40] <stakkars> arigo: go ahead
[21:40] <cfbolz> how do i revert?
[21:41] <stakkars> seemstobe done already
[21:41] <hpk> cfbolz: just kill the section
[21:42] <cfbolz> ok.
[21:42] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[21:43] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: Client Quit
[21:43] <stakkars> arigo: ah, and a bad one, thanks. You got one of these, too, changed it: """``int`` with ``int_add``""" s/int/sub/
[21:43] <stakkars> s/sub/add/ waah
[21:45] <cfbolz> ok. I fixed the errors and reverted the real news.txt
[21:45] <arigo> thanks :-)
[21:47] <hpk> arigo: can you look at codespeak:/tmp/makerelease/make-0.6.sh ?
[21:47] <hpk> and see if that looks about right?
[21:49] Action: arigo looks
[21:50] <arigo> hpk: do we want to produce nice Windows end-of-lines in the zip ?
[21:50] <hpk> arigo: ah, good point, yes
[21:50] <arigo> svn export --eol-style CRLF
[21:50] <arigo> svn export --native-eol CRLF
[21:50] <arigo> sorry
[21:50] <hpk> screen
[21:51] <arigo> i'm there
[21:51] <arigo> TARGET= too late
[21:52] <hpk> arigo: you can edit, you know :-)
[21:52] Action: arigo still marginally afraid of vi :-)
[21:53] <hpk> cfbolz: i streamlined your news-0.6 a bit
[21:53] <cfbolz> thanks.
[21:53] <arigo> hpk: re-enable svn cp at the end?
[21:53] <arigo> btw, svn/pypy/tag is cluttered already
[21:54] <hpk> arigo: i wanted to activate the links in the getting-started doc first
[21:54] <arigo> would it make sense to have svn/pypy/tag/release/0.6 ?
[21:54] <hpk> then have us download the stuff, see if it all works
[21:54] <arigo> hpk: ok
[21:55] <hpk> arigo: personally, i'd prefer svn/pypy/release/0.6
[21:55] <arigo> I'm fine with that
[21:56] <arigo> hpk: samuele's checking in a fix yet...
[21:57] <hpk> after having run the tests i guess
[21:57] <arigo> (we might want a no-change-to-the-code-base policy that extends over several days, for the next time)
[21:57] <arigo> hpk: yes
[21:57] <hpk> arigo: yes
[21:58] <arigo> ok, checked in.
[21:58] <pedronis> done
[21:59] <hpk> i am changing the getting-started to contain the links
[22:01] <arigo> ouap, testresults have gone down a few percents
[22:02] <arigo> test_sets fails in pickle issue :-( :-(
[22:03] <hpk> you want to delay the release? :-)
[22:03] <hpk> anyway, getting-started contains the links now
[22:03] Action: hpk is trying the bz2 one
[22:03] <arigo> I'd like to see if we can at least understand the new errors
[22:04] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[22:04] <hpk> arigo: if you can meanwhile download one of the packages and test it on your machine, that would be good :-)
[22:04] <hpk> stakkars, aleale: would also be good if you did that on win32
[22:04] <arigo> while I'm trying to understand test_sets? hum
[22:05] <hpk> well, ok, i thought we were just now trying to release
[22:05] <cfbolz> sorry, can't really try it, downloading takes a bit long for me
[22:06] <arigo> well, as I said we should at least see if the failure is in any way critical
[22:06] <hpk> sure, go ahead
[22:07] <cfbolz> ok. then my download _has_ the chance to finish ;-)
[22:09] <aleale> trying zip file
[22:10] <arigo> pickle is completely broken :-( :-(
[22:11] <hpk> hum, fix it?
[22:11] <hpk> i mean i can easily take away the links
[22:11] <hpk> then we have all the time in the world (if i don't fall asleep)
[22:12] <arigo> remove them for now, yes
[22:12] <hpk> arigo: shall i? it's np
[22:12] <arigo> please... :-(
[22:12] Action: cfbolz starts to snore
[22:12] <arigo> class X: pass; pickle.dumps(X()) ----> crash
[22:12] <arigo> help welcome :-)
[22:12] <stakkars> what's up?
[22:13] <ericvrp> I think there is a problem with printing the documentation from Firefox on Windows. It thinks there is just one page. IE prints all pages! Firefox does print all apges on other sites.
[22:13] <stakkars> pickle is broekn, how?
[22:13] Action: hpk would like all other commits to stop except for fixing pickle
[22:14] <arigo> yes and yes
[22:15] <hpk> ericvrp: hua, strange
[22:15] <hpk> ericvrp: do you have any idea why? (not that i think it would be easily fixable)
[22:16] <cfbolz> hpk: ouch, got the same problem under linux with firefox
[22:17] <cfbolz> sorry, there are even more problems :-(
[22:17] <cfbolz> if get
[22:17] <cfbolz> from pypy.tool import pytestsupport
[22:17] <cfbolz> ImportError: cannot import name pytestsupport
[22:18] <cfbolz> when running python test_all.py
[22:18] <ericvrp> hpk: I will run Firefox DeveloperTools extension to see if the pagelayout is somehow broken
[22:19] <stakkars> hpk: when testing test_pickle, I get
[22:19] <stakkars> E self.kind = d['nodekind'] == 'directory' and 'dir' or d['nodekind']
[22:19] <stakkars> > KeyError: 'nodekind'
[22:19] <stakkars> [D:\pypy\dist\py\path\svn\wccommand.py:465]
[22:19] <stakkars> [testcode : D:\pypy\dist\lib-python\2.3.4\test\test_pickle.py:2147483648]
[22:19] <hpk> ericvrp: sounds good, thanks
[22:19] <hpk> stakkars: what are you invoking in which directory?
[22:20] <hpk> cfbolz: what are you invoking in which directory?
[22:20] <cfbolz> python test_all.py in pypy-0.6/pypy
[22:21] <hpk> cfbolz: very odd
[22:21] <stakkars> arigo: you mean the _reconstructor thingie?
[22:21] <arigo> stakkars: look at _cache/objecttype_xxxx.py
[22:21] <arigo> yes
[22:22] <arigo> we are bringing parts of copy_reg into _cache/objecttype_xxxx.py :-( :-(
[22:22] <stakkars> har har now I remember *one* reson why we wanted to switch imports on and off. uuhuu
[22:22] <hpk> cfbolz: where did the import error happen?
[22:22] <hpk> cfbolz: i mean which file/lineno
[22:22] <aleale> It works for me on win2k (from the zip file)
[22:22] <aleale> tests finished: 1330 passed, 88 skipped in 438.31 seconds
[22:22] <stakkars> sure, copy_reg must go into thespecial cases!
[22:23] <arigo> no
[22:23] <stakkars> nope
[22:23] <arigo> that's insane
[22:23] <arigo> :-/
[22:23] <stakkars> makes not even sense
[22:23] <cfbolz> stakkars: argh! there is a defect conftest.py hanging around in some parent directory
[22:23] <cfbolz> I mean hpk:
[22:23] <cfbolz> hpk: My mistake
[22:24] <hpk> cfbolz: ah, that makes sense :-)
[22:24] <hpk> pytestsupport is long gone
[22:24] <cfbolz> no it looks good.
[22:24] <cfbolz> even the llvm-tests pass
[22:25] <hpk> ok, that makes pickle and the web-page issue the problems
[22:25] <cfbolz> why don't we have tests for that?
[22:26] <cfbolz> something else:
[22:26] <cfbolz> Skipped in /home/cfbolz/pypy-0.6/pypy/lib/test2/test_codeccallbacks.py:2
[22:26] <cfbolz> reason: Skipped: this test module doesn't belong here
[22:26] <cfbolz> huh?
[22:26] <hpk> cfbolz: tests for pickle you mean?
[22:27] <cfbolz> yes
[22:27] <hpk> we have only compliance tests
[22:27] <hpk> and arigo noticed it failing, therefore he came to it
[22:27] <hpk> from the test result page
[22:28] <cfbolz> ok.
[22:28] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: "Chatzilla 0.9.68.3 [Firefox 1.0.4/20050511]"
[22:28] <stakkars> arigo: sure. This piece of code is no longer RPython, because of the eraly imports.
[22:28] ericvrp (~chatzilla@ericvrp.demon.nl) joined #pypy.
[22:29] <hpk> ericvrp, cfbolz : what i find odd, is that the web pages validate as xhtml1.0 transitional
[22:29] <stakkars> I think to either mark it as NOT_RPYTHON,or we need a registry which modules are considered objects
[22:30] <hpk> ericvrp, cfbolz: but i can imagine that the CSS somehow causes problems
[22:30] <cfbolz> hpk: I'm testing other browser at the moment
[22:30] <ericvrp> hpk: I am looking at /pypy/default.css right now, but I'm no CSS expert
[22:31] <hpk> ericvrp: if you simply copy the html-source locally, scrap the CSS and then try printing?
[22:31] <ericvrp> hpk: My printer jammed. :( Will try it...
[22:32] <stakkars> arigo: pickle is fixed
[22:32] <cfbolz> you can do print preview
[22:33] <ericvrp> hpk: <table border class="table"> .... border seems out of place here
[22:33] <stakkars> checked in
[22:34] <hpk> ericvrp: which file exactly? coding-guide?
[22:34] <cfbolz> getting_started
[22:34] <ericvrp> doc
[22:34] <ericvrp> all
[22:34] <cfbolz> :-)
[22:35] <hpk> hum, that is generated by ReST :-(
[22:35] <hpk> but anyway, i think that without a CSS it does print (on my mozilla/linux)
[22:35] <arigo> stakkars: you're not really solving the problem globally.
[22:35] <arigo> stakkars: the same problem occurs at other places, I fear
[22:36] <arigo> stakkars: we can as well disable geninterp for the release :-(
[22:36] <arigo> stakkars: (not suggesting we actually do, mind)
[22:36] <stakkars> arigo: why?
[22:36] <cfbolz> hpk: yes, without the css it works
[22:36] <arigo> stakkars: did you review all imports in all applevel helpers?
[22:36] <hpk> cfbolz: so that gets tough
[22:37] <hpk> cfbolz: i am tempted to try to walk through the CSS-validation problems ...
[22:37] <stakkars> arigo: no but that's a quick task. I had them all in my fingers, of course,but notfor that
[22:37] <cfbolz> yes, but it doesn't really affect the release: the css file is not included
[22:37] <arigo> stakkars: e.g. stringobject.py, import codecs
[22:37] <arigo> stakkars: but don't do that please
[22:38] <hpk> cfbolz: sure not, it's just likely that people will want to print out stuff
[22:38] <stakkars> arigo: I'm checking right now.
[22:38] <arigo> stakkars: please don't
[22:38] <stakkars> arigo: and copy_reg is a special case, because it strangely depends on what is where
[22:39] <stakkars> arigo: alternatively, I could quickly go back to the distinction we made?
[22:39] <arigo> stakkars: please hold on
[22:39] <cfbolz> hpk: ok
[22:40] <hpk> cfbolz: i am working on the assumption that with a valid CSS it will print nicely
[22:40] <cfbolz> hpk: well, the problem seems pretty firefox-centric: opera and konqueror work fine
[22:40] <hpk> konqi as well
[22:40] <hpk> firefox sucks anyway
[22:41] <hpk> IE is much better
[22:41] <stakkars> arigo: I did not want to produce a problem, but heard you saying you don't like so different modes. Today we learned that it is necessary and should probably be enabled for genc level, too? I think it can do it now
[22:41] <cfbolz> hpk: links, lynx and dillo, too (but they don't support css anyway :-) )
[22:41] <cfbolz> what does the css validator say?
[22:43] <hpk> cfbolz: it's mostly valid and now i have to guess what the unit for padding/margin is likely to be
[22:43] <ericvrp> hpk, cfbolz: # Line: 144 Context : div.project_title
[22:43] <hpk> ericvrp: already fixed :-)
[22:44] <hpk> ericvrp: i copied the failing CSS definitions from somehwere and now the question is if they are EMs, px's or what
[22:44] <ericvrp> hpk: what that title thing. or the printing issue?
[22:44] <hpk> ericvrp: title thing
[22:45] <hpk> cfbolz: if you feel like finding that: one can define a different CSS for printing
[22:45] <hpk> cfbolz: but i don'T know it by hard, that would shortcut the problem
[22:45] <hpk> likely
[22:45] <cfbolz> i removed the css and printing worked, so it is certainly a css issue
[22:46] <hpk> cfbolz: yip, same here
[22:47] <stakkars> arigo:no,bookkeeper is unhappy if I want to import
[22:48] <hpk> cfbolz, ericvrp: i think i fixed the problem on the live pages
[22:49] <ericvrp> cool
[22:49] <aleale> yes it seems to work here. What did you do?
[22:49] <hpk> i won't tell, you gotta keep some business secrets
[22:50] <hpk> media="screen"
[22:50] <hpk> so the css does not get evaluated for media="print"
[22:51] <aleale> cool
[22:51] <ericvrp> hpk: verified here aswell. Great stuff!
[22:51] <hpk> the print-pages look slightly odd though, but i don't exactly care at the moment
[22:51] <aleale> nice trick
[22:52] fredrik (fredrik@c83-248-135-181.bredband.comhem.se) left irc: "http://fredrikj.net";
[22:52] <hpk> wasn't test_cpickle one of those really-long-running tests?
[22:53] <cfbolz> works for me too
[22:53] <hpk> like >60 minutes?
[22:54] <ericvrp> hpk: do you get paid by the minute ;)
[22:54] <hpk> :-)
[22:54] <hpk> arigo, stakkars: is anything concluded?
[22:54] <cfbolz> no, but the release gets really delayed then
[22:54] <arigo> hpk: about to test...
[22:55] <hpk> cfbolz: not really, if the simple example armin posted earlier works, then we can release
[22:55] <cfbolz> ok
[22:55] <ericvrp> success with the release. I have to catch some sleep, goodnight
[22:55] <cfbolz> I don't really want to wait much longer.
[22:55] <hpk> ericvrp: thanks and even more so for helping!
[22:56] <hpk> cfbolz: sure, don'T worry, i am past the point of falling asleep
[22:56] <hpk> and there is nothing left except the pickling issue
[22:57] <arigo> stakkars: sorry about hurrying up a bit here
[22:59] ericvrp (~chatzilla@ericvrp.demon.nl) left irc: "Chatzilla 0.9.68.3 [Firefox 1.0.4/20050511]"
[22:59] <hpk> so it appears we are getting there?
[23:00] <cfbolz> hpk: No, I want to stand this through now (is this germanism?)
[23:00] <arigo> hpk: yes, running some last tests...
[23:00] <hpk> i guess that you still want to run some translate targets
[23:00] <hpk> cfbolz: isn't germanism a germanism itself? i mean at application level?
[23:01] <arigo> hpk: yes
[23:01] <stakkars> arigo: how do we fix it no is the question. Later we will have the question what is correct for the future, well.
[23:01] <stakkars> s/no/now/
[23:01] <cfbolz> hpk: no idea. there are many, so a word for it is useful :-)
[23:02] <hpk> cfbolz: are you noticing how much i am trying to avoid any smileys lately? times are tough, anyway.
[23:03] <cfbolz> hpk: is this part of the game or is your mood that bad?
[23:04] <arigo> stakkars, hpk: fixed
[23:04] <hpk> cfbolz: it's as fine as it can get after this week
[23:04] <hpk> arigo: my playing with pickle seems ok too
[23:05] <cfbolz> hpk: which means quite bad
[23:05] <hpk> cfbolz: nah
[23:05] <hpk> does anybody mind if i quickly remove README.22?
[23:05] <arigo> stakkars: sorry again, for doing that in a hurry -- it's just easier to speak and work out a solution face-to-face...
[23:05] <arigo> hpk: not me
[23:06] <stakkars> you reverted to having the distinction, right?
[23:06] <cfbolz> hpk: I would remove it
[23:06] <arigo> stakkars: yes
[23:07] <hpk> cfbolz: done, and sorry for being unclear, i am in a somewhat funny mood, let's put it this way
[23:07] <arigo> stakkars: that was the only quick solution if we didn't want to drop lots of NOT_RPYTON here and there :-/
[23:07] <arigo> stakkars: the rule now is basically that imports inside app-level helper functions are always lazy
[23:07] <arigo> stakkars: and we moved "import _formatting" out of the function, to make it eager.
[23:08] <stakkars> arigo: but ah I see. Hmm
[23:08] <arigo> indeed, Hmm about the whole subject around here...
[23:09] <cfbolz> hpk: I'm feeling a bit strange myself (and I really don't want to think about the stuff I still have to learn till Tuesday)
[23:10] <stakkars> arigo: I have some code ready that allows the same behavior for genc, too. We can discuss it later
[23:10] <arigo> ok
[23:11] <arigo> hpk: we are release-ready as far as I can see
[23:11] <hpk> ok, i am re-packaging then after committing the links (which i just tested for bz2 at least)
[23:13] <arigo> hold on...?
[23:13] <hpk> ?
[23:14] <stakkars> arigo: just tested: I can import in genc, too.
[23:14] <arigo> no, sorry, ok (after I removed my _cache)
[23:14] <arigo> stakkars: nice
[23:15] <stakkars> added __import__ to builtin, did somespecialcasing to *not* pass the globals dict on,and fine.
[23:15] <stakkars> so the specialcasing is needed to avoid picklingofglobals, but it can be unconditional if we want
[23:16] <stakkars> I guess I should not check that in right now?
[23:16] <arigo> guess not :-)
[23:16] <hpk> let me reinforce the guess
[23:17] <arigo> champagne !
[23:17] <hpk> cheers!
[23:17] <hpk> http://codespeak.net/svn/pypy/release/
[23:17] <cfbolz> announcements and news items?
[23:17] <stakkars> klingt wie ne Schiffstaufe
[23:17] <aleale> can I download it now
[23:18] <hpk> but wait, bin/py.py crashes
[23:18] <hpk> cfbolz: you see where missing smileys can lead to?
[23:19] <stakkars> and people will not like browsing things like translation.txt, and they get 500 byte long lineswhich don't wrap
[23:19] <cfbolz> hpk: I would have nearly drawn a gun :-
[23:19] <cfbolz> )
[23:20] Action: arigo reads how to kick hpk out of the channel
[23:20] Action: hpk added the news item
[23:21] <cfbolz> ban is really so much more effective
[23:21] <cfbolz> where do we send the mails?
[23:21] #pypy: mode change '+o arigo' by ChanServ!ChanServ@services.
[23:21] hpk kicked from #pypy by arigo: arigo
[23:21] <arigo> works :-)
[23:21] <stakkars> yieks
[23:22] the_monster (~hpk@merlinux.de) joined #pypy.
[23:22] #pypy: mode change '-o arigo' by ChanServ!ChanServ@services.
[23:22] <the_monster> i am back and even worse than ever!
[23:22] <arigo> :-)
[23:24] <the_monster> i guess i should disable svn access for everyone to avoid svn/pypy/dist getting broken in a couple of minutes
[23:24] Action: the_monster looks into the apache docs
[23:24] <cfbolz> is someone sending announcement mails?
[23:25] <the_monster> not yet
[23:25] <the_monster> i guess
[23:26] <the_monster> now we get to roll dices who does it
[23:26] <cfbolz> is it an honor or a punishment?
[23:26] <aleale> Congratulations everybody. I think it is great work - thanks
[23:26] <the_monster> cfbolz: mixed honour/punishment level
[23:27] <aleale> Goodnight and good weekend
[23:27] aleale (~redorlik@cpe.atm0-0-0-129140.0x3ef2fa3a.bynxx3.customer.tele.dk) left irc: "BitchX sucks"
[23:27] <arigo> anyone got a crypto-safe internet-based dice rolling program?
[23:28] <the_monster> so whom do we mail apart from c.l.py?
[23:28] <cfbolz> pypy-dev?
[23:28] <cfbolz> python-announce?
[23:28] <cfbolz> LLVM-dev?
[23:28] <the_monster> i guess python-announce and c.l.py are enough
[23:28] <cfbolz> (they probably are interested)
[23:28] <the_monster> llvm-dev sounds good
[23:29] <the_monster> you should send that mail
[23:29] <arigo> I guess pypy-dev is very interested too
[23:29] <the_monster> cfbolz: i guess you can even give the release announce a personal note at the beginning if you like
[23:30] <the_monster> cfbolz: like "i have been working on a LLvm backend for pypy, so i thought i share ..." whatever
[23:30] <cfbolz> yes, I planned that, too
[23:30] <stakkars> arigo: always late importing passes all tests. (had to modify one in flow/test/test_objspace)
[23:30] <the_monster> arigo: yes, so you mail to python-dev and i to c.l.py?
[23:31] <arigo> I'm fine with that
[23:32] <pedronis> stakkars: is getting late to have that kind of discussion
[23:32] <stakkars> hpk, the_monster: I like you new nickname. Will be hard to get rid of it, again.
[23:32] <stakkars> forgot the :-)
[23:32] <the_monster> it's all arigo's fault anyway
[23:32] <pedronis> and eager importing is typically what we want with genc
[23:34] <stakkars> pedronis: ok, anyway it is possible to switch behavior, I'll leave it in as option.
[23:34] <arigo> stakkars: makes sense.
[23:34] <stakkars> until we really know better
[23:35] Action: arigo sends release-0.6.txt to python-dev and pypy-dev. (ok?)
[23:36] <the_monster> arigo: ok
[23:36] <the_monster> arigo: wait
[23:36] <the_monster> arigo: i am the monster so i can't be ok with that, obviously especially since you turned me into one
[23:36] <cfbolz> just sent the one LLVM-dev
[23:36] <cfbolz> to
[23:37] Action: the_monster sent it to c.l.py and if the mailman gods don't notice it will go through
[23:38] <cfbolz> it allready did
[23:38] <arigo> er, what about python-announce?
[23:38] <the_monster> i am about to send it
[23:38] <arigo> ok
[23:38] <stakkars> what about checkin in, is that ok again? Or do we need tocopy stuff?
[23:38] <arigo> the_monster: how do you read an external file from vi?
[23:38] <the_monster> are
[23:39] Last message repeated 1 time(s).
[23:39] <the_monster> damn xchat
[23:39] <the_monster> 'r'
[23:39] <stakkars> poor monster
[23:39] Action: pedronis logging out, g'night and congrats
[23:39] pedronis (pedronis@ratthing-b246.strakt.com) left irc: "Chatzilla 0.9.67 [Firefox 1.0.2/20050325]"
[23:39] <arigo> 'r' doesn't work?
[23:40] <arigo> :r filename?
[23:40] <the_monster> yes
[23:42] Action: stakkars would like to check something in and go home.possible?
[23:43] <cfbolz> the_monster: you made new packages? I'm still getting an error (but I could also have used the old archive. asking here is faster than downloading again)
[23:43] <the_monster> cfbolz: error for what?
[23:44] <cfbolz> pickle
[23:44] <stakkars> We invite you to head over to to the
[23:45] <the_monster> cfbolz: you mean arigo's example of test_pickle?
[23:45] <arigo> stakkars: yes, check in, at this point :-)
[23:46] <cfbolz> the_monster: yes
[23:46] <stakkars> what exactlymust I do to correct the news page?
[23:46] <cfbolz> I don't want to cause fuss, just answer my question :-)
[23:46] <the_monster> stakkars: i am doing it already
[23:46] <arigo> cfbolz: hpk did new packages, yes
[23:46] <cfbolz> It's very possible that i used the old archive
[23:47] <stakkars> that's no answer because next time I want to do it :-)
[23:47] <arigo> cfbolz: the release != the one you downloaded, as far as I can tell
[23:47] <cfbolz> ok.
[23:47] <arigo> good night and cheers.
[23:47] arigo (~arigo@ratthing-b409.strakt.com) left irc: "[x]chat"
[23:47] <the_monster> stakkars: well, modify website/news.txt
[23:47] <the_monster> and checkin
[23:47] <the_monster> stakkars: i'd like to know at some point if you autonosying-in on issue really means that you are reading everything, btw
[23:47] <stakkars> sorry, thanks
[23:48] <stakkars> the_monster: no, I have to admit that I didn't read everything, but I promise to improve
[23:49] <the_monster> cfbolz: thanks for your great help for this release btw!
[23:49] <cfbolz> you're most welcome
[23:50] Action: the_monster feels like it can morph back now that arigo is gone
[23:50] Nick change: the_monster -> hpk
[23:50] <cfbolz> hpk is so much easier to type
[23:50] <hpk> you don't tell me you don't have completion?
[23:51] <cfbolz> probably. Don't know how to use it
[23:51] <cfbolz> ok, tab works :-)
[23:51] <hpk> anything that doesn't complete on <tab> sucks
[23:51] <cfbolz> who would have guessed
[23:51] <cfbolz> :-)
[23:52] <hpk> :-)
[23:52] <stakkars> hpk: what dod you say I forgot toread,something specific?
[23:52] <cfbolz> yay, a smiley!
[23:52] <hpk> no, i just had the impression that you had asked a couple of questions that i described in issues already, nevermind
[23:53] <cfbolz> he probably meant the part where he explained how to change news
[23:53] <stakkars> sorry about that. I have mailer problems and don'tpoll the website all the time.
[23:54] <hpk> i also had mailer problems in the form of overflows (i guess we easily had >100 mails on issues+svn alone each day)
[23:54] <cfbolz> not that much
[23:55] <cfbolz> hpk: re helping: I guess I identify with PyPy pretty strongly :-)
[23:55] <hpk> cfbolz: you are quite likely not the only one :-)
[23:56] <cfbolz> right :-)
[00:00] --- Sat May 21 2005