Camelia, the Perl 6 bug

IRC log for #mojo, 2010-11-22

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

All times shown according to UTC.

Time Nick Message
00:20 spleenjack left #mojo
01:33 Alias joined #mojo
01:33 ltriant left #mojo
01:34 ltriant joined #mojo
03:42 arpadszasz joined #mojo
04:03 tholen42 joined #mojo
04:05 tl left #mojo
04:28 su-bzero_ joined #mojo
04:28 su-bzero left #mojo
04:33 daviddelikat1 left #mojo
05:10 yakudzo left #mojo
05:21 Alias left #mojo
05:42 arpadszasz left #mojo
06:28 y0j joined #mojo
06:29 kvorg left #mojo
06:30 ltriant left #mojo
06:36 MojoGuest357 joined #mojo
06:36 MojoGuest357 From: http://www.google.co.th/search?hl=th&client=​firefox-a&hs=Y65&rls=org.mozilla%3Aen-US​%3Aofficial&channel=np&q=websocket+demo&​amp;aq=f&aqi=&aql=&oq=&gs_rfai= (1 hits)
06:36 MojoGuest357 hi
06:36 MojoGuest357 left #mojo
06:37 koban joined #mojo
06:41 yakudzo joined #mojo
07:04 ysyrota joined #mojo
07:13 * sri yawns
07:51 fhelmber_ joined #mojo
08:07 und3f[A] is now known as und3f
08:18 crab "We want it to loo like a web site." me: "What do you mean? It *is* a web site." "But it doesn't *look* like a web site!"
08:20 kvorg joined #mojo
08:25 sri how does a web site look like?
08:40 crab sri: i guess it's missing some <blink> tags, maybe an animated gif or two
09:02 memememomo_ is now known as memememomo
09:03 memememomo is now known as memememomo_
09:17 mfontani left #mojo
09:32 otaviof joined #mojo
09:33 otaviof left #mojo
10:02 kaare joined #mojo
10:09 kvorg left #mojo
10:17 und3f hello
10:18 und3f sri, what about adding ANY request to resolver? It is same as A query
10:18 und3f sri, also latest resolver improvements allows to use ANY request results
10:27 nuclon joined #mojo
10:27 nuclon morning
10:28 nuclon did somebody start mojo application with Plack?
10:28 nuclon MOJO_HOME=/home/nuclon/workspace/Project plackup ./script/project_web.pl
10:28 nuclon Error while loading ./script/project_web.pl: Can't locate object method "new" via package "Project::Web" at /usr/local/share/perl/5.10.1/Mojo/Server.pm line 23.
10:28 nuclon Any ideas?
10:28 und3f same problem here
10:33 und3f nuclon, seems like problem with script/start_script need to update it
10:33 und3f nuclon, to work with plackup
10:37 und3f Also tried to generate .psgi file and run it
10:38 und3f got same error, now in browser
10:38 rhaen hep!
10:46 und3f yes, it works with latest startup script
10:51 und3f except json_config plugin
10:52 und3f It tries to load plackup.json
10:53 MojoGuest985 joined #mojo
10:53 MojoGuest985 From: http://www.google.com/search?q=websocket+demo&am​p;hl=zh-TW&amp;tbas=0&amp;source=lnt&amp;sa=X&am​p;ei=hUvqTIDBGYuevgOMtYDCCA&amp;ved=0CAYQpwU4Cg (1 hits)
10:53 MojoGuest985 ?
10:53 MojoGuest985 anyone online?
10:53 MojoGuest985 left #mojo
11:05 spleenjack joined #mojo
11:11 sri yea, plackup is annoying
11:11 sri makes proper detection very very hard
11:18 sri good rule of thumb is not to use any of the detection features in mojolicious with plack
11:30 sri i tried to get some env vars to make detection easy into the psgi spec, but it got blocked
11:40 perlrocks left #mojo
11:40 perlrocks joined #mojo
11:50 sri guess this will be a slow week :)
11:51 rhaen hm.
11:51 rhaen ok.
11:52 sri not much activity on twitter today
11:57 sri oh, but a new futurama episode!
11:57 rhaen no - however, I finished my config example for membase/mojo user auth on the weekend
11:57 sri now thats a good way to start a week
11:58 rhaen it works - sorta
11:58 rhaen I am not happy with it - so far.
11:59 rhaen btw. I had more troubles figting membase than understanding the mojo concepts
11:59 sri too few features in membase?
11:59 rhaen no, I wrote a "model" class which does the job you
11:59 rhaen basically you provide a key and your data in structure
11:59 rhaen it will encode it to a json structure and store it.
12:00 rhaen query possibilities are limited (of course) as in every pure key/value store
12:00 rhaen but that's perfectly ok
12:00 sri should go very well with IndexedDB
12:01 rhaen true :)
12:01 sri thats of course something it has going for it
12:01 sri IndexedDB doesn't have many features either
12:02 rhaen to be true - are they needed for user auth? I don't think so.
12:02 sri i'm glad sqlite has been removed from html5
12:02 rhaen You know the username - you want the password, bio, email, just a basic profile
12:02 sri IndexedDB is bound to the domain
12:02 rhaen for more sophisticated things you should check other options
12:03 rhaen hm. SQLite is nice, however it has an interesting locking model
12:03 sri nosql forever!
12:03 rhaen like - I lock the file for everyone to ensure correct data
12:04 rhaen pah! nosql evangelista - inside this channel. HEEELP!
12:04 rhaen :)
12:04 sri i kinda want to work on a javascript framework
12:04 sri got some fun ideas
12:05 sri but not for node.js
12:05 sri i want to go client side
12:05 rhaen *brrr* sri - will you leave us and join the dark side again?
12:05 sri never
12:06 sri mvc on the client side
12:06 sri possibly with optional mojolicious backend
12:07 sri truly taking advantage of html5
12:07 rhaen oh.
12:08 GabrielVieira joined #mojo
12:08 rhaen Just embedd flash to do everything more complicated!
12:08 rhaen *hides*
12:08 * sri throws shoes after rhaen
12:10 rhaen SILVERLIGHT!
12:11 sri did a shoe hit your head too hard? Oo
12:11 rhaen yup.
12:12 rhaen One day I will call the zergs and rush you
12:12 sri my thors are waiting!
12:18 su-bzero_ left #mojo
12:18 rhaen pah!
12:19 rhaen I will use my banelings to crush your wall in and my zerglings will rush you
12:19 rhaen easy!
12:40 sri wonder if Mojolicious::Plugin::POE would be a good idea
12:42 omega the world is getting so confusing, with what wraps what, what can be embedded where etc
12:44 sri but at leats stuff works together
12:57 nuclon left #mojo
13:07 kvorg joined #mojo
13:11 kvorg left #mojo
13:15 otaviof joined #mojo
13:15 otaviof left #mojo
13:18 MojoGuest842 joined #mojo
13:18 MojoGuest842 From: http://www.google.com/search?q=websocket+demo&​amp;ie=utf-8&amp;oe=utf-8&amp;aq=t&amp;rls=org​.mozilla:en-US:official&amp;client=firefox-a (1 hits)
13:18 MojoGuest842 left #mojo
13:23 daviddelikat joined #mojo
13:26 perlrocks Twitter: "RT @kraih: If there is anything we could do to make Mojo::IOLoop work better with common #perl event loops please let me know. #mojolicious" --insurgentsoft http://twitter.com/insurgent​soft/status/6699938209595392
13:39 und3f sri, so what about ALL DNS query? It have same format as 'A'
13:44 koban left #mojo
13:47 Akron joined #mojo
13:50 kaare left #mojo
13:55 und3f sri, checkout pull request
13:58 kvorg joined #mojo
14:03 arthas joined #mojo
14:12 sri und3f: where in the spec is ANY defined?
14:16 und3f sri, http://tools.ietf.org/html/rfc1035 page 12, record name '*'
14:19 sri hmm
14:19 sri adding it is ok, but i want it differently
14:23 kaare joined #mojo
14:25 sri und3f: the test is also way too weak
14:35 GitHub197 joined #mojo
14:35 GitHub197 mojo: master Sebastian Riedel * 9b0da82 (3 files in 3 dirs): added "*" query support to the Mojo::IOLoop resolver - http://bit.ly/dag0aU
14:35 GitHub197 left #mojo
14:36 sri und3f: strictly following the spec is better imo
14:37 und3f sri, i just followed wireshark
14:37 und3f sri, also you can easly allow to query any record type
14:39 sri http://search.cpan.org/~bfg/Mo​joX-Run-0.11/lib/MojoX/Run.pm # wonder if that works on windows
14:40 und3f windows? Who uses that?
14:46 und3f sri, also one record type we need to support is SOA
14:47 und3f sri, ANY requests returns SOA record
14:51 sri we ignore unknown records
14:51 sri so it's all fine
14:54 und3f sri, what about handling DNS messages SERVFAIL and NOTHING FOUND?
14:55 sri whats the point?
14:57 und3f sri, to handle cases when DNS didn't returned any data because of some conditions (like too many requests or recursion request time exceeded)
14:57 und3f and reresolve host
14:57 sri how does that help the resolver to do its job?
15:01 sri we don't support reresolving
15:01 sri so there is no actual value for us
15:02 sri recursion opens a huge can of worms that requires tons of tests, which i don't want to write :)
15:07 kvorg left #mojo
15:11 sri i'm all for making the resolver better, don't get me wrong, but it needs to be maintainable
15:16 und3f sri, i am asking just to return status code of DNS response to callback
15:16 und3f nothing else
15:16 sri find a clean solution and make it testable
15:18 y0j left #mojo
15:27 und3f sri, what about supporting TCP DNS query for truncated DNS replies?
15:28 sri huuuuuuuge can of worms
15:30 sri i'm happy with the resolver in its current state
15:31 sri all additional features need to be really really well done patches
15:31 sri especially such hard ones
15:40 koban joined #mojo
15:42 arpadszasz joined #mojo
16:08 koban left #mojo
16:14 y0j joined #mojo
16:59 ysyrota left #mojo
17:00 arpadszasz left #mojo
17:02 esskar joined #mojo
17:08 esskar left #mojo
17:09 su-bzero joined #mojo
17:10 Kulag left #mojo
17:10 Kulag joined #mojo
17:15 GabrielVieira left #mojo
17:26 arthas left #mojo
18:03 perlrocks Twitter: "Blog updated. - [Perl] Mojolicious try to specify a class template - http://ow.ly/3dLkE" (ja) --hid_tgc http://twitter.com/hid_tgc/status/6769659739119617
18:03 perlrocks Twitter: "RT @ hid_tgc: Blog updated. - [Perl] Mojolicious try to specify a class template - http://ow.ly/3dLkE" (ja) --AE_PerlIsm http://twitter.com/AE_PerlI​sm/status/6769662729658368
18:05 perlrocks Twitter: "RT! Hid_tgc: Blog updated. - [Perl] Mojolicious try to specify a class template - http://ow.ly/3dLkE" (ja) --perlism http://twitter.com/perlism/status/6770146261606400
18:29 arpadszasz joined #mojo
19:04 spleenjack left #mojo
19:10 s0va sri...
19:11 s0va i'm having really, really weird problem.
19:12 s0va sri: i'm using MojoX::Run to spawn external programs. Everything works fine in mojo 0.9xxxx46
19:12 s0va i mean 0.9xxx36
19:12 s0va but works partially with mojo 0.9xxxx41
19:13 s0va i get stdout/stderr output when i spawn command ls -al /etc
19:13 s0va but i don't get anything when i spawn command i would really like to use :/
19:14 s0va handle's read_cb is NEVER called even if i'm 100% sure that program creates some output
19:14 s0va on_hup is called.
19:15 sri no idea really, not much changed, just some cleanup
19:16 s0va i'll check if it works with .37, .37. etc
19:16 s0va ok
19:17 sri but thats the risk of doing untested stuff
19:18 s0va sri: you added also $socket->connected check somewhere in ioloop
19:18 stephen left #mojo
19:18 s0va that's why i needed to wrap IO::Handle into IO::HandleRun
19:18 s0va https://github.com/bfg/mojox-run/b​lob/master/lib/MojoX/HandleRun.pm
19:18 sri possible, core only deals with sockets
19:18 sri thats what ioloop is designed for
19:19 s0va yep, i know.
19:19 s0va is it possible that read would fail if setsockopt would fail on this fake "socket" handle?
19:20 sri possible
19:22 rhaen hola!
19:22 rhaen we need a collectd plugin
19:22 rhaen is there a possibility to make a request with mojo client and get the time for the execution back?
19:23 sri my $before = time; ... my $after = time - $before;
19:24 rhaen *ugh!*
19:24 rhaen ok :)
19:25 s0va sri: last working version: 0.999940
19:25 s0va xxx41 doesn't work.
19:26 s0va ok, let's see how ioloop differs in between last two versions...
19:29 MadManchegan joined #mojo
19:30 s0va sri: can i ask you something... why you didn't reuse any existing ioloop, let's say, anyevent... lot's of things you implemented (http client, dns resolver) are already there.
19:30 s0va just curious.
19:30 sri which one?
19:30 s0va AnyEvent
19:31 sri i hate the anyevent "community"
19:31 s0va http://search.cpan.org/~mlehmann/AnyEvent-5.28/
19:31 sri i love the POE community but not the api
19:32 s0va poe is really nice, but fucking slow if running lot's of sessions.
19:32 sri thats the problem, picking an event loop is like picking a religion
19:32 s0va in my expirience it scales really badly above 500 concurrent sessions.
19:33 sri so my goal now is to work equally well with all of them
19:34 sri ioloop will never be a general purpose event loop
19:34 sri it's a tradeoff, but everything is
19:34 sri i had to implement http myself anyway
19:35 sri thats the real problem
19:35 s0va jep.
19:35 s0va fuck, does anyone know how to create diff between two git tags?
19:36 sri github does that
19:36 sri https://github.com/blog/612-in​troducing-github-compare-view
19:40 sri https://github.com/facebook/torna​do/blob/master/tornado/ioloop.py # main inspiration for ioloop btw
19:41 Akron left #mojo
19:42 sri not sure why everybody assumes it is meant to be a general purpose event loop :S
19:43 sri maybe i need to remove pod from it
19:43 s0va sri: http://pastebin.com/SH9dmvgQ <== i guess this is all that has been changed between 0.40 and 0.41
19:43 s0va no no no :)
19:44 s0va on_read is still there...
19:44 sri you have 42 mixed in there
19:45 sri most changes in the diff are actually 42
19:45 s0va ahem... this is from my cloned master against revision tagged for 0.40
19:46 sri oh, you've actually found a bug :)
19:46 GitHub115 joined #mojo
19:46 GitHub115 mojo: master Sebastian Riedel * deb2357 (1 files in 1 dirs): fixed typo - http://bit.ly/bFhiZz
19:46 GitHub115 left #mojo
19:48 sri s0va: quite sure thats only 42 changes, missing everything from 41
19:48 sri CNAME was added for example
19:51 s0va damm iz.
19:51 s0va not having time to figure out why stuff broke, i will just bundle version xxx40 in my app..
19:52 s0va sri: i used ioloop jst becouse POE blocks when executing external commands using poe::wheel::run.
19:52 s0va i used ioloop->on_tick()
19:52 s0va :)
19:53 sri that would be a POE bug
19:54 sri we can only trust POE not to block on POE::Kernel->run_one_timeslice
19:54 s0va sri: have you ever opened poe::kernel source code?
19:55 s0va it's a fucking disaster...
19:55 sri not yet
19:55 s0va i think that is possible to implement very nifty anyevent-like event loop on top of your ioloop.
19:56 s0va it would be a bummer imho.
19:56 s0va you have http infrastructure built in, timers, epoll, kqueue...
19:56 sri i would have designed a general purpose loop different
19:57 s0va the only thing missing are so called "poe sessions" and poe- like filters. howevery, anyevent doesn't implement any of them.
19:57 s0va how?
19:57 s0va curious
19:57 sri start with event watchers and put actual read/write code in separate modules
19:58 sri the loop would only add/remove handles from poll/kqueue/epoll
19:58 sri and fire readable/writable events
19:58 s0va sri: this is how poe is designed
19:59 s0va sri: async socket/file reading writing is implemented on few low level functions
19:59 s0va $poe_kernel->select_read and ->select_write()
20:00 s0va http://search.cpan.org/~rcaputo/POE-1.294/​lib/POE/Kernel.pm#select_read_FILE_HANDLE_[,_EVENT_NAME_[,_ADDITIONAL_PARAMETERS]_]
20:00 s0va "current session" is just some kind of abstraction ...
20:01 sri i would use a good looking high level api ;p
20:02 s0va well, mojo only needs some kind of priority queue implementation for executing subs, packed in a nice api...
20:02 s0va and voila...
20:02 s0va you have much cleaner and neater stuff than poe is.
20:03 s0va i'm really tempted to  start a project like this and hook it into ->on_tick(), heh
20:05 sri you would also need io watchers for normal handles
20:06 sri my $id = $loop->add($handle, on_readable => sub { my ($loop, $id, $handle) = @_; });
20:06 sri or so
20:06 sri if you were going to turn ioloop around
20:08 spleenjack joined #mojo
20:08 s0va sri: if everything works (well at least until 0.999940, heh) for spawned process pipes, it should work also for normal files.
20:09 s0va i also implemented writing to spawned process'es stdin...
20:09 s0va therefore async normal file writing should also be possible.
20:09 sri file writing is never async
20:09 s0va i've seen that you try to read 4MB blocks in ioloop's _read()
20:09 s0va well "async"
20:09 sri you would need IO::AIO
20:10 sri which works through MojoX::AIO already
20:11 xantus[] ;)
20:11 sri yea, 4mb is mostly the maximum people set their kernel send buffers to on tuned servers
20:11 vel joined #mojo
20:13 vel left #mojo
20:13 sri or in this case receive buffer
20:13 vel joined #mojo
20:15 vel left #mojo
20:16 vel joined #mojo
20:25 GitHub45 joined #mojo
20:25 GitHub45 mojo: master Sebastian Riedel * 412e451 (1 files in 1 dirs): cleanup - http://bit.ly/fTEB4X
20:25 GitHub45 left #mojo
20:25 sri connected check is actually not neccessary anymore
20:26 xantus[] sri: IOS 4 is out on ipad now
20:26 xantus[] sri: +WebSockets
20:26 sri downloading 4.2 right now :)
20:27 sri lots of silence on the hybi list, no good sign
20:31 su-bzero left #mojo
20:34 fhelmber_ left #mojo
20:42 ysyrota joined #mojo
20:43 vel left #mojo
20:43 MadManchegan left #mojo
20:45 ysyrota sri: is it possible to define an i18n package with %Lexicon in separate file?
20:50 vel joined #mojo
20:55 arpadszasz left #mojo
20:56 xantus[] just load Lexicon from a file
20:56 tholen joined #mojo
20:59 ysyrota xantus[]: I'd like perl did this work instead of me. I can do this if I comment out one line in M:Plugin:I18n. Just wonder if it is possible to do it w/o editing Mojolicious :)
21:02 GitHub174 joined #mojo
21:02 GitHub174 mojo: master Sebastian Riedel * 096c6f0 (1 files in 1 dirs): cleanup - http://bit.ly/hD1wHb
21:02 GitHub174 left #mojo
21:03 vel1 joined #mojo
21:05 MojoGuest829 joined #mojo
21:05 ltriant joined #mojo
21:06 MojoGuest829 left #mojo
21:06 ysyrota left #mojo
21:06 sri xantus[]: i can connect with dev.xantus.org but not say anything
21:06 sri same as normal safari :/
21:07 vel left #mojo
21:09 MojoGuest477 joined #mojo
21:09 MojoGuest477 why does it work in webkit, but not safari? :S
21:09 MojoGuest477 left #mojo
21:09 xantus[] probably cause its old mojo code
21:10 sri ?
21:10 xantus[] I think I left it @ the point before you removed draft 75 support
21:10 sri likely unrelated
21:11 sri my safari is draft 76
21:11 xantus[] hmm
21:11 sri it receives, but can't send
21:11 MojoGuest936 joined #mojo
21:11 sri test
21:11 sri yea, receives fine
21:12 xantus[] odd
21:12 sri just can't say anything
21:12 xantus[] js bug?
21:12 xantus[] probably
21:12 sri well, it works in webkit
21:12 sri broken exactly the same in mobile safari on ios
21:13 sri no sending, just receiving
21:13 xantus[] try http://aws2.xantus.org/
21:13 sri works
21:13 sri flawless
21:14 xantus[] thats recent mojo
21:14 sri don't think it is mojo
21:14 xantus[] nope, I don't either
21:14 sri maybe some js glitch in safari thats doesn't show in webkit
21:15 xantus[] yea
21:15 xantus[] safari mac?
21:15 sri yea, latest
21:15 sri and ios 4.2.1 safari
21:15 xantus[] I'll update tonight and play with it
21:16 MojoGuest936 left #mojo
21:22 MojoGuest852 joined #mojo
21:22 MojoGuest852 From: http://www.google.co.za/search?hl=en&a​mp;client=safari&amp;&amp;sa=X&amp;ei=​-d7qTM6XMomj4Qac7-j4Ag&amp;ved=0CBUQvw​UoAQ&amp;q=websocket+demo&amp;spell=1 (1 hits)
21:23 MojoGuest852 left #mojo
21:25 stephen joined #mojo
21:27 a|newkirk[assoc] joined #mojo
21:27 a|newkirk[assoc] left #mojo
21:31 GitHub126 joined #mojo
21:31 GitHub126 mojo: master Sebastian Riedel * e65d95e (2 files in 1 dirs): cleanup - http://bit.ly/i6YtW6
21:31 GitHub126 left #mojo
21:32 sri s0va: your code might work again, no guarantees though
21:35 marcus left #mojo
21:36 GitHub23 joined #mojo
21:36 GitHub23 mojo: master Sebastian Riedel * 4426e5d (1 files in 1 dirs): cleanup - http://bit.ly/fiOeqf
21:36 GitHub23 left #mojo
21:38 sri $loop->connect(handle => $pipe) might now just work
21:39 sri it's not a feature, but the code is a little more neutral now
21:41 tholen left #mojo
21:48 MojoGuest158 joined #mojo
21:49 MojoGuest243 joined #mojo
21:49 MojoGuest243 left #mojo
21:50 spleenjack left #mojo
22:04 xantus[] nice
22:10 kaare left #mojo
22:19 GitHub33 joined #mojo
22:19 GitHub33 mojo: master Sebastian Riedel * 9f00bd5 (1 files in 1 dirs): fixed typo - http://bit.ly/g2RvuN
22:19 GitHub33 left #mojo
22:33 MojoGuest158 left #mojo
22:34 GitHub52 joined #mojo
22:34 GitHub52 mojo: master Sebastian Riedel * 7c5c7c4 (1 files in 1 dirs): fixed typo - http://bit.ly/hjaQx1
22:34 GitHub52 left #mojo
22:43 daviddelikat left #mojo
23:13 und3f is now known as und3f[A]
23:17 daviddelikat joined #mojo
23:36 vel1 left #mojo
23:45 MojoGuest874 joined #mojo
23:45 MojoGuest874 From: http://news.ycombinator.com/item?id=1277067 (55 hits)
23:50 xantus[] hi
23:52 xantus[] grr, another moose based lib
23:52 xantus[] AnyEvent::Riak
23:53 MojoGuest445 joined #mojo
23:53 MojoGuest445 From: http://news.ycombinator.com/item?id=1277067 (56 hits)
23:53 xantus[] it's fairly simple to do without moose
23:54 MojoGuest445 left #mojo

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