Camelia, the Perl 6 bug

IRC log for #mojo, 2011-01-30

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:42 daviddelikat1 joined #mojo
00:47 daviddelikat left #mojo
00:58 su-bzero is now known as su-bzero[A]
02:12 daviddelikat joined #mojo
02:17 daviddelikat1 left #mojo
04:04 tl joined #mojo
04:06 tholen42 left #mojo
04:47 jwang joined #mojo
05:39 kaare joined #mojo
06:45 kaare left #mojo
06:48 cside_ joined #mojo
07:11 cside_ is now known as Cside
07:15 Cside left #mojo
07:16 perlrocks Twitter: "@ Hidekiy sudo cpan Mojolicious worked on. Thank you!" (ja) --nekogigi_now http://twitter.com/nekogigi_​now/status/31611621486895104
08:39 sri good morning sunshines
08:42 marcus MOOO
09:17 daviddelikat1 joined #mojo
09:19 daviddelikat left #mojo
09:58 perlrocks Twitter: "mojolicious-lexicon 0.91 by UNDEF - http://frepan.org/~undef/mojolicious-lexicon-0.91/" --cpan_new http://twitter.com/cpan_ne​w/status/31652396581003264
10:27 su-bzero[A] is now known as su-bzero2
10:49 stephan48 when will --reload pickup template changes?
10:49 sri instantly
10:50 stephan48 even when they are not in programmfile?
10:50 sri always
10:50 sri if it doesn't make sure to update
10:50 sri there was a fix recently
11:15 otaviof left #mojo
11:46 perlrocks Twitter: "mojolicious-lexicon 0.991 by UNDEF - http://frepan.org/~undef/m​ojolicious-lexicon-0.991/" --cpan_new http://twitter.com/cpan_ne​w/status/31679592498470912
11:52 sri grrrr
11:53 sri forwardever screwed up big time with the refactoring :S
12:03 otaviof joined #mojo
12:14 stephan48 does it make things faster?
12:22 alnewkirk joined #mojo
12:29 sri unlikely
12:29 sri maybe slower
12:30 stephan48 ;)
12:39 sri it does make url_for caching easier though
12:50 GitHub129 joined #mojo
12:50 GitHub129 mojo: master forwardever * e510ae4 (4 files in 3 dirs): decouple Mojolicious::Routes::Match from controller object
12:50 GitHub129 mojo: master Sebastian Riedel * 835b431 (4 files in 3 dirs): cleanup
12:50 GitHub129 mojo: master commits 294baa8...835b431 - http://bit.ly/g56gFN
12:50 GitHub129 left #mojo
12:52 sri the api did get a bit cleaner imo
12:56 perlrocks Twitter: "This should prove that we are interested in making even the high level components of #mojolicious reusable. ;) http://t.co/1hJYHcq #perl" --kraih http://twitter.com/kraih/status/31697249499090944
12:56 marcus left #mojo
13:05 GitHub6 joined #mojo
13:05 GitHub6 mojo: master Sebastian Riedel * 46eafaf (1 files in 1 dirs): pod cleanup - http://bit.ly/dUlvoz
13:05 GitHub6 left #mojo
13:09 ispy_ left #mojo
13:09 perlrocks Twitter: "@f055 You mean like a cookbook recipe? :) http://t.co/GwiAZQq" --kraih http://twitter.com/kraih/status/31700642170933248
13:19 su-bzero2 url_for doesn't work now? ;)
13:22 GitHub64 joined #mojo
13:22 GitHub64 mojo: master Sebastian Riedel * 396e97d (1 files in 1 dirs): test cleanup - http://bit.ly/i7MhxD
13:22 GitHub64 left #mojo
13:22 sri no i just removed url_for and broke every single app out there...
13:22 sri *cough*
13:23 sri thats of course not what i did
13:46 perlrocks Twitter: "You could theoretically reuse the #mojolicious routes implementation for your own frameworks already. :) http://t.co/9eT4fFz #perl" --kraih http://twitter.com/kraih/status/31709931333746689
13:51 sri http://cpanratings.perl.org/dist/Mojolicious # wtf
13:53 sri http://cpanratings.perl.org/user/blue # awesome guy
13:53 yko ahaha lol
13:53 yko should be author of Tie-Hash-Indexed
13:54 abra left #mojo
13:54 yko sri: actually, he gave Mojolicious 2 stars, not 1 as he usually do
13:54 sri yea lol
13:57 perlrocks Twitter: "Guess a two star rating for #mojolicious from this guy is actually an achievement. :D http://t.co/qVnYVsI #perl" --kraih http://twitter.com/kraih/status/31712640564727808
13:58 perlrocks Twitter: "One more reason to get voting if you love #mojolicious! http://t.co/zeMQ9dv #perl" --kraih http://twitter.com/kraih/status/31712924078706688
14:38 GitHub145 joined #mojo
14:38 GitHub145 mojo: master Sebastian Riedel * 32b1144 (1 files in 1 dirs): cleanup - http://bit.ly/f2RPRz
14:38 GitHub145 left #mojo
14:48 forwardever joined #mojo
14:52 forwardever sri: reviewing your cleanup, in the logs, you wrote there were problems with my commit, could should provide some details
14:52 sri yes, i was little disappointed
14:52 forwardever why
14:52 sri there were many problems
14:53 forwardever like
14:53 forwardever ?
14:53 sri you did stuff like $r->{_via}
14:53 sri thats a total no go
14:53 sri the pod was all wrong
14:53 sri many things were not cleaned up
14:54 yko there's no method to remove event in IOLoop, i should do it manually?
14:54 sri yko: what do you mean?
14:55 yko my $id = $loop->on_tick(sub { }); $ioloop->remove_event($id) or so
14:56 sri thats what ->drop is for
14:56 forwardever sri: what is wrong with $self->{_via}, e.g. there is also $self->{_name}
14:56 sri why would you ever remove an on_tick though?
14:57 sri forwardever: not $self, $r
14:57 sri you were poking inside foreign objects
14:58 yko sri: thx
14:58 sri yko: i don't know if it works for on_tick
14:58 yko ill check
14:58 sri thats why i'm asking for a reason
14:58 yko thanks anyway
14:58 sri if you want temporary events you can use zero timers
14:59 forwardever sri: use an accessor in these cases?
14:59 sri forwardever: of course
15:00 sri you don't ever touch anything marked private that doesn't belong to the current class/instance
15:00 forwardever yea
15:01 sri if it's not documented you're not allowed to use it
15:01 forwardever what was wrong with pod?
15:01 sri url_for for example was all wrong
15:01 sri the is_websocket accessor was also very bad
15:01 mohan joined #mojo
15:01 sri it was set on url_for
15:02 sri well, you can look through my commit, there was more i don't remember
15:03 forwardever well, url_for, I think I didn't change anything there
15:03 forwardever in the pod
15:03 forwardever "path_for" makes much more sense though
15:05 sri url_for changed completely
15:05 sri you left it in the version where it returned Mojo::URL objects
15:06 forwardever yea, true
15:06 sri yko: oh, turns out ->drop works for on_tick :)
15:06 yko ye, i see
15:07 yko i'ts not one-time event, it's presistant, but i need to remove it in some cases
15:09 perlrocks Twitter: "I'm just now looking Teng.pm the POD. Not good or bad, CGI:: App and Mojolicious:: Lite familiar with the "File 1 = 1CGI" or when compatibility is, ORM is also "a class-oriented database to be like a" coming out in or Data:: Model A's would be nice?" (ja) --tsukamoto http://twitter.com/tsukamot​o/status/31730807810822144
15:24 sri forwardever: what's your next target?
15:24 sri btw. http://www.gitready.com/advanced/2009/0​2/10/squashing-commits-with-rebase.html
15:24 forwardever will look at Mojolicious::Routes and try to understand more of it :)
15:24 sri rewriting history is the best part of git
15:25 sri yea, the dispatcher part will be very very tricky
15:25 forwardever i read exactly that article yesterday, but without success, maybe problems on windows and editor
15:26 stephan48 sri: i thought rewrite repo history is bad?
15:26 sri only after you made it public
15:27 sri i actually expect multiple commits to be rewritten to a single one before i get a pull request
15:28 sri it's everyday git stuff :)
15:28 stephan48 okey
15:30 yko sri: in Client is it possible to drop pending connections?
15:30 yko i don't think that $loop->drop($_) for keys %$cs; from outside of Mojo/Client.pm is a good idea...
15:31 stephan48 btw, i ended up using Mojo::Home directly for my Config Class and Mojo::Base for accessors :)
15:31 sri yko: it is
15:32 sri the wonders of non blocking connect
15:32 sri yko: no idea what you're talking about
15:36 yko at some point i want to stop client and drop all not-finished connections
15:36 yko problem is that i don't want to destroy client instance and client's ioloop may be shared with app, so stopping it also bad idea
15:37 sri thats way beyond the scope of Mojo::Client i'm afraid
15:37 sri if you want that much control you have to write your own
15:37 yko hmm.ok, htanks
15:38 sri why don't you want to destroy the Mojo::Client instance?
15:38 yko it may contain user configuration.
15:39 sri ok, better to step back a bit.... what do you actually want to achieve? :)
15:39 yko i think i can do something like my $new = $old->new; undef $old; and pray
15:39 sri $client->clone is not bad
15:39 yko it doesn't copy ioloop
15:40 sri back to the question
15:40 yko ye, answering
15:43 yko okay. im building crawler that recieves urls from a list throwing N requests at once and keeping number of request in queue
15:43 yko at some point i want to be able to stop the client hardly. if it uses ioloop that wasnt running at startup it's ok - i just stop ioloop.
15:43 yko but if ioloop was running than somebody else depends on it, so i want just to drop all not finished connections
15:45 sri like my $new = $old->clone; $new->ioloop($old->ioloop); undef $old;?
15:45 yko ye i'm thinking in that way
15:45 yko because my $new = $old->new is not an option. it will copy {_cs}
15:45 yko thanks again
15:46 sri i was expecting that you wanted a hard timeout or so
15:46 sri would have been good since i still want that as a core feature :)
15:47 sri but i'm too lazy to think of a test case
15:48 yko hmm. hard timeout?
15:48 sri $client->timeout(5) and nothing is allowed to take more than 5 seconds
15:49 sri connection would be dropped and request aborted early
15:50 yko umm. keep_alive_timeout doesn't do so?
15:50 sri only for inactivity
15:50 yko ah
15:51 sri it mostly does just the right thing
15:51 yko now i get it. well, looks like i don't need it atm :p
15:51 sri aww
15:51 sri :,(
15:51 * yko giggles
15:52 sri a way to abort requests could be fun too
15:52 sri $client->on_start(sub { my $id = pop; ->timer(5 => sub { $client->abort($id) }) }) or so
15:53 yko :)
15:53 ispy_ joined #mojo
15:53 yko gonna using on_start btw
15:53 sri on_start makes a lot of hacks possible
15:54 yko ye, like changin User-Agent :D
15:54 sri you could use it for your own connection management
15:55 yko that's why i'm going using it
15:55 sri :)
15:55 sri ah then you already can have hard timeouts
15:55 yko overloading _connect sux so much
15:56 sri $ioloop->timer(5 => sub { shift->drop($id) })
15:56 sri the client should catch that and end the request gracefully
15:56 yko hm. nice
15:56 sri since we throw hup events all the time now
15:57 sri thought about renaming on_hup to on_close or so
15:57 sri on_end
15:57 perlrocks Twitter: "Cloud mojolicious cute favicon" (ja) --yasutaka http://twitter.com/yasutak​a/status/31742918238670848
15:58 yko cute tweet --^
15:58 hexa sri: Do you know what Continuity is?
15:58 purl it has been said that Continuity is available at http://thelackthereof.org/​projects/perl/continuity/, or http://continuity.tlt42.org/ or a continuation-based web-programming thingy
15:58 sri :)
15:59 hexa Hehe
15:59 sri yes i do
15:59 hexa Is it a standalone thing?
15:59 hexa Used with fcgi etc
16:00 sri ye
16:00 hexa What do you think of it?
16:00 sri i prefer the original
16:01 sri the whole style is just a nice toy though
16:01 sri it just doesn't fit well into the web of today
16:01 hexa It looks good for doing quick and dirty applications
16:01 sri since it is antiREST
16:01 hexa I see
16:02 sri if you're into that continuation style you should check out http://www.seaside.st/
16:02 sri thats the original
16:05 sri it was rather popular for a short time before the whole RESTful wave kicked in
16:06 * hexa has a look
16:12 perlrocks Twitter: "script / mojolicious Where did the> <anyone tell me if I wonder unusual" (ja) --maa_zoooooo http://twitter.com/maa_zoooo​oo/status/31746664955908097
16:22 tempire_ joined #mojo
16:22 tempire_ Mojolicious: wonder of the unusual
16:23 sri i think he is wondering where the "mojolicious" script went
16:24 tempire_ I haven't used on_hup, but it seems more obvious than on_close or on_end
16:24 sri i'm not too keen on renaming anyway
16:26 tempire_ I really dig ioloop.  Wish I had a reason to use it for my current projects.
16:28 sri ioloop should become a spinoff project with a hip name
16:28 sri like node.pl!
16:32 tempire_ I wonder what vti will call his framework.  I vote for trololo.
16:33 tempire_ I would have to start using it, just based on the awesome name.
16:37 tempire_ left #mojo
16:37 sri irc on a ipad Oo
16:37 sri *+n
16:38 yko sri: https://github.com/kraih/mojo/blob/​master/lib/Mojo/Client.pm#L629-630 Client passes cert && key but doesn't clone it in ->clone or ->async
16:38 yko it's intentional?
16:38 sri nope
16:38 yko i mean that async client doesn't inherit tsl stuff
16:40 su-bzero2 is now known as su-bzero[A]
16:41 tempire sri: it's true.  there doesn't seem to be any way to not display that away message.
16:41 tempire annoying.
16:43 tempire I never looked @ clone when doing the tls stuff.  I presume it's just a matter of adding ->cert & ->key to 'sub clone' ?
16:44 sri ye
16:44 sri there is more missing
16:45 yko oh great! found it.... :)
16:54 GitHub157 joined #mojo
16:54 GitHub157 mojo: master Sebastian Riedel * 11c493e (3 files in 3 dirs): fixed Mojo::Client cloning - http://bit.ly/fPGf48
16:54 GitHub157 left #mojo
16:54 sri yko: fixed
16:54 yko wow :)
16:54 yko sri: can lines 601 and 642 in Client.pm be moved down to L649?
16:55 yko they are doing just the same...
16:55 yko oops now they have different numbers
16:55 sri was about to say it
16:56 yko 608 and 649 to 656: $self->{_cs}->{$id} = {cb => $cb, tx => $tx};
16:57 yko that would prevent anyone like me who want to set up connection manually touching ->{_cs}
16:57 sri hmm, i know a few months ago it was not possible
16:58 yko ohm. now i'm cofused a bit. need to check something
16:58 sri nope, things break badly
16:59 yko oh. just forget it
16:59 sri thats rather complicated code, you better test stuff before asking
16:59 yko ye, sorry
16:59 sri i would have to read up on it too, which is annoying :/
17:00 sri connection handling is one of the most complicated pieces of code we have actually
17:01 sri due to keep alive, redirects and proxy connects
17:01 yko i feel it :)
17:01 sri :)
17:01 yko hmm. i can not stop transaction in on_start, yea?
17:01 sri feel free to clean it up too
17:02 sri dunno
17:02 sri you're not intended to at least
17:04 yko well, if i'm going to set up connection manually i need to drop tx if couldn't connect at least
17:04 sri i'm afraid you might have to find out for yourself
17:04 yko _connect is called right after on_start. and there's no any checks there.
17:04 yko thanks. i think ill just die
17:04 sri i really don't want to poke around in that code :/
17:05 tempire yko: if I understand correctly, you can just use a timer.
17:05 sri oh right
17:06 sri even easier maybe
17:06 sri no wait
17:06 yko tempire: timer? how it can help me if i fail to set up connection in on_start?
17:06 * sri needs a beer
17:06 tempire I suppose it depends on what you mean by 'fail'
17:06 tempire but just using the timer to set a timeout for say, n seconds
17:08 sri he has no connection id yet
17:08 yko tempire: i'm setting up connection for tx in on_start. i need something like https://github.com/kraih/mojo/blob/​master/lib/Mojo/Client.pm#L642-646 but in on_start callback
17:08 sri no wait...he starts the connection himself... ignore me
17:08 * sri really really needs beer
17:08 yko let's close this discussion, it's mine problem not yours and i think you all are tired of me for today
17:09 sri i'm not tired of you, just tired in general :)
17:09 yko :)
17:09 * tempire is awake
17:10 sri it's a valid discussion, i'm just too tired
17:12 tempire hmm
17:13 tempire it just cilcked how neat on_start is.
17:24 kaare joined #mojo
17:31 non joined #mojo
18:13 RidasJim joined #mojo
18:15 forwardever left #mojo
18:16 RidasJim left #mojo
19:27 ispy_ left #mojo
19:47 bart joined #mojo
19:58 zipkid 18:14  * sri really really needs beer   <- sri should come to the Fosdem beer event next friday! :-)
20:04 bart left #mojo
20:45 ispy_ joined #mojo
20:46 non left #mojo
20:54 bart joined #mojo
20:55 ispy_ left #mojo
21:16 bart left #mojo
21:34 maja joined #mojo
21:35 maja is now known as chuckderbiber
21:40 chuckderbiber left #mojo
21:43 kaare left #mojo
21:53 marcus joined #mojo
21:55 marcus durr
21:55 marcus colloqoy isn't a very awesome IRC client
21:55 marcus ccolloquy even
21:55 zipkid indeed
21:55 marcus feels like the ui was made by a 14-year old
21:55 zipkid i never use it
21:56 * marcus tries limechat
21:56 marcus that works well for me on the ipad
21:56 marcus otherwise I guess I will have to buy linkius
22:04 marcus oh noes
22:04 marcus hmm. Actually, this client is pretty nice.
22:06 forwardever joined #mojo
22:06 forwardever sri: lastest commit not on github?
22:07 tempire limechat is nice?
22:07 marcus yeah, it looks allright
22:08 marcus http://nordaaker.com/docs/skitch/%28marcus%29​_%23mojo_%2880%2C%2Bn%29-20110130-230801.jpg
22:08 marcus wow, it even shows a preview of that image
22:08 marcus =)
22:10 hexa left #mojo
22:31 bart joined #mojo
22:40 bart /?
22:40 purl / is probably not missing. it's not intended to be the url (though maybe it should be...)
22:49 bart /?
22:49 purl somebody said / was not missing. it's not intended to be the url (though maybe it should be...)
22:50 bart /?
22:50 purl it has been said that / is not missing. it's not intended to be the url (though maybe it should be...)
22:54 bart left #mojo
23:17 GitHub165 joined #mojo
23:17 GitHub165 mojo: master Sebastian Riedel * 42eb293 (2 files in 2 dirs): various routes tweaks - http://bit.ly/h93eE4
23:17 GitHub165 left #mojo
23:17 sri forwardever: wanted to clean it up some more
23:22 sri forwardever: how are you measuring performance?
23:25 bart joined #mojo
23:26 bart /?
23:26 purl / is probably not missing. it's not intended to be the url (though maybe it should be...)
23:32 bart was kicked by sri: bart
23:34 MojoGuest536 joined #mojo
23:34 MojoGuest536 left #mojo
23:37 MojoGuest132 joined #mojo
23:37 MojoGuest132 left #mojo
23:52 forwardever sri: haven't measured anything in this case, just multiple sub calls to get the same information didn't make lot of sense

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary