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

IRC log for #mojo, 2015-08-28

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

All times shown according to UTC.

Time Nick Message
00:35 didge joined #mojo
00:51 diopow joined #mojo
00:52 diopow $c->render(data => $bytes, format => 'png');
00:52 diopow how do i load a png file in $bytes ?
01:10 Adurah joined #mojo
01:12 pink_mist why would you do it like that? just use this way of doing it: https://metacpan.org/pod/Mojolicious::Plugin::DefaultHelpers#reply-asset
01:14 pink_mist or possibly this one: https://metacpan.org/pod/Mojolicious::Plugin::DefaultHelpers#reply-static
01:15 Adura joined #mojo
01:17 didge joined #mojo
01:18 diopow i don't want to display the file name, i suppose thsi method would reveal file name?
01:18 pink_mist noooo
01:18 pink_mist why would you think so?
01:21 diopow 'cause that's the usual behavior? the browser gets such information if its just being served
01:23 diopow well, i will try
01:30 Grinnz oh, he's gone
01:50 didge joined #mojo
02:15 kaare joined #mojo
02:17 noganex joined #mojo
02:22 PryMar56 joined #mojo
04:09 damaya joined #mojo
04:16 didge joined #mojo
04:18 zivester joined #mojo
04:46 mrallen1 joined #mojo
05:18 jb360 joined #mojo
05:28 inokenty-w joined #mojo
05:39 jwang joined #mojo
06:18 didge joined #mojo
06:39 dod joined #mojo
06:44 dod joined #mojo
06:45 bobkare joined #mojo
07:04 romel joined #mojo
07:10 dod joined #mojo
07:15 didge joined #mojo
07:18 Vandal joined #mojo
07:29 tomboh left #mojo
07:32 memowe joined #mojo
07:33 memowe_ joined #mojo
07:34 AndrewIsh joined #mojo
07:34 memowe joined #mojo
07:35 memowe joined #mojo
07:35 memowe \o/
07:53 jontaylor_ joined #mojo
07:55 eseyman joined #mojo
08:16 damaya joined #mojo
08:24 dod joined #mojo
08:32 cpan_mojo CallBackery-0.2.0 by OETIKER https://metacpan.org/release/OETIKER/CallBackery-0.2.0
08:38 cpan_mojo CallBackery-0.2.1 by OETIKER https://metacpan.org/release/OETIKER/CallBackery-0.2.1
08:45 dod joined #mojo
08:45 sri /o/
08:45 sri \o\
08:46 damaya joined #mojo
08:56 didge joined #mojo
09:02 cpan_mojo CallBackery-0.2.2 by OETIKER https://metacpan.org/release/OETIKER/CallBackery-0.2.2
09:04 sri at first i saw CallBakery
09:04 sri guess i need breakfast :S
09:10 damaya joined #mojo
09:19 sri anyone disagree? https://github.com/kraih/minion/issues/13#issuecomment-135706238
09:23 sri <shameless-promotion> Minion needs 5 more votes to enter the top 100! https://metacpan.org/pod/Minion</shameless-promotion>
09:25 sri we also have no consensus on #835 yet https://github.com/kraih/mojo/pull/835
09:26 meshl joined #mojo
09:29 cpan_mojo Minion-1.16 by SRI https://metacpan.org/release/SRI/Minion-1.16
09:32 jontaylor_ joined #mojo
09:36 memowe joined #mojo
09:49 cpan_mojo CallBackery-0.2.3 by OETIKER https://metacpan.org/release/OETIKER/CallBackery-0.2.3
09:52 Skylos_ joined #mojo
09:55 melo joined #mojo
10:02 jontaylor joined #mojo
10:07 sri argh
10:07 sri i hate it when people wait with opening an issue until after i've made a new release
10:18 Skylos_ joined #mojo
10:26 bd can I access a dbm::deep file storage from multiple processes? e.g. one producer and one consumer?
10:27 Skylos_ left #mojo
10:28 sri yes, the minion backend does that
10:29 sri but you need to do some locking
10:30 sri Minion::Backend::File uses a scope guard for that
10:32 didge joined #mojo
10:42 pink_mist why not use an sqlite db instead?
10:46 bd pink_mist: because i want to save a hash
11:05 neilhwatson joined #mojo
11:24 damaya joined #mojo
11:36 sh4 joined #mojo
11:39 dod joined #mojo
12:03 Skylos_ joined #mojo
12:11 batman is there anything in the "cpan toolchain" that reads "cpanfile"? just curious if i want to ship that when i make modules or not.
12:12 batman currently i use cpanfile to generate Makefile.PL :)
12:14 jberger batman: Module::Build::Tiny does
12:14 Skylos_ Some of the packagers like Carton use them.
12:15 Skylos_ #toolchain people know more about that.
12:15 jberger batman: I don't think there's much reason to include it if you are using eumm or mb
12:20 batman yeah, really looking forward to Module::Build::Tiny
12:20 batman jberger: ok. thanks
12:23 jberger I have a couple modules that use it
12:23 batman which one? (one example is good enough)
12:25 jberger https://github.com/jberger/Mojolicious-Plugin-Kavorka
12:25 jberger Use App::ModuleBuildTiny with it
12:29 LV joined #mojo
12:31 damaya joined #mojo
12:34 * sri finds all this build system stuff so boring
12:34 batman jberger: where does it get Module::Build::Tiny from? (trying to find it here https://metacpan.org/source/JBERGER/Mojolicious-Plugin-Kavorka-0.01)
12:34 LordVorp joined #mojo
12:36 * sri can barely look at #835 without snoozing off
12:38 Skylos_ joined #mojo
12:51 ajr_ joined #mojo
12:51 meshl joined #mojo
12:52 jberger batman: what do you mean?
12:52 LordVorp joined #mojo
12:53 batman jberger: Module::Install works because it is bundled with the module inside ./inc/. eumm works because it is bundled with core.
12:53 jberger batman: so do you think that Module::Build is dead since it is removed from core?
12:54 jberger https://metacpan.org/source/JBERGER/Mojolicious-Plugin-Kavorka-0.01/META.yml#L7
12:54 jberger configure requires in the meta files
12:54 batman aha. so META.yml is read first, stuff is done and then "perl Build.PL" is run?
12:55 batman btw: i don't think anything. i'm trying to understand how it all works.
12:56 jberger yes, that's exactly how the modern toolchain works
12:57 jberger it was created for Module::Build but since it took forever for the toolchain to catch up MB was put in core so it would be usable
12:57 batman thank you jberger
12:57 genio ah.  I was wondering about that too.
12:57 jberger you should have been at Chicago.pm last night, this was our topic :-)
12:58 genio I was looking into making an Alien:: thingy and M::B was making me not like things since I didn't understand it
12:58 batman then i see no reason why i wouldn't switch to Module::Build::Tiny :)
12:58 jberger for Alien:: authors, Alien::Base should mostly be configuration-only
12:59 jberger batman: give it a try on a simple dist
12:59 jberger its still pretty spartan
12:59 batman i only have simple dists :)
13:04 didge joined #mojo
13:09 LordV joined #mojo
13:16 hernanGOA joined #mojo
13:17 plicease joined #mojo
13:23 cpan_mojo CallBackery-0.2.4 by OETIKER https://metacpan.org/release/OETIKER/CallBackery-0.2.4
13:24 neilhwatson In mojo::pg is there an equivalent to dbi's quote_identifier (https://metacpan.org/pod/DBI#quote_identifier)?
13:32 sri no
13:32 PryMar56 joined #mojo
13:33 neilhwatson :(
13:35 sri why would there be?
13:36 sri not like it's used that much it needs a shortcut
13:38 Skylos_ <-- having lots of fun at swiss perl workshop.  I should have gone to perl confs before.
13:39 pink_mist olten?
13:40 Skylos_ Yo.
13:40 Skylos_ (that's yes in swiss-german)
13:40 pink_mist hope you're having lots of fun =)
13:40 sri also yes in northern germany
13:41 Skylos_ yess.  :D
13:42 sri we say it more like youuuuuu though
13:42 mrallen1 joined #mojo
13:42 Skylos_ swiss talk french and german in ways that annoy the french and the germans, it seems.  This amuses me.
13:43 sri indeed
13:43 * sri waves fist
13:44 * Skylos_ giggles.
13:45 * genio glares angrily at Skylos.
13:45 genio I have a song stuck in my head now and it's all your fault
13:45 * Skylos looks suitably ashamed.
13:46 genio "How do you pronounce that nick, I wonder?  sky-lo? skee-lo?  Skee-lo? ohh, I remember that song.  'I wish I was a little bit taller...'"  and it's gone downhill from there
13:46 gryphon joined #mojo
13:46 Skylos Genio - I pronounce it skai- low-sss.  the greeks cannot be dissuaded from pronouncing it ski-low-ss
13:46 genio https://www.youtube.com/watch?v=ryDOy3AosBw  Now you can get it stuck too
13:47 * Skylos cannot play vids out loud onhis notebook during talks!
13:47 Skylos I'll savei t for later though.
13:58 berov joined #mojo
13:59 damaya joined #mojo
14:02 jberger that song was essentially the theme song for a friend of mine
14:02 jberger no joke he basically hit puberty at age ~24
14:06 eitz joined #mojo
14:19 dvinciguerra joined #mojo
14:21 howitdo joined #mojo
14:30 damaya joined #mojo
14:43 asarch joined #mojo
14:52 damaya joined #mojo
14:59 neilhwatson Mojo::Reactor::EV: Timer failed: Mojo::IOLoop::Delay: Can't call method "query" on an undefined value the line is:
14:59 neilhwatson $mdbh->query( $query => $delay->begin );
15:00 jberger that says that $mdbh is undefined
15:00 neilhwatson What value is undefined? $query and $delay are both defined.
15:01 neilhwatson $mdbh is defined
15:01 jberger that's what the error is saying
15:02 Skylos maybe there's a code execution context you're not thinking of.
15:02 Skylos and its defined ins ome but not all.  That happens when doing async.
15:06 neilhwatson Here's mdbh: https://gist.github.com/neilhwatson/70c91768c1da0cfe69a0 Is something missing?
15:06 CromeDome joined #mojo
15:07 damaya joined #mojo
15:08 batman neilhwatson: show us the surrounding code instead.
15:14 neilhwatson Here: https://gist.github.com/neilhwatson/ca7da1fe72142f159ef2
15:17 Skylos mdbh isn't declared here.  hm.
15:19 btyler_ a question I was asked during my talk on delays at SPW: what's the best way to figure out which step bailed out when you're in the ->catch() sub? Just look at the $d->remaining? Make sure any exceptions are unique per step so you can look at the contents of $err?
15:19 neilhwatson Declared elsewhere. arn "mdbh = ". Dumper( $mdbh ); in the same sub returns the first gist, so it does not appear undefined.
15:19 neilhwatson s/arn/warn/
15:20 Skylos does mdbh get cleared at some point?  set to undef or such, in other code place?
15:21 Skylos when doing asynchronous actions, sometimes things run in a different order than you might expect.
15:22 neilhwatson How do you gaurd against that? Localize it at the start of the sub?
15:23 Skylos passing it in puts it on the stack, instead of depending on global/closure outers context
15:23 jberger normally if you close over an object you create it should be there
15:23 jberger the way that it wouldn't would be if you weaken it somewhere
15:23 jberger (or undef it)
15:23 Grinnz_ yeah, unless you weaken it, it'll hold a strong reference to it
15:24 Grinnz_ are you using $mdbh anywhere else?
15:24 Skylos the reference is held so not auto colelcted, hence the thought if he undeffed it.
15:24 didge joined #mojo
15:24 Grinnz_ that particular $mdbh, that is
15:27 neilhwatson Thanks I will look into this.
15:37 ZadYree joined #mojo
15:51 didge joined #mojo
15:56 cpan_mojo Mojolicious-Plugin-TagHelpers-Pagination-0.06 by AKRON https://metacpan.org/release/AKRON/Mojolicious-Plugin-TagHelpers-Pagination-0.06
16:18 didge joined #mojo
16:27 Flying_Squirrel joined #mojo
16:51 cpan_mojo Minion-1.17 by SRI https://metacpan.org/release/SRI/Minion-1.17
16:58 lluad joined #mojo
17:08 Corgi joined #mojo
17:14 didge joined #mojo
17:22 eitz joined #mojo
17:23 damaya joined #mojo
17:31 damaya joined #mojo
17:40 sri hmm, the phoenix framework 1.0 announcement reminds me, we are still pretty weak when it comes to sending emails
17:47 jberger cpan can do it, why should Mojo have that integrated?
17:47 jberger or do you mean some Mojo::SMTP?
17:48 Grinnz_ there is Mojo::SMTP::Client
17:48 Grinnz_ that won't use the system's sendmail facility though, direct SMTP
17:49 sri jberger: what is the right cpan solution?
17:49 sri how do you integrate it with mojolicious?
17:49 jberger https://metacpan.org/pod/Email::Sender
17:49 dod joined #mojo
17:50 Grinnz_ Email::Sender doesn't have a non-blocking interface, though
17:52 sri so, should it be used from minion tasks or doesn't it block long enough usually?
17:53 Grinnz_ it depends on how many people you send to, i've had it block for a full minute
17:53 sri a writeup would be nice
17:53 lluad Sending mail can block for quite a while. Multiple seconds, easily.
17:53 jberger off hand, I would probably do it via minion
17:53 sri how to use different transports for testing too
17:54 jberger it depends on what you want from a client perspective though
17:54 jberger if I wanted to inform a web client of success I might use ForkCall
17:54 jberger but minion is nice for the retry/audit stuff
17:54 sri password reset should be the most common task
17:55 sri an example app or blog post about it would be great
17:55 jberger that would make a nice example
17:55 Grinnz_ hmm
17:55 jberger I'll see if I can mock one up this weekend, I've been bad about blogging lately
17:56 lluad Often there's not much you can do in the way of immediate feedback - as the local smarthost will accept the message and there's no feedback at all other than an synchronous bounce.
17:56 sri jberger++
17:56 Grinnz_ yeah tbh i've never had sendmail "fail", even though i am using try_to_sendmail
17:56 lluad Unless I'm doing something fancy I usually treat email as "lob it over the wall to SMTP and forget about it".
17:56 lluad To do it *right* you also need to be handling bounce management and that's a much bigger area to try and get right.
17:57 sri yes, i guess fire and forget is the most common use
17:57 * lluad <-- works in email
17:57 jberger If someone else would prefer to do such a post, I won't object
17:57 jberger but it does sound like a worth-while example
17:58 lluad Yeah.  "how to do a fire and forget password reset mail" would be useful.
18:00 jberger oh! I could do Mojo+Minion+Email::Sender+Bcrypt+Mojo::JWT
18:01 lluad I've been using Mojolicious::Plugin::Mail.
18:01 lluad Something Email::Sender based would probably be better.
18:03 jberger hmmm MIME::Lite usually gets laughed at now right?
18:03 lluad Oh, yeah.
18:03 jberger mstpan is quiet on the matter: http://shadow.cat/blog/matt-s-trout/mstpan-15/
18:03 sri perhaps a new plugin is needed
18:03 lluad I mean, MIME::Lite works just fine, but it's not a great API, and the code quality in Email::* tends to be better.
18:03 sri if only for documentation and steering people into the right direction
18:03 damaya joined #mojo
18:04 jberger Mojo::SMTP::Client is Oleg
18:04 jberger that's a good start
18:04 sri didn't we just come to the conclusion that Email::Sender is the right choice? Oo
18:05 bobkare joined #mojo
18:05 lluad Mojo::SMTP::Client looks like a decent SMTP level implementation, but it's just SMTP. Message compostion (MIME etc) is a big part too.
18:05 sri i wouldn't want to recomment raw smtp
18:06 lluad Email::Sender does composition, and sending via a variety of transports (including SMTP and SUBMIT, which are the important ones).
18:06 jberger Email::Sender can to pluggable transport
18:07 * sri is totally fine with recommending Email::Sender and a job queue
18:07 ZoffixWork joined #mojo
18:07 jberger so I imagine that a plugin for Mojo::SMTP::Client would be interesting
18:07 jberger but yeah, I'm with sri for now
18:07 ZoffixWork What tool did you guys use to generate that ASCII file list? Or was it written by hand? http://mojolicio.us/perldoc/Mojolicious/Guides/Growing#Differences
18:07 lluad Yeah, for an app that's not email-centric Email::Sender would be the way I'd want to go, I think.
18:07 sri ZoffixWork: hand
18:07 ZoffixWork aw :(
18:07 ZoffixWork Need to write one myself, but lazy :)
18:08 * sri needs to go running again, but lazy
18:09 lluad "tree" is the traditional tool to do that sort of directory tree, on unix-ish oses.
18:09 jberger http://linux.die.net/man/1/tree
18:09 ZoffixWork jberger, yeah, MIME::Lite is not recommended by its maintainer: https://metacpan.org/pod/MIME::Lite#WAIT
18:09 jberger lluad: you keep beating me to it
18:10 ZoffixWork jberger++ # I looked to see whether I had a "tree" command but didn't think of searching the repos :)
18:11 ZoffixWork lluad, thanks
18:14 ZoffixWork Heh. Went to look for a tree maker, submitted two PRs to Text::FileTree, found a tree maker... decided I didn't need a tree >_<
18:17 jontaylor joined #mojo
18:20 Grinnz Email::Sender does sendmail too, which is important, because it's the only one that will "just work" out of the box
18:20 damaya joined #mojo
18:23 lluad Yeah, Grinnz. That's the easy one, so everyone supports it. Having good SMTP/SUBMIT support isn't quite as universal. Email::Sender does both, quite nicely.
18:24 Grinnz of course, the best transport is Email::Sender::Transport::DevNull
18:24 Grinnz that one always works!
18:25 asm35 joined #mojo
18:26 lluad A plugin for mocking email delivery for testing would be very nice too.
18:37 Skylos_ joined #mojo
18:42 kaare joined #mojo
18:46 batman https://metacpan.org/pod/distribution/Mojolicious-Plugin-Mail/README.pod#TEST-MODE ?
18:48 lluad Yeah. For testing I really want the mail to go ... somewhere ... so my tests can look at the mail sent and make sure it was sent properly.
18:49 lluad I don't have a good answer for that - for some apps I have a switch that causes delivery to go to a database table, and for others my test suite spins up a private SMTP server to watch the traffic the app sends.
18:49 lluad Being able to null route it rather than send it is much better than nothing, though.
18:49 batman sri, jberger: do you have any input on this issue? https://github.com/jhthorsen/mojolicious-plugin-assetpack/issues/55
18:49 batman not sure how to do Cache-Control for assetpack the "right way"
18:51 ajr_ joined #mojo
18:52 asm35 joined #mojo
18:54 batman (or anyone else for that matter)
18:55 jberger I can't say that I would have any useful input
18:57 batman ok
19:10 sri batman: don't forget about development
19:10 sri cached assets during developent can result in a lot of trouble
19:11 sri this is why mojolicious doesn't set Cache-Control for static files
19:12 sri just the safe stuff, ETag/Last-Modified
19:13 ashimema joined #mojo
19:14 sri that's how Cache-Control became the example for nested helpers http://mojolicio.us/perldoc/Mojolicious/Guides/Rendering#Adding-helpers
19:15 * sri still likes how $c->cache_control->five_minutes looks
19:15 batman sri: it doesn't matter, since even under dev mode you get a new md5 sum
19:16 sri so the path changes?
19:16 batman yeah
19:16 sri ah, should be fine then
19:16 batman if the content of the css/js/whatever change, the path change
19:16 batman but an "after_static" hook is ok?
19:18 sri i don't know how you serve assets
19:19 batman it's done with the default Mojolicious::Static object, since the output assets are stored in public/packed
19:20 amon joined #mojo
19:20 batman maybe i should create my own Mojolicious::Static and server them under a route get "/packed" ?
19:20 sri ah, ->reply->asset
19:21 sri then no, after_static should not work
19:21 mgrimes joined #mojo
19:21 sri you already have your before_routes hook
19:22 sri you should not need anything else
19:22 batman i meant, it's done with https://metacpan.org/source/SRI/Mojolicious-6.17/lib/Mojolicious.pm#L96
19:22 batman i only use before_routes for in memory assets. not for the assets saved to disk.
19:22 sri now i'm confused about https://github.com/jhthorsen/mojolicious-plugin-assetpack/blob/master/lib/Mojolicious/Plugin/AssetPack.pm#L150
19:22 sri do you handle some assets differently?
19:23 sri that sounds problematic
19:23 batman no. i handle two cases: 1) you can write to ./public/packed 2) you cannot write to ./public/packed
19:24 sri oh wow, i think i see a security issue
19:24 batman in case #2 i inject the before_routes hook, which will serve the assets from memory
19:24 batman please tell
19:25 sri you never canonicalize the path or check for ..
19:25 sri for case #2
19:25 sri is there a way for it to reach the file system?
19:26 sri (i've not followed all the code)
19:26 batman no. the input path is just used as lookup in a %hash
19:27 sri oh, ok
19:27 sri false alarm then :)
19:27 batman or... wait a sec.
19:27 batman i will write a test :P
19:28 sri after all the trouble we've had i get nervous when i see $c->req->url->path being used verbatim ;p
19:28 batman yeah.
19:29 sri for anyone wondering, this is the relevant code in core https://github.com/kraih/mojo/blob/master/lib/Mojolicious/Static.pm#L31-L34
19:29 sri just ->canonicalize and check if the first part is '..'
19:30 sri but if it doesn't hit the file system you should be fine
19:30 sri also reminds me of the fun we've had with /.../
19:35 damaya joined #mojo
19:39 sri does the application specific plugin part fit into the description, or has it become too bloated? https://metacpan.org/pod/Minion#DESCRIPTION
19:56 Grinnz_ i'm not sure it would make sense in a section on its own, it's kind of a continuation of the previous parts
19:56 Grinnz_ maybe it could
20:04 neilhwatson FYI, I had good luck with my mojo:pg tests today. Thanks for making it.
20:04 mrallen1 joined #mojo
20:18 batman about Cache-Control: https://github.com/jhthorsen/mojolicious-plugin-assetpack/commit/84776c4d8b463f6a9d1676d37870d750ef06d1ed
20:40 talexb_ I'm having trouble getting parameter passing to work when I do a POST: http://pastie.org/10382348
20:41 talexb_ Doing GETs with embedded parameters works fine.
20:41 batman talexb_: you're posting "json", not "form"
20:41 batman json is available from $c->req->json;
20:42 diopow joined #mojo
20:43 diopow Hi
20:44 talexb_ batman, Hmm .. I tried changing it to form, and that didn't work, but I'll try your suggestion of $c->req->json.
20:45 diopow $self->defaults(storablething => ''); If the app change defaults value with some action, will this new value be exclusive for one proccess/user or will it be shared among everyone? I tried with morbo and it was shared, but i think morbo does not fork?
20:48 talexb_ batman++ Awesome! Thanks. :) I appreciate your time.
20:48 meshl joined #mojo
20:49 batman talexb_: you're welcome.
20:49 batman but if you just replaced "json =>" with "form =>" then it should work with param()
20:50 talexb_ Hmm .. I'm going to clean up and commit, then go back and test that, because eventually it's going to be a web form that's doing this.
20:51 talexb_ And form makes more sense.
20:52 Grinnz_ diopow: correct, in prefork or hypnotoad you can only affect the application within the current process if you change something within an action
20:53 Grinnz_ you should set application-wide settings in startup, this is run before the fork
20:57 diopow cool
20:58 diopow i have to setup something production for test to avoid those doubts, but there are so many options.
20:59 diopow is there any benchmark on performance among different production settings?
20:59 Grinnz_ what do you mean "production settings"?
21:00 diopow i mean deplyment, i.e. hypnotoad + nginx, fast cgi etc
21:00 Grinnz_ aha
21:00 Grinnz_ nginx proxy to hypnotoad is the most popular, and gives you the most options
21:01 Grinnz_ so if that is a possibility i would always choose that
21:01 diopow how does it compare to FCGI?
21:03 diopow the plack/fcgi option
21:03 Grinnz_ going through PSGI means you cannot use the Mojo IOLoop, which is critical for websockets or delayed rendering
21:04 Grinnz_ for a simple site it works fine, but it is just extra complication really
21:05 jberger sri / batman / lluad: how does this look for an example: https://gist.github.com/jberger/2d77bdf4dbb239eafc20
21:05 diopow i'm mostly concerned about performance
21:06 Grinnz_ hmm maybe the Cookbook deployment section could clarify the limitations of the plack method, or indicate that proxy to hypnotoad is preferred
21:06 Grinnz_ i remember being confused by the options at first as well
21:06 disputin joined #mojo
21:07 diopow i've read it all, but it does not say anythign about performance
21:07 lluad jberger: Looks like a pretty decent example.
21:07 Grinnz_ you can tune hypnotoad settings based on whether your requests are mostly blocking or non-blocking
21:08 jberger diopow: hypnotoad is comparable to Starman
21:08 batman jberger: never used minion, so i don't know :/
21:08 jberger in capabilities
21:08 Grinnz_ but using either the built-in Poll reactor or the optional EV reactor are very performant for non-blocking
21:08 jberger batman: as far as minion goes, just think of it as a complex fork_call
21:08 batman i know what it is :)
21:09 Grinnz_ diopow: https://metacpan.org/pod/Mojo::Server::Hypnotoad#clients
21:09 batman jberger: example for what..?
21:10 jberger batman: sorry, I thought you were involved in that conversation, maybe you weren't
21:10 batman ok. no worries :)
21:11 jberger batman: you did make one comment, but it was in with your work on AssetPack, so I'm not surprised that you mentally moved on
21:11 batman fg
21:12 batman oops. looking at the tv and typing at the same time:P #fail
21:12 diopow left #mojo
21:12 jberger ./batman &
21:17 juikuen joined #mojo
21:32 mrallen1 joined #mojo
21:35 sri jberger: seems a little too big for a lite app
21:35 sri diopow: performance is not that simple
21:35 asm35 joined #mojo
21:35 jberger sri: it grew of course ;p
21:40 sri hypnotoad is actually sooooo close to starman now even with small concurrency
21:55 PryMar56 joined #mojo
21:56 bpmedley sri: Perhaps adding a BEST PRACTICE section to Minion would allow for the DESCRIPTION section to be smaller and easier to process?
22:13 romel_ joined #mojo
22:23 romel joined #mojo
22:27 sri bpmedley: so you think it is not good enough now?
22:29 stephan48 sri++ # thanks for the addendum of application specific plugins to the Minion POD
22:29 bpmedley sri: Content is fantastic.  From the perspective of a newbie, is it a lot to read now?  If so, then perhaps a new section would help.
22:29 sri that's the question, is it too much, or does it work? ;p
22:30 stephan48 yea, maybe something like GROWING or so?
22:30 bpmedley My personal belief is that a new section is needed.
22:30 stephan48 i overread the plugin stuff at first and then backtracked
22:35 sri this works better? http://mojolicio.us/perldoc/Minion#GROWING
22:36 stephan48 yea
22:39 good_news_everyon joined #mojo
22:39 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vG39c
22:39 good_news_everyon mojo/master 20a8efa Sebastian Riedel: Merge pull request #835 from kraih/meta-spec-2...
22:39 good_news_everyon left #mojo
22:41 good_news_everyon joined #mojo
22:41 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vG39D
22:41 good_news_everyon mojo/master 38f5fe4 Sebastian Riedel: fix indentation
22:41 good_news_everyon left #mojo
22:43 absolut__ joined #mojo
22:44 bpmedley http://pastie.org/10382491 <-- Thoughts on this approach?
22:45 Skylos_ left #mojo
22:46 bpmedley SYNOPSIS is a copy 'n pastable example.  However, in the pastie USAGE is somewhat out of place.
22:50 sri i don't think that's necessary
22:51 sri and i do not like the USAGE block at all
22:51 Grinnz_ it's a little unorthodox for pod to have a whole section that's just code other than the SYNOPSIS
22:52 bpmedley I agree on USAGE.  Are you saying that SYNOPSIS should not necessarily be copy 'n pastable?
22:53 cpan_mojo Mojolicious-Plugin-AssetPack-0.63 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Mojolicious-Plugin-AssetPack-0.63
22:56 sri which synopsis is copypasta-able?
22:56 ajr_ joined #mojo
22:57 bpmedley I believe the one here: http://pastie.org/10382491#82
22:58 bpmedley I didn't test it.
22:58 sri umm, no, i mean out in the real world
22:58 bpmedley Oh!  One sec.
22:59 sri anyway, i do not think a synopsis should be copypasta-able
23:00 sri synopsis is a quick overview of the capabilities of a module, not an example app
23:01 bpmedley https://metacpan.org/pod/IO::Tee and https://metacpan.org/pod/Mojo::APNS
23:01 bpmedley Understood.  Let me reread your documentation with that in mind.
23:04 bpmedley The GROWING section works well, in my opinion.  I can't think of anything else.
23:05 sri Grinnz_: are you sure your Makefile.PL patch actually works?
23:05 sri oh, nevermind
23:05 sri i looked into MYMETA.yml
23:06 sri seems it is always version 1.4
23:06 Grinnz_ yes, I ran a build to diff the META.json
23:06 Grinnz_ the yml ones are stuck at 1.4
23:06 sri i assume .json is 2.0 and .yml 1.4 for backcompat
23:06 Grinnz_ right
23:06 Grinnz_ all CPAN tools that are up to date will use .json
23:07 sri wait a minute
23:07 sri https://api.metacpan.org/source/SRI/Mojolicious-6.17/META.json
23:07 sri this is already version 2
23:07 sri why do we manually include the version?
23:07 Grinnz_ EUMM parses the META_MERGE as version 1.4 if you don't
23:07 sri "'meta-spec'    => {version => 2},"
23:07 Grinnz_ and then turns that into version 2 for the META.json
23:08 Grinnz_ but you can't set version 2 keys unless you specify that it's to be parsed as version 2
23:08 sri hmm
23:15 absolut_todd joined #mojo
23:18 Grinnz_ i just noticed the .perltidyrc in the repo sets the line max to 79 chars not 80
23:18 Grinnz_ o_0
23:20 sri for reasons
23:20 sri which i don't remember
23:20 asm35 left #mojo
23:22 damaya joined #mojo
23:28 sri hmm, i really don't remember :S
23:28 sri jberger: do you?
23:29 Grinnz_ i just vaguely remember you (i think) changed the line lengths on the entire project haha
23:29 sri for docs
23:29 Grinnz_ ah yes
23:29 sri not sure why not for code
23:35 meshl joined #mojo
23:37 good_news_everyon joined #mojo
23:37 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vGseh
23:37 good_news_everyon mojo/master a6382fd Sebastian Riedel: use consistent line length
23:37 good_news_everyon left #mojo
23:38 good_news_everyon joined #mojo
23:38 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vGsvq
23:38 good_news_everyon mojo/master bf9473e Sebastian Riedel: use consistent line length everywhere
23:38 good_news_everyon left #mojo
23:38 Grinnz_ haha, that commit
23:42 Grinnz_ you just dropped 125 lines of code :O
23:44 sri :o
23:44 Grinnz_ mostly in test files though heh
23:45 sri just 10 lines in lib
23:52 jnbek joined #mojo

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