Camelia, the Perl 6 bug

IRC log for #mojo, 2012-08-10

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

All times shown according to UTC.

Time Nick Message
00:07 daniel_ joined #mojo
00:15 jnap joined #mojo
00:20 amirite joined #mojo
00:30 jnap left #mojo
01:05 laouji joined #mojo
01:06 scott joined #mojo
01:06 vervain joined #mojo
01:15 xaka joined #mojo
01:17 bird joined #mojo
01:59 d4rkie joined #mojo
02:38 Psyche^ joined #mojo
02:57 noganex_ joined #mojo
03:08 xaka joined #mojo
03:23 GitHub53 joined #mojo
03:23 GitHub53 [mojo] kraih pushed 1 new commit to master: http://git.io/ZssjTA
03:23 GitHub53 [mojo/master] fixed 1xx, 204 and 304 response support (closes #369) - Sebastian Riedel
03:23 GitHub53 left #mojo
03:24 xaka joined #mojo
03:27 yko joined #mojo
03:29 asarch joined #mojo
03:38 GitHub99 joined #mojo
03:38 GitHub99 [mojo] kraih pushed 2 new commits to master: http://git.io/eJZRAA
03:38 GitHub99 [mojo/master] test finish event - Sebastian Riedel
03:38 GitHub99 [mojo/master] test 100 response with events - Sebastian Riedel
03:38 GitHub99 left #mojo
03:42 yko joined #mojo
04:21 GitHub134 joined #mojo
04:21 GitHub134 [mojo] kraih pushed 2 new commits to master: http://git.io/uBOeCQ
04:21 GitHub134 [mojo/master] renamed no_body attribute to skip_body - Sebastian Riedel
04:21 GitHub134 [mojo/master] renamed has_no_body method to is_empty - Sebastian Riedel
04:21 GitHub134 left #mojo
05:27 ovnimancer joined #mojo
05:41 bird joined #mojo
05:41 Britzel joined #mojo
05:48 sri gotta love rfc 2616
05:48 sri The response MUST NOT include an entity.
05:49 sri in the description of the 205 status code
05:49 sri for comparison
05:50 sri Any response message which "MUST NOT" include a message-body (such as the 1xx, 204, and 304 responses and any response to a HEAD request)
05:50 laouji_rouge joined #mojo
05:51 sri so i've actually searched through the http-wg list to find an explanation
05:51 sri turns out entity in that context means that a 205 response needs Content-Length: 0, always
06:37 kmx joined #mojo
06:44 spleenjack joined #mojo
06:58 Vandal joined #mojo
07:06 pau4o joined #mojo
07:08 batman joined #mojo
07:10 taryk joined #mojo
07:33 Leandr joined #mojo
07:50 tabbi joined #mojo
07:53 jmmills joined #mojo
07:53 cosmincx joined #mojo
07:57 sri hmm, wonder if there's a better name for Mojo::Message::Response->is_empty
08:14 batman i vote for is_not_unempty() !
08:14 batman :)
08:14 * batman brews some coffee
08:14 * sri brews some perl
08:14 batman hehe
08:18 bjoernfan batman: Or a proper pure-ftpd-style with isnt_not_unempty()
08:19 batman bjoernfan: we could also do it the php way: isnt_NotUn_empty()
08:20 sri it doesn't have to be about empty/unempty... http://mojolicio.us/perldoc/M​ojo/Message/Response#is_empty
08:20 sri it just checks if the current status code belongs to the group of responses that can't have a message body
08:21 batman sri: sorry for talking crap. it's just that i can't see when i want to call that method :/
08:21 sri you won't ever call it ;p
08:22 batman are you sure? i'm doing Mojo::IOLoop->server directly ;)
08:22 sri well, that one is a lot more versatile
08:23 batman versatile?
08:24 sri it's a non-blocking TCP server, what could be more versatile? :)
08:26 tabbi hi, how can i investigate a problem in a huge mojolicious sub not closing the connection after the sub is completed? (the last command in the sub was already executed, but some seconds later i got an error in firefox saying, the connection is lost).
08:26 tabbi the log also says 200 OK
08:32 tabbi any ideas?
08:35 rem_lex| joined #mojo
08:35 batman sri: is "versatile" the same as "generic"?
08:36 batman tabbi: tcpdump :)
08:36 sri kinda, i guess
08:36 sri versatile = has many uses
08:37 batman ok. i've heard versatile many times but i've never understood the word. just the context :(
08:37 batman tried to look it up even...
08:37 sri to me generic is more primitive, while something versatile might be more sophisticated
08:38 batman so... the "drain" event in ioloop-stream will never get fired if the client close the connection right? but then "close" will be fired instead?
08:38 sri english is not my first language either though ;p
08:38 batman :)
08:40 batman aiai! need to see a doctor. *bbl*
09:17 marcus_ hay
09:18 marcus_ Response->is_full
09:18 * sri waves
09:18 * sri facepalms
09:22 batman am i right about ioloop-stream?
09:23 baton8 joined #mojo
09:24 sri try and see!
09:25 batman it's so much easier to ask :P
09:25 sri doubt anyone knows the exact ioloop semantics
09:26 batman hehe... no... events live their own lives :)
09:29 batman mojolicious rocks: i've thrown away Moose and now i'm using Mojo::Base -base; instead
09:29 batman i mostly only need has() anyway
09:31 sri minimalism :)
09:31 batman yup.
09:31 batman i know that word. yay!
09:31 batman :)
09:32 batman using Applify (my module) and Mojo::Base -base; has made my scripts start on under 200ms instead of 900
09:32 batman could we add Sub::Name as an optional module?
09:32 batman would like to have the actual sub name instead of __ANON__ in my stacktrace...
09:34 * sri is not a fan of more XS
09:35 batman but if it's optional?
09:36 sri it's still a gazillion new special cases that need to be unit tested and maintained
09:36 sri that's always the problem, we add it to core, we have to maintain it
09:37 batman could i make a module Mojo::SubName which mokeypatch Mojo::Base?
09:37 batman i know i could, but do i want to?
09:37 sri we also don't know yet in which form it will be added to Perl 5.18
09:39 fhelmber_ joined #mojo
09:39 pau4o left #mojo
09:40 pau4o joined #mojo
09:43 daxim joined #mojo
09:43 arpadszasz joined #mojo
09:52 sri batman: anyway, why do you think we need Sub::Name?
09:52 batman to make stacktraces readable
09:52 sri we compile custom perl anyway
09:53 sri no need to monkey patch
09:53 batman "custom perl" ?
09:54 sri i've customized Mojo::Base to generate code like this, and it appears to work http://hastebin.com/raw/deximidice
09:55 sri in fact, i've found two attributes that had missing documentation :D
09:55 sri since Test::Pod::Coverage found them now
09:56 batman wow! didn't know Test::Pod::Coverage could detect the attributes :)
09:56 batman cool
09:56 sri it couldn't before
09:56 sri since i assume they were anon
09:57 batman oh. i guess i need to upgrade then
09:58 tabbi left #mojo
09:58 sri it's not committed
09:59 batman ok
10:35 GitHub30 joined #mojo
10:35 GitHub30 [mojo] kraih pushed 1 new commit to master: http://git.io/ICsuKQ
10:35 GitHub30 [mojo/master] improved Mojo::Base to assign names to generated subroutines - Sebastian Riedel
10:35 GitHub30 left #mojo
10:35 sri batman: that should do the trick
10:36 batman i like it
10:36 batman sri++
10:36 sri there's still quite a bit of anon stuff going on though
10:37 sri https://github.com/kraih/mojo/blob​/master/lib/Mojo/UserAgent.pm#L31
10:37 batman that's ok. i don't think you should try to fix $self->respond_to(json => sub{}, ...);
10:37 sri https://github.com/kraih/mojo/blob​/master/lib/Mojo/ByteStream.pm#L10
10:37 batman ah
10:37 sri https://github.com/kraih/mojo/bl​ob/master/lib/Mojo/Headers.pm#L8
10:37 batman i see
10:38 batman i suppose it could be fixed with eval as well...
10:38 batman but no worries
10:56 jamesw joined #mojo
11:01 sri so, i guess Mojo::Message::Response->is_empty stays
11:26 batman sri: why do you use recurring instead of just setting a timer? https://metacpan.org/source/SRI/Mojolici​ous-3.27/lib/Mojo/IOLoop/Stream.pm#L120
11:27 batman is it so you change active instead of doing remove($timer_id) and setting a new timer all the time?
11:27 sri yes
11:27 batman so it's more efficient?
11:27 sri yes
11:28 batman that's a bit strange i think...
11:28 sri if you can make it more efficient (and back it up with profiling data) knock yourself out ;)
11:29 batman nono... hehe. i'm just surprised
11:30 batman that's awesome btw. it solves a problem i got in my code :)
11:31 sri https://github.com/kraih/mojo/blo​b/master/lib/Mojo/IOLoop.pm#L229
11:31 sri that one might surprise you even more
11:32 batman how?
11:32 sri you'll know if you can figure out what it does :D
11:33 batman it removes connections which is finished gracefully?
11:34 sri *a lot* more, but don't worry if you can't figure it out ;p
11:38 batman ok
11:38 sri if you trace all side effects you'll notice that it's responsible for pretty much all performance relevant tasks, load balancing between workers and the like ;)
11:41 d4rkie joined #mojo
12:03 GitHub135 joined #mojo
12:03 GitHub135 [mojo] kraih pushed 1 new commit to master: http://git.io/LV7_KA
12:03 GitHub135 [mojo/master] warn if attributes are redefined - Sebastian Riedel
12:03 GitHub135 left #mojo
12:13 batman recurring is my friend :)
12:15 mire joined #mojo
12:15 GitHub178 joined #mojo
12:15 GitHub178 [mojo] kraih pushed 1 new commit to master: http://git.io/mZRZnA
12:15 GitHub178 [mojo/master] test response progress - Sebastian Riedel
12:15 GitHub178 left #mojo
12:23 marcus_ Mojolicious all day today #ohgod
12:26 pau4o joined #mojo
12:49 GabrielVieira joined #mojo
12:58 Lucas1 joined #mojo
13:00 CromeDome joined #mojo
13:09 jnap joined #mojo
13:28 marty joined #mojo
13:56 labrown joined #mojo
14:07 jamesw joined #mojo
14:07 preflex_ joined #mojo
14:09 gryphon joined #mojo
14:39 vel joined #mojo
14:52 mire joined #mojo
15:19 ZadYree joined #mojo
15:23 reezer joined #mojo
15:43 Britzel joined #mojo
16:00 xaka joined #mojo
16:04 zivester joined #mojo
16:07 arpadszasz joined #mojo
16:09 pau4o left #mojo
16:13 juanjose joined #mojo
16:15 Leandr joined #mojo
16:17 xaka1 joined #mojo
16:27 asarch joined #mojo
17:11 batman this is so weird... my Mojo::IOLoop based app keeps on eating memory, but i can't figure out where
17:11 batman i'm using Deve::Size on $ioloop, but it has a stable size
17:11 batman *Devel
17:12 batman and my objects seems to be cleaned up... not sure where to look :/
17:14 gryphon joined #mojo
17:28 Leandr joined #mojo
17:32 diegok Is there a better way to make same url to dispatch to different controller/action based on some request condition (something on the session) than using a router condition?
17:39 lisantra joined #mojo
18:01 jnap joined #mojo
18:10 mire joined #mojo
18:26 alexeiras joined #mojo
18:37 diegok sri: I've found a problem mixing bridged route and conditional route and was able to reproduce it here -> https://github.com/diegok/mojo-route-wtf
18:38 diegok sri: It's a bug or I'm just missing something?
18:38 diegok marcus_ tempire ^
19:11 juanjose left #mojo
20:19 xaka joined #mojo
20:55 GitHub66 joined #mojo
20:55 GitHub66 [mojo] kraih pushed 1 new commit to master: http://git.io/6Iumzw
20:55 GitHub66 [mojo/master] document and test multiple conditions - Sebastian Riedel
20:55 GitHub66 left #mojo
20:59 GitHub39 joined #mojo
20:59 GitHub39 [mojo] kraih pushed 1 new commit to master: http://git.io/gLXCCA
20:59 GitHub39 [mojo/master] show all over use cases - Sebastian Riedel
20:59 GitHub39 left #mojo
21:14 sri diegok: multiple files and you're not even mentioning the expected result, just "fails"... sorry but i'm not gonna look into that
21:22 * tempire is really happy with his recent vim plugin additions
21:22 tempire http://bit.ly/OQgwv0
21:24 dwierenga tempire: spiffy! what plugin is that?
21:25 tempire neocomplcache and syntastic
21:25 dwierenga cool thanks
21:26 tempire indent-guides is pretty cool too
21:29 marcus_ tempire: looks sharp
21:30 tempire though I would trade a limb for some rounded corners
21:31 * tempire wonders when light table will be released
21:35 marcus_ tempire: sublime text.. ;)
21:36 tempire it's too mousy
21:37 marcus_ I use it with vintage and the ex stuff
21:37 tempire oh you can add your own bindings
21:38 marcus_ yepp
21:38 tempire is there one package I can download?
21:38 marcus_ hmm?
21:39 marcus_ you can download package control, and then add other packages through it. ;)
21:42 noganex joined #mojo
21:44 * tempire is confused
21:44 tempire downloaded vintageex through package control
21:44 tempire apparently I need to do something else
21:45 tempire oh it needs enabling
21:55 tempire this really is pretty slick
22:09 sri it is
22:10 sri make sure to install the Mojolicious and PerlTidy packages ;)
22:22 xaka tempire: wow, syntastic is pretty cool, thx, but i wonder why they choose to use -w instead of -W to check the file
22:28 gryphon joined #mojo
22:36 daniel_ joined #mojo
22:38 daniel__ joined #mojo
22:38 * marty is packing up for a 30 day vacation.  Whoopie!!!
22:49 daniel_ joined #mojo
22:56 daniel__ joined #mojo
23:00 daniel___ joined #mojo
23:03 noganex joined #mojo

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