Upgrade to YAML::Syck 0.85 before smoking | pugscode.org | sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse (show (scanl (*) 1 [1..] !! 4)) | "Perl 6 Today" video from YAPC::Asia: xrl.us/vw3p
Set by avar on 27 April 2007.
00:07 jisom joined 00:23 lumi_ joined 00:30 perlbot joined 00:44 perlbot joined 01:09 nperez joined, dolmans joined 01:14 amnesiac joined 01:25 chylli joined 01:38 takesako___ joined 01:46 thoughtpolice joined 01:47 chylli left 02:03 mako132_ joined
TimToady @tell ruz .{@keys}:p is currently specced to do what you want, though it may end up as .:pairs{@keys} instead. 02:03
lambdabot Consider it noted.
TimToady @tell ruz however, it's not yet implemented either way
lambdabot Consider it noted.
TimToady @moosages
lambdabot cmarcelo said 9h 6m 58s ago: Just read chromatic meeting minutes. And I'm stuck at understanding Haskell->P5 bridge yet. Maybe I should play directly with PCR... Makes me wonder, Pugs used PGE/PCR
instead of Parsec because convenience or because Parsec limitations?
cmarcelo said 9h 5m 3s ago: Also in the minutes you hint about some clear goal/objective for people impl. S05. Could you explain this better?
02:07 thoughtp1lice joined 02:08 thoughtp1lice is now known as thoughtpolice 02:21 dolmans joined 02:24 REPLeffect joined 02:40 rho joined, CindyLinz joined
rho hi all. newbie question ahead: 02:41
is it possible to embed Haskell programs into Perl6?
03:05 DoctorMach joined
spinclad rho: no, neither in Perl6 spec-land nor in pugs (haskell additions to pugs would happen among its haskell internals, not the language it parses). 03:27
though the pugs team is getting ready to reimplement much of it in Perl6, this still won't mix haskell and Perl6 code directly, IIUC. 03:29
i think with some effort this could change, as with v6 you can now switch back and forth between Perls 5 and 6, but this is pure speculation on my part. 03:33
03:36 jisom joined
rho spinclad: thx, so this will have to be in a module. 03:44
03:58 scsibug joined
gaal spinclad: you can, in fact, do this in pugs if you have hs-plugin: say eval 'concat ["m", "oo", "se"]' :haskell 04:10
but there's no "real" ffi 04:11
04:20 SubStack joined 04:26 DoctorMach left
spinclad gaal: thanks, now i know (or better, now i'll look and then i'll know) 04:33
04:40 SubStack joined
spinclad right, i've even seen that before. maybe i won't forget so soon this time. 04:41
rho: see the brief note in INSTALL. 04:44
after that, ask around again. 04:45
05:08 rissy joined 05:19 rfordinal joined
dolmans Perl6's OO system has a large collection of MOP feature, but i want to ask: dose i still can manipulate the sysmbol table, scalar, hash, array, code? 05:21
i think yes. 05:23
Aankhen`` You can access any package's symbol table like you would a hash, IIRC.
dolmans good. 05:24
Aankhen`` Ah, here we go. See S02, under "Names", starting from the second bullet point. 05:25
Or you could search directly for this: To do direct lookup in a package's symbol table without scanning, treat the package name as a hash:
dolmans sound's good. so even not using the WHAT, HOW, etc., i can still do some reflection jobs, like in 5. In language design, if you provide some utility in kernel, then you must leak some flexibility. IMO 05:29
rho spinclad: ad loading Haskell modules: thx, I will try that. 05:30
05:33 Alias__ joined
Aankhen`` Sure. 05:35
Although I don't know why anyone wouldn't want to twiddle with metaclasses. <G>
gaal dolmans: doing reflection via symhash manipulation may not work in classes that use a different metamodel. 05:36
Aankhen`` Different metamodel? Okay, now I'm scared. :-( 05:37
gaal sure, some of the point of having a Real MOP<tm> was to make this stuff pluggable
Aankhen`` And hi!, BTW, gaal. ^_^
gaal there are at least two application targets for this 05:38
hello! };)
dolmans gaal: don't understand that yet
gaal I'll explain:
Aankhen`` I'm still a novice to metaobjects, and my understanding is basically in terms of Moose, so... would that be like changing the metaclasses used in a Moose class, or at a higher meta-level?
gaal first, consider use java:some.java.thing 05:39
Aankhen`` considers.
gaal that loads a java class; the bridging to the foreign methods and objects is via the MOP
Aankhen`` Right.
gaal naturally, you can't just munge a symhash there 05:40
Aankhen`` Uh huh.
dolmans ok, i got this, symtables cannot cope with alien object.
gaal the other application is desmonstrated by a database mapper
*demonstrated
suppose I want to implement something like Class::DBI
dduncan oh, pray tell
Aankhen`` Hehe. 05:41
gaal what I can do in Perl 6 is do put most of my logic in the metamodel.
Aankhen`` watches dduncan's ears perk up.
dduncan I'm interested in anything related to databases, up to some point
gaal for example, the accessors for members trigger database updates
but of course I want to make this efficient
Aankhen`` Aye.
gaal so I need freshness flags 05:42
dduncan what kind of efficient?
gaal dduncan: simplest kind, for the purpose of this explanation
eg.
dduncan eg, efficient for a programmer to write?
gaal not do a db get on every member read if it's fresh
I'm not focusing on database here really, bear with me a sec 05:43
I want for starters this code:
dduncan well, in any event, I'm going back to other things ...
gaal $studentcount = $classroom.students; $studentcount = $classroom.students; to only hit the DB once 05:44
Aankhen`` nods.
dduncan makes sense
gaal one way to do this is for Classroom to have a private %FRESH memeber
but with flexible metamodel, I could make every memeber that maps to a db column have a freshness property of itself 05:45
forgive the clumsiness of terminology here -- better talk with nothingmuch or stevan for the correct names, but suppose we have 05:46
"has Int $students"
dduncan it could also help if there are smarts such that, eg, if we know the class size is small, then just fetch all the student data right away, then we don't have to query each time whether it is fresh
gaal instead of creating a simple %Classroom::<students> slot
create an object with .value and .fresh attribs 05:47
05:47 scsibug joined
gaal dduncan: certainly that can be done there too 05:47
dduncan but a more general application of what I'm saying ...
gaal also foreign data loading, and perhaps the kind of join optimizations I actually know nothing about :-) 05:48
dolmans i want to ask is a normal private attribute still rides in symhash?
dduncan is that it can be useful to fetch more than was asked for if the smarts are in place to guess there is a good chance that some closely related data may soon be asked for too ... it is more efficient to fetch it all in one chunk rather than several smaller ones
gaal so if you're following: any class that has MyGrandORM for its metamodel does not just have a value in the slot field
dduncan sort of like a buffering issue 05:49
gaal it instead has an object that encapsulates the value and anything else my ORM wants to know about it
so direct, naive symhash manipulation would be wrong!
the thing is that although metamodels are pluggable, there is a common interface to them 05:50
dduncan actually, if this whole discussion is just to teach about the metamodel, I'll shut up as what I"m saying is irrelevant then
gaal so for regular programming, $obj.fld calls the appropriate accessors (get/set)
dolmans gaal: partially understood. i conclude that i cannot use naive symhash to really reveal the object's interface, but all the things is still in symhash?(alien object excl.) 05:54
05:54 gaal_ joined
dduncan speaking of the metamodel, I look forward to actually using it, so hopefully audreyt will soon get the time needed to integrate it into Pugs ... I think the statement was a solid 2 days required 05:55
gaal_ and $obj.methods is more robust than p5 grep { defined lassroom::{$_}{CODE} } keys lassroom::
hehe, cgiirc++ but it's eating chars
that should be $obj.meta.methods but .methods a shortcut. 05:56
dolmans: not necessarily! 05:57
TimToady $obj.HOW.method or $obj.^methods
dolmans: packages are for public interface only 05:58
gaal_ if for some reason MyGrandORM decided not to use the symhash, it could
oops, thanks for the correction, TimToady
dolmans TimToady: that mean that the private attribute `has $!attr' is not in symhash? 05:59
TimToady that is correct
(and it isn't in P5 either)
if by symhash you mean symbol table hash
dolmans in P5 i even don't know private 06:00
gaal_ hee, could be abbreviated even further to sysh
TimToady in P6 the standard layout is "P6opaque", and not even the class itself is supposed to worry about the layout
dolmans yes, i got the symhash word from gaal
gaal_ which has "sheesh" for a mnemonic
I'd better grab some coffee and get [back] to work 06:01
moose!
TimToady anyway, this is spelled out in S12 in gruesome detail
most of the time users of P6 won't even know there's a metamodel behind it. 06:02
gaal_ and to perhaps give a good answer to your original question, dolmans, reflection is now *easier*. 06:05
dolmans gaal_: i know that, but my view is: if the designed provide some easier but non-general mechanics, you lose some flexibility. see p5, it has no kernel level's mop, but through symtables, you can do many fantastic things. 06:08
06:21 rissy joined
gaal_ dolmans: you can do many fantastic things in Perl 6 as well :) if there's something specific you think is impossible, do tell 06:22
06:34 goban joined
dolmans gaal_: i don't know, i just guess, i think i should draw back the stupid guess. 06:35
06:37 iblechbot joined 06:48 devogon joined 07:05 franck__ joined 07:18 gaal joined 07:32 BooK joined
Aankhen`` Am I correct in deducting from S12 that $class.bless(...pairs...) creates a P6opaque object? 07:36
07:52 elmex joined 07:53 buetow joined 08:09 amnesiac left 08:21 the_dormant joined 08:24 pbuetow joined 08:40 goban left, dduncan left 08:53 the_dormant joined 08:56 andara joined 09:00 REPLeffect joined 09:22 ayrnieu joined 09:35 polettix joined 09:38 macroron joined 10:07 buetow joined 10:10 rindolf joined 10:25 ruoso joined 10:51 penk joined 10:54 Loro_ joined 11:04 penk joined 11:06 nothingmuch joined 11:07 ecl joined 11:09 ecl left 11:15 DarkWolf84 joined 11:52 larsen_ joined 11:53 mako132_ joined 12:13 andreum joined 12:18 iblechbot joined, Manaus joined 12:24 jerrylee joined, sstipic joined 12:27 sstipic left 12:29 the_dormant joined 12:31 VanilleBert joined 12:32 jferrero joined, Manaus left 12:34 cognominal joined 12:36 Grrrr joined 12:39 idiotmax joined 13:00 turrepurre joined 13:10 larsen_ joined
offby1 has his employer do all his deducting 13:18
13:19 larsen__ joined
rindolf Hi all. 13:21
QtPlatypus Hi rindolf. 13:23
Juerd Who maintains trac on feather? 13:24
rindolf Hi QtPlatypus
Juerd It needs to be optimized.
Yaakov Juerd: Is the eval bot still MIA? 13:29
Juerd I have no idea 13:30
I admin the server, not the individual project related services
Juerd has disabled trac (dev.pugscode.org) until he knows what it's used for, and how it's going to be fixed 13:33
CGI-scripts that use up to a minute of CPU time are broken. 13:34
kolibrie Juerd: audreyt set up trac for project management
Juerd audreyt: ping 13:35
kolibrie pugs.blogs.com/pugs/2007/02/pugscodeorg_ser.html 13:38
lambdabot Title: Pugs: Pugscode.org services.
Juerd Ah, useful. 13:39
I've been killing off the backup process because trac.cgi's collectively used up 80% of the CPU, and it slowed down the backups so much that they started overlapping. 13:40
That has to stop, because backups are way more important than any single service.
(Suggestions for lighter backups are welcome, as long as things are compressed on feather's end and stored compressed on a remote ssh server) 13:41
13:57 deq` joined 14:10 [particle] joined 14:20 VanilleBert left 14:26 svnbot6 joined 14:27 vel joined
diakopter wb svnbot6 14:30
diakopter hugs devbot6 14:31
14:38 electrogeek joined 14:53 specbot6 joined 14:56 evalbot_r16148 joined 15:01 araujo joined 15:02 veritos joined
diakopter wb, specbot6 15:08
wb, evalbot
15:09 [particle] joined 15:20 REPLeffect joined 15:25 ruoso joined 15:27 jerje joined 15:29 iblechbot joined 15:42 bonesss joined 15:47 rlb3_work joined 16:04 kanru joined 16:22 veritos joined
veritos morning 16:25
16:26 bones_ joined
veritos Juerd: i think i accidentally infinite-looped my bashrc :( 16:30
can you restore it since i can't log in?
never mind that 16:33
just logged in with dash instead
16:35 justatheory joined 16:39 franck__ joined 16:41 justatheory_ joined 16:42 buetow joined 16:46 andara left 16:50 jhorwitz joined 17:11 veritos joined 17:29 lisppaste3 joined 17:31 rindolf joined 17:32 riffraff joined
riffraff hi 17:32
veritos heya
riffraff sorry the ubernoob question: but how do I do a replace-not-in-place in perl6 ? 17:33
I mean,. a sytring substitution
jrockway ?eval my $foo = "world"; "Hello, $foo."
veritos do { local $str = $str; $str =~ s/yada/foo/; say $str }? 17:34
evalbot_r16148 "Hello, world."
jrockway evalbot_r16148: awake?
veritos ?eval do { local $str = $str; $str =~ s/yada/foo/; say $str }? 17:35
evalbot_r16148 Error: ā¤Unexpected "?"ā¤expecting statement modifier, term postfix or operator
PerlJam riffraff: the same way as in perl5
riffraff PerlJam, this would be a useful info I actually knew perl5 :) 17:36
[particle] subst
PerlJam riffraff: or use the OOP interface. $copy = $str.substr($pat,$replacement);
riffraff the only way I understand is making a copy and use that
but that changes $str in place
PerlJam riffraff: ($copy = $str) =~ s/foo/bar/; # perl5
riffraff I mean. substr
Tene riffraff: maybe give an example of what you want to do?
lambdabot Tene: You have 2 new messages. '/msg lambdabot @messages' to read them.
PerlJam riffraff: It shouldn't. 17:37
Tene lambdabot: thanks
riffraff mh maybe I misrecall, I'll check again
oh and I should update opugs so I'll check in tewnty minutes, it seem :) 17:38
17:39 allbery_b joined
PerlJam riffraff: $str.=subst($par,$rep); # is the one that will do the in-place substitution. 17:40
rindolf Hi PerlJam
PerlJam Is it just me or is "subst" far too close to "substr"?
rindolf PerlJam: from my impression you seem very protective of Perl 6.
PerlJam rindolf: protective? 17:41
I'll take that as a compliment ;-)
17:41 ruoso joined
rindolf PerlJam: yes, whenever someone says something bad about it. 17:41
PerlJam: you tell him that everything's alright.
PerlJam rindolf: oh. I'm just trying to bring balance to the universe. There are far too many people nay-saying perl6.
rindolf PerlJam: BTW, is Parrot now GPL+Artistic2? 17:42
riffraff perljam: I think one of the two was not implemented last time I checked, which is two weeks ago, but I'm compiliung and will whail later :)
PerlJam riffraff: entirely possible. I haven't kept up with pugs. (pshaw! Who needs an actual implementation anyway? ;-) 17:43
rindolf: I don't think it's Artistic2 just yet. 17:44
rindolf PerlJam: perl.net.au/wiki/Perl_6_-_Common_Criticisms
PerlJam: hmmm...
Lately I daresay I'm a bit unhappy from the edits people make to the articles I start on perl.net.au
lambdabot Title: Perl 6 - Common Criticisms - PerlNet
Title: Main Page - PerlNet
17:45 crashmatrix joined 17:47 Psyche^ joined
PerlJam rindolf: do you know of any die-hard perl5 fans who are also die-hard perl6 naysayers? 17:47
rindolf PerlJam: I don't know.
jjore-w PerlJam: They exist. What of it? 17:50
PerlJam jjore-w: I'd like to interview them.
avar perl5-porters :)
PerlJam avar: heh
jjore-w Go look on perlmonks.org then. 17:51
Post a "News" thing requesting that they contact you.
17:52 cognominal joined
PerlJam I've been loathe to do that for fear of the fanatics (too much noise) 17:53
I'm looking for clear, strong voices with good arguments against perl6.
(well, *I* don't have to think they're good, just that they need to and have reasonable evidence to support their position) 17:54
jjore-w So filter out the noise.
17:55 sapir joined, bones_ is now known as bonesss 17:56 lichtkind joined 17:58 Psyche^ is now known as Patterner, cognominal joined 18:04 cognominal joined 18:11 cognominal joined 18:15 bonesss joined, deq`_ joined
qmole are there pugs binaries for os x anywhere? 18:20
TimToady not officially. historically audreyt has made some available from time to time, but lately I think most of the osx folks have been rolling their own from the sources 18:25
I'm not an osx folk myself... 18:26
and it may be that most of the osx folks who hang out here are asleep right now. 18:27
qmole ah ok
i've found something at www.unobe.com/packages/, giving that a shot 18:28
lambdabot Title: Index of /packages/
TimToady only about 270 revisions behind... :) 18:29
qmole that's ok, i'm at least 5000 revisions behind :) 18:31
TimToady always something to be said for improvement...
18:32 rhandom joined, justatheory joined 18:35 deq` joined
geoffb Speaking of audreyt, anyone know if the current coding hiatus is likely to be a long one? 18:48
18:49 fglock joined
fglock recent hacks: I tried out the several regex engines around, and I found that working on PCR is the most productive way for me to get STD parsed; 18:52
PCR is easily ported to mp6 after it works
TimToady and then backend to P6?
fglock yes, s/mp6/p6/ 18:53
TimToady that is, to P6/pugs
fglock I think that if I were starting directly with p6, I would get lost in parsing algorithms
there is a parser module in ext/, but I couldn't make it work 18:54
TimToady geoffb: I don't think anyone knows. audreyt was making some hopeful noises last week, but my guess is that the family obligations are somewhat open-ended. 18:55
geoffb That sucks. :-(
geoffb wishes audreyt a speedy recovery from family. ;-)
fglock we can get a usable parse tree with PCR 18:56
18:57 jisom joined
TimToady meanwhile the rest of us are peddling as hard as we can, but it seems that we all make better progress when audreyt is able to hang out 18:58
fglock so I'm kind of implementing 'cheat' into PCR/v6.pm
geoffb audreyt seems to be an active catalyst
kolibrie fglock: it's nice to know that all the parsers you wrote are coming in so handy
18:59 the_dormant joined
TimToady I think perhaps fglock does something like genetic programming. :) 18:59
fglock "TimToady"
jrockway hey, btw... how come so few Perl6 talks at YAPC::NA? 19:00
fglock kp6 is much cleaner, but v6.pm is a year ahead
jrockway: $$$ for me 19:01
jrockway no audreyt or timtoady talks though :(
TimToady I'll be there
geoffb Personally, I just want to have some $p6_impl able to do what pugs does + packed arrays. It sounded like Pugs was closest to getting there, because fglock's genetic designs had a few more generations to go. But it seems like no one wants to work on Pugs until audreyt's new AST stuff lands for real. Which means we had the familial equivalent of a bus error on Pugs. Sigh.
jrockway oh, ok
fglock later &
we need more mutant implementors 19:02
19:03 fglock left
geoffb I probably qualify as a mutant implementor, except that -Ofun for me involves implementing a higher layer ... 19:04
jrockway needs twice as much time (or caffeine) 19:07
geoffb I find that twice as much caffeine is pretty brutal on the guts
jrockway exactly 19:08
19:08 Daveman joined 19:23 REPLeffect_ joined
TimToady caffiene, schedule, guts--pick two 19:23
lumi ?eval <caffeine schedule guts>.pick(2) # ? 19:27
evalbot_r16148 ("guts", "caffeine") 19:28
lumi 1x caffeine, then
19:37 HiTech69 joined
diakopter will request audreyt work on pugs Tue-Fri, but I can't guarantee it will occur. :) quod licet Iovi non licet bovi. 19:37
19:37 the_dormant joined
avar ?eval <good fast cheap>.pick(2) 19:40
evalbot_r16148 ("good", "cheap")
TimToady that's us all right... 19:43
gaal ?eval <moose moose moose>.pick(1) 19:45
evalbot_r16148 ("moose",)
gaal ?eval <black black black>.pick(1) # henry ford 19:46
evalbot_r16148 ("black",)
diakopter ?eval <pugs self>.pick(3) 19:48
evalbot_r16148 ("self", "pugs")
pasteling "evalbot6_r16148" at 65.89.234.66 pasted "Pugs build failure" (326 lines, 19.6K) at sial.org/pbot/24622
Tene ?eval ("moose" xx 90).pick(1) 19:49
evalbot_r16148 ("moose",)
19:49 evalbot6_r16148 joined
gaal i111.photobucket.com/albums/n141/wi...ival58.jpg 19:49
lambdabot tinyurl.com/29tbaf
geoffb very nice 19:50
19:51 benny_ joined 19:54 dduncan joined 19:55 devogon joined
TimToady I wonder if it can see the fjords from there... 19:57
gaal I don't know about fjords, but these may be pines: i111.photobucket.com/albums/n141/wi...ival47.jpg 20:00
lambdabot tinyurl.com/2dxcgy
20:12 limp` joined 20:15 vel6608 joined 20:29 benny_ is now known as benny
kolibrie jrockway: I was feeling a little lonely looking over the list of talks 20:32
fglock: you can help me with my talk :) 20:33
jrockway heh, there are like 4 catalyst talks this year
which is good
20:36 veritos joined 20:39 explorer joined 20:48 REPLeffect_ left 21:12 metaperl joined
metaperl Juerd: ping? 21:12
Juerd metaperl: pong 21:14
metaperl Juerd: are you getting my private msgs? 21:15
Juerd Yes 21:16
I'm not identified to nickserv at the moment 21:17
So I cant send replies
metaperl oh 21:18
maybe we can join #juerd and talk 21:19
Tene But what if other people spy on you?! 21:20
metaperl :)
beppu: hey!
moritz Juerd: if somebody /msgs you, your replies will work 21:21
Juerd: unless the other one /unqueries prior to reply, afaict
moritz is not registered eitehr 21:22
Juerd Apparently not 21:23
geoffb Apparently someone beat me to registering my preferred nick here, but so far as I know we've never tried to be in the same channels, so my only worry is that the other geoffb might be some jerk and people think I'm him.
Juerd I get this when I reply to metaperl: 21:24
23:15 [freenode] -!- Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( freenode.net/faq.shtml#privmsg )
lambdabot Title: freenode: frequently-asked questions
geoffb I've considered registering a new nick, but I haven't had enough push to overcome laziness
moritz Juerd: did you try /msg nickserv register ?
geoffb: same for me ;)
Juerd moritz: I am registered, but it's not accepting my password. 21:25
moritz I got kicked twice because the "owner" of my nick joined, but he was gone after a few minutes
Juerd I don't care enough to do anything about it.
geoffb moritz: ditto.
Juerd I generally just hate freenode for things like this.
I used to hate it for the wallops flood too, but that stopped when lilo died. 21:26
geoffb lilo *died*? Wow, I wondered why the wallops went down so much, but I just assumed lilo got a life or something 21:27
avar heh, I was just wondering how many users didn't know about lilo but just noticed a reduction in wallop spam:)
apperently at least one:) 21:28
geoffb Amazing the impact people have ...
moritz
.oO(People are overrated ;)
21:29
21:29 ruoso joined
geoffb "A person is smart. People are dumb, panicky, dangerous animals and you know it." 21:30
moritz "the source of all unhappiness is other people. the sooner you learn to thonk of other people as noisy furniture, the you will be happy" - read somwhere on the dilbert cartoons ;) 21:34
[particle] sartre wrote, "hell is other people." 21:35
my views differ. 21:36
21:39 larsen_ joined 21:46 rissy joined, cmarcelo joined 21:48 mdiep_ joined 21:54 Limbic_Region joined 21:56 cmarcelo joined 21:59 cmarcelo joined, rhandom left 22:29 SubStack joined
cj is the fancy test report at smoke.pugscode.org/ created using Test::TAP::HTMLMatrix? 22:33
lambdabot Title: Pugs Smoke Reports
avar cj: yes 22:35
jrockway avar, long time no see 22:39
avar jrockway: wrarr 22:40
22:42 Aankhen`` joined
cj jrockway: good call 22:50
jrockway cj: no problem :) 22:56
i need to get this set up at work too... glad someone reminded me
23:05 gravity joined
cmarcelo is it possible that the trac abnormal resource usage was caused by bots scanning the irclog and getting to diff pages generated by trac via the magic r1234 links? 23:11
TimToady when the load average was 12, there were two bots using trac, googlebot and msnbot
that was several days ago that I looked 23:12
geoffb Why not just block those bots?
"We didn't know how to do that in Trac" is a valid answer ....
cmarcelo was that abnormal usage happening before moritz++ irclog? 23:14
jrockway we had similar problems with catalyst trac 23:15
we switched to fastcgi + postgres and that solved everything
trac does dumb locking on sqlite databases
cmarcelo jrockway: sqlite support something else than locking the entire DB? (sorry, I'm not very familiar with SQLite..) 23:16
jrockway i don't think so 23:17
but i'm pretty sure trac was spinning out of control while waiting for the lock
cmarcelo geoffb: I don't know how to do this with trac :(
jrockway while(locked){ acquire_lock } or something
i didn't look at the code, so don't yell at me if i'm off :)
python looks like line noise
geoffb *chuckle* 23:18
TimToady just run it through py2p6 and then it'll run slow all the time...
jrockway haha 23:19
jrockway recruits interested parties to join #killtrac on irc.perl.org :)
cmarcelo jrockway: what's killtrac? 23:20
jrockway perl glue for RT / bugzilla + SVN + etc.
basically trac, but not as bad
also mojomojo for the wiki
cmarcelo oh.. seems nice
jrockway i haven't written much though
only KT::Timeline, which is like trac's timeline
cmarcelo is there a webpage somewhere? 23:21
jrockway but with CPAN Testers reports, and other nice stuff in addition to tickets, commits, etc.
killtrac.bsdro.org i think
it's totally a work in progress
we really need people that care enough to write some code
we all want it, but nobody wants to write it :(
not sure why
geoffb jrockway: perhaps because many of us have to do that kind of stuff for $day_job and want to spend our limited fun time on something completely different 23:22
jrockway yeah, that could be it 23:23
geoffb Then again, there are people who are lucky enough not to have a $day_job yet. :-)
23:29 lichtkind joined
cmarcelo jrockway: just a thought: VCS independent is a little bit harder if you want to do fancy integration. some VCS have different ways to identify revisions or even doesn't have concept of revisions (darcs), for example. 23:42