Camelia, the Perl 6 bug

IRC log for #mojo, 2010-11-14

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

All times shown according to UTC.

Time Nick Message
00:17 ash__ joined #mojo
00:19 ash___ joined #mojo
00:19 ash__ left #mojo
00:19 ash_ left #mojo
00:19 ash___ is now known as ash_
01:32 yuki joined #mojo
01:49 yuki left #mojo
01:52 yuki joined #mojo
02:09 yuki left #mojo
02:32 Akron left #mojo
04:03 tholen42 joined #mojo
04:05 tl left #mojo
04:07 MojoGuest490 joined #mojo
04:07 MojoGuest490 From: http://news.ycombinator.com/item?id=1277067 (49 hits)
04:08 MojoGuest490 left #mojo
05:13 memememomo_ is now known as memememomo
05:32 memememomo is now known as memememomo_
07:57 MojoGuest88 joined #mojo
07:57 MojoGuest88 From: http://www.google.ca/search?hl=en&client​=firefox-a&hs=TCY&rls=org.mozilla%3A​en-US%3Aofficial&q=websocket+demo&aq​=f&aqi=&aql=&oq=&gs_rfai= (1 hits)
07:58 MojoGuest88 left #mojo
08:45 perlrocks Twitter: "[Mojolicious] / Mojolicious Japanese translation project documents - livedoor Wiki (Wiki) http://htn.to/T6nwLP" (ja) --shiba_yu36 http://twitter.com/shiba_yu​36/status/3730154043998208
09:16 daviddelikat joined #mojo
09:19 daviddelikat1 left #mojo
09:24 und3f[A] is now known as und3f
09:28 perlrocks Twitter: "Mite Teru mojolicious use. Let's try Xslate" (ja) --shiba_yu36 http://twitter.com/shiba_yu​36/status/3740927336120320
09:51 tholen joined #mojo
10:36 MojoGuest576 joined #mojo
10:36 MojoGuest576 From: http://www.google.com.ph/search?hl=tl&cl​ient=firefox-a&hs=Wga&rls=org.mozill​a%3Aen-US%3Aunofficial&q=html5+irc&a​q=f&aqi=&aql=&oq=&gs_rfai= (1 hits)
10:36 MojoGuest576 left #mojo
10:43 und3f sri, will you add PTR resolving to Mojo::IOLoop if i'll write it?
11:11 spleenjack joined #mojo
11:25 und3f sri, checkout pull request. Also fixed some dns bugs
11:33 gabb I will officially quit programming now that I have seen this: http://www.youtube.com/watch?v=WAr-xYtBFbY
11:33 gabb I am not worthy :S
11:52 spleenjack left #mojo
12:16 MojoGuest303 joined #mojo
12:16 MojoGuest303 From: http://www.google.com/url?sa=t&source=web&a​mp;cd=3&sqi=2&ved=0CCUQFjAC&url=htt​p%3A%2F%2Fdev.xantus.org%2F&rct=j&q=htm​l5%20irc&ei=IdPfTPSTKIbtOdvx2P0O&usg=AF​QjCNGw6nZv0wO42uzUnHhWcD0vcKNULg&cad=rja (1 hits)
12:18 MojoGuest303 left #mojo
12:22 s0va hello :)
12:22 s0va i've got question regarding logging...
12:23 s0va i've created MojoX::Run, which is a nice interface for running external programs without blocking external process.
12:23 s0va server process...
12:23 s0va i'm using Mojo::Log object for displaying debugging messages.
12:24 s0va now debugging messages show up even if i start server with --mode production
12:25 memememomo_ is now known as memememomo
12:25 memememomo is now known as memememomo_
12:25 und3f s0va, you should setup loglevel
12:25 und3f s0va, $app->log->level('fatal')
12:25 und3f for example
12:27 s0va und3f: loglevel is set by daemon, isn't it?
12:27 s0va --mode development => debug, otherwise info
12:28 und3f loglevel by default is debug
12:28 s0va i'm seeing debugging messages only from my MojoX::Run module
12:28 s0va in "production" mode
12:28 und3f s0va, does your MojoX::Run uses other Mojo::Log instance than Mojolicious?
12:29 s0va und3f: in Run.pm: my $_log = Mojo::Log->new();
12:29 s0va should i use anything else?
12:29 und3f ...
12:30 und3f s0va, if you want same log with Mojolicious you should get Mojo::Log from mojoliciouys
12:30 su-bzero_ und3f: Will MojoX::Run suitable for our ping?
12:31 und3f su-bzero_, ??/
12:31 su-bzero_ For running ping utility on servers ;))
12:32 und3f s0va, can you show your module?
12:32 s0va su-bzero_: yesteday i tried with ab2 with 500 concurrent clients...
12:32 s0va response time for 50% percentile: 400msec
12:33 su-bzero_ And share the software, you'll be free... ;))
12:34 s0va und3f: will pastebin be ok?
12:35 und3f isn't it a free software?
12:36 s0va http://pastebin.com/HYWAZQdm
12:36 s0va i don't use github.
12:36 s0va and it lacks documentation.
12:36 s0va but, it works :)
12:38 s0va su-bzero_: time /tmp/z.pl
12:38 s0va real    0m0.005s
12:40 s0va su-bzero_: and via mojo lite webapp handler: http://pastebin.com/i1hnm4k3
12:40 s0va su-bzero_: it supports streaming reads from stdout/stderr and streaming writes to stdin.
12:41 und3f s0va, add __PACKAGE__->attr(log => sub { Mojo::Log->new })
12:41 s0va su-bzero_: you can run more than one process at once
12:41 und3f s0va, in mojolicious app create this module with log => $app->log  parameter
12:41 s0va und3f: and then use $self->log->debug("msg") ?
12:41 und3f s0va, right
12:42 und3f s0va, github is nice
12:43 s0va und3f: no difference. i still see debugging messages from Run.pm in production mode.
12:44 s0va hm, Mojo::Server uses $self->app->log()->info/debug()....
12:45 und3f s0va, show how do you use MojoX::Run in mojolicious
12:45 MojoGuest245 joined #mojo
12:45 MojoGuest245 left #mojo
12:48 s0va http://pastebin.com/EkG7qWbL
12:49 und3f s0va, [14:42] und3f> s0va, in mojolicious app create this module with log => $app->log  parameter
12:49 s0va und3f: now i copied attr "app" from Mojo::Server.pm to my Run.pm and it works! no more debugging messages in "production" mode
12:49 und3f good boy :)
12:52 s0va damn, this thing is really fast: http://pastebin.com/HTdrVTCT
12:55 dotan joined #mojo
12:55 s0va und3f: if i create proper perldoc, should i beg sri to include this code into mojolicious or should i create my own cpan username?
12:55 s0va i've heard, that MojoX:: namespace is obsolete...
12:56 su-bzero_ To github at first, then to CPAN.
12:56 su-bzero_ und3f: Yes?
12:57 s0va one more thing.
12:58 s0va i've read mojolicious coding style; "no feature without test."
12:58 s0va well, can anyone point to some simple document how to create perl module tests?
12:59 s0va yeah, it's embarassing...
13:00 dotan s0va - you can look at the tests in the t/ directory of the mojolicious distribution, but you probably want to read the perldoc of Test::Simple and Test::More for an intro to how testing works in perl.
13:01 s0va dotan: thanks. there are so many Test::XXX modules in cpan; i will check this out.
13:03 dotan s0va: Test::More is pretty much the only one you need unless you want to do something really fancy. Also look at Test::Mojo  - part of the Mojolicious dist - it's used everywhere
13:06 s0va dotan: yeah, looks quite simple...
13:06 s0va just git-cloned mojo dist from github
13:26 su-bzero_ s0va: I get 'stdout' => '' every time. For example for /bin/date.
13:33 avkhozov joined #mojo
13:34 s0va did you defined stdout_cb?
13:34 s0va if you do, it's called with data chunk
13:35 s0va if stdout_cb is not defined, then is appended to stdout buffer.
13:35 su-bzero_ O, ok.
13:35 s0va su-bzero_: currently working on perldoc, heh
13:38 avkhozov Hello everyone!
13:38 avkhozov How I can get only POST parameters in controller?
13:38 avkhozov If I write
13:38 avkhozov my $request = $self->req->method('POST');
13:38 avkhozov $login = $request->param('login');
13:38 avkhozov I get parameter in both cases. (GET and POST)
13:38 su-bzero_ s0va: Nice module.
13:41 s0va su-bzero_: thanks, i needed it (i'm using mojo to expose REST functionality which calls external commands which take ages to complete (10+ seconds))
13:41 s0va without this the whole server is blocked.
13:45 dotan avkhozov - to get POST params only, try $self->req->body_params
13:45 dotan (It's a method in Mojo::Message, which is a base class for Mojo::Message::Request)
13:48 avkhozov dotan: thanks, it works!
13:49 dotan avkhozov: :-)
14:01 sri und3f: without tests there is no chance
14:03 sri you can't fix something without a test, thats bullshit
14:06 und3f sri, bullshit is that url "test.1.1.1.1.com" is ipv4 address
14:07 sri s0va: depends on how clean the code is, portability, tests, documentation and how useful it would be for the majority of our users
14:07 sri und3f: then why didn't you test that?
14:07 sri without a test it does not exist
14:08 sri und3f: about PTR records, can't they be affected by compression?
14:08 und3f sri, mojo::log doesn't exists?
14:09 und3f sri, what compression?
14:09 sri dns compression
14:09 sri see spec
14:10 sri und3f: are you telling me it is ok to fix bugs without tests because Mojo::Log is lacking tests anyway?
14:11 und3f sri, as you wish. fixing bugs is not ok.
14:11 sri ?
14:12 sri without a test case it is just a random code change
14:12 sri but will never be a fix
14:13 sri i can't read minds
14:14 sri the only reason i'm a little annoyed is because i've been telling you that for the last few weeks, releatedly :(
14:14 sri *repeatedly
14:14 und3f sri, perl -MMojo::URL -le 'print "test.1.1.1.1.com"=~$Mojo::URL::IPV4_RE'
14:14 und3f here is your test
14:14 sri add it to the patch please
14:18 sri und3f: about compression see RFC1035 4.1.4
14:21 und3f sri, btw, [0::127.0.0.1] is right ipv4 address?
14:21 und3f und3f@kotepc ~ % ping 0::127.0.0.1
14:21 und3f ping: unknown host 0::127.0.0.1
14:22 sri i'm sure there are specs to answer that
14:25 und3f sri, checkout pull
14:27 sri und3f: why did the ptr code change?
14:27 sri fix is also gone?
14:28 und3f yes, fix was with bug
14:28 und3f sri, ptr not changed, just in other commit
14:28 sri what about compression?
14:28 und3f dunno
14:28 sri thats bad
14:30 sri we need something that can decode names with compression before i'll apply ptr/mx/cname etc patches
14:31 und3f can't TXT be compressed? :|
14:31 sri don't think so
14:31 sri see spec i pointed to before
14:32 sri it's about names, mx might be a good test case
14:32 sri lets say you request mx records for foo.com
14:33 sri answer is mail.foo.com
14:33 sri you'll get one element "mail" and a pointer to the first "foo.com"
14:34 und3f i see
14:34 und3f we don't have DNS tests at all?
14:34 sri live tests
14:34 sri t/mojo/client_online.t
14:35 und3f nice, i'll add PTR to them
14:35 sri ok, but there is no chance i can apply the patch before the compression problem is solved
14:38 sri i suggest adding MX too and write a proper decoder for name records
14:39 sri with that we could actually support pretty much all record types i think
14:42 dotan sri: does mojo go crazy (as in, churn CPU cycles) if doing blocking IO? (I'm using perl 5.8.7 which may be part of the problem)
14:43 sri dotan: mojo is doing nothing while you block of course
14:44 su-bzero_ Are there any solutions to use Mojo::IOLoop + AnyEvent?
14:44 sri su-bzero_: on_tick
14:52 su-bzero_ sri: Could you explain a bit detailed?
15:00 dotan su-bzero_: http://irclog.perlgeek.de/​mojo/2010-11-08#i_2976478
15:00 alnewkirk left #mojo
15:01 su-bzero_ dotan: Thanks. Need to try it.
15:09 s0va omfg...
15:10 s0va created my project with module::starter
15:10 s0va now pod-coverage test complains about sub new() being undocumented...
15:10 s0va but i have =head1 OBJECT CONSTRUCTOR block above it...
15:10 s0va wtf...
15:14 s0va looks like object constructor must be documented with =head2 in order to make Test::Pod::Coverage happy.
15:19 JonChisciotte1 joined #mojo
15:20 JonChisciotte1 is now known as JonChisciotte
15:21 avkhozov left #mojo
15:31 * szabgab is just wondering if there is already some interest in setting up a team for PLat_Forms ?
15:34 sri szabgab: don't think so
15:37 sri und3f: you changed TXT record handling without any explanation too
15:38 sri und3f: thats exactly what i'm complaining about, you don't give any explanations at all
15:40 und3f sri, yeah )
15:41 und3f sri, but didn't you understand what that join does?
15:41 sri not at all
15:41 sri don't ever assume the code speaks for itself
15:41 sri it never does
15:44 * kthakore hugs sri!
15:44 * kthakore runs
15:44 * sri needs a hug
15:44 sri :(
15:45 und3f sri, oh, i broke it too
15:45 * und3f hugs sri
15:45 und3f I got so much problems with Mojo last time...
15:46 und3f and i easy assume that Mojo doesn't work property
15:46 ash_ left #mojo
15:46 und3f debugging of check-host servers and client screw me up..
15:47 * kthakore hugs sri again
15:47 sri :)
15:47 kthakore :D
15:47 ash_ joined #mojo
16:09 s0va sri: can i ask you about your opinion regarding code cleaness in mojox-run at https://github.com/bfg/mojox-run ?
16:09 s0va sri: beware: tab is tabulator :)
16:10 sri tab indent is a red flag for me ;p
16:11 s0va shit :)
16:11 s0va ok, let's try to teach my eclipse editor to use 4 space tab delimiter :)
16:12 sri same goes for inline pod :)
16:12 sri you should just use perltidy
16:12 sri mojolicious has a .perltidyrc included thats very good
16:13 sri based on best practices most perl hackers agree on
16:14 sri lots of non perlisms
16:15 sri we tend not to use camel case method names in perl
16:16 s0va sri: changed tab...
16:16 Akron joined #mojo
16:17 s0va sri: that's why only private methods use camelcase :)
16:17 sri long time perl hackers don't do that either ;p
16:17 s0va hm... what's wrong with inline pod?
16:19 sri thats a trickier topic, half of all perl hackers thing it's bad the other half loves it
16:19 sri *think
16:19 tempire_ joined #mojo
16:20 sri i'm in the first camp now, used to be the second
16:20 s0va hm...
16:20 sri for consistency in mojolicious no inline pod is a rule
16:20 s0va why did you changed camp?
16:20 s0va ok, no problem, your project, your rules.
16:21 sri stricter separation between documentation for users and developers
16:21 s0va i'm coming from java, which has really neat javadoc inline documentation system.
16:21 sri comments are for developers, pod for users
16:21 s0va ah.
16:21 sri i also tend to comment very extensively
16:22 sri there's no real best practice though
16:22 sri more matter of taste
16:22 * tempire_ tests tests tests!
16:22 sri i highly recommend the book Perl Best Practices
16:23 sri thats mostly what our .perltidyrc is based on
16:24 und3f sri, i have a question about this line in Mojo::IOLoop https://github.com/kraih/mojo/blo​b/master/lib/Mojo/IOLoop.pm#L672
16:24 sri und3f: if a dns query times out we assume the dns server is dead
16:24 * tempire_ stands defiantly with -et=4
16:24 sri and fallback toblocking
16:25 und3f sri, is it really so necessary? because is i got internet problem on some of servers of check-host.net the resolve will not work anymore.
16:26 sri und3f: i don't know
16:26 s0va sri: thanks; reformatted code with mojo's perltidyrc; so you think it's best to create my own cpan distribution and submit my module there?
16:26 sri s0va: for sure
16:27 s0va sri: i've heard something about MojoX:: namespace deprecation... is this only in your distribution or mojo extensions should be put to Mojo:: namespace?
16:28 sri s0va: thats something i'm unsure about
16:28 sri a lot of stuff should rather be Mojolicious::Plugin::*
16:28 sri in your case MojoX might be appropriate
16:30 s0va becouse i implemented something that doesn't interfer with request lifecycle? Correct me if i'm wrong: plugins are extensions executed in different parts of request lifecycles?
16:30 sri nono, they can be anything
16:31 s0va oh
16:31 sri it's arguaable for your module though
16:32 Akron Is there a platform planned for community plugins?
16:32 sri do whatever you like more :)
16:32 sri Akron: community plugins?
16:36 Akron sri: A Plugin directory with better hierarchy than CPAN provides, with a stricter policy, with more discussions. For example, I made some Plugins specific for usage in share hoster environments - I think, environment specific plugins will need more discussions, recipes etc.
16:36 sri Akron: cool idea
16:36 sri nothing planned yet though
16:36 tempire_ It would be great to have that on the mojolicious site.  Something like jquery does.
16:37 Akron tempire_: Yeah!
16:38 tempire_ Though, I think if we want that, we'll have to make it.  sri is worshipping the toad at the moment.
16:38 sri indeed
16:40 tempire_ I've been thinking more and more about having a place to submit it mojolicious apps, like an app store with nifty icons.
16:40 Akron tempire_: Sure! I think, it would be a good idea to start that beginning with a little directory in the Wiki.
16:41 JonChisciotte left #mojo
16:42 s0va toad? :))
16:42 perlrocks Twitter: "More troublesome because you Mojolicious good." (ja) --hisaichi5518 http://twitter.com/hisaichi5​518/status/3850361961975808
16:44 dotan tempire_: there is (of course) a vti implementation of a mojolicious app index site somewhere... :)
16:44 tempire_ s0va: Hypnotoad, a hot deployable web application server that sri is working on.
16:45 Akron left #mojo
16:45 tempire_ dotan: I remember vti mentioning that, but never saw anything.
16:47 dotan tempire_: Here it is: https://github.com/vti/mojolicious-apps
16:47 s0va 3 weeks for CPAN pause registration?!
16:47 s0va wtf?!
16:47 s0va sri must be huge futurama fan, heh
16:48 sri you have no idea :)
16:48 tempire_ s0va: understatement
16:51 Akron joined #mojo
16:53 tempire_ left #mojo
16:55 Akron Latest CPAN Mojolicious seems to fail on installation for WindowsXP/StrawberryPerl. Stays at 1/16 of websocket_tls_proxy_lite_app.t ... hmmm ...
16:56 yko joined #mojo
16:58 und3f sri, lets assume that we have system with two dns servers in /etc/resolv.conf, second dns doesn't work. Most of software works right in that environment  (because gethostbyname tries all possible dns) but not mojo, right? Just after trying second dns it will stop trying to resolve anything, thats the right logic?
16:59 sri und3f: could be
16:59 sri Akron: is IO::Socket::SSL broken in starberry?
16:59 ash_ left #mojo
16:59 und3f don't you think this is bad way?
16:59 sri *strawberry
17:00 sri und3f: you came up wth that
17:00 und3f no
17:00 sri und3f: yes, you did resolv.conf detection
17:00 Akron sri: It worked properly with the latest Mojo version and I did not update.
17:00 und3f sri, i tried all DNS
17:00 und3f sri, one by one
17:00 und3f sri, but not just last one
17:01 ash_ joined #mojo
17:01 und3f sri, https://github.com/und3f/MojoX-DNS​/blob/master/lib/MojoX/DNS.pm#L124
17:07 flamey joined #mojo
17:08 flamey hi guys
17:08 flamey trying to figure out how to use MojoX::Auth::Simple
17:08 flamey to authenticate user
17:08 dotan left #mojo
17:09 sri und3f: must have been some bug then, don't remember
17:09 sri there were some leaks i think
17:09 flamey I do $auth->log_in(); , and check $auth->is_logged_in() right after that and it returs nothing
17:09 und3f sri, i am just telling that ->resolve doesn't work like gethostbyname. Actually it stop working after first unreachable dns server...
17:10 sri und3f: umm, gethostbyname does a lot more
17:10 und3f sri, and first think that it does - it works.
17:10 sri duplicating it's full funtionality is beyond the scope of our resolver
17:12 und3f sri, so you don't plan to try all dns in resolv.conf?
17:13 sri maybe
17:13 sri depends on how well it can be done and tested
17:14 und3f sri, could you at least delete https://github.com/kraih/mojo/blo​b/master/lib/Mojo/IOLoop.pm#L672
17:14 sri in the end it is my responsibility to keep it working, so i'm very reluctant about adding stuff that can't be tested
17:14 dotan joined #mojo
17:15 sri und3f: thats a possibility
17:16 und3f sri, there are better way to indicate that DNS server doesn't work, i guess.
17:16 sri i'm all ears
17:17 sri und3f: btw. any progress on dns compression?
17:17 und3f i didn't worked on it
17:17 sri oh
17:18 und3f sri, now i am trying to fix some bugs in check-host...
17:21 perlrocks Twitter: "Today Mojolicious, Xslate, DBIx:: Skinny had to try several on. I know that you can meet quite a lot, sleep." (ja) --shiba_yu36 http://twitter.com/shiba_yu​36/status/3860195130216448
17:22 su-bzero joined #mojo
17:23 dotan left #mojo
17:24 vel joined #mojo
17:26 s0va question: webapps without accessing to some kind of database are mostly useless...
17:26 su-bzero_ left #mojo
17:26 sri doesn't look like a question
17:26 s0va mojo daemon is singlethreaded process in which nothing should block.
17:27 s0va what is the best way to query databases?
17:27 s0va dbi->connect(), prepare, execute...
17:27 sri DBI is mostly blocking
17:28 s0va is there any special way of doing this or we may just hope that operation will not block for too long
17:28 sri you would have to use an external process pool
17:28 und3f sri, also here are some more tests https://github.com/und3f/mojo/commit/6e​97321688e744dfee83ed9ea2730908e4e5bfb3
17:28 und3f sri, Mojo::Client interrupted by keep_alive_timeout returns success status
17:29 und3f sri, if interruption happens after header were received by client
17:29 und3f HTTP header i mean
17:29 sri ok, so?
17:29 und3f sri, it is a bug
17:29 s0va maybe i will need to teach mojox-run  to run perl coderefs in separate process...
17:30 sri i dunno, is it?
17:30 und3f sri, it isn't????
17:30 sri und3f: you tell me, can it be handled differently?
17:31 und3f sri, for sure request isn't success
17:31 sri patche welcome
17:31 sri *+s
17:31 und3f left #mojo
17:33 ash__ joined #mojo
17:35 ash___ joined #mojo
17:35 ash__ left #mojo
17:36 sri there are many more implications here
17:36 sri like long polling
17:37 und3f joined #mojo
17:37 sri what side effects would such a change have?
17:37 ash_ left #mojo
17:37 ash___ is now known as ash_
17:43 yko left #mojo
17:44 s0va und3f: are you planning to imeplement hosts support in mojox-dns?
17:45 sri mojox-dns is obsolete
17:46 sri the functionality is now provided by ioloop natively
17:46 s0va oh, that's great.
17:46 s0va does it support /etc/hosts?
17:46 sri no
17:46 s0va some resolvers don't do this.
17:48 sri that should be handled by the blocking fallback
17:49 s0va ok.
17:50 s0va sri, just one more thing.
17:50 s0va logging.
17:50 s0va if i use Mojo::Log directly in my module
17:50 s0va most of calls are $_log->debug()
17:50 s0va i see debug output in webapp daemon even if i run it with --mode production
17:51 sri use the applications logger
17:51 sri thats where the plugin api comes in handy
17:51 s0va that's why i added: app attribute
17:51 s0va https://github.com/bfg/mojox-run/​blob/master/lib/MojoX/Run.pm#L19
17:51 sri thats ugly
17:51 sri very
17:51 s0va i agree
17:52 sri wrap it in a plugin to get the right logger in mojolicious apps
17:52 s0va now i'm creating tests (without app) and i'm getting this: t/mojox-run-ls.t ......... Can't locate object method "app_class" via package "MojoX::Run" at /export/home/bfg/development/M​ojoX-Run/blib/lib/MojoX/Run.pm line 27.
17:52 s0va hm, ok, i'll try to do this
17:53 s0va shit, shit, shit!
17:53 sri outside mojolicious the logger is all your responsibility
17:53 sri if you go low level you better be prepared to do a lot yourself
17:54 tempire hrm.
17:54 sri actually i don't suggest leaving mojolicious at all unless you really know what you're doing
17:54 tempire vti's mojolicious apps thing dumps out a template exception.  was hoping it would just work.
17:55 s0va well, i'm not :)
17:57 s0va sri: so, i just need to do the following thing: use base 'Mojolicious::Plugin', then create register method and set $_log = $app->log(); otherwise leave it alone?
17:57 sri und3f: i have a solution for dns compression :)
17:58 und3f nice to hear
17:58 und3f i want to get PTR records in next cpan release :)
17:59 sri it is rather simple actually
17:59 sri make PTR work for v4 and v6 and we can talk :)
17:59 sri compression will be solved by me
18:00 sri i'll add MX as a proof of concept
18:00 und3f sounds like a deal
18:00 sri gmail.com is a good test case
18:00 sri lots of pointers
18:07 yko joined #mojo
18:09 Akron Got a problem with static file dispatching (downloads): http://pastebin.com/jNZhCjv7 | this creates 3 redefinement errors in Mojolicious::Controller.
18:10 sri root is an absolute path
18:10 sri or supposed to be
18:11 sri and why didn't you just use $self->render_static?
18:11 Akron sri: Because I need to set up the filename.
18:11 sri and?
18:12 Akron sri: Thanks - changed it to abs, but nothing changes. There are still redefinement errors after dispatching favicon.ico
18:12 sri nopaste errors
18:13 sri and use $self->render_static
18:15 Akron Okay ... I do - but still errors. What do you mean with "nopaste errors"?
18:15 sri or pastebin or whatever
18:17 Akron With errors: http://pastebin.com/82xt4WQ8
18:19 sri umm, thats not an error
18:19 sri just a warning
18:21 Akron sri: Okay - a warning. But it's a redefinition in Base - I think, this means there is some re-init of the Module ... that should not happen ... or should it?
18:22 sri reloadr causes it i guess
18:22 sri *reloader
18:22 sri gonna disable it
18:22 sri absolutely harmless
18:23 sri could be a side effect of the new more aggressive reloader
18:23 Akron sri: Thanks - yes, it was reload!
18:25 und3f sri, can you tell me some ipv6 address to get it's PTR?
18:25 sri ipv6.google.com
18:25 sri they made a special ipv6 only test case
18:27 und3f sri, seems like they didn't made PTR record for ipv6 address
18:28 sri not nice
18:30 GitHub102 joined #mojo
18:30 GitHub102 mojo: master Sebastian Riedel * 03a48c2 (4 files in 3 dirs): added MX record type support to the Mojo::IOLoop resolver - http://bit.ly/9BSSWo
18:30 GitHub102 left #mojo
18:30 und3f sri, ipv6tools.org has..
18:30 sri Akron: warning should be gone with github master
18:31 sri und3f: use _parse_name
18:31 und3f sri, what do you mean by that?
18:31 sri see commit
18:32 Akron sri: Thank you!
18:42 OpossumPetya joined #mojo
18:43 GitHub59 joined #mojo
18:43 GitHub59 mojo: master Sebastian Riedel * 2f8a849 (4 files in 3 dirs): cleaned up tests - http://bit.ly/ck3mve
18:43 GitHub59 left #mojo
18:43 sri und3f: all additional resolver tests go into t/mojo/ioloop_online.t
18:47 flamey left #mojo
18:47 Akron left #mojo
19:01 und3f sri, have you looked at tests for Mojo::URL?
19:01 sri und3f: not yet
19:02 und3f sri, i am planning to rewrite them by new PTR resolver (merged into your latests commits) so please do it
19:03 yko left #mojo
19:12 GitHub147 joined #mojo
19:12 GitHub147 mojo: master Sebastian Riedel * 6d2ed0b (3 files in 3 dirs): fixed IPv4 address detection bug in Mojo::URL - http://bit.ly/d2MRFH
19:12 GitHub147 left #mojo
19:19 OpossumPetya .
19:19 OpossumPetya left #mojo
19:19 OpossumPetya joined #mojo
19:20 OpossumPetya left #mojo
19:20 flamey joined #mojo
19:20 GitHub78 joined #mojo
19:20 GitHub78 mojo: master Sebastian Riedel * 2ada2b9 (1 files in 1 dirs): cleanup - http://bit.ly/bEHE1T
19:20 GitHub78 left #mojo
19:21 flamey i'm trying to install Mojolicious via cpan on Windows, and get stuck on
19:21 flamey t/mojolicious/websocket_tls_proxy_lite_app.t .. 1/16
19:21 sri flamey: github master has that test disabled for windows
19:22 sri i suspect IO::Socket::SSL is broken in strawberry
19:22 flamey i don't know how to install from Github
19:22 flamey this in AS
19:22 flamey is Active State, i meant
19:22 flamey 5.10
19:22 sri hmm, maybe IO::Socket::SSL is broken on windows in general
19:23 sri anyway, fixed in github master then
19:23 sri i think cpanm can install from github
19:26 flamey latest download package on github is 0.999937 ?
19:26 sri no
19:26 sri you need master
19:27 und3f sri, the only IPv6 address i know is address of ipv6tools.org
19:27 und3f sri, include it in tests?
19:27 sri und3f: sure
19:27 flamey # Branch: master
19:28 sri https://github.com/kraih/mojo/tarball/master
19:29 flamey kraih-mojo-v0.999937-38-g2ada2b9.zip -- confusing
19:30 s0va hm...
19:31 s0va is there a way to send binary files from controller without reading entirely to ram before sending with $self->render_data($buf)?
19:32 und3f sri, have you looked at Mojo::URL tests? I am ready to overwrite it
19:33 sri und3f: already fixed
19:33 sri s0va: yes, see how Mojolicious::Static does it
19:34 sri a Mojo::Asset::FIle can be used to stream files chunk wise
19:35 s0va sri: static only serves from public/ if i am not mistaken...
19:35 s0va my files are not there :)
19:35 sri just look how it does what it does
19:36 s0va k :)
19:37 ash_ left #mojo
19:38 s0va omg.
19:39 flamey how to do user auth + session management (store sessions in db) ? are there any complete examples out there? can't figure out MojoX::Auth::Simple :(
19:39 sri cookie based sessions are built in
19:39 s0va so i just need to create new asset with my file and apply it to $self->res->asset();
19:39 sri as shown in the lite tutorial
19:41 und3f sri, why you deleted all resolver tests?
19:42 sri und3f: see above...
19:42 und3f oh..
19:42 und3f fuck
19:49 und3f github deleted all repositories???
19:50 und3f sri, will push when github come up
19:50 sri wtf
19:50 sri forks: 0
19:51 und3f sri, i actually can't see any of repositories
19:51 und3f sri, also github where dead some moments ago
19:54 und3f again 500..
19:54 s0va is there option to get catch exception message in exception.html.ep and display it?
19:55 s0va @_ is empty.
19:56 sri s0va: see the one generated by mojo generate app
19:59 perlrocks Twitter: "Looks like @github is having huge problems, all 62 #mojolicious forks are gone. :S #perl" --kraih http://twitter.com/kraih/status/3899860893827073
20:00 und3f sri, http://twitter.com/github
20:00 s0va great! sri: someone will need to write book about mojolicious...
20:00 sri s0va: totally
20:00 s0va almost everything is there, but there is no good tutorial for beginners.
20:01 sri writing that kind of tutorials is an art
20:02 sri sadly i'm not good enough for it
20:04 und3f sri, btw, in my commit i also fixed bug in parsing question )
20:04 und3f DNS question section
20:05 sri und3f: what kind of bug?
20:05 und3f sri, you will se that. just a small typo
20:05 sri ok
20:08 s0va sri: i guess... your perldoc is complete and good, but it is quite difficult to see the big picture...
20:08 perlrocks Twitter: "RT @kraih: Looks like @github is having huge problems, all 62 #mojolicious forks are gone. :S #perl" --kovnsk http://twitter.com/kovnsk/status/3902181681922048
20:09 sri s0va: suggestions and patches are very welcome
20:13 sri (in case github comes back)
20:13 sri otherwise we are all screwed and open source is dead
20:15 DaTa ah well, i push to gitorious and a private server also :)
20:15 * sri can't stand gitorious
20:17 tholen left #mojo
20:17 * baton8 plus one
20:20 und3f sri, otherwise we will send patches by email like good old times
20:25 Akron joined #mojo
20:31 vel left #mojo
20:36 sri oh, yko has released a cpan uploader
20:37 vel joined #mojo
20:46 esskar joined #mojo
20:48 kvorg moin
20:48 esskar left #mojo
20:48 kvorg i seem to have pulled sucessfully, tests pass
20:48 yko joined #mojo
20:49 yko sri: long time ago. maybe 2 month
20:49 yko just fixed some tests and path()
20:49 yko btw Mojo::URL::path does not accept relative path anymore?
20:49 sri yko: seems quite useful, you should have said something ;p
20:50 sri of course it does
20:50 yko vti were saying about it when there was a release
20:50 kvorg sri: i think we all assumed you knew, i thought yko even tweeted it.
20:50 yko hehe, ye. i'm pretty sure i twitted about before going down
20:50 kvorg (or perhaps it just got mixed in with the tweeterbot moment)
20:51 yko hmm. strange. i will recheck about path, but some tests were broken between .99927 and .99936
20:51 yko whatever, it worksforme at the moment (rly usefull stuff, ye)
20:52 yko oh, wanna ask something.
20:52 perlrocks Twitter: "Perl на ya.ru: Mojolicious и sessions в DB? http://ff.im/tKglC" --planetperlru http://twitter.com/planetper​lru/status/3913126965673984
20:52 sri if you find a bug it really helps to report them
20:52 yko sure, but only if i rly find it. ill recheck now
20:52 JonChisciotte joined #mojo
20:53 yko at the moment there's no direct way to pass any params to Client's ioloop when connecting (get, post, etc)
20:53 sri why would you want to do that?
20:54 yko because of LocalAddr
20:55 sri why?
20:55 yko i had to use different IP's on single machine, so i just subclassed and redefined  _connect
20:55 yko why? because i didn't want to got banned :)
20:55 yko 1k of 1k requests, you know...
20:55 sri hmm
20:56 sri you can also make your own socket
20:56 sri $tx->connection($socket)
20:56 yko sure it's some cheaty stuff, but while Mojo::Client is a pretty stuff to make tons of requests, it's a goot point to have ability to switch between ips ;)
20:57 kvorg left #mojo
20:59 yko well... you are right. it's possible. but i just prefer to have subclass with $self->localaddr attribute. or something like that. anyway thank you
20:59 perlrocks Twitter: "# Mojolicious officially in # debian (unstable, testing), Version 0.999938 - http://goo.gl/Oqfrv" (ru) --chernomyrdin http://twitter.com/chernomyr​din/status/3915050930667520
21:00 sri yko: thats ugly too
21:01 yko well, ye. but every time compose transaction by hands... hate that :D
21:01 yko one more question. maybe i'm little behind the times... is there's some way to know how many requests client process at the moment?
21:02 yko main idea is to keep client work with N connections at once and push requests from on_tick when queue decreases
21:04 yko at the moment i do it like $counter++; $client->get($url, sub { $counter-- } )->start;
21:09 tempire dernit.
21:09 tempire I never realized how much I rely on github until now.
21:09 * tempire wimpers
21:09 sri yko: no, i think thats beyond teh scope of Mojo::Client
21:10 yko left #mojo
21:10 sri yea, please come back octocat :(
21:11 yko joined #mojo
21:11 yko hmm. maybe then ioloop? or the best way is to use counter method?
21:12 sri way beyond the scope of ioloop
21:12 sri i think it's a higher level
21:12 yko hmm. so counters are correct way you think?
21:12 sri ye
21:13 yko thanks again
21:13 sri you couls also turn it into a neat queued client object or so
21:13 sri *could
21:14 yko hmm. maybe, but i think i left it as is at se moment. But looking in the future, maybe similar task will appear again
21:14 yko thanks for advices
21:17 yko hehe, btw, releasing MojoX::CPAN::Uploader worth it just for have it first package uploaded to cpan. With itself, sure. Nice recursion
21:27 sri oh, github is back
21:28 und3f sri, checkout pull request
21:30 und3f And dead again )
21:30 ash_ joined #mojo
21:31 ltriant joined #mojo
21:32 JonChisciotte1 joined #mojo
21:32 sri und3f: @parties? Oo
21:35 JonChisciotte left #mojo
21:35 JonChisciotte1 is now known as JonChisciotte
21:42 Akron left #mojo
21:47 sri und3f: the ipv6 code doesn't look good
21:48 und3f patches are welcome
21:48 und3f :P
21:48 und3f sri, why?
21:49 sri thats the problem, i don't want to patch it
21:50 sri for example i don't see v4 compat support
21:50 und3f i don't understand you
21:51 sri i think you know exactly what i mean
21:51 und3f whats wrong? There are tests, some comments and code is perltydied
21:51 sri "ip6 address (must be in full format)"
21:51 sri thats your comment
21:51 und3f oh, i forgot
21:51 JonChisciotte Hi, is it possibile with an html form to make a "delete" request? I've read to add an hidden control named "_method" but it seems it doesn't work...
21:51 und3f sri, we need some ip utils in mojo::url
21:52 und3f sri, i think it is stupid to do URL transformation in MOjo::IOLoop...
21:52 sri und3f: agreed
21:52 sri something to turn ip addresses into binary form
21:53 und3f after we got method to transform any ipv6 format to ipv6 everything will be just fine
21:53 und3f sri, so that wasn't my part of work - thats why i didn't it
21:53 yko left #mojo
21:54 sri und3f: well, it is required for ptr support
21:54 und3f ptr supported - check out tests :P
21:55 sri it is not complete without proper ipv6 handling
21:55 sri i would have to complete it
21:55 und3f ok. Will you write ipv6 transforming code?
21:55 sri don't think so
21:56 und3f btw, mojo don't support ipv6
21:56 und3f according to latest commits
21:56 und3f so why we need to work with ipv6 ptr records?
21:57 sri because at some point we will support ipv6 again
21:58 snearch joined #mojo
21:58 sri actually ipv6 address transformation might allow that
21:58 und3f oh, thats around 10 minutes of programming, i guess...
21:58 sri maybe
21:59 und3f where to put that?
21:59 sri i would expect something like inet_pton
21:59 sri Mojo::URL->pton?
21:59 sri takes ipv4 or ipv6 address and turns it into their binary form
22:00 und3f sri, can you write tests for that method and i'll write method?
22:00 und3f according to tests
22:00 sri hmm
22:01 und3f sri, you better know how it should look like. Just send me a diff and i'll make a method.
22:02 sri actually thats easy :D
22:02 sri do you have perl 5.12?
22:02 und3f sri, yes
22:03 sri perfect, i'll give you examples
22:06 snearch left #mojo
22:08 sri perl -E'use Socket qw/inet_pton AF_INET AF_INET6/; say unpack "H*", inet_pton(AF_INET6, "::1"); say unpack "H*", inet_pton(AF_INET, "127.0.0.1")'
22:08 sri und3f: thats all
22:08 sri we really just need a portable inet_pton
22:09 sri actually it would fit in very well into Mojo::Util i guess
22:09 und3f why not to use inet_pton?
22:09 sri 5.12 only
22:09 und3f copy paste :)
22:10 sri c code
22:10 und3f ah...
22:10 sri well, ok, we might be able to make ipv6 a 5.12+ only feature
22:11 sri hmm
22:11 und3f thats much better than rewrite existing functions
22:23 Alias left #mojo
22:28 sri think thats what we'll do
22:28 sri but sockets need to be built manually
22:30 und3f vim sucks on text with long lines
22:31 und3f It took around 5 seconds to create new line :|
23:19 OpossumPetya joined #mojo
23:20 GitHub198 joined #mojo
23:20 GitHub198 mojo: master Sebastian Riedel * eeefe82 (3 files in 3 dirs): added PTR record type support to Mojo::IOLoop resolver - http://bit.ly/9Y1rQ0
23:20 GitHub198 left #mojo
23:21 OpossumPetya left #mojo
23:21 GitHub143 joined #mojo
23:21 GitHub143 mojo: master Sebastian Riedel * 6b04386 (1 files in 1 dirs): cleanup - http://bit.ly/9ynEgE
23:21 GitHub143 left #mojo
23:24 JonChisciotte left #mojo
23:24 flamey left #mojo
23:35 und3f sri, you didn't included my name to changelog
23:39 und3f thats not fair
23:44 und3f is now known as und3f[A]
23:47 GitHub94 joined #mojo
23:47 GitHub94 mojo: master Sebastian Riedel * 320bb27 (1 files in 1 dirs): cleanup - http://bit.ly/9naxpN
23:47 GitHub94 left #mojo
23:47 sri und3f[A]: i'm distracted thinking about the next step regarding ipv6
23:48 sri the amount of complaints recently is giving me flashbacks from catalyst :D

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