The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2015-01-14

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

All times shown according to UTC.

Time Nick Message
00:18 Averna joined #mojo
00:42 good_news_everyon joined #mojo
00:42 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/TBqmCQ
00:42 good_news_everyon mojo/master 4dc7887 Sebastian Riedel: more consistent route examples
00:42 good_news_everyon left #mojo
00:45 average joined #mojo
00:46 average left #mojo
01:32 zivester joined #mojo
01:41 neyasov joined #mojo
01:56 asarch joined #mojo
02:33 klapperl joined #mojo
02:58 davido__ joined #mojo
03:05 sri rofl http://readwrite.com/2015/01/13/io-js-fork-node-js-version-1-0
03:06 sri gotta love how the project for experimenting is the one that hits 1.0
03:06 sri ¯\_(ツ)_/¯
03:12 Grinnz haha
03:12 Grinnz well, a more focused project can stabilize better, perhaps
03:20 sri still bugs me a little that we have dependencies that are not 1.0 :S
03:20 Grinnz heh
03:21 Grinnz everyone interpretation of what 1.0 means is a little different, though
03:21 sri yea, and that should be documented
03:22 sri although, there is really no excuse for "stable" 0.x releases imo
03:22 Grinnz perhaps "not tested enough"?
03:22 Grinnz in real-world usage
03:24 sri then it's simply not stable
03:26 genio node's been around for way too long to not have a > 0.x release
03:27 noganex_ joined #mojo
03:27 Grinnz true, node is used in lots of production systems
03:27 sri ¯\_(ツ)_/¯
03:30 neyasov joined #mojo
03:56 Grinnz woo cpantesters is back!
04:03 Grinnz by the way sri, looking at EV and POE's usages of poll, it seems the practice is to send both read and write events on POLLERR | POLLHUP
04:05 Grinnz https://metacpan.org/source/RCAPUTO/POE-1.366/lib/POE/Loop/IO_Poll.pm#L322
04:05 sri Grinnz: the layer above doesn't actually check for errors on write... so... meh
04:05 jberger yeah, I should bump a bunch of my modules to 1.0 for the semantic meaning
04:05 jberger my thing is that I've never thought of many of my modules as "production" level until recently
04:05 Grinnz i worked around it, just letting you know how its handled in other places
04:06 sri Grinnz: if you can make stuff "better" and prove it with tests, be my guest ;)
04:06 Grinnz i may take a look :P
04:06 sri (checking for errors on write literally makes things a lot worse though for http)
04:07 Grinnz i still don't know how to force a POLLERR or POLLHUP though
04:07 sri the famous connection closed on huge file upload problem
04:08 sri example https://github.com/kraih/mojo/issues/717
04:09 sri there you get a write event with an error before you've been able to read the whole 4xx/5xx response
04:10 sri by only handling errors on read you can work around that, and make sure the connection close only arrives after reading the whole response
04:10 sri fun problem
04:10 sri only took me 5 years to kinda get right
04:10 Grinnz heh
04:11 Grinnz makes it a little harder to use the reactor for generic filehandle watching though :P
04:11 Grinnz but i'm fine with that for now
04:13 Grinnz what happens when you're using the EV backend though? especially if it's using poll, it will send both events
04:13 sri like i said, the layer above doesn't care
04:14 Grinnz so then couldn't the poll reactor be changed to send both as well, without changing anything else?
04:14 sri consistency between the reactors is a worthy goal
04:14 Grinnz EV already introduces enough headaches for me :P
04:15 Grinnz it messes with signals, it messes with exceptions
04:15 sri my warning was in case you were planning to test with the stream api
04:16 sri muhahaha... io.js officially calls it 1.0.0 (unstable), and node.js 0.10.35 (stable)
04:17 Grinnz seriously, $EV::DIED is the only way to handle exceptions in EV watchers, it's idiotic
04:17 Grinnz so if say you use my POE backend which has to set that, you can't set your own as well
04:17 Grinnz i should probably add that to the docs
04:18 Grinnz 1.0.0 unstable? lol
04:22 Grinnz_ joined #mojo
04:29 irq joined #mojo
04:31 dotandimet joined #mojo
04:50 Grinnz anyway what i was looking into EV for in the first place... i was wondering what the port backend was for, and apparently it's for solaris 10, so lol
04:51 Grinnz apparently perl does build on solaris though?
04:53 preaction yep
04:58 Grinnz ... if the port backend is for solaris 10, why does it work on my linux boxes?
04:58 sri iocp is actually also a windows thing
04:59 sri but EV doesn't support it there :(
05:00 sri one of the early big additions to libuv was iocp on windows
05:00 Grinnz windows is stuck with select on EV eh
05:01 sri it does actually have poll()
05:01 sri :)
05:01 sri of course that's only a little bit less shitty
05:01 Grinnz ev docs say poll is "everywhere except windows" lol
05:01 sri tell that to IO::Poll!
05:02 Grinnz but ev docs also say port is for solaris
05:02 sri mojolicious literally doesn't install without poll()
05:03 sri select() is quite pathetic really, you can't watch a file descriptor above 1024
05:04 sri operating system gives you fd 1025? TOUGH LUCK!
05:05 Grinnz heh
05:21 neyasov joined #mojo
05:48 disputin joined #mojo
06:10 hshong joined #mojo
06:16 Eddy joined #mojo
06:29 dotan joined #mojo
06:59 Lee joined #mojo
07:01 McA joined #mojo
07:05 juikuen joined #mojo
07:07 neyasov joined #mojo
07:10 tempire ok
07:10 tempire I'm hooked on postderefs now
07:10 tempire or
07:11 tempire at least, natural with them
07:12 irq joined #mojo
07:15 dotan joined #mojo
07:23 reneeb joined #mojo
07:37 rem_lex joined #mojo
07:44 rem_lex|pivo joined #mojo
07:46 hesperaux joined #mojo
07:47 basiliscos joined #mojo
07:52 Shaeto joined #mojo
07:56 sugar joined #mojo
08:03 Lee joined #mojo
08:05 crab what is a postderefs
08:06 Adura joined #mojo
08:14 eseyman joined #mojo
08:16 trone joined #mojo
08:20 Vandal joined #mojo
08:23 neyasov joined #mojo
08:23 reneeb crab: http://search.cpan.org/dist/perl-5.19.5/pod/perldelta.pod#Experimental_Postfix_Dereferencing
08:27 basiliscos joined #mojo
08:28 Grinnz i am having a hell of a time recreating a POLLHUP | POLLERR response on a socket without POLLIN or POLLOUT
08:29 Grinnz it happens in the POE test because the connection fails immediately, but with IO::Socket::INET the connection happens in the same call that it creates the handle to monitor... i can't get a handle, then try to connect
08:46 irq joined #mojo
08:56 sugar_ joined #mojo
08:59 Grinnz well, i think i have a workable test case anyway
09:07 Grinnz sri, of course, EV passes my tests when under the poll and epoll backends, and fails it under select
09:08 Grinnz so i don't think adding these tests to reactor_ev.t is a good idea...
09:09 Grinnz but i still think the change is worthwhile to be consistent with other IO::Poll usage
09:23 neyasov joined #mojo
09:26 fhelmberger joined #mojo
09:29 Grinnz i'm going to see if i can come up with a better test case before submitting a pr
09:46 denny joined #mojo
10:26 irq joined #mojo
10:27 ashimema joined #mojo
10:35 Vandal joined #mojo
11:07 dod joined #mojo
11:12 neyasov joined #mojo
11:14 basiliscos1 joined #mojo
11:15 denis_boyun joined #mojo
11:40 alnewkirk joined #mojo
12:13 neyasov joined #mojo
12:27 hesperaux joined #mojo
13:00 neilhwatson joined #mojo
13:23 rem_lex joined #mojo
13:25 asarch joined #mojo
13:45 neyasov joined #mojo
14:03 dvinciguerra joined #mojo
14:06 D4RK-PH0ENiX joined #mojo
14:25 amon joined #mojo
14:30 zivester joined #mojo
14:38 marty joined #mojo
14:47 arthas joined #mojo
15:00 arpadszasz joined #mojo
15:05 davido__ joined #mojo
15:10 good_news_everyon joined #mojo
15:10 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/T_0F9w
15:10 good_news_everyon mojo/master f8776a3 Sebastian Riedel: fixed typo in Mojo::Template example
15:10 good_news_everyon left #mojo
15:12 sri Grinnz: make sure the test is also portable... os x, linux, *bsd and especially windows
15:13 Grinnz sri, i don't have such test environments.. but another dilemma, if it's not consistent between EV backends should it even be something we test for?
15:14 Grinnz although, the POE test works on all EV backends, so there must be a way
15:19 sri sounds like you simply can't expect consistency there
15:20 batman Haha! This is so true https://twitter.com/iamdevloper/status/555027652986552320
15:22 Ptolemarch joined #mojo
15:22 Grinnz sri, exactly
15:23 Grinnz sri, but it is still what every other poll backend seems to do... but i think i will just drop it
15:23 Grinnz not worth my time
15:27 sri oh, http/2 last call expired... so it's going to be an rfc in a few days
15:27 sh4 joined #mojo
15:33 neyasov joined #mojo
16:13 gryphon joined #mojo
16:32 absolut_todd joined #mojo
16:57 Eddy joined #mojo
16:58 Ptolemarch joined #mojo
17:01 mib_0cgpu4 joined #mojo
17:06 jkramer joined #mojo
17:06 jkramer Ahoy
17:07 jkramer I have a question about validating placeholders within the route. I'd like to make sure one placeholder is a date formatted like YYYY-MM-DD.
17:08 jkramer Now the docs say I can use regex, but not ^ or $. So I guess my regex \d{4}-\d{2}-\d{2} would also match 2015-01-14foobar?
17:08 jkramer Or is there another way to make this work?
17:09 Grinnz_ you could use \b so that it at least needs a word boundary... but that would still match 2015-01-14-asdf
17:10 jkramer And while I'm here, another question: can I combine default values and validation? E.g. get '/:date' => [ date => qr/.../ ] => { date => '1970-01-01' } => sub ...
17:10 Grinnz_ maybe just validate the parameter in the action ;)
17:10 jkramer Hmm yes, but I found the idea nice to do this in the route :)
17:11 Grinnz_ seems overcomplicating it to me, but if there is a way sure
17:11 sri you could just try
17:11 jkramer I like the idea that an invalid date immediately returns a 404 instead of calling my sub and then returning some error
17:12 Grinnz_ the sub can return 404 much the same way
17:12 Grinnz_ not as clean maybe, but beats an overcomplicated regex
17:13 sri you two are making guesses and treating them as facts
17:13 sri perl -Mojo -E 'a("/:foo" => [foo => qr/\d{4}-\d{2}-\d{2}/] => {text => "works!"})->start' get /2015-01-14foobar
17:13 sri this is what we have one-liners for
17:14 Grinnz_ oh, its already bounded :)
17:14 jkramer Heh, nice
17:14 jkramer Didn't know about -Mojo
17:15 sri http://mojolicio.us/perldoc/ojo
17:15 jkramer Yeah, NOW I know it exists :)
17:16 jkramer Might be overkill, but can I use subs for validating? [ date => sub { is_date(@_) } ] or something?
17:16 ribasushi jkramer: in my personal opinion `perl -Moo` is much much better
17:16 sri btw. i would still like to find a better solution for this https://github.com/kraih/mojo/blob/master/lib/Mojolicious/Commands.pm#L28
17:17 sri just popped up on #perl
17:17 jkramer ribasushi: Could be confused with http://search.cpan.org/~haarg/Moo-1.006001/lib/Moo.pm
17:17 Grinnz_ Moo already has -Moo
17:18 Grinnz_ i think thats what he was referring to
17:18 ribasushi jkramer: https://metacpan.org/source/HAARG/Moo-1.006001/lib/oo.pm
17:18 sri when you want to do app->start('daemon', '-l', ...), and handle a few flags yourself before that, a -h would override the daemon command
17:18 mst jkramer: that's what -Moo is for
17:19 mst jkramer: there's an oose.pm in Moose for -Moose too
17:20 denny heh
17:20 jberger ribasushi: how dare you question ojo!
17:20 jberger I demand satisfaction!
17:20 jberger pistols at dawn
17:20 purl hand grenades at half past!
17:20 jberger hahahahha
17:20 sri and that kills the serious discussion i wanted to have...
17:21 Grinnz_ sri: i would expect -h to override any further arguments (if thats what you meant)
17:21 Grinnz_ most applications do that
17:23 jberger sri: sorry about that, I missed that you hard steered the topic a slightly different direction
17:23 neyasov joined #mojo
17:24 jberger can you summarize what was asked on #perl?
17:24 jberger I don't understand the context
17:24 jberger (is #perl logged?)
17:24 Grinnz_ not publicly
17:24 Grinnz_ i can't find the original question
17:25 gryphon joined #mojo
17:26 Grinnz_ jberger: http://fpaste.org/169638/21256420/
17:29 Grinnz_ found the question: http://fpaste.org/169639/56577142/
17:29 dod joined #mojo
17:30 Grinnz_ (referring to App::SimpleHTTPServer
17:30 Grinnz_ )
17:31 Grinnz_ sri: so i see the problem here is you can't specify your own -h
17:32 Grinnz_ or you can, and this app just doesn't
17:32 Grinnz_ ?
17:33 sri this app just doesn't
17:35 sri allright
17:35 sri here we have an app https://gist.github.com/anonymous/dfaff6f9de0b20ec9c6e
17:35 sri you can start it with "perl app.pl"
17:36 sri but if you do "perl app.pl -h" it shows the daemon help
17:36 Grinnz_ hmm
17:36 sri same for --mode and --home, both are global
17:36 sri all the command specific options do not work, app->start(...) overrides those
17:37 sri the root of the problem is that we need --mode at compile time
17:38 sri https://github.com/kraih/mojo/blob/master/lib/Mojolicious/Commands.pm#L41
17:38 sri it happens there
18:08 irq joined #mojo
18:16 basiliscos joined #mojo
18:20 dotandimet joined #mojo
18:24 sri perhaps a warning
18:24 purl it has been said that perhaps a warning is in order when these special words are used in routes
18:24 sri "Note that the options C<-h>/C<--help>, C<--home> and C<-m>/C<--mode>, which are shared by all commands, will always be extracted from C<@ARGV> during compile time."
18:28 good_news_everyon joined #mojo
18:28 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/gNWqfA
18:28 good_news_everyon mojo/master 45ea7b0 Sebastian Riedel: @ARGV is not ignored
18:28 good_news_everyon left #mojo
18:31 Grinnz_ yeah explaining it in the docs would help
18:31 sri i'm not so sure
18:31 sri bad explanation can make it worse
18:32 Grinnz_ well right now, it's just a surprise :P
18:33 Grinnz_ i haven't done enough with commands to know how to explain it
18:34 ente if I write a client for a TCP based service, is it a good idea to use Mojo::IOLoop::Client as a base class?
18:35 mst I'd stash an object of that as an attribute rather than subclassing
18:35 ente I've been thinking about both of these options and now I'm wondering what the "proper way" is
18:35 sri not a good idea
18:35 sri Mojo::IOLoop::Client creates a Mojo::IOLoop::Stream object
18:36 mst has-a is correct, isa will be a disaster
18:36 mst basically, trying to subclass the IOLoop bits is like trying to have sex down a nine foot metal pipe - it's theoretically possible, but it isn't clever, it won't be comfortable, and it wasn't designed for that.
18:38 crab how wide is the pipe?
18:38 mst crab++
18:39 neyasov joined #mojo
18:43 ente alright, thanks!
18:56 disputin joined #mojo
18:57 disputin joined #mojo
18:59 Ptolemarch joined #mojo
19:08 ente and how do I write to a Mojo::IOLoop::Client instance? Just ->write?
19:09 ente or do I need to write to the Stream instance?
19:10 sri http://mojolicio.us/perldoc/Mojo/IOLoop#SYNOPSIS
19:10 ente ah, that's where the docs are. I only looked at the perldoc for Client
19:11 good_news_everyon joined #mojo
19:11 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/POjRcA
19:11 good_news_everyon mojo/master 0283ada Sebastian Riedel: mention which options are parsed during compile time
19:11 good_news_everyon left #mojo
19:18 sri btw. those http/2 criticisms i linked to have been resolved by ignoring them until after the last call deadline
19:18 Grinnz_ hehe
19:19 Ptolemarch joined #mojo
19:24 trone joined #mojo
19:24 jberger sri: hasn't "ignore criticism" basically been the MO of the working group anyway though?
19:25 dotandimet joined #mojo
19:27 juikuen joined #mojo
19:30 ente the phk post?
19:31 ente he's the only vocal http/2 critic I know about from the top of my head
19:37 denis_boyun joined #mojo
19:47 tempire hmm
19:47 tempire wonder if postbox is a good replacement for mail
19:47 Grinnz_ as in the CLI "client"?
19:48 Grinnz_ mailx?
19:48 purl well, mailx is a recommends
19:48 Grinnz_ wat
19:48 Grinnz_ i use mutt lately, if i need one
19:48 tempire no, gui
19:48 tempire http://www.postbox-inc.com/
19:49 Grinnz_ site doesn't work...
19:49 tempire it's just slow
19:50 tempire it's part of the macheist bundle
19:50 Grinnz_ there it goes... chrome said it wasnt available, then loaded it :P
19:50 Grinnz_ is mail a mac thing?
19:50 tempire yes
19:50 Grinnz_ that makes a lot more sense, then
20:02 dotandimet joined #mojo
20:09 neyasov joined #mojo
20:23 Grinnz_ i want to make a command that runs some code, and then runs the existing "get" command... but using the same Mojo::UserAgent as the previous code
20:23 Grinnz_ any simple way i could do this?
20:24 Grinnz_ (for authentication)
20:32 jberger ente: maybe my view is biased because I get most of my info from sri, but it seems like there has been a steady undercurrent of criticism of it, mostly (1) only helps big companies (2) doesn't address known problems of the current spec
20:33 rem_lex joined #mojo
20:33 jberger Grinnz_: hehe, that's why I built my Basic Auth infrastructure the other week
20:34 jberger Grinnz_: I would suggest building your own command
20:34 Grinnz_ jberger: yeah basic auth doesn't quite cut it for this.
20:34 jberger its not that had
20:34 jberger hard
20:34 rem_lex joined #mojo
20:34 Grinnz_ jberger: my next idea was just to copy the entire get command, and stick my stuff at the beginning
20:35 jberger Grinnz_: totally valid
20:35 Grinnz_ sure... just rings my "duplicate code" warning bells :P
20:35 jberger http://mojolicio.us/perldoc/Mojolicious/Command/get#DESCRIPTION
20:36 jberger you can "delegate" to commands, but I'm not sure you can do so with the same attributes
20:36 jberger actually ....
20:36 Grinnz_ well the problem is the get command actually creates the Mojo::UserAgent it uses
20:37 Grinnz_ so i can't have it use an existing one, or set a cookie in it
20:41 jberger if you wanted to play that game, submit a proposal to sri (and the team) to make that 'has ua' so that you could just create an instance of the get command class and delegate to that
20:42 jberger or
20:43 jberger make a proposal to be able to serialize that cookie jar so that it can be reused between runs
20:43 jberger that would take more work, but would be similar to wget and curl etc
20:44 sri ente: what jberger said, there's been criticism every other week for years
20:44 sri ente: but also this one from last week https://lists.w3.org/Archives/Public/ietf-http-wg/2015JanMar/0106.html
20:44 Grinnz_ for my needs, the ua attribute would be much easier... the idea is for the user not to have to touch authentication
20:44 sri some important points regarding security
20:44 sri completely ignored
20:45 irq joined #mojo
20:47 Grinnz_ more regarding "https is inconvenient", but he has good points
20:48 Grinnz_ https is certainly not newbie-friendly to set up
20:49 sri http:// addresses will never use http/2, just think about that
20:49 sri what kind of premise is that for the most important protocol on the internet?
20:49 Grinnz_ personally i think all communication should be encrypted, but that https/tls is not an ideal method
20:50 genio "If someone clicks an https:// link, but their browser does not
20:50 genio support https"
20:50 genio wha?!
20:51 Grinnz_ that browser wouldn't support http/2 either, so that point is moot
21:03 Grinnz_ hmm.. what's the correct way to get my application to recognize a new command? i.e. to set the namespaces for Mojolicious::Commands
21:03 Grinnz_ would i do that in the application script?
21:03 Grinnz_ before it calls start_app
21:03 sri http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#Adding-commands-to-Mojolicious
21:04 Grinnz_ aha, $self->commands
21:04 Grinnz_ thanks
21:05 dotandimet joined #mojo
21:05 * sri was just wondering if the get command could use $self->app->ua
21:06 Grinnz_ i could work with that i think
21:06 sri doesn't appear to work though
21:08 sri i'm -1 on an attribute just to make subclassing easier
21:08 sri if someone gets $self->app->ua working, that would be fine
21:08 sri and consistent with ojo
21:09 Grinnz_ not even subclassing, i would use it to dispatch to the command
21:09 Grinnz_ to subclass i have to override all of run() which is what i'm doing anyway
21:10 sri i think you have two options, a) make $self->app->ua work, or b) fork the get command
21:10 Grinnz_ the problem with $self->app->ua is it modifies $ua, which the app may not expect
21:11 Grinnz_ although its really just max_redirects it looks like
21:12 Grinnz_ well, and the $ua's cookie jar etc,
21:12 Grinnz_ i dunno if that's a good idea
21:13 human39 joined #mojo
21:13 sri we will never support serializing the cookie jar
21:14 sri btw. this is the one-liner you need to get working
21:14 sri perl -Ilib -Mojo -E 'a(sub { my $c = shift; Mojo::IOLoop->timer(3 => sub { $c->render(text => "works") }) })->start' get /
21:15 Grinnz_ actually if $ua was an attribute i could just do sub run { my $self = shift; ... do my stuff with $ua ... $self->SUPER::run(@_) }
21:16 Grinnz_ which is basically the same without subclassing it anyway
21:17 Grinnz_ ignore that statement, i realized halfway through it that's what i was already trying to do
21:19 dod joined #mojo
21:20 Grinnz_ hrm... its not finding my command
21:20 Grinnz_ its definitely trying to load the module though
21:20 Grinnz_ derp... PEBCAK
21:29 dotandimet joined #mojo
21:34 Grinnz_ sri: did you say someone added -f for JSON form data to the get command before? cause that would be really helpful
21:35 Grinnz_ (that's a thing in ojo, right?)
21:36 Grinnz_ maybe not... dunno what i was thinking of
21:38 marmez joined #mojo
21:42 dotandimet joined #mojo
21:48 disputin joined #mojo
21:49 jberger Grinnz_: sri proposed that but I don't think it went anywhere
21:50 jberger having the get command use the app's ua is a good idea though
21:50 jberger Grinnz_: I would tell you to make a PR for that
21:50 * jberger gets on the road
21:50 jberger o/
21:53 Grinnz_ i will later today
21:53 Grinnz_ at least, try it out
21:58 neyasov joined #mojo
22:04 sri i have a working patch
22:05 sri https://gist.github.com/anonymous/f0aec8589b1771180d69
22:07 Grinnz_ nice
22:09 Grinnz_ new problem... i want my custom get command to print a newline once it is done streaming (if there is no selector)
22:09 Grinnz_ i set $\ = $/ and that works, but it screws up prints within the app...
22:10 Grinnz_ oh, i guess i could just set that inside the read event
22:10 Grinnz_ callback
22:10 purl callback is called and the user assumes the object is faulty and can no longer be used
22:10 Grinnz_ no, that doesnt work either, it could be chunked
22:13 denis_boyun joined #mojo
22:13 good_news_everyon joined #mojo
22:13 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/hc_Waw
22:13 good_news_everyon mojo/master fe192a0 Sebastian Riedel: improved get command to use the user agent of the application
22:13 good_news_everyon left #mojo
22:18 Grinnz_ i suppose the drain event isnt helpful for this, that's for writing to content, right?
22:19 Grinnz_ eh... i'll just use $\ = $/ in the read callback, it's not like it will matter if there is a random newline in my test output
22:22 ente sorry I'm asking so many questions here, I'm trying to get started, but
22:23 ente I have a Mojolicious::Lite application and I want to connect somewhere using IOLoop, how do I start the IOLoop?
22:23 Ptolemarch joined #mojo
22:24 sri http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#REAL-TIME-WEB
22:30 neyasov joined #mojo
22:33 sri i guess Mojolicious::Commands is now our ugliest class
22:34 sri because of the option hacks
22:36 ente ha, finally. thanks
22:39 ente is it safe to pass $loop and $stream out of the client into my class?
22:39 ente hm, guess I don't need to
22:39 ente sorry, I'm slightly confused :|
22:40 ente I should think first and ask second
22:42 sri sometimes it helps to think loud
22:42 ente it does, but I feel like I'm making too much noise :)
22:42 sri not yet
22:59 juikuen joined #mojo
22:59 sri jberger, marcus, batman, tempire, crab: maybe one of you has an idea for improving Mojolicious::Commands?
23:04 stl joined #mojo
23:05 good_news_everyon joined #mojo
23:05 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/qRj0hw
23:05 good_news_everyon mojo/master 5f39e3d Sebastian Riedel: the argument did not serve a purpose
23:05 good_news_everyon left #mojo
23:08 good_news_everyon joined #mojo
23:08 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/qvV8UA
23:08 good_news_everyon mojo/master 181a9e2 Sebastian Riedel: more environment detection tests
23:08 good_news_everyon left #mojo
23:10 d4rkie joined #mojo
23:11 good_news_everyon joined #mojo
23:11 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/OAkOhg
23:11 good_news_everyon mojo/master e4caa1b Sebastian Riedel: order of methods is not significant
23:11 good_news_everyon left #mojo
23:20 * sri wonders if maybe we should just get rid of --mode and --mode options
23:20 trone joined #mojo
23:21 sri s/--mode/--home/
23:22 sri --help is no problem, that can be done later too
23:22 sri but --mode and --home are what makes the whole option handling ugly
23:42 Grinnz_ personally i only use --mode for morbo, which is handled through a separate script anyway
23:42 Grinnz_ but i'm hardly a representative of the whole :P
23:42 sri right, this only affects commands
23:43 sri specifically daemon and prefork i suppose
23:43 Grinnz_ i could see --mode=production on a get command if you had some mode-specific functionality perhaps

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