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

IRC log for #mojo, 2017-08-10

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

All times shown according to UTC.

Time Nick Message
00:36 gizmomathboy joined #mojo
00:39 dhg joined #mojo
01:12 kavuria joined #mojo
01:16 kavuria Curious if Mojo::* stack has any non-blocking loggers? Mojo::Log, per my understanding, does blocking logging (as my Mojo app hung when the filesystem filled up this morning and I had to make some space and had to restart the app).
01:17 kavuria How does others do logging in a non-blocking way (especially when the target filesystem to which Mojo app is slow and you want Mojo app to return results relatively way faster)?
01:17 marty joined #mojo
01:24 preaction a non-blocking I/O will often block when the buffers are full
01:33 aborazmeh joined #mojo
01:51 ilbot2 joined #mojo
01:51 Topic for #mojo is now � nom nom | http://mojolicious.org | http://irclog.mojolicious.org | http://code-of-conduct.mojolicious.org
02:11 noganex joined #mojo
03:08 itaipu joined #mojo
03:20 tchaves joined #mojo
03:22 aborazmeh joined #mojo
04:07 veryrusty joined #mojo
05:19 aborazmeh joined #mojo
05:31 tyldis kavuria: We do not log to disk, but ship UDP to logstash/elasticsearch
05:38 stryx` joined #mojo
06:09 trone joined #mojo
06:13 veryrusty joined #mojo
06:32 aborazmeh joined #mojo
06:48 inokenty joined #mojo
07:33 * ashimema uses tyldis approach too
07:37 mr_trousers joined #mojo
07:55 sh14 joined #mojo
07:56 * sri is switching to using systemd for logging
08:20 Vandal joined #mojo
08:31 inokenty joined #mojo
08:38 dhg joined #mojo
09:12 marcus You're live at YAPC ;p
09:13 bc547 joined #mojo
09:17 marcus Sri: mooo
09:18 sri marcus: ?
09:19 marcus Sri: I put #mojo on the projector until my talk in 7 minutes. Want to say something to the nice people?
09:19 bc547 joined #mojo
09:19 sri hi everyone! :)
09:20 marcus ????
09:23 erry joined #mojo
09:27 ashimema love it.
09:30 sri \o\
09:30 sri /o/
09:30 erry where are the slides for the talk?
10:03 vroomanj_ joined #mojo
10:05 marty joined #mojo
10:09 petru joined #mojo
10:33 marcus erry: http://marcus.nordaaker.com/MojoliciousIntroduction/#/
10:34 marcus and source at https://github.com/marcusramberg/MojoliciousIntroduction
10:35 erry thanks :)
10:36 erry marcus: does mojo do automatic routing like catalyst?
10:52 dotan_convos erry: don't know Catalyst, but see here:
10:52 dotan_convos http://mojolicious.org/perldoc/Mojolicious/Guides/Routing#Special-stash-values
10:54 dotan_convos You can easily setup a generic route like /:controller/:action/ which will automatically dispatch any match to the given action in the specified controller, or render the template in the matching path.
11:01 dhg joined #mojo
11:07 karjala joined #mojo
11:11 veryrusty joined #mojo
11:21 erry dotan_convos: cool
11:34 marcus What dotan said :)
11:47 itaipu joined #mojo
11:57 petru joined #mojo
11:57 sri dotan_convos: re with_roles patch, just make everything as minimalistic as possible
11:58 sri you have until tomorrow, then i'll have time to finish it
11:59 sri one way or another, by next week we'll support roles
12:23 dotan_convos OK, I'll rewrite the test to use simple dummy classes instead of messing with real ones.
12:23 dotan_convos I can/should put the packages in t/mojo/lib, right?
12:26 sri seems unnecessary to make new files
12:27 sri have two roles, and make two objects, one with one role applied, the other with two roles applied
12:27 sri should cover everything
12:27 sri don't overthink it, just make it simple
12:44 VVelox joined #mojo
12:54 gizmomathboy joined #mojo
12:55 bwf joined #mojo
12:59 CandyAngel That moment OOM-killer goes on a movie-worthy rampage on your server :\
13:04 gryphon joined #mojo
13:23 marty joined #mojo
13:36 orev joined #mojo
14:33 mcsnolte joined #mojo
14:33 jacoby joined #mojo
14:59 htaunay joined #mojo
15:06 * batman just submitted a talk :)
15:06 batman 45 minutes live demo. I've been told I'm crazy.
15:07 sri you're crazy
15:07 jberger the crazy we deserve
15:08 jberger marcus: how'd it go?
15:08 dotan_convos The inline packages in my test use Role::Tiny, and this crashes in compile-time if the module isn't installed. :(
15:09 dotan_convos Is it ok if I just wrap the use Test::More; skip_all => ... bit in a BEGIN block? That seems to work
15:09 PryMar56 joined #mojo
15:09 pink_mist ehh, the 'use' statement should mean that all of that is *already* in a BEGIN block
15:09 pink_mist oh wait, you didn't put the skip_all in the use statement
15:09 pink_mist never mind
15:11 Grinnz you can pass skip_all to Test::More as an import arg for this purpose
15:11 Grinnz often used like "use if ..., 'Test::More', skip_all => $reason;"
15:11 Grinnz if you do that, the whole line is evaluated at compile time
15:11 Grinnz dont actually use a variable $reason or you'llhave to set that at compile time too :P
15:15 Grinnz use if !Mojo::Base->can_roles, 'Test::More', skip_all => 'Role::Tiny 2.000001+ required for this test!';
15:15 Grinnz should work
15:15 Grinnz but you also need to use Test::More normally afterward
15:16 Grinnz dotan_convos: ^
15:17 itaipu joined #mojo
15:22 sri we already have boilerplate for tests with optional deps...
15:22 dotan_convos Grinnz: use if is core? https://metacpan.org/changes/distribution/if
15:23 Grinnz yep
15:23 sri just do what the other tests do
15:23 dotan_convos Looks like I have it in perl 5.10.1 (plenv), so I guess it's legit
15:23 pink_mist perldoc if
15:23 purl i think perldoc if is about the pragma
15:24 pink_mist botsnack
15:24 purl thanks pink_mist :)
15:24 dotan_convos sri: I don't think the other tests explicitly use an optional module, they depend on it being used by some Mojo lib
15:27 Grinnz heres one, but it depends on the require occurring within the line at compile time, which your check wouldn't do: https://github.com/kraih/mojo/blob/master/t/mojo/reactor_ev.t
15:35 sri hmm, guess it has to be inconsistent
15:41 itaipu joined #mojo
15:43 marcus jberger: It went quite well, thanks for the great slide stack!
15:44 marcus jberger: I added a few slides after I saw you spent about 40 minutes, and ended up with 3 minutes for questions :)
15:47 jberger perfect
15:47 jberger marcus++
15:49 jberger ah, the validation stuff
15:49 jberger nice
15:50 jberger and Mojo::File, good point I hadn't added that yet
15:51 stryx` joined #mojo
15:53 petru joined #mojo
16:16 karjala joined #mojo
16:16 karjala Does Mojolicious plan to implement a GraphQL handler?
16:17 pink_mist I would extremely much doubt that
16:18 karjala And why would that be, if I may ask?
16:18 karjala Looks useful and good
16:19 karjala Maybe as a plugin
16:19 pink_mist sure, as a 3rd party plugin anybody could implement it
16:20 Grinnz because it's easily something that could be put on CPAN, and not very central to web apps
16:20 karjala But we trust sri more than anyone
16:20 pink_mist but the core mojolicious has absolutely no reason to implement it
16:21 dotan_convos If I squash all the commits in my pull request and force push the branch, this doesn't mess with the discussion, because the commits are still there, right?
16:22 pink_mist uh, if you squash the commits, they aren't there, no 0_o
16:22 pink_mist as far as I'm aware at least
16:24 Grinnz the commits would be removed, github would show "comments on outdated code" or so
16:24 Grinnz the comments will still be there, but i'm not sure if the code snippets will be
16:24 pink_mist dotan_convos: any time you need to force push, you're losing commits afaik
16:26 dotan_convos I think you lose branch history, because there's a single new squashed commit rebased on the branch point that replaces the old commits, but the old commits are still floating about in git, unconnected but addressable by their hash
16:26 Grinnz yes
16:27 Grinnz but they won't be visible in the PR either
16:27 Grinnz it doesn't affect the comments on lines
16:27 pink_mist pretty sure github does garbage collection pretty aggressively
16:33 marcus I'm kinda interested in GraphQL as well, would like to see some nice perl implementation of it.
16:34 marcus second mojo-related lightning talk in a row \o/
16:35 marcus https://metacpan.org/pod/GraphQL seems rather incomplete so far.
16:36 batman I'm glad I've done the demo before. Makes me more confident that it will actually work :)
16:37 * sri does not care about GraphQL at all
16:41 dotan_convos GraphQL looks like an awful lot of boilerplate
16:42 dotan_convos Use Role::Tiny->VERSION in the version command or stick to $Role::Tiny::VERSION like the rest?
16:42 sri dotan_convos: i'll merge your pull request once the two comments are addressed
16:42 Grinnz always use ->VERSION
16:42 Grinnz numerical comparisons of versions are incorrect
16:43 marcus https://medium.freecodecamp.org/rest-apis-are-rest-in-peace-apis-long-live-graphql-d412e559d8e4 I found this vaguely interesting despite the boilerplate.
16:43 Grinnz oh, in the version command, nevermind
16:44 marcus I meant to say despite the hyperbole.
16:44 Grinnz it's not a comparison, so $Role::Tiny::VERSION as the others have is fine
16:45 dotan_convos Doh
16:48 dotan_convos sri: I'm a bit mystified by the indentation thing in version.pm, so I'm assuming perltidy will get me what you want?
16:48 dotan_convos I also aligned the paren in the output template
16:49 Grinnz i usually do `perltidy path/to/file.pm > path/to/file.pm.new; diff -u path/to/file.pm*`
16:49 Grinnz then replace the file if it's ok
17:09 marcus NetworkGlue another mojo mention from the lightnng talks
17:11 dhg joined #mojo
17:14 petru joined #mojo
17:27 tchaves joined #mojo
18:28 pink_mist https://about.gitlab.com/2017/08/10/gitlab-9-dot-4-dot-4-released/
18:33 * jberger watches the lightning talk video
18:45 Grinnz the roles section just added to the PR for the testing guide is great, gives me some ideas for testing at work
18:45 Grinnz scottw++ dotan_convos++
18:53 sri dotan_convos: the version numbers are not aligned
19:06 sri scott is doing a great job
19:06 sri the role example is not quite ideal though
19:07 good_news_everyon joined #mojo
19:07 good_news_everyon [mojo] kraih pushed 2 new commits to master: https://git.io/v7DF8
19:07 good_news_everyon mojo/master c213395 Dotan Dimet: Mojo::Base::with_roles patch...
19:07 good_news_everyon mojo/master 492e6c4 Sebastian Riedel: Merge pull request #1120 from dotandimet/with_roles...
19:07 good_news_everyon left #mojo
19:15 jberger dotan_convos++
19:24 good_news_everyon joined #mojo
19:24 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/v7DNJ
19:24 good_news_everyon mojo/master 5e46ffa Sebastian Riedel: update Changes
19:24 good_news_everyon left #mojo
19:53 Pyritic joined #mojo
19:56 PopeFelix How do I finagle $c->reply->exception to return JSON, instead of doing (e.g.) $c->render( json => { error => 1 } )
19:56 PopeFelix ?
19:57 * PopeFelix would like to use Grinnz 's Mojolicious::Plugin::Future if he could
19:57 jberger PopeFelix: http://mojolicious.org/perldoc/Mojolicious/Guides/Rendering#Rendering-exception-and-not_found-pages
19:57 Grinnz PopeFelix: last example in https://metacpan.org/pod/Mojolicious::Guides::Rendering#Rendering-exception-and-not_found-pages
19:58 PopeFelix thans
19:58 PopeFelix thanks
20:01 jberger nicomen: what did you say during your lightning advert? the echo on the video is making this hard to understand
20:21 jberger LeoNerd's lightning talk was pretty funny
20:25 arcanez jberger: where are the videos?
20:25 purl the videos are okay, I just wish I could get something to convert them from quicktime to something (anything) else. or acceptable
20:25 arcanez purl: forget the videos
20:25 purl arcanez: I forgot videos
20:29 jberger pirateFinn: did I hear your talk say that Mojo and strawberry perl don't agree?
20:31 VVelox joined #mojo
20:32 jberger arcanez: https://www.facebook.com/theperlconference
20:32 jberger these were the livestream
20:32 jberger single shot, speaker and slides
20:33 jberger room audio isn't so easy to hear always
20:34 arcanez interesting screens
20:36 arcanez is that mdk?
20:41 jberger talking about Glasgow yes
20:55 sri i guess that this works it not really obvious https://github.com/kraih/mojo/pull/1118#discussion_r132563810
21:00 jberger a really easy example is the ->d method of Test::Mojo::Role::Debug
21:05 sri ok, i think we have the final version https://github.com/kraih/mojo/pull/1118
21:05 sri please proofread
21:05 ribasushi jberger: envy me: https://twitter.com/ribasushi/status/895752254917103616
21:06 sri i'll merge tomorrow
21:06 jberger ribasushi: I DO!
21:06 jberger damn, now I have to watch it again
21:06 ribasushi ( the entire museum is amazing fwiw, 250 year old programs, all doing exactly what they are supposed to do ;)
21:07 sri how unrealistic...
21:07 jberger oooh and opening youtube I see that YAPC::NA posted Damian's talk: https://www.youtube.com/watch?v=ob6YHpcXmTg
21:07 jberger embargo released I suppose
21:07 Grinnz where were the managers to add feature requests
21:08 preaction embargo off!
21:09 jberger ribasushi: didn't he say that they made it work enough to play a scale?
21:09 arcanez ribasushi: do they run on perl 5.8?
21:10 ribasushi jberger: they did initially, and you can see the scale "program" in place on the wheel, but the marble loader keeps jamming, and they gave up
21:11 jberger ah
21:11 Grinnz did they at least open a bug report
21:11 jberger Grinnz: they are in the middle of a total rewrite :-P
21:11 arcanez jberger and ribasushi having a serious conversation and Grinnz and I trolling.
21:11 ribasushi jberger: apparently during the original filming the machine never could work for more than 15 seconds, and they stitched together ~40 takes to make the final video ( though the sound you hear at any time belongs to the take you are watching )
21:12 jberger I knew it was a lot of takes
21:12 ribasushi jberger: there was supposed to be a video of all the failures posted, but they never got to it ( yet at least )
21:18 arcanez https://www.youtube.com/watch?v=IvUU8joBb1Q looks like he ends up with a few marbles on the floor
21:29 arcanez sri: suse let you work on mojo or is this all done in your "spare" time?
21:31 sri arcanez: during work time i only work on stuff that's work relevant, like the keep-alive problem last week
21:32 sri contributing back upstream to open source projects is heavily encouraged though
21:33 jabberwok left #mojo
21:38 arcanez that's cool
21:46 sri definitely is, this year mojolicious and minion got dozens of features that way
21:55 Lee joined #mojo
22:12 PopeF joined #mojo
23:02 gordonfish- joined #mojo
23:08 gordonfish- Hi, I'm trying to figure out why hypnotoad seems to take up so much cpu for the worker processes? https://f.perlbot.pl/p/gp4al6
23:09 gordonfish- Each one appears to be taking around 25%
23:14 preaction probably because they just started up. %cpu from ps seems to be "CPU time used" divided by "real time running"
23:15 gordonfish I'm seeing the same thing in just `top` too

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