Camelia, the Perl 6 bug

IRC log for #mojo, 2012-05-01

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

All times shown according to UTC.

Time Nick Message
00:06 mire joined #mojo
00:22 rem_lex|pivo joined #mojo
01:02 jnap joined #mojo
01:44 xaka1 joined #mojo
02:03 d4rkie joined #mojo
02:14 SmokeMachine joined #mojo
02:28 irq_ joined #mojo
02:29 mattastrophe joined #mojo
02:34 asarch joined #mojo
03:23 rem_lex| joined #mojo
04:06 yko joined #mojo
04:12 yko joined #mojo
04:19 yko joined #mojo
04:35 inokenty joined #mojo
04:36 Foxcool joined #mojo
04:38 yko joined #mojo
04:58 rem_lex|pivo joined #mojo
05:24 yko joined #mojo
05:28 Htbaa joined #mojo
05:32 Foxcool joined #mojo
06:03 SmokeMachine joined #mojo
06:42 rem_lex| joined #mojo
06:43 Vandal joined #mojo
07:15 cosmincx joined #mojo
07:38 Foxcool joined #mojo
08:33 arpadszasz_ joined #mojo
09:22 arpadszasz joined #mojo
09:29 ladnaV joined #mojo
09:56 cosmincx joined #mojo
09:58 phips joined #mojo
10:00 batman joined #mojo
10:01 xaka1 joined #mojo
10:01 phipsy joined #mojo
11:07 batman joined #mojo
11:55 fibo joined #mojo
12:08 lbr MisterHatt: if you want really fast json, you probably don't want to use Mojo::JSON
12:09 lbr JSON::XS is about 40 times faster on my laptop
12:10 lbr (of course, i use Mojo::JSON lots of places, because that's not a real bottleneck most of the time)
12:13 inokenty There is also Mojo::JSON::XS
12:13 * lbr did not know that :)
12:13 inokenty Faster than Mojo::JSON but slower than JSON::XS
12:25 MisterHatt lbr: it's all a perl interface though
12:25 MisterHatt in the end I actually went for neither perl nor ruby
12:25 MisterHatt backbone.js was the green light instead
12:38 lbr so your python is javascript ;)
12:44 arpadszasz_ joined #mojo
12:49 Akron joined #mojo
12:57 Akron I've got a simple M::L app https://gist.github.com/2567746 When I start it with hypnotoad -f app.pl, I don't hit different workers each time but each worker 25 times in a row and then switch to another worker. What am I doing wrong?
12:58 Akron Mojolicious 2.91, Perl Version 5.12.4 on Linux. On another machine it seems to work fine.
13:00 inokenty If there is the javascript for server side now, I want the Perl on client side =) <script type="text/perlscript">say 'Hello world'</script>
13:04 MisterHatt inokenty: and the browser hacks with use Perl or die;
13:04 batman joined #mojo
13:13 sri Akron: nothing i guess
13:14 Akron sri: Broken Perl?
13:14 sri no idea, or flock just behaving slightly different on one box
13:15 sri Akron: https://github.com/kraih/mojo/blob/mas​ter/lib/Mojo/Server/Hypnotoad.pm#L317
13:15 Akron sri: Okay, thank you. I think it works fine here a couple of days ago so I thought maybe I screwed up the config or something.
13:15 sri that is the code that decides which worker gets the next connection
13:16 sri or wait
13:16 sri do you make all requests over the same connection? :)
13:16 sri keep alive *cough*
13:16 Akron Ah! Yeah, that maybe the case!
13:17 sri http://mojolicio.us/perldoc/Mojo/Se​rver/Hypnotoad#keep_alive_requests
13:17 sri try setting that to 1
13:17 Akron That's it! Thank you!
13:17 sri :)
13:17 sri 1 is very slow though
13:19 Akron sri: This is just for testing issues - thank you!
13:19 sri and might not work perfectly, since we multi accept too
13:19 sri http://mojolicio.us/perldoc/M​ojo/Server/Hypnotoad#clients # set this to 1 too
13:19 sri then you basically turn hypnotoad into a prefork apache :)
13:20 sri (a very bad one... since apache does keep alive)
13:21 * sri wonders if he should make multi accept more configurable too
13:23 sri prolly not, or does anyone here grok https://github.com/kraih/mojo/blo​b/master/lib/Mojo/IOLoop.pm#L238 ?
13:23 Akron sri: Nice - thanks. It's for testing database connection pools and caching etc ... I was running FastCGI exclusively until now - and starting to love Hypnotoad.
13:25 sri nginx only has an on/off switch for multi-accept, hmmm
13:37 jnap joined #mojo
13:46 asarch joined #mojo
14:09 Leandr joined #mojo
14:19 batman joined #mojo
14:42 fibo joined #mojo
15:37 xaka1 joined #mojo
15:42 bowtie joined #mojo
15:42 bowtie hi
15:46 bowtie 1 Q if you don't mind
15:46 bowtie 2 http://mojolicio.us/perldoc​/Mojolicious#BUNDLED_FILES
15:46 bowtie 3 Mojolicious ships with a few popular static files bundled in the public directory.
15:46 bowtie 4 ie jQuery
15:46 bowtie 5
15:46 bowtie 6 so how do I access jquery the mojolicious way?
15:46 bowtie 7 because if I mojo generate all Thing
15:46 bowtie 8 it dose not have jquery.js in public
15:46 bowtie 9
15:47 sri 1 hi
15:47 sri 2 the bundled jQuery is for internal use
15:48 sri 3 i'm gonna change the description to make it less misleading
15:48 bowtie sri, ok, so I just do as I would normally then, thanks
15:48 Psyche^ joined #mojo
15:48 sri 4 yes
15:49 bowtie sri, yes, POD needs a kick :)
15:51 bowtie sri, info only 'Topic for #mojo is Mojolicious real-time web framework 🐫 http://mojolicio.us 🐫 http://irclog.perlgeek.de/mojo/today' yuck charters
16:04 GitHub119 joined #mojo
16:04 GitHub119 [mojo] kraih pushed 1 new commit to master: http://git.io/wJpy_w
16:04 GitHub119 [mojo/master] documentation tweaks - Sebastian Riedel
16:04 GitHub119 left #mojo
16:04 sri that should make it more clear
16:07 bowtie sri,  it's now 2012 :)
16:08 sri ?
16:08 bowtie sri, i look at your POD tweak
16:08 * sri doesn't follow
16:09 bowtie +  Copyright (C) 2011, John Resig.
16:09 bowtie but it's now 2012
16:10 sri you don't update a copyright year when the code didn't change
16:10 sri the google notice below even says 2006
16:11 d4rkie joined #mojo
16:11 bowtie sri, ok, thanks for that
16:21 qbit left #mojo
16:27 brambles joined #mojo
17:38 bowtie left #mojo
18:39 arpadszasz_ joined #mojo
19:40 irq_ joined #mojo
19:44 * tempire snaps
19:55 ryanc joined #mojo
20:12 * sri growls
20:19 Leandr joined #mojo
20:29 marcus https://twitter.com/shadowcat_m​st/statuses/197378336840892418 tempted to call something bloated... *sits on hands*
20:34 tempire probably best
20:38 marcus Unless my hands turn blue and fall off.
20:57 tempire bloated hands
20:58 tempire the complexity of your hands is stupid
21:02 marcus yeah, should just replace them with axe heads
21:02 marcus might complicate operating a keyboard tho
21:04 tempire or boxing gloves
21:13 jwang joined #mojo
21:14 jnap joined #mojo
21:22 mire joined #mojo
21:23 xaka1 joined #mojo
21:31 batman joined #mojo
21:50 lukep joined #mojo
22:26 fibo joined #mojo
22:26 marcus jazz nyan might be the death of me. it's simply just not annoying enough for me to turn it off.
22:27 tempire I feel that way about all nyans
22:27 tempire I could honestly listen to them for hours
22:28 tempire but songify has replaced it currently
22:28 tempire in particular, this has been on loop for at least a week: http://www.youtube.com/watch?v=bFEoMO0pc7k
22:28 tempire also, the new sketch is the fantasmo
22:28 tempire love. it.
22:30 marcus neat
22:31 marcus Ain't nobody got time for that
22:31 marcus is it ported to ios yet? :)
22:31 tempire it's a far!
22:31 * sri too
22:41 atmh joined #mojo
22:42 atmh left #mojo
22:43 atmh_ joined #mojo
22:45 atmh_ hi there, I was contemplating using Mojolicious for a WS broker service and was wondering how well does Mojo cope with concurrent connections
22:46 atmh_ WS=websocket
22:47 marcus I always thought it was Web Service.
22:47 marcus :-P
22:47 marcus atmh_: it should be doing pretty well. Did you have trouble?
22:48 atmh_ no no, I didn't, I was just wondering .. I'll be coding away with a test tonight to see how much it'll handle
22:50 atmh_ mind you, I'm actually new to Mojolicious but think it's pretty darn impressive.
22:51 marcus :)
22:51 mire joined #mojo
22:55 sri memory cost per websocket connection might still be a bit high, since we keep the handshake transaction around, but that could be optimized away if it turns out to be a problem
22:55 marcus memory is dirt cheap tho
22:57 atmh_ hmm, okay, I'll have to actually write some code to mimic some 100 simultaneous connections to see that with my own eyes
22:57 atmh_ yes, 100 only for now
22:57 sri 100 is nothing ;p
22:57 atmh_ :)
22:59 marcus sri: I seem to remember you saying you'd tested it with 80k connections on a single machine?
23:00 sri don't think so, could have been xantus
23:01 sri the limiting factor will always be memory though
23:02 sri parsing and building websocket messages is dirt cheap and perl i/o quite speedy
23:02 sri with EV backend you also get epoll on linux
23:05 sri epoll gets important with thousands of concurrent connections since websockets tend to be idle most of the time
23:08 atmh_ cool, if any one has got upto 80k, it's quite impressive
23:58 d4rkie joined #mojo

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