IRC Logs

2008 9
Mo Tu We Th Fr Sa So
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          

13. 09 2008

[00:11:02] * aconbere has joined #pocoo
[00:30:16] * stifal has quit IRC
[00:41:57] * aconbere has quit IRC
[00:55:18] * aconbere has joined #pocoo
[01:01:30] * davidcramer has quit IRC
[01:01:52] * davidcramer has joined #pocoo
[01:03:13] * davidcramer has joined #pocoo
[01:08:44] * izibi has quit IRC
[01:21:19] * leche has quit IRC
[01:26:42] * Tik-Tok has quit IRC
[01:40:18] * pete__c has joined #pocoo
[01:45:17] * pete__c_ has quit IRC
[02:15:00] * izibi has joined #pocoo
[03:19:01] * M3ntor5 has quit IRC
[03:58:50] * izibi_ has joined #pocoo
[04:15:09] * Tik-Tok has joined #pocoo
[04:15:16] * izibi has quit IRC
[06:29:23] * Tik-Tok has left #pocoo
[07:07:43] * izibi_ has quit IRC
[07:08:04] * izibi has joined #pocoo
[07:15:18] * izibi has quit IRC
[08:47:07] * EnTeQuAk has joined #pocoo
[09:02:58] * EnTeQuAk has quit IRC
[09:11:51] * leifkb_ has joined #pocoo
[09:27:30] * leifkb has quit IRC
[09:31:31] * luks has quit IRC
[09:31:57] * luks has joined #pocoo
[09:46:27] * leifkb has joined #pocoo
[09:53:02] * leifkb has quit IRC
[09:57:32] * leifkb_ has quit IRC
[09:58:41] * leifkb has joined #pocoo
[10:05:49] * leifkb has quit IRC
[10:10:54] * leifkb has joined #pocoo
[10:23:08] * leifkb has quit IRC
[10:23:39] * leifkb has joined #pocoo
[10:49:41] * leifkb has quit IRC
[10:50:22] * leifkb has joined #pocoo
[11:03:34] * leifkb has quit IRC
[11:07:38] * leifkb has joined #pocoo
[11:11:20] * stifal has joined #pocoo
[11:20:54] * leifkb has quit IRC
[11:20:59] * leifkb has joined #pocoo
[11:30:46] * aconbere has quit IRC
[11:34:15] * leifkb has quit IRC
[11:34:36] * leifkb has joined #pocoo
[11:45:55] <CIA-49> python-doctools: georg.brandl * r66443 /doctools/trunk/tests/root/autodoc.txt: Add missing file.
[11:47:05] <CIA-49> python-doctools: georg.brandl * r66444 /doctools/trunk/ (5 files in 4 dirs): Add test for coverage builder and fix an AttributeError in it.
[11:47:22] * leifkb has quit IRC
[11:47:55] * leifkb has joined #pocoo
[12:13:23] * leifkb has quit IRC
[12:18:27] * leifkb has joined #pocoo
[12:28:39] * leifkb has quit IRC
[12:33:47] * leifkb has joined #pocoo
[12:42:58] * leifkb has quit IRC
[12:44:53] * stifal has quit IRC
[12:48:06] * leifkb has joined #pocoo
[12:56:16] * leifkb has quit IRC
[12:58:15] * leifkb has joined #pocoo
[13:18:27] * stifal has joined #pocoo
[13:22:26] * davidcramer has left #pocoo
[14:00:35] * pete__c has quit IRC
[14:00:40] * pete__c has joined #pocoo
[14:10:07] * leche has joined #pocoo
[14:22:37] * leche has quit IRC
[14:24:19] * leche has joined #pocoo
[14:25:11] * stifal has quit IRC
[15:17:49] * mikeyumatov has joined #pocoo
[16:24:35] * vmx has joined #pocoo
[16:45:05] * EnTeQuAk has joined #pocoo
[18:16:56] * delroth has quit IRC
[18:17:05] * delroth has joined #pocoo
[18:43:09] * leche has quit IRC
[18:53:52] * M3ntor5 has joined #pocoo
[19:03:39] * stifal has joined #pocoo
[19:12:55] <Leonidas> mitsuhiko: you're right, bitbucket is nice. There are some rough edges yet, but still. Maybe I can get jespern to rename my account and to make it case-insensitive.
[19:13:26] <mitsuhiko> also the quotas are extremely generous
[19:16:45] <Leonidas> 100MB? Thats the same as github.
[19:17:10] <aa_> except if you know jespern eh :)
[19:17:21] <steg> ali!
[19:17:24] <Leonidas> heh :)
[19:17:25] <aa_> steg: !
[19:17:37] <aa_> steg: I was chatting to a guy for half an hour about glashammer
[19:17:40] <aa_> steg: (at pycon)
[19:17:41] <steg> nice
[19:17:46] <steg> i thought of plugging it in #pylons again
[19:17:48] <aa_> steg: he is a Pylons user
[19:17:51] <steg> but then i decided it would be shameless
[19:17:56] <Leonidas> I should move some of my projects to bitbucket, the issue tracker looks useful.
[19:18:03] <aa_> steg: anyway, I was astounded by his pylons knowledge
[19:18:07] <steg> aa_: who was it?
[19:18:17] <steg> did you exchange IRC nicknames? :P
[19:18:18] <Leonidas> Ben Bangert? ;)
[19:18:20] <aa_> steg: and he seemed to know about wsgi, and the technical level of the conversation rose..
[19:18:30] <aa_> steg: James Gardner :)
[19:18:32] <steg> ahh!
[19:18:39] <aa_> didn't realise at all
[19:18:54] <steg> i don't think i could meet james gardner without asking him about why authkit is so shit
[19:18:55] <aa_> steg: what an awfully friendly and good looking posh chap
[19:19:10] <steg> it'd be the elephant in the room
[19:19:17] <aa_> steg: he is working on stuff that he hopes will be pylons 2
[19:19:27] * M3ntor5 has quit IRC
[19:19:30] <steg> "say james i was trying to write a Or permission and ran into some trouble..."
[19:19:35] <steg> aa_: cool
[19:19:36] <CIA-49> Jinja 2: mitsuhiko trunk * 565:293a22033564 /jinja2/loaders.py: Disabled iron python support until new lexer is implemented.
[19:19:46] <steg> aa_: are you at pycon now?
[19:19:50] <mitsuhiko> Leonidas: nope. not the same as github
[19:19:59] <prencher> Leonidas - you get 150mb for one, but on top of that.. github has a limit on number of repos and contributors and such stuff.. bb only limits space and private repos
[19:20:00] <aa_> steg: yes]
[19:20:04] <mitsuhiko> github gives you 100mb, bitbucket gives you 150mb per repository
[19:20:06] <aa_> steg: well, at the hotel
[19:20:08] <steg> aa_: how is it going? when are you talking? are you going to see gjhiggins ?
[19:20:12] <aa_> steg: knackered after giving my talk
[19:20:13] <prencher> mitsuhiko - too slow, f00l
[19:20:17] <birkenfeld> bitbucket has an issue tracker?
[19:20:19] <aa_> steg: I saw higgins
[19:20:19] <steg> aa_: nice! how did it go?
[19:20:23] <mitsuhiko> prencher: pah. i was fixing stuff :P
[19:20:25] <prencher> birkenfeld - it's rather basic, but yeah
[19:20:28] <aa_> steg: ok
[19:20:31] <birkenfeld> nice touch
[19:20:34] <prencher> i prefer tracs still
[19:20:43] <prencher> but its got some nice features
[19:20:45] <aa_> steg: I was timetabled against "James Gardner - Introduction to Pylons"
[19:20:49] <steg> ouch!
[19:20:56] <mitsuhiko> it's basic but it's a good beginning
[19:21:09] <mitsuhiko> and heck, it's way superior to sourceforge
[19:21:13] <steg> you know next semester they have timetabled computational complexity and types and languages into the same slots :(
[19:21:17] <steg> i want to take both
[19:21:35] <steg> aa_: did you get much interest?
[19:21:41] <steg> aa_: were you talking about pida?
[19:21:56] <steg> "crazy nutcase ali expounds his mad philosophies"
[19:22:00] <aa_> nah pygtk and kiwi
[19:22:06] <aa_> a bit of interest, yeah
[19:22:13] <steg> nice
[19:22:19] <steg> i wish i could have afforded to come
[19:22:27] <steg> but for some reason i wasn't paid much over the summer
[19:23:05] <aa_> well, perhaps because you did fuck all?
[19:23:22] <steg> true
[19:23:24] <steg> well, not technically true
[19:23:49] <steg> i did quite a lot of research
[19:24:11] <aa_> you broke a bed researching
[19:24:24] <steg> it's still on the back burner :P
[19:24:25] <aa_> now that's a field I want to get into eh!
[19:24:29] <steg> there's some neat ideas there
[19:24:32] * EnTeQuAk has quit IRC
[19:24:48] <aa_> steg: am still very interested
[19:25:14] <steg> ok
[19:25:25] <steg> there's a semantic web course i was considering
[19:25:34] <aa_> cool
[19:25:36] <steg> i think i'd find it fairly simple now :P
[19:25:54] <steg> but it might be interesting
[19:33:56] <ronny> yo steg
[19:33:58] <ronny> sup?
[19:34:02] <steg> hiya ronny
[19:34:19] <steg> not much, i have so much work to do i am currently shivering in a corner screaming
[19:34:25] <steg> how about you?
[19:37:53] <mitsuhiko> what the.... http://www.youtube.com/watch?v=gBWPf1BWtkw ?
[19:47:25] <ronny> steg: im going insane on algebra exams due my bad algebra training
[19:47:41] <ronny> steg: oh, and i gutted pida vairois times
[19:47:47] <vmx> ronny: any news on braklet? i'm in need of a wiki parser and tekisuto looked nice (though it seems to be like a "finished experiment", but you were always talking about braklet (in irc logs)
[19:48:39] <steg> ronny: nice
[19:50:17] <ronny> vmx: thats ages ago, i forgot where i put the src (i had bugs in nested lists and table handling i couldnt figure)
[19:50:39] <ronny> vmx: but there is the creoleparser lib
[19:50:45] <ronny> vmx: which is a complete wiki parser
[19:51:23] <vmx> i had a quick look, and it looked ... huge, but probably wikiparsers need to be that huge
[19:52:24] <ronny> vmx: wiki parsing is compley and a pain
[19:52:34] <ronny> *complex
[19:53:00] <ronny> creoleparser is fairly simple
[19:53:41] <vmx> i need to take a second look at it
[19:57:34] <aa_> mitsuhiko: there was a python blogging "BoF" session. I pimped Zine.
[19:58:11] <vmx> ronny: it looks nice indeed, perhaps i had a look at it in the beginning of my wiki parser research (before i found out how complex/ugly wiki parsing is. thanks for the hint
[19:58:45] <ronny> vmx: you writing a wiki ?
[19:59:14] <ronny> that reminds me, i wanted to finish my distributed wiki some months ago ;P
[19:59:43] <ronny> bbl
[19:59:45] <vmx> i'm thinking about it. i know there are sooo many out there, but (as everybody's) mine will be different
[19:59:57] <vmx> a bit like ikiwiki in python
[20:00:40] <vmx> and a bit of "i edit my website like a wiki"
[20:01:01] <vmx> distributed wiki sounds nice, i think i have to talk to you later :)
[20:06:51] <birkenfeld> mitsuhiko: have a moment?
[20:08:34] <mitsuhiko> birkenfeld: yep
[20:10:01] <birkenfeld> mitsuhiko: I'd like to make Sphinx notice when a config value changes with that only the HTML output must be regenerated
[20:10:25] <birkenfeld> however, that would mean to have some file containing the old config in the output directory
[20:10:41] <mitsuhiko> pickle :)
[20:11:05] <aa_> or just a timestamp
[20:11:13] <mitsuhiko> or pickle the config and sha1 it
[20:12:10] <birkenfeld> aa_: won't help
[20:12:41] <aa_> birkenfeld: ok, but I can't see why
[20:12:42] <birkenfeld> mitsuhiko: I know, but having that file in the output dir is ugly
[20:12:51] <birkenfeld> aa_: because I only want to rebuild when necessary
[20:13:15] <birkenfeld> so I need the whole old config somewhere
[20:13:40] <mitsuhiko> birkenfeld: you can write the pickled config as data into the project logo png ;)
[20:13:47] <birkenfeld> hehe
[20:14:14] <mitsuhiko> birkenfeld: _build/html/.sphinx
[20:14:16] <aa_> birkenfeld: oh I see
[20:14:18] <mitsuhiko> and drop binary crap there :D
[20:14:28] * vmx has quit IRC
[20:14:33] <birkenfeld> hmm
[20:16:23] <birkenfeld> mitsuhiko: the alternative is storing the config on the environment, one per output dir
[20:16:37] <mitsuhiko> bad idea
[20:16:55] <birkenfeld> why?
[20:17:23] <mitsuhiko> i can imagine situations where the output dir changes daily (nightly docs into folders)
[20:17:35] <mitsuhiko> then the env would grow and grow and nobody knows why
[20:17:51] <birkenfeld> yes, right
[20:18:51] <mitsuhiko> birkenfeld: write it into the makefile :D
[20:19:07] <birkenfeld> I just don't want to leave crap behind for people who directly serve the output dir
[20:19:12] <birkenfeld> which most do I guess
[20:20:23] <mitsuhiko> ugh. ike's over texas
[20:20:32] <birkenfeld> messy?
[20:21:48] * EnTeQuAk has joined #pocoo
[20:22:58] <mitsuhiko> birkenfeld: the build parent folder
[20:23:10] <mitsuhiko> _build/html.state _build/html
[20:23:49] <aa_> why not just save the config file?
[20:24:02] <mitsuhiko> aa_: the location is the problem :)
[20:24:37] <aa_> this is per-build right
[20:24:51] <aa_> and you can't guarantee the output directory?
[20:25:57] * M3ntor5 has joined #pocoo
[20:28:52] * Tik-Tok has joined #pocoo
[20:43:20] <mitsuhiko> i wonder why 2.6 is that much slower
[21:24:47] * M3ntor5 has quit IRC
[21:30:18] <birkenfeld> mitsuhiko: it's the ast module
[21:30:41] <mitsuhiko> birkenfeld: haa haa ;)
[21:30:48] <birkenfeld> :)
[21:30:49] <mitsuhiko> or has the ast runtime implications now?
[21:31:00] <birkenfeld> seriously, is it "that much" slower?
[21:31:12] <birkenfeld> it's very specific benchmarks only
[21:32:13] <birkenfeld> there are some that are faster as well
[21:32:31] <mitsuhiko> hmm
[21:32:33] <birkenfeld> and the overall is faster by 5%
[21:32:44] <mitsuhiko> it seems like object allocation is slower
[21:33:01] <birkenfeld> seems from what?
[21:33:23] <mitsuhiko> loading a zeml tree from a dump
[21:33:37] <mitsuhiko> zeml == secret world domination zine subproject
[21:33:48] * mitsuhiko is building the latest rc now, and comparing then
[21:33:56] <birkenfeld> another persistence format? :)
[21:34:16] <birkenfeld> or markup language?
[21:34:18] <EnTeQuAk> „<mitsuhiko> zeml == secret world domination zine subproject“ <-- that sounds mysterious
[21:34:28] <mitsuhiko> EnTeQuAk: zeml == zine extensible markup language
[21:34:37] <mitsuhiko> and it's actually just a simple html form
[21:34:41] <mitsuhiko> easy to parse and to extend
[21:34:50] <EnTeQuAk> not again a new markup language
[21:34:59] <mitsuhiko> EnTeQuAk: it's just replacing the beautifulsoup hack
[21:35:17] <mitsuhiko> there is just no upgrade path which is why it's still not pushed
[21:35:19] <EnTeQuAk> well, that an excuse ;)
[21:36:18] <mitsuhiko> EnTeQuAk: beautiful was never a long term solution, it was just a temporary solution because i was too lazy to write an html parser
[21:37:16] <EnTeQuAk> i know – but the extensionable api remains persistent?
[21:37:34] <mitsuhiko> you mean the fragment.py stuff?
[21:37:39] <mitsuhiko> zeml would replace fragment
[21:38:00] <EnTeQuAk> fragment.py yea, that was a cool module ;)
[21:38:19] <EnTeQuAk> just copied some ideas from there to dmlt ;)
[21:38:20] <mitsuhiko> why cool?
[21:38:24] <mitsuhiko> ah
[21:38:36] <mitsuhiko> well. it was based on the broken dom concept, zeml is based on element tree
[21:39:21] <EnTeQuAk> yea, but the query method was just easy to extend ;)
[21:39:52] <mitsuhiko> birkenfeld: latest rc seems to be nearly as fast as 2.5 here for zeml, slightly slower but that could be different compilation options
[21:40:26] <mitsuhiko> EnTeQuAk: i'm not sure how and when zeml is integrated. it will happen before the 0.1 release, that's for sure
[21:40:35] <mitsuhiko> and it should provide a query api too
[21:41:01] <EnTeQuAk> then it's possible to port my plugins, yea ;)
[21:52:56] <birkenfeld> mitsuhiko: sounds good
[21:53:49] <mitsuhiko> hum
[21:53:50] <mitsuhiko> ListSlicing: 207ms 130ms +58.6% 228ms 147ms +55.4%
[21:54:50] <birkenfeld> your own test?
[21:55:03] <mitsuhiko> yes
[21:55:19] <mitsuhiko> actually, i *ran* it, it's not *my* test ;)
[21:55:38] <birkenfeld> yes :)
[21:55:51] <birkenfeld> I can't run benchmarks ATM
[21:56:04] <birkenfeld> since my laptop constantly throttles the CPU to 50% and back to 100%
[21:56:48] <mitsuhiko> but i don't get why i get +58% there where the bench on py-dev shows -32
[21:56:55] <mitsuhiko> (same compiler options)
[21:57:06] <birkenfeld> run it again...
[21:57:36] * EnTeQuAk has quit IRC
[21:57:53] * mitsuhiko does. but actually, i think it doesn't matter :
[21:57:54] <mitsuhiko> :)
[21:58:15] <birkenfeld> we'll see
[22:00:10] <xorAxAx> birkenfeld: cpufreq?
[22:00:14] <birkenfeld> xorAxAx: no
[22:00:32] <xorAxAx> C* state throttling because of cooling issues?
[22:00:39] <birkenfeld> xorAxAx: I think yes
[22:00:43] <xorAxAx> fun
[22:01:34] <mitsuhiko> birkenfeld: btw. what's the proposed equivalent to weakdict.keys() / weakdict.values() now?
[22:01:35] <birkenfeld> actually, it's T* states
[22:01:46] <birkenfeld> mitsuhiko: no idea
[22:02:38] <mitsuhiko> .keys() in py3 is no longer atomic and can raise runtime errors if the dict is modified, for weak dicts this makes them nearly impossible to use i fear
[22:03:44] <birkenfeld> mitsuhiko: well, ask python-3000
[22:04:17] <mitsuhiko> i'll do later
[22:06:40] * M3ntor5 has joined #pocoo
[22:10:54] * davidcramer has joined #pocoo
[22:12:31] <mitsuhiko> birkenfeld: okay. seems like .keyrefs() has to be used instead
[22:12:38] <mitsuhiko> at least i guess so
[22:13:31] <mitsuhiko> na. that also doesn't work
[22:15:21] <davidcramer> hey mitsuhiko
[22:15:38] * davidcramer notes that full tracebacks are gonna be useful in this debug tool
[22:15:42] <davidcramer> ./sw/lib/python2.5/pprint.py:292 (_safe_repr); rep = repr(object)
[22:15:47] <davidcramer> does jinja do anything like that?
[22:15:51] <davidcramer> use pprint or whatever it is
[22:15:55] <davidcramer> trying to figure out where this sql query is :)
[22:16:21] <mitsuhiko> jinja1 or jinja2?
[22:16:24] <mitsuhiko> jinja1 maybe
[22:16:38] <davidcramer> jinjua1
[22:16:45] <davidcramer> let me dig around some more
[22:16:51] <davidcramer> im sure django's magically doing it somewhow :P
[22:16:58] <mitsuhiko> jinja1 uses pprint for the pprint filter and the debug tag i think
[22:17:01] <davidcramer> gonna go add better tracebacks to my debugger tho, that'll find it
[22:17:06] <davidcramer> hrmm, not using either so maybe its not that
[22:17:38] <davidcramer> I'm trying to build in everything like django's error handler, so you can get full traces on cache calls, sql queries, etc.
[22:17:42] <davidcramer> to really find those bottlenecks quickly
[22:19:27] <davidcramer> this could probably be used outside of django with a few changes,m you'd just need to monkey patch different spots instead of django's spots
[22:19:37] <davidcramer> e.g. the mysql cursor's directly
[22:20:39] <mitsuhiko> birkenfeld: sent
[22:21:32] <mitsuhiko> davidcramer: do you really think that's that useful?
[22:21:39] <davidcramer> I find it EXTREMELY useful
[22:21:46] <mitsuhiko> what exactly does it solve?
[22:21:48] <davidcramer> when i want to figure out where a query is coming from, and it has inline EXPLAIN stuff
[22:21:57] <davidcramer> well, it gives you a nice overview of "this is how this page is doing"
[22:22:03] <davidcramer> # of cache hits/misses, # of queries (time executed, batch groups, etc)
[22:22:11] <davidcramer> a lot of random meta information which can save time debugging
[22:22:14] <davidcramer> it already has for me :)
[22:22:37] <davidcramer> i find it especially useful if you're not writing SQL, because it can be hard to figure out where thats happening when you use middleware, context procs, template tags, etc.
[22:22:55] <mitsuhiko> i personally would like to see a remote console for development servers that hooks up with a firefox extension
[22:23:18] <davidcramer> that would be interesting
[22:23:26] <mitsuhiko> that forwards the queries, logging statements and an interface to a python shell and the python debugger right into a firebug tab
[22:23:39] <davidcramer> ya
[22:23:49] <davidcramer> we're going to integraet this wiht the logging package too
[22:23:57] <davidcramer> right now you can just add panels to it, but its still very rough
[22:24:09] <davidcramer> and it floats on top of the page
[22:48:16] <davidcramer> turns out the query i needed to debug, is happening because of django's debugging
[22:48:20] <davidcramer> or jinja's one of the two
[22:49:01] <davidcramer> ah ya it's django's stuff
[22:49:14] <davidcramer> its doing repr's on the field values in the traceback which is causing... "omg wtf" queries :)
[22:49:20] <davidcramer> e.g. select * from table
[22:53:43] <mitsuhiko> birkenfeld: seems to be a heating problem here too. cpu temperature jumps to 95°C according to smcFanControl
[23:19:05] * Tik-Tok has left #pocoo
[23:25:30] * prencher^ has joined #pocoo
[23:43:55] * prencher has quit IRC