Camelia, the Perl 6 bug

IRC log for #mojo, 2011-09-25

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

All times shown according to UTC.

Time Nick Message
00:30 andrefs joined #mojo
01:05 MojoGuest821 joined #mojo
01:05 MojoGuest821 From: http://news.ycombinator.com/item?id=1277067 (3 hits)
01:10 jwang joined #mojo
01:23 y0j_ joined #mojo
01:34 andrefs joined #mojo
02:09 MojoGuest333 joined #mojo
02:09 MojoGuest333 From: http://www.google.com/url?sa=t&source=w​eb&cd=4&ved=0CC4QFjAD&url=http%​3A%2F%2Fdev.xantus.org%2F&rct=j&q=w​ebsockets%20irc&ei=rIx-TqPiFIm2twfoo6lw​&usg=AFQjCNGw6nZv0wO42uzUnHhWcD0vcKNULg (1 hits)
02:47 andrefs joined #mojo
02:48 perlrocks Twitter: "MCPC: Mojolicious on Google Chrome Web Browser:: http://t.co/yIsx1lvZ to view Twitter streams in real time from the Lite server connected WebSocket" (ja) --zipangulog http://twitter.com/zipangulo​g/status/117792474306584576
04:01 perlrocks Twitter: "HAP is quite useful exchange but I write in Mojolicious. Templates and best" (ja) --wareohji http://twitter.com/wareohji​/status/117810979731013632
04:11 tempire I don't have any better names than trigger.
04:12 tempire I don't think promise/future are suited for it, because the syntax doesn't immediately imply that functionality.
04:12 tempire and those terms aren't widely known.
04:12 tempire but "trigger" makes sense as a layer on top of an event emitter.
04:18 mohan joined #mojo
05:27 kaare joined #mojo
06:22 Foxcool joined #mojo
06:27 Vandal joined #mojo
06:52 sromanov joined #mojo
07:06 perlrocks Twitter: "Hacked together a native AnyEvent backend for #mojolicious, anyone feel free to turn it into a #cpan module. :) http://t.co/QE5ZFdNj #perl" --kraih http://twitter.com/kraih/status/117857398491852800
07:35 marcus looks like a pwned a server by running hypnotoad hot deployments every minute in a while loop :-/
07:37 sri that does not sound like such a good idea
07:40 perlrocks Twitter: "Testing with Mojo is great fun. The Mojolicious framework provides an easy way to test your web applications ... http://t.co/3tgxYtv5" --Lucillapzcsp http://twitter.com/Lucillapzc​sp/status/117866001458135040
07:40 perlrocks Twitter: "Test::Mojo – Intro into testing with CSS3 selectors - pkgbox: Testing with Mojo is great fun. The Mojolicious ... http://t.co/MvewqvrE" --Rickipvocj http://twitter.com/Rickipvoc​j/status/117866014049452032
07:43 sri hmm
07:44 * sri wonders why graceful_timeout and upgrade_timeout are the same by default
07:48 GitHub155 joined #mojo
07:48 GitHub155 [mojo] kraih pushed 1 new commit to master: http://git.io/1tdKsw
07:48 GitHub155 [mojo/master] changed default upgrade timeout of Hypnotoad from 30 to 60 seconds - Sebastian Riedel
07:48 GitHub155 left #mojo
07:53 sri the default values are still pretty optmistic, for real production setups maybe minutes would be better
08:00 sri marcus: if you really want to cron hot deployment i would suggest giving it an hour
08:01 sri with sensible timeouts in minutes
08:03 sri i'm certain there are still a few rare but possible race conditions in hypnotoad that can be avoided that way
08:40 arthas joined #mojo
08:41 arthas joined #mojo
08:49 driller_work joined #mojo
08:56 marcus sri: seems it ended up with 6500 zombie processes
08:56 marcus that broke the server
08:56 crab i guess the server wasn't zombie-hardened
09:10 marcus guess zo
09:10 driller_work It's the new thing; ubuntu comes with zombies out-of-the-box
09:13 marcus it freebsd
09:13 marcus it is
09:13 driller_work did the zombie processes really *break* your server?
09:14 driller_work thought the kernel would failsafe and shutdown in those type of scenarios; if not the kernel then at least the hardware
09:31 sri marcus: weird
09:34 sromanov joined #mojo
09:38 henq joined #mojo
09:40 henq sunday
09:47 sri marcus: it does sound like failed upgrades
09:48 sri which might actually get fixed by my commit earlier, in case you're using the default settings
09:48 sri once the upgrade timeout runs out the new parent gets a SIGKILL from the old one
09:49 sri all children of the new parent would then commit suicide (if there are any)
09:49 sri which could leave behind zombies i suppose
09:52 sri then again, something else would have to go wrong too
09:53 sri since once the old parent gets a SIGQUIT from the new one the upgrade timeout is disabled
09:53 sri your app startup time would have to be pretty close to the limit
09:57 sri i'm just speculating though
10:01 sri could just as well be something freebsd specific we don't handle correctly
10:02 sri i've just been while 1 restarting a hypnotoad on os x for 15 minutes at 500 rps without problems
10:02 sri so who knows
10:07 sri btw. i don't suggest while 1 restarting :)
10:11 sri oh wait
10:11 sri marcus: how long was the server running?
10:21 sri the number of zombies should be equal to the numer of restarts
10:22 marcus sri: only like 10 hours
10:22 marcus not sure what happened
10:23 sri hmm
10:23 sri that would be just 600
10:26 sri marcus: can you test something?
10:27 marcus yeah?
10:27 purl totally!
10:27 sri http://pastie.org/2588426
10:27 sri it's just a hunch
10:32 marcus sri: not able to reproduce zombies atm, but this was a bit strange - http://paste.scsys.co.uk/144435
10:33 marcus note #5 there
10:33 sri marcus: that's perfectly fine
10:33 marcus ok
10:34 sri it's the split second when the pid gets deleted and recreated by the new master
10:35 sri looking at the code again… my patch shouldn't really have an effect :S
10:35 sri the restart exec should trigger setsid anyway
10:37 sri marcus: just to be sure, you can't create zombies with and without the patch, right?
10:38 marcus sri: right
10:39 marcus sri: might have been something hardware related on the server that made it generate a lot of zombies
10:39 sri ok, we'll call it a fresk accident then :)
10:39 sri *freak
10:43 sri the new upgrade timeout is good though
11:06 noganex_ joined #mojo
11:17 janus hi
11:18 janus suggestion: replace jquery with d3 - it seems to be a more natural fit in many aspects :)
11:20 janus revolution++ # if you like ;)
11:22 Kulag joined #mojo
11:34 crab why bother?
11:42 preflex_ joined #mojo
12:54 sri marcus: how did you restart every minute?
12:55 marcus sri: while true do; git pull; hypnotoad --config /app/hypnotoad.conf /app/script/app; sleep 60; done
13:00 sri ok
13:03 sri that's how i test too
13:04 sri usually without sleeping though
13:38 PerlNinja joined #mojo
13:38 PerlNinja g'day mojo-ers
13:39 andrefs joined #mojo
13:45 driller_work Hi!
13:50 * sri waves
13:59 PerlNinja anyone have any comet examples that involve Mojolicious? Specifically things that involve lots of happy non-blocking action?
14:00 sri https://gist.github.com/1130253 # i recently wrote this to demonstrate mojolicious+anyevent+comet
14:01 sri http://blog.kraih.com/mojolicious-​hack-of-the-day-html5-eventsource # and of course there's this recent blog post
14:03 PerlNinja the latter one is my favorite but i can't use it since.. *sigh* there's the need for IE compatibility, and IE doesn't support server-side events
14:03 PerlNinja unless that happened to change in the last month or so
14:05 PerlNinja oh... wow
14:05 PerlNinja caniuse.com got itself hacked it seems
14:09 mohan_ joined #mojo
14:12 perlrocks Twitter: "/ It's hard to flow like smiling videos "MCPC: Mojolicious on Google Chrome Web Browser:: Display smiling like a Twitter stream video in real time from the Lite server connected WebSocket" http://t.co/X9O2UMir" (ja) --CLCLCL http://twitter.com/CLCLCL/​status/117964705833828352
14:21 perlrocks Twitter: "Mojolicious-Plugin-Authentication 1.19 by MADCAT - http://t.co/REMbrNrz" --cpan_new http://twitter.com/cpan_new​/status/117966934510800897
14:22 PerlNinja there done my good deed for today
14:42 perlrocks Twitter: "MCPC: Mojolicious on Google Chrome Web Browser:: http://t.co/1qJxc143 Show smiling like a Twitter stream video in real time from the Lite server connected WebSocket" (ja) --wPerlnews http://twitter.com/wPerlnew​s/status/117972136768978944
14:43 smpb joined #mojo
14:54 crab perlninja: does http://toroid.org/ams/etc/​mojolicious-http-streaming count?
14:58 abra joined #mojo
14:59 andrefs joined #mojo
15:11 PerlNinja hey crab, yeah that gets close
15:11 GitHub94 joined #mojo
15:11 GitHub94 [mojo] kraih pushed 1 new commit to master: http://git.io/Si3sTA
15:11 GitHub94 [mojo/master] improved connection_timeout method in Mojo::IOLoop by allowing it to be called as a class method - Sebastian Riedel
15:11 GitHub94 left #mojo
15:11 PerlNinja I'm actually trying to build a bridge between JS and RabbitMQ ... I could cheat and use something like Meteor in the middle but I'd rather have a single web app that takes care of being said bridge, since there's an AnyEvent::RabbitMQ and Mojolicious does AnyEvent... etc. etc. etc. :D
15:28 sri Mojo::IOLoop::AnyEvent is still up for adoption *hint hint*
15:29 sri s/IOLoop/IOWatcher/
15:37 sri guess everybody is happy with the EV backend :)
15:38 PerlNinja I haven't done anything with it to be honest ... having a bit of a time wrapping my head around asynchronous stuff :P I'm still a member of the oldschool IO::Select club :/
15:38 sri :o
15:38 sri async is hard, lets go shopping!
15:40 driller_work argh, cant find the multi language function for mojolicious in the docs, was called something with i8..:?
15:41 driller_work someone please refresh my memory :(
15:43 sromanov driller_work, https://github.com/kraih/mojo/blob/m​aster/lib/Mojolicious/Plugin/I18N.pm may be what you're looking for
15:43 driller_work exactly what I was looking for! many thanks sromanov
15:46 preflex_ joined #mojo
15:50 PerlNinja sri: good idea, you buying? :P I'm in dire need of a MM Stingray Classic 4 :P
15:50 sri PerlNinja: no, you are!
15:51 PerlNinja oh... well damn :(
15:52 sri \o/
15:52 PerlNinja what's a guy gotta do to get a new bass guitar?!
15:52 PerlNinja :P
15:55 perlrocks Twitter: "Mojolicious:: 2 that I shed on Twitter Streaming API Lite + WebSocket + GoogleChrome (0:52) http://t.co/8aDKFzhE" (ja) --CLCLCL http://twitter.com/CLCLCL/​status/117990682366447616
16:12 GitHub87 joined #mojo
16:12 GitHub87 [mojo] kraih pushed 1 new commit to master: http://git.io/xUIaGA
16:12 GitHub87 [mojo/master] added another connection test - Sebastian Riedel
16:12 GitHub87 left #mojo
16:34 crab what is an mm stingray classic 4? sounds like an suv
16:40 grim_fandango joined #mojo
16:40 PerlNinja crab: a bass guitar
16:41 PerlNinja correction... a really damn nice bass guitar
16:42 PerlNinja that my wife won't let me buy since she feels dropping about a grand and a half on a piece of wood and 4 metal strings is silly :(
16:44 driller_work hmm.. how do I do a ->to and stash something in one line? would: $r->route('/this/:that')->to('this#that', my $that = $self->stash('that')); do it? or is there a prettier way?
16:48 PerlNinja why would you want to do that?
16:48 PerlNinja also don't know if there's a prettier way, explain what you want to do :D
16:49 driller_work interesting, I think I am overcomplicating things for me now when you mention it. If I specify :that in the route wouldn't that convert to a variable ($that) I can use inside my controller, or am I mixing things up again?
16:51 driller_work I want to route /page/:uri to my controller, say ->to('page#index'), and within the controller see if :uri is a valid page (and if it exists), if it does, then render the requested template
16:52 driller_work but, I am not sure if I can pick up the :uri within the controller, how would I see what :uri contains? does it convert to a variable I can use?
16:53 driller_work err, I think I need to fallback to the video tutorials... so sure this was explained multiple times there
16:55 andrefs joined #mojo
16:58 PerlNinja driller, like this :)
16:58 PerlNinja $r->route('/page/:uri')->to('page#index', uri => 'index');
16:58 PerlNinja inside your controller
16:59 PerlNinja my $uri = $self->stash('uri'); # or $self->stash->{'uri'} ... both work
17:08 PerlNinja then either render with: $self->render(text => 'Ohnoes!', status => 404) for a not-found, or $self->render(template => ....) ... or however you want to handle that bit
17:28 sri Contagion was not as bad as i though
17:45 Eugene joined #mojo
17:46 GitHub144 joined #mojo
17:46 GitHub144 [mojo] kraih pushed 1 new commit to master: http://git.io/ho7S6g
17:46 GitHub144 [mojo/master] fixed typo - Sebastian Riedel
17:46 GitHub144 left #mojo
17:54 kaare joined #mojo
18:02 GitHub197 joined #mojo
18:02 GitHub197 [mojo] kraih pushed 1 new commit to master: http://git.io/HZz6eQ
18:02 GitHub197 [mojo/master] added experimental accepts attribute to Mojo::IOLoop::Server - Sebastian Riedel
18:02 GitHub197 left #mojo
18:13 driller_work PerlNinja: many thanks :)
18:22 andrefs joined #mojo
19:05 PerlNinja driller_work, no problem :)
19:08 andrefs joined #mojo
19:14 perlrocks Twitter: "We just won Oslo Startup Weekend with http://t.co/BvzTrCbG developed in Mojolicious #swoslo \o/" --marcusramberg http://twitter.com/marcusramb​erg/status/118040587550195712
19:16 sri marcus: congrats :)
19:45 PerlNinja woo!
19:48 andrefs joined #mojo
19:50 tempire hooray marcus!
19:51 tempire what does http://letmeinapp.com/ do?
19:52 tempire the line for the hong kong apple store is ridiculous
19:55 MojoGuest434 joined #mojo
19:55 MojoGuest434 From: http://t.co/xbW19y9q (8 hits)
20:07 y0j_ joined #mojo
20:29 MojoGuest743 joined #mojo
20:29 MojoGuest743 From: http://news.ycombinator.com/item?id=1277067 (4 hits)
20:31 MojoGuest743 left #mojo
20:33 MojoGuest790 joined #mojo
20:33 MojoGuest790 From: http://news.ycombinator.com/item?id=1277067 (5 hits)
20:45 mattastrophe joined #mojo
21:29 perlrocks Twitter: "A #perl 6 Mojolicious *would* be compelling, as it has few CPAN dependencies. That strategy works in my favor there." --chromatic_x http://twitter.com/chromatic​_x/status/118074551098802177
21:32 PerlNinja wait what
21:32 purl Sure, I can wait. I've got plenty of time. It's you mortal meat bags that need to hurry
21:33 PerlNinja this coming from the "omg mojolicious why u reinvent wheel" guy?
21:33 * PerlNinja goes to buy a lottery ticket
21:37 sri at some point he criticized me for reinventing a wheel i invented :)
21:42 PerlNinja Gotta love that stash :D
21:48 PerlNinja dumdedum... you are in a twisty little maze of mojolicious frontends, mongodb shards, rabbitmq triggered services and some random javascript, all alike...
21:52 elb0w` joined #mojo
21:52 elb0w` What should I name a plugin for DBIx::Connector? Mojo::Plugin::DBIx::Connector ?
22:36 PerlNinja Mojolicious:: ...
22:52 tempire elb0w: Mojolicious::Plugin::AFittingName
22:53 driller_work tempire: that's a nice site you linked earlier, your creation?
22:54 driller_work something like that would make an excellent demonstration app for mojolicious... I remember that some of you smoke about that earlier hehe
22:55 Kulag joined #mojo
23:23 tempire driller_work: not sure what site you're referring to, but most of my time is spent researching, so probably not mine.
23:27 driller_work http://letmeinapp.com/ <-
23:31 tempire that's marcus' company.
23:51 preflex_ joined #mojo

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