Camelia, the Perl 6 bug

IRC log for #mojo, 2011-04-26

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

All times shown according to UTC.

Time Nick Message
00:01 alnewkirk left #mojo
00:08 pranq left #mojo
00:15 keith left #mojo
00:55 metaperl joined #mojo
00:55 perlrocks Twitter: "@smallgiantsbuzz Hi guys! Got my xtra mojo pins today! ThankU! Fun to get something so clinky, cool & mojolicious from you in mail today!" --AneseCavanaugh http://twitter.com/AneseCavan​augh/status/62681136937181184
00:55 metaperl_ joined #mojo
01:41 sigue left #mojo
02:12 sigue joined #mojo
02:42 jamesw joined #mojo
02:42 sigue left #mojo
02:42 sigue joined #mojo
03:15 sigue left #mojo
03:35 sh4 joined #mojo
03:37 sh4 left #mojo
03:59 sigue joined #mojo
04:07 perlrocks Twitter: "Vuln: Mojolicious Directory Traversal Vulnerability (SecurityFocus Vulnerabilities):  http://bit.ly/gXNWOv #ISC" --b0rn2Frag http://twitter.com/b0rn2Fra​g/status/62729397710487552
04:08 perlrocks Twitter: "Vuln: Mojolicious Directory Traversal Vulnerability (SecurityFocus Vulnerabilities): http://bit.ly/hAvqrK #ISC" --b0rn2Frag http://twitter.com/b0rn2Fra​g/status/62729840058576897
04:16 perlrocks Twitter: "Vuln: Mojolicious Directory Traversal Vulnerability (SecurityFocus Vulnerabilities) http://www.securityfocus.com/bid/47402" (es) --D3Seguridad http://twitter.com/D3Segurid​ad/status/62731657744097280
04:29 sigue left #mojo
04:42 sigue joined #mojo
04:43 kaare joined #mojo
05:07 sigue left #mojo
05:19 koban joined #mojo
05:25 koban left #mojo
05:25 koban joined #mojo
05:57 sigue joined #mojo
06:14 AmeliePoulain joined #mojo
06:25 perlrocks Twitter: "1.21 Mojolicious http://ow.ly/1clsiP" (sw) --campuscodi http://twitter.com/campusco​di/status/62764101847367680
06:37 Christian joined #mojo
06:37 Christian morning
06:39 ysyrota joined #mojo
07:03 koban left #mojo
07:03 koban joined #mojo
07:13 yakudzo left #mojo
07:13 yakudzo joined #mojo
07:24 Sugar joined #mojo
07:48 fhelmber_ joined #mojo
07:54 sigue left #mojo
07:54 sigue joined #mojo
07:54 cosmincx joined #mojo
07:59 sigue left #mojo
07:59 sigue joined #mojo
08:04 ajgb joined #mojo
08:06 sigue left #mojo
08:06 sigue joined #mojo
08:27 Sugar left #mojo
08:27 Sugar joined #mojo
08:30 Sugar left #mojo
08:31 Sugar joined #mojo
09:02 Foxcool joined #mojo
09:03 Foxcool left #mojo
09:04 Foxcool joined #mojo
09:12 sigue left #mojo
09:13 marcus sri: HALP
09:14 marcus sri: I have a client that uses the old Mojo::Client to do parallel requests. It uses build_tx and queue to set up a set of connections with callback, then ->start to process all of them blocking.
09:14 marcus sri: Not sure how I can do the same using UserAgent
09:17 marcus just start->tx( => $cb) then Mojo::IOLoop->start to block?
09:17 sh4 joined #mojo
09:17 marcus ($tx=>$cb) even
09:17 yegor joined #mojo
09:18 sigue joined #mojo
09:20 yegor left #mojo
09:31 perlrocks Twitter: "Fedora update for perl-Mojolicious http://bit.ly/eimsrV" (ja) --vulnow http://twitter.com/vulnow/status/62810931104718849
09:33 yko damn!
09:36 * yko waves
09:39 yko marcus: looks like that impossible without doing hackish stuff
09:39 cosmincx1 joined #mojo
09:44 cosmincx left #mojo
09:52 marcus yko: :(
09:52 yko :)
09:53 marcus yeah. looks like Mojo::Client->start doesn't block
09:53 yko i think that problem is deeper
09:53 koban left #mojo
09:54 yko if you force kraih to make some non-compatible changes in UA i'll curse you with all slavic curses i know!
09:54 yko hehe
09:56 marcus yko: I'd rather not.
09:56 marcus but I'm not sure how I can get this use case to work
09:57 marcus looks like there's no way to do parallel requests in a blocking fashion now :(
09:58 yko i think it's possible by putting us in non-blocking mode, setting IOLoop manually to custom (not singleton) and than starting tx's with $ua->_start_tx($tx => $cb);
09:59 yko but, as i said before, it's dirty way and will be broken by changes without any warnings
10:00 sigue left #mojo
10:00 yko s/us/ua/
10:00 sigue joined #mojo
10:06 marcus yeah, that sounds pretty bad
10:07 yko maybe it's good that you can not do blocking - non blocking requests. too much freedom is bad too
10:07 marcus heh
10:08 yko wanna do bad things - create fork, separate process or so
10:08 cosmincx1 left #mojo
10:08 marcus I have an app built around a web service
10:08 cosmincx joined #mojo
10:08 marcus not being able to do parallel requests is pretty bad for this.
10:09 yko but wait... i don't understand you from this point
10:09 yko you have web app, running in deaemon mode i guess
10:09 yko is it right?
10:09 marcus it's actually running in fastcgi mode now, but I plan to move it to hypnotoad
10:10 yko so if you are in daemon mode yoyu just don't need to start ioloop.
10:10 yko you do $c->render_later; and in last response you render answer to user
10:11 marcus so, I have an action which does something like $self->stash(ev => $self->stash('ws')->events); $self->stash(ent=>$self->stash('ws->entities));
10:11 yko about a week ago i had similar gist
10:11 marcus and then I want to block in the template when I do $ev->next()
10:11 marcus and fetch both of the requests
10:12 marcus if I do the render in the callbacks, how can I make sure I've got the results from both the requests?
10:12 yko use counter
10:13 marcus hrm, this sounds pretty horrible :(
10:13 yko well, i don't think so :p
10:15 yko well, i didn't found gist with many requests in action, but that one: https://gist.github.com/a3ee562cd495996b684a
10:16 yko i'm sure if you change line 28 to something like "render() if ++$counter >= 2" you'll have what you need
10:16 yko and app won't block while serving requests
10:16 marcus I guess I will have to rewrite my ws api completely
10:16 marcus to use callbacks
10:17 yko dunno, it's your app not mine
10:17 marcus actually at this point it might be better to just use the hacks above
10:17 yko i was affected by ua much too and didn't reqrite all my apps yet
10:17 yko so some of them just use 1.12
10:17 yko i'd rather rewrite half of app or just don't update
10:19 yko using hacks is pretty cool thing while you need just here, just now run something one time, one day, maybe one week
10:20 yko but my sad experience tells me - don't use hacks on module/framework that's not maintained by you
10:20 yko or kill maintainer. so nobody will make changes that brakes your hacks D
10:20 yko *breaks
10:22 * yko should go to dentist again.
10:22 marcus mmm
10:22 marcus yko: been in a fight?
10:22 yko no. not being at dantist for seven years
10:22 yko *dentist. damn
10:23 marcus ouch
10:23 marcus I go every year
10:24 yko good for you. i had hard times and i'm afraid of dentists till being child
10:24 marcus it probably makes it worse ;)
10:24 yko s/till/from?/ # admn, i forget how to speak english at all
10:24 marcus since
10:25 yko damn, you are right
10:25 marcus since my childhood
10:25 yko vacations made me weak
10:25 marcus yepp
10:25 marcus no more vacation for you!
10:25 sigue left #mojo
10:26 * yko sadly goes back to the alligator pit.
10:26 sigue joined #mojo
10:38 * marcus gets a headache from trying to solve this shit
10:38 marcus not a good task for first day after vacation. rewrite ws from resultset style to callbacks
10:39 marcus actually, I would really like to rewrite it to sax style chunk parsing
10:39 marcus but it seems very hard with this api
10:41 marcus think maybe I should stash this away for now and focus on user value
10:43 spleenjack joined #mojo
11:03 sigue left #mojo
11:03 sigue joined #mojo
11:07 Foxcool left #mojo
11:08 Foxcool joined #mojo
11:08 Sugar|2 joined #mojo
11:13 Sugar left #mojo
11:15 Foxcool left #mojo
11:29 Foxcool joined #mojo
11:45 ispy_ left #mojo
11:46 sigue left #mojo
11:46 sigue joined #mojo
11:47 jamesw left #mojo
11:48 jmmills left #mojo
11:52 dotan joined #mojo
11:53 sigue left #mojo
11:53 sigue joined #mojo
11:57 fmerges joined #mojo
12:00 Foxcool left #mojo
12:00 jamesw joined #mojo
12:06 alnewkirk joined #mojo
12:08 yegor joined #mojo
12:25 Foxcool joined #mojo
12:26 ispy_ joined #mojo
12:31 * sri yawns
12:31 sri marcus: there is no blocking style non-blocking api anymore, that was the whole point of ua
12:32 sri so if you want non blocking you have to do it right
12:32 sri http://mojolicio.us/perldoc?Mojolic​ious/Guides/Cookbook#Non2DBlocking
12:33 marcus I don't really want non-blocking tho
12:33 marcus I just want parallel requests
12:34 sri that's the same
12:35 sri the ioloop is responsible for making it look blocking
12:35 sri you could build a wrapper around the example to build a new blocking parallel client
12:36 sri (blocking looking)
12:36 marcus not sure I understood that
12:36 marcus but it sounds like what I want
12:37 sri well, the primitives are still there, just the blocking parallel request api is gone
12:43 Sugar joined #mojo
12:44 marcus so I just need to use a separate ioloop and call ->stop in the last request?
12:44 sri exactly
12:46 Sugar|2 left #mojo
12:48 sri actually, i think it is not that easy
12:48 sri you might need "local $Mojo::IOLoop::Loop = Mojo::IOLoop->new;"
12:52 sri "local $Mojo::IOLoop::LOOP = Mojo::IOLoop->new;" actually
12:53 sri since ua now always uses the singleton if there's a callback
12:54 kaare left #mojo
12:59 abra joined #mojo
12:59 PerlNinja joined #mojo
13:03 yegor_ joined #mojo
13:05 marcus will try
13:07 yegor left #mojo
13:08 yegor_ is now known as yegor
13:09 Sugar|2 joined #mojo
13:12 abra left #mojo
13:13 Sugar left #mojo
13:16 metaperl_ left #mojo
13:16 metaperl left #mojo
13:19 abra joined #mojo
13:25 kaare_ joined #mojo
13:49 metaperl joined #mojo
13:49 metaperl_ joined #mojo
13:52 vel joined #mojo
13:54 metaperl left #mojo
13:55 amoore joined #mojo
14:03 yegor_ joined #mojo
14:04 yegor left #mojo
14:07 AmeliePoulain left #mojo
14:09 yegor_ is now known as yegor
14:09 marcus my app is just hanging :/
14:13 sh4 left #mojo
14:15 fhelmber_ left #mojo
14:15 fhelmber_ joined #mojo
14:19 PerlNinja marcus: whip it and tell it to stop being lazy
14:20 * sri cracks the whip
14:21 PerlNinja o_O
14:21 PerlNinja so much whipping going on in here
14:23 Foxcool left #mojo
14:24 marcus guess I've broken the main ioloop or something :)
14:25 sri don't do that
14:26 marty left #mojo
14:27 marcus but it's so fun!
14:27 marcus or something
14:32 PerlNinja bad marcus! bad! no cookie for you
14:33 marcus I give up
14:33 marcus fuck this with a shovel
14:33 PerlNinja o_O
14:33 PerlNinja beer?
14:33 purl somebody said beer was fluessiges Brot or the opiate of the working classes
14:33 marcus bikes
14:33 PerlNinja muscle powered?
14:34 marcus yes
14:36 Sugar|2 is now known as Sugar
14:37 marty joined #mojo
14:38 PerlNinja marcus, those are scary :P
14:38 yko Haha! I'm one tooth bigger now!
14:40 ajgb left #mojo
14:41 sri marcus: not sure what your problem is, works flawless for me
14:47 marcus sri: probably something stupid I am doing. :/
14:48 sri marcus: https://gist.github.com/942395
14:52 marcus sri: thanks. will try to adjust my stuff
14:58 Foxcool joined #mojo
14:58 sri feel free to turn it into a plugin or so
14:58 sri i suppose with a bit of clever hacking you can also make it benefit from keep alive
15:00 Christian i wish all of you a nice evening
15:00 * sri waves
15:00 * moritz particles
15:00 purl hmmm... particles is fine
15:01 sri purl: stupid bot
15:01 purl stoopid hyooman
15:02 Christian left #mojo
15:03 yegor_ joined #mojo
15:07 yegor left #mojo
15:07 mattastrophe joined #mojo
15:09 sri http://www.railshotline.com/ # lol
15:16 metaperl_ $self->param->to_hash or $self->to_hash # is the way to get the form data in a hashref?
15:17 metaperl_ I guess I could grep the test suite to answer this
15:19 sri it would be documented if it was
15:19 sri $self->req->params->to_hash
15:22 cosmincx left #mojo
15:22 yko left #mojo
15:23 * PerlNinja starts a mojolicious hotline that plays a message telling people to just come on IRC
15:31 Sugar =))
15:32 sri text2speech!
15:35 dotan left #mojo
15:38 ispy_ left #mojo
15:44 yko_ joined #mojo
15:45 yko_ is now known as yko
15:47 Ranguard joined #mojo
15:47 Ranguard left #mojo
16:04 yegor__ joined #mojo
16:04 elb0w fuck paypal
16:05 yegor_ left #mojo
16:05 tabbi joined #mojo
16:14 tabbi left #mojo
16:16 kmx left #mojo
16:29 PerlNinja elb0w: yes
16:29 PerlNinja also: fuck mongodb at the moment
16:33 kmx joined #mojo
16:41 ysyrota left #mojo
16:46 ilbot2 joined #mojo
16:46 Topic for #mojo is now join the revolution ♥ http://mojolicio.ushttp://irclog.perlgeek.de/mojo/today
16:55 Sugar left #mojo
16:55 yegor__ left #mojo
16:55 * mateu has been happily using mongodb, but hasnt' messed w/ sharding.
16:57 fmerges left #mojo
17:12 PerlNinja sharding is nice, easy :)
17:12 PerlNinja sharding to shards that are replica sets.. still nice and easy
17:13 PerlNinja the perl driver for some reason deciding that halfway during an operation it will bomb out stating "can't connect to db' while a cursor is trying to get a next value from the query... now that really ticks me off
17:15 yegor joined #mojo
17:16 yegor left #mojo
17:17 yegor joined #mojo
17:21 Kallisti joined #mojo
17:31 fhelmber_ left #mojo
17:41 sh4 joined #mojo
17:51 spleenjack left #mojo
17:54 tabbi joined #mojo
18:04 yegor_ joined #mojo
18:06 yegor left #mojo
18:23 perlrocks Twitter: "Thinking about a Socket.IO plugin for #mojolicious" --perlninja http://twitter.com/perlninj​a/status/62945001570713600
18:24 sri think xantus had one
18:29 PerlNinja :D
18:29 PerlNinja not on cpan he doesn't
18:29 sri yea, he didn't like socket.io
18:29 PerlNinja mmm me neither
18:29 PerlNinja but with that websocket spec being fluid as hell still
18:29 sri neither do i
18:30 PerlNinja it's the closest thing to being able to do websocket-like stuff without actually needing a real websocket
18:30 sri well, the latest spec changed only minimal, browsers should update again soon
18:30 sri firefox minefield already has ietf-07 websockets
18:31 PerlNinja myeah chrome stable hasn't got it
18:31 PerlNinja can't rely on it if chrome, safari and ff don't all work on the same spec
18:31 sri chrome stable still has ancient websockets, the ones with security holes
18:31 PerlNinja yeh
18:32 sri you can't rely at all on websockets yet
18:32 sri your best bet is plain old long polling
18:32 sri which we already do
18:33 PerlNinja long polling makes my testicles itchy :/
18:33 sri i doubt very much that you could reliably use socket.io in production
18:33 Foxcool left #mojo
18:33 PerlNinja mm even if I could I don't think I'd want to
18:34 sri ye, just go with long polling
18:34 PerlNinja the node.js crowd's all over it which means it's gotten covered in ick
18:34 sri i've only seen toy apps using it so far
18:34 sh4 left #mojo
18:36 PerlNinja heheh
18:36 PerlNinja yeah dunno, right now i'm just firing ajax requests on a timer
18:36 PerlNinja since i have this regular thing that needs updated, long polling would mean i'd have to block a connection for a while so that just eats up workers
18:37 PerlNinja or go that whole comet route
18:37 sri huh
18:37 sri why would a connection eat up workers?
18:38 sri hypnotoad workers accept 1000 connections each by default
18:39 sri if you need to scale for something serious you could also just use nginx with the comet module
18:39 PerlNinja eh.. workers.. not that.. damn my head is full of junk
18:40 PerlNinja it eats up connections and i dont like pushing nginx past 4000 connections
18:40 sri http://pushmodule.slact.net/
18:40 PerlNinja where it might end up
18:40 sri would be just the same with websockets
18:41 PerlNinja that reminds me there was something else i needed to figure out on nginx...
18:41 PerlNinja yeah true.. .. wtf am i thinking
18:41 PerlNinja someone stole my brain!
18:41 sri and it's not like idle keep alive connections are much less expensive
18:42 sri regular ajax pull via keep alive connection can be just as costly
18:42 sri possibly worse with connect overhead
18:44 PerlNinja yea it is
18:45 PerlNinja that's the problem with building an app where the process itself has taken close to a year - in that year, technology tends to pass you by. Already made a big leap of faith at one point when I got fed up with Catalyst and ported the whole thing to Mojolicious ... which worked out great :D .. but don't think my partners are ready for me to go off on another "let's replace X with Y" tangent :)
18:46 sri i would most likely not use any realtime features in a production web app yet unless it really depends on it
18:46 PerlNinja it will, eventually
18:46 sri if ajax pull is good enough it is good enough
18:46 PerlNinja collaborative bits and pieces that aren't there because websockets aren't ready
18:47 PerlNinja ajax pulls now are just for pulling in notifications and other events that need to do something to the UI
18:48 tabbi left #mojo
19:04 yegor__ joined #mojo
19:09 yegor_ left #mojo
19:22 yegor__ left #mojo
19:25 vividsnow joined #mojo
19:32 perlrocks Twitter: "decided to learn Mojolicious" --girishbhat http://twitter.com/girishbh​at/status/62962207276802048
19:59 sri https://github.com/blog/844-​forking-with-the-edit-button # github keeps getting better and better
20:05 vividsnow_ joined #mojo
20:09 sugar joined #mojo
20:10 vividsnow left #mojo
20:10 vividsnow_ is now known as vividsnow
20:30 vividsnow left #mojo
21:00 sugar left #mojo
21:02 amoore left #mojo
21:16 perlrocks Twitter: "Mojolicious-Plugin-Mongodb 1.07 by MADCAT - http://frepan.org/~madcat/Moj​olicious-Plugin-Mongodb-1.07/" --cpan_new http://twitter.com/cpan_ne​w/status/62988427712397312
21:16 vel left #mojo
21:44 kaare_ left #mojo
22:03 metaperl_ Regarding http://search.cpan.org/~kraih/Mojoliciou​s-1.21/lib/Mojolicious/Guides/Cookbook.p​od#Making_Your_Application_Installable   ... wouldnt a layouts directory also need to be moved under lib/MyApp?
22:03 sri it's templates/layouts
22:04 metaperl_ i see
22:15 perlrocks Twitter: "Mojolicious-Plugin-Bundle 0.001 by SIDD - http://frepan.org/~sidd/Mojo​licious-Plugin-Bundle-0.001/" --cpan_new http://twitter.com/cpan_ne​w/status/63003284595679233
22:16 sigue left #mojo
22:16 sigue joined #mojo
22:18 perlrocks Twitter: "Mojolicious-Plugin-Bundle 0.002 by SIDD - http://frepan.org/~sidd/Mojo​licious-Plugin-Bundle-0.002/" --cpan_new http://twitter.com/cpan_ne​w/status/63004057542991873
22:20 metaperl_ left #mojo
22:28 sigue left #mojo
22:28 sigue joined #mojo
22:57 PerlNinja rawrrgh
23:01 * PerlNinja pokes the mongodb perl driver in the eye with a few sharp sticks before nailing the damn thing to the wall
23:04 sigue left #mojo
23:17 sigue joined #mojo
23:34 Kulag left #mojo
23:43 sigue left #mojo
23:43 sigue joined #mojo
23:47 Kulag joined #mojo
23:51 sigue left #mojo
23:51 sigue joined #mojo

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