Camelia, the Perl 6 bug

IRC log for #mojo, 2010-08-02

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

All times shown according to UTC.

Time Nick Message
00:01 Alias_ joined #mojo
01:08 ashleydev joined #mojo
02:33 MojoGuest973 joined #mojo
02:34 MojoGuest973 left #mojo
02:35 janus joined #mojo
03:16 Foxcool joined #mojo
03:43 tempire Is the format no longer detected from the url & put into $self->stash ?
03:59 Alias It's curious that Mojo, which has no non-core deps, is 50% larger than Dancer, which does
03:59 Alias 16meg vs 11meg
03:59 Alias Just an FYI, not a complaint
04:03 tl joined #mojo
05:27 MojoGuest785 joined #mojo
05:42 mvuets joined #mojo
06:25 ysyrota joined #mojo
06:30 garu Alias: Mojo does a whole lot more than Dancer at this point :)
06:31 Alias garu: I like to look at it the other way around
06:32 Alias garu: Mojo can ONLY do a lot more than Dancer :)
06:32 Alias And you pay for it whether you use it or not
06:34 garu Alias: I see your point... but, could this be a time-memory tradeoff?
06:35 Alias If it is, it's only relevant under one particular constraint scenario
06:35 Alias Where time is the most important thing
06:36 garu like, say, web development? :)
06:37 garu I mean, if Mojo required instead of used things like websockets or ipv6 or json, a simple load would grab less memory, but it could potentially be slower during runtime
06:40 garu I'm not saying it's not a valid analysis at all, I'm just trying to think why/how/if it would be relevant for Mojo to try and reduce it's footprint
06:40 garu *its
07:13 vti sri: http://github.com/xoma/Russian-​translate-of-Mojolicious-guides !
07:13 garfield [ xoma's Russian-translate-of-Mojolicious-guides at master - GitHub ]
07:13 garfield http://xrl.us/bhuyjy
07:23 Alias garu: CGI vs mod_perl modality
07:23 Alias garu: Go look at the prefork.pm module
07:23 Alias It explains what I'm talking about
07:23 Alias It's one of the few features I'd actually love to have added to the core
07:23 Alias Some hint indicators that suggest the resource stance of the program
07:24 Alias So you can have lots of code change the way it loads up depending on if you want to conserve memory or CPU or pre-fork/post-fork costs etc
08:00 Alias_ joined #mojo
08:14 MojoGuest923 joined #mojo
08:14 MojoGuest904 joined #mojo
08:17 MojoGuest985 joined #mojo
08:18 alex joined #mojo
08:28 MojoGuest678 joined #mojo
08:30 MojoGuest888 joined #mojo
08:30 MojoGuest888 fsdfsd
08:30 MojoGuest757 joined #mojo
08:30 MojoGuest888 wwwwwwadasd
08:31 MojoGuest757 /me says wow!
09:17 MojoGuest988 joined #mojo
09:30 MojoGuest663 joined #mojo
09:41 sri Alias: mojo has way more features than dancer, they are not int he same league
09:43 sri Alias: about cgi, as long as we are fast and small enough, we are fast and small enough
09:56 sri tempire: format didn't change
10:01 vti sri: ask contacted me and said that he will release tt soon :)
10:01 vti *would
10:01 sri :)
10:01 ask Did it a couple of hours ago.  :)
10:02 vti ask++!
10:03 sri \o/
10:04 vti now tt supports data section and all new features :)
10:20 Alias_ joined #mojo
10:24 ask http://search.cpan.org/~ab​h/MojoX-Renderer-TT-0.40/
10:24 garfield [ Ask Bj�rn Hansen / MojoX-Renderer-TT-0.40 - search.cpan.org ]
10:24 garfield http://xrl.us/bhuyuf
10:53 MojoGuest472 joined #mojo
10:53 MojoGuest472 Hello
10:55 arthas joined #mojo
11:09 Foxcool joined #mojo
11:31 arthas joined #mojo
11:35 wsconsole joined #mojo
11:36 vti wsconsole: test
11:36 wsconsole wsconsole foo
11:36 sri heh
11:37 wsconsole chatting from wsconsole
11:37 wsconsole WebSockets!!!
11:38 sri websockets!
11:39 sri wonder if wsconsole works on windows
11:40 vti i really don't know how irc works
11:40 vti can't chat easily :D
11:41 wsconsole test
11:41 wsconsole switching to select isn't hard i hope
11:42 wsconsole {"msg":"hello"} :)
11:43 vti i joined another channel via ws :)
11:43 vti not secure! :)
11:44 wsconsole юникод
11:44 vti hm... :)
11:45 wsconsole joined #mojo
11:46 wsconsole те☁
11:46 wsconsole joined #mojo
11:47 wsconsole
11:47 sri :o
11:47 wsconsole Unicode!!! ☀
11:49 sri
11:49 sri i'm not sure if stdin/out can be non blocking on windows
11:49 wsconsole we can do while(1) :)
12:06 yko joined #mojo
12:08 MojoGuest841 joined #mojo
12:32 arthas_ joined #mojo
12:38 marcus  forever
12:38 sri \o/
12:45 dotan joined #mojo
13:05 vti sri: i have some problems with websockets on early start... i don't receive a message from a server when i connect, i will try to write a test, but i remember that you have one, maybe my subrequest is causing that problem
13:05 sri yes, i think that case has a test
13:06 vti sometimes it works, sometimes not
13:06 vti first i thought that the order of receive_messag callbacks matters, but not
13:06 sri need test case :)
13:06 vti i know..
13:06 vti :)
13:25 ashleydev joined #mojo
13:26 GitHub79 joined #mojo
13:26 GitHub79 mojo: master Sebastian Riedel * b89c997 (1 files in 1 dirs): added test - http://bit.ly/bWBB8P
13:26 GitHub79 left #mojo
13:26 sri http://www.youtube.com/watch?v=GzyWKWu4TLs # wow, one of the best songs ever
13:26 garfield [ YouTube - Puscifer - Rocket Man (Rocket Mantastic) ]
13:29 vti sri: are your tests async?
13:30 sri everything is async, we just emulate sync behavior
13:31 sri each test runs in its own loop though
13:32 vti well, ye.. but i think the problem is somewhere there... because here it doesn't work even when i send a message and close websocket connection, sometimes i get that message and sometimes not
13:33 arthas joined #mojo
13:33 vti tests work flawlessly
13:33 sri it's the exact same code, no idea what could be different
13:33 sri maybe try parallel websocket connections?
13:35 vti you mean test them? i have one simple websocket server and wsconsole
13:42 vti sri: http://core.org.ua/~vti/tmp/websockets.mov
13:43 vti 'foo' message appears randomly
13:44 vti it's not a test... but :)
13:45 stephan48 how you make such movies?
13:46 vti stephan48: screenflick
13:47 sri vti: nothing i could work with
13:49 vti maybe at least a bit of empathy? :)
13:49 * vti sighs deeply
13:49 * sri hugs vti
13:52 BlueT_ joined #mojo
14:03 und3f joined #mojo
14:15 sri vti: i might have something
14:15 sri the handshake seems to sometimes eat the first websocket message
14:16 sri (maybe)
14:16 vti how can we test it?
14:16 sri i'm not even sure what could trigger it
14:18 sri oooh
14:18 sri of course
14:21 sri the handshake transaction buffers the first message sometimes as a leftover, which we don't handle yet
14:21 sri easy to fix
14:25 vti nice find
14:25 sri there are some timing issues with the tests i'm unsure about
14:26 vti too fast or too slow?
14:26 sri slow events
14:26 sri but thats in the nature of an event loop
14:26 sri and the randomness of the websocket finish
14:27 sri you never know which peer closes the connection first
14:27 sri finish callbacks might be a few ticks late
14:28 sri it's not really tragic since you expect async behavior, but you might miss an event if you ->stop the loop early
14:29 vti hm
14:29 vti call finish in a timer?
14:30 sri Oo
14:30 vti few ms later :)
14:30 sri thats not how it works
14:31 vti i mean from a client side
14:31 * sri can't follow
14:32 vti or maybe when you finish and there is some data, just wait until it comes?
14:33 sri ???
14:33 sri there are multiple ways the connection can end
14:33 sri ideally a handshake, but it could also just close
14:41 vti i just should stop saying something stupid :D
14:48 nordicdyno joined #mojo
14:53 ispeak joined #mojo
15:04 sri vti: try reading the spec ;p
15:09 vti sri: so basically there is no solution? :-[
15:09 sri vti: wrong question
15:09 sri "is there actually a problem?"
15:12 vti oh
15:13 vti so what is the solution to the scenario when i want to report and error and close the connection? or server shoudn't close connection in any reason?
15:13 keith joined #mojo
15:16 ashleydev joined #mojo
15:16 sri what?
15:16 sri you can't prevent a connection close
15:17 vti of course not, but when i have to close a connection but want to be sure that user received my last message what should i do?
15:18 sri you can't guarantee anything
15:19 sri you would have to build your own protocol on top of websockets
15:23 sri think i have a good pattern to stop ioloops in async tests btw. ->idle_cb(sub { shift->stop })
15:24 vti and what about the first message just after connect? the same issue?
15:24 sri that will result in a clean finish handshake
15:24 sri what issue?
15:24 sri i'm having a hard time following you :/
15:25 vti the issue "you can't guarantee anything"
15:26 sri how do you guarantee that your http response reached the client?
15:27 vti i thought there was a tcp somewhere
15:28 sri thats only from kernel to kernel...but there are multiple buffers in between
15:28 sri the client could just not read at all
15:30 sri it's not udp, but things might still go wrong
15:31 vti ok, i got your point
15:34 dotan left #mojo
15:44 GitHub85 joined #mojo
15:44 GitHub85 mojo: master Sebastian Riedel * 51bd149 (3 files in 3 dirs): fixed a few websocket bugs and added tests - http://bit.ly/b5Dbqd
15:44 GitHub85 left #mojo
15:44 sri vti: that should solve your problems
15:45 sri also added MOJO_CLIENT_DEBUG=1
15:46 sri lets you see the raw messages that go over the wire
15:46 vti sri: yes it did, thank you
15:47 sri and the last really ugly method in Mojo::Client is finally clean :)
16:13 yko joined #mojo
16:22 MojoGuest922 joined #mojo
16:31 Akron joined #mojo
17:07 dotan joined #mojo
17:07 dotan left #mojo
17:44 crab is a new CPAN release coming soon?
17:50 ask joined #mojo
17:54 dotan joined #mojo
17:56 xantus_ morning
17:59 crab hi xantus.
18:07 MojoGuest667 joined #mojo
18:07 MojoGuest667 test
18:10 xantus_ crab: hi
18:11 xantus_ I have a prototype twitter irc bot if anyone would like to help with it
18:11 xantus_ its POE + Mojo + Couchdb
18:20 Akron left #mojo
18:40 xantus_ seen esskar
18:41 xantus_ ...bummer
18:41 xantus_ we need a bot :)
18:44 marcus 10:38 < esskar> np
18:45 xantus_ here?
18:45 marcus Day changed to 31 Jul 2010
18:45 marcus yes
18:45 xantus_ ah
18:45 marcus on the 30th
18:46 xantus_ k, maybe I'll catch him some time
18:46 xantus_ marcus: how's it going?
18:46 xantus_ went to dublin recently I see
18:46 marcus xantus_: It's going OK. Just started work properly again today
18:46 marcus had to stay with eva due to daycare being closed last week
18:47 marcus Watching boondocks atm :)
18:47 marcus Yeah, we were in Ireland for a week. Dublin and Cork.
18:47 marcus Very pretty.
18:48 marcus xantus_: and with you? Life's good?
18:50 xantus_ cool
18:50 xantus_ its getting better
18:50 xantus_ I bought a car, dumped the crazy russian
18:51 xantus_ and I'm seeing my old gf
18:51 marcus smart moves :)
18:51 marcus cool. The south american one?
18:51 xantus_ yeah, she's good to me
18:51 marcus excellent :)
18:51 marcus think the russian one was too much of a gold digger :p
18:52 xantus_ yeah she was
18:53 marcus xantus_: do you play words with friends?
18:54 xantus_ I have it on my ipad, but haven't played
18:54 xantus_ I've been playing borderlands on 360 lately
18:55 marcus cool. A friend of mine has burnt a lot of hours on that.
18:55 marcus Been considering star craft ii
18:55 spleenjack joined #mojo
19:05 ask joined #mojo
19:09 xantus_ I hear there are some sc2 players here
19:15 Htbaa joined #mojo
19:39 xantus_ http://gist.github.com/505189
19:39 garfield [ gist: 505189 - MojoX::JSON- GitHub ]
19:40 xantus_ MojoX::JSON->singleton
19:40 xantus_ auto uses JSON if available, or hands you Mojo::JSON
19:40 xantus_ or MojoX::JSON->new if you like
19:43 dotan joined #mojo
19:43 dotan left #mojo
19:52 jakoubek joined #mojo
20:07 ashleydev joined #mojo
20:08 ashleydev joined #mojo
20:51 MojoGuest407 joined #mojo
21:54 y0j joined #mojo
21:55 ltriant joined #mojo
22:14 ask joined #mojo
22:28 stephen joined #mojo
23:07 abqar joined #mojo
23:08 ask joined #mojo
23:27 abqar joined #mojo
23:37 abqar joined #mojo

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