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

IRC log for #mojo, 2014-04-01

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

All times shown according to UTC.

Time Nick Message
00:04 sri i'm on the edge about it... there's a lot of value for teaching... but i'm not so sure about real use cases
00:07 sri damn... april 1st... time to abandon the internet
00:17 d4rkie joined #mojo
00:31 jack joined #mojo
00:39 laouji joined #mojo
00:41 sri wow, mariadb now supports sharding natively... maybe worth taking another look
01:10 dvinciguerra joined #mojo
01:44 klapperl_ joined #mojo
01:48 sri ok, the more i play with it the less i like result in minion
01:49 sri it would get used wrong... waiting for results just doesn't work since you can't guarantee responsiveness of workers
01:50 sri while workers could listen on a pub/sub channel as well for new jobs being queued... if you have 1000 idle workers waiting, they would all crush the database trying to be the first to claim the new job
01:51 sri i actually have a working variant of that
01:51 sri http://pastie.org/8984271 # for completeness sake
01:52 sri pub/sub in both directions is kinda fun small scale
01:53 sri good news for jberger, minion won't ever be a replacement for forkcall ;)
01:53 sri was a fun experiment though
02:13 shmuel joined #mojo
02:20 btyler joined #mojo
02:44 laouji joined #mojo
02:52 s joined #mojo
02:55 * jberger high fives self
02:55 jberger also I have posted in several place about the X-Powered-By header
02:55 jberger The lack thereof
03:02 s ooh
03:02 s that's a nice metaphor
03:27 jhg joined #mojo
03:32 robinsmidsrod joined #mojo
03:46 laouji joined #mojo
05:47 alienresidents joined #mojo
06:00 dpetrov_ joined #mojo
06:07 ver joined #mojo
06:08 denis_boyun joined #mojo
06:14 Eke- joined #mojo
06:23 sujithm joined #mojo
06:29 mire joined #mojo
06:35 Vandal joined #mojo
06:36 punter joined #mojo
06:38 s I need to change the rhythm of a bunch of request/responses
06:39 s like one at a time over a few seconds, and maybe not more than a few seconds worth
06:39 dod joined #mojo
06:52 dod joined #mojo
06:54 sujithm joined #mojo
07:03 shmuel why doesn't '/app/plugin/semuel/mt-plugin-entryfiledrop/save' route match? :-(
07:03 basiliscos joined #mojo
07:04 shmuel I have / and +/app and +/plugin/:repos_user/:repos_name and another +/ and +/save.
07:04 shmuel what I'm missing?
07:05 sujithm joined #mojo
07:15 eseyman joined #mojo
07:16 Bravo13 joined #mojo
07:22 trone joined #mojo
07:38 moritz shmuel: show some code?
07:39 shmuel wait, I'll upload some to nopaste.
07:39 shmuel but this is the first time I try placehoders.
07:39 shmuel maybe I'm doing something wrong.
07:41 moritz also note that you can get 404 errors from missing templates, not just from failures to match routes
07:41 sujithm joined #mojo
07:41 shmuel http://paste.scsys.co.uk/338011
07:42 Dandre left #mojo
07:42 shadowpaste joined #mojo
07:43 shmuel moritz: what do you think?
07:43 purl I think shmuel should try flossing more often!
07:43 shmuel flossing?
07:44 moritz shmuel: ignore the stupid bot
07:44 moritz shmuel: dumb question, is it actually a POST request you're doing?
07:45 shmuel yes, look at the poorly format error: "None of these routes matched your POST request"
07:45 moritz shmuel: and do the other routes work?
07:45 shmuel yes, everything else works.
07:45 shmuel I just added the /save thing.
07:46 moritz nothing obviously wrong to me
07:47 shmuel I'll try and re-post the error screen, better formatted.
07:50 Lee joined #mojo
07:51 dexteruk joined #mojo
07:52 shadowpaste Someone at 217.168.150.38 pasted "formatted error screen" (19 lines) at http://paste.scsys.co.uk/338014
07:52 shmuel is there a way to debug the routing process?
07:54 shmuel to make it say why each route was skipped? the order of observation?
07:54 stephan48 does bridge_load return true?
07:55 fhelmber_ joined #mojo
07:55 shmuel bridge_load is not called.
07:55 shmuel I put a print STDERR in the beginning to find that out
07:56 stephan48 maybe trs "$r->bridge(...)->..." instead of "$app->bridge(...)->..."?
07:56 stephan48 try*
07:56 stephan48 :)
07:56 stephan48 oh i see $app is declared further above... what does maybe_user and must_user say?
07:57 edestler joined #mojo
07:57 shmuel find out if there is a user logged in, (maybe_user)
07:57 shmuel and block if the user is not logged in. (must_user)
07:58 stephan48 does they both return true?
07:59 shmuel I looked, and they both are called.
08:00 shmuel and must_user return 1
08:01 shmuel the problem is in $plugin bridge definition. somehow.
08:01 shmuel Ahh! found it.
08:02 shmuel Plugin.pm is a file that I just created,
08:02 shmuel and I copy-pasted the package+controller base def from other file
08:03 stephan48 and didn't change the package line?
08:03 Dandre joined #mojo
08:03 shmuel yes. so mojo didn't find the method.
08:04 shmuel uffff.
08:08 s how can I get stacktrace error messages?
08:20 moritz doesn't the default dev exception template include a stacktrace by default?
08:20 moritz also, Carp::Always
08:32 laouji joined #mojo
08:35 laouji joined #mojo
08:39 denisboyun joined #mojo
08:40 maxhq joined #mojo
08:41 shmuel moritz: no, it display the location of the error, not stacktrace.
08:54 moritz then Carp::Always helps
09:03 arthas joined #mojo
09:03 sujithm joined #mojo
09:19 sujithm joined #mojo
09:24 laouji joined #mojo
09:31 laouji joined #mojo
10:04 rem_lex| joined #mojo
10:09 sujithm joined #mojo
10:17 s is there an easy way to fork off a chunk of my mojo application and communicate with it? it may not be that responsive
10:18 s it being a Devel::ebug process that might be busy running for a while and then respond
10:18 laouji_ joined #mojo
10:22 laouji__ joined #mojo
10:35 sri of course it includes a stacktrace
10:36 tholen joined #mojo
10:38 denisboyun joined #mojo
10:45 s hmm
10:45 gregf joined #mojo
11:07 * sri is glad nobody cares about the result feature in minion
11:13 mire joined #mojo
11:13 * sjn reads https://www.dnorth.net/2012/03/17/the-port-0-trick/ and wondering how Mojolicious does in this regard
11:15 sri https://github.com/kraih/mojo/blob/master/lib/Mojo/IOLoop/Server.pm#L42
11:18 stephan48 result feature?
11:18 sri https://github.com/kraih/minion#minion-
11:19 sri opinions don't matter anymore though, it's going away again
11:21 stephan48 what was its purpose anyway? to feed back informations to the app which generated the job?
11:28 * nic cares about the result feature but can't look at it till tomorrow
11:30 nic I think my needs are like async exception handling
11:31 nic if you see job TypeA failed cos platform A has crashed, don't spawn more (instances of) jobs of that type
11:32 nic but if a job failed due to a timeout or deadlock, requeue with a brief random delay
11:46 sujithm joined #mojo
11:49 sri at least ->auto_perform(1) will make testing easier
11:49 sri $t->app->minion->auto_perform(1) and all jobs run automatically in your tests
11:53 jack joined #mojo
11:55 stephan48 ah
11:56 stephan48 does this just run all job "directly" in the app without passing it to workers?
12:07 tholen joined #mojo
12:09 sri it runs a worker in the current process and forks every time a job is actually performed
12:09 sri and back to the old example without results https://github.com/kraih/minion#minion-
12:10 sri all you have to do for tests https://github.com/kraih/minion/blob/master/t/minion_lite_app.t#L40
12:12 sujithm_ joined #mojo
12:15 fhelmber_ joined #mojo
12:23 Negor|3 joined #mojo
12:23 Negor|4 joined #mojo
12:23 Negor|41 joined #mojo
12:30 Negor joined #mojo
12:30 Bravo13 joined #mojo
12:31 Negor|2 joined #mojo
12:31 toebu joined #mojo
12:50 neilhwatson joined #mojo
12:54 dp_ joined #mojo
13:03 DuMu joined #mojo
13:10 rj11 [A
13:11 DuMu Does Test::Mojo work with Mojolicious::Lite apps?
13:14 DuMu apparently it does. I have found it in the docs. sorry.
13:24 jrbeilke joined #mojo
13:33 fhelmber_ joined #mojo
13:37 sugar joined #mojo
13:45 d4rkie joined #mojo
13:46 btyler joined #mojo
13:53 mattastrophe joined #mojo
13:54 disputin joined #mojo
13:58 gryphon joined #mojo
14:00 d4rkie joined #mojo
14:03 bluescreen joined #mojo
14:08 rihegher joined #mojo
14:09 sh4 joined #mojo
14:10 s http://pastebin.com/EM8nmwhF
14:10 s websocket client, server
14:11 s this needs a graphical interface
14:16 s anyway yeah I've got a Mojo::Transaction::HTTP when I should have a Mojo::Transaction::WebSocket?
14:20 s couldn't get that other thing working either
14:37 bluescreen_ joined #mojo
14:41 sri http://mojolicio.us/perldoc/Mojolicious/Guides/FAQ#What_does_Premature_connection_close_mean
14:51 sugar joined #mojo
14:57 aggrolite joined #mojo
15:00 rihegher 2014-04-01T19:00:00+0900
15:00 rihegher left #mojo
15:01 mire joined #mojo
15:11 nicomen that timestamp is actually wrong
15:12 nicomen or undefined
15:12 purl undefined is false, but so are some defined values
15:12 nicomen +09:00
15:23 Dandre joined #mojo
15:29 sujithm joined #mojo
15:40 cpan_mojo Mojolicious-Plugin-CHI 0.07 by Nils Diewald - http://metacpan.org/release/AKRON/Mojolicious-Plugin-CHI-0.07
16:01 Dandre joined #mojo
16:01 beyondcreed joined #mojo
16:10 denis_boyun joined #mojo
16:16 mattastrophe joined #mojo
16:17 sri and job restarting added to minion
16:19 marty joined #mojo
16:40 gryphon joined #mojo
16:46 rihegher joined #mojo
16:47 rihegher oops sorry about the timestamp
16:48 rihegher I mean the Isotime
16:49 rihegher Does anybody know who is the trainer for the day of training during the mojoconf?
16:53 dotan rihegher: I suspect it's tempire
16:54 dotan (suspect = it was mentioned on the channel that he's busy prepping for the training)
16:54 dotan tempire = mojocasts.com dude.
16:56 sri it is tempire
16:57 dod joined #mojo
16:57 * sri can't think of anyone more qualified to do the training
16:57 DuMu why doesn't Test::Mojo have a 'json_isnt' method?
16:58 sri why doesn't Test::More have a "isnt_deeply" method?
16:59 DuMu right... pointers don't represent unique keys?
16:59 DuMu sorry, I'm new to JSON
17:00 DuMu I'm trying to test that a JSON pointer doesn't return a certain value
17:02 sri let me rephrase that, patches welcome
17:05 sri on a related note... we need more volunteers, tickets are adding up https://github.com/kraih/mojo/issues?page=1&state=open
17:05 DuMu I would be delighted to contribute... but I can't fit it in this month.
17:05 DuMu Well I would be honoured to participate.
17:06 DuMu I am new to Mojo too ... at some point I wd like to be in a position to help : )
17:08 dotan sri: tempire++
17:10 sri seriously, if i was planning to learn mojolicious, that training would be my first choice
17:12 DuMu I can only afford to read the docs
17:12 DuMu left #mojo
17:12 sri hmmmmm... minion api seems mostly complete now... just need something good for job listing
17:13 sri my $first_ten_jobs = $minion->list('finished', 10, 0) doesn't look that great
17:14 sri (10, 0 is limit and offset for paging)
17:17 sri perhaps different methods for states... $minion->finished(10, 0)
17:17 * sri shrugs
17:27 alnewkirk joined #mojo
17:29 genio I really can't wait for Perl 5.20 to be released.  sub signatures make me happy
17:29 sri me too
17:29 sri 50 days to go
17:30 rihegher left #mojo
17:31 Dandre joined #mojo
17:33 Eke- joined #mojo
17:43 disputin joined #mojo
17:53 sujithm joined #mojo
18:02 meshl joined #mojo
18:19 bjoernfan If you guys who know a bit more about mojo internals comment on issues it might be easier for others like myself to submit a patch. Just something simple like if you agree with the submitter or not. Like the one about _ for localization, if you just say "Yeah wasn't supposed to happen" so I don't start working on a patch for something you intentionally removed (for whatever reason). Just a thought. :)
18:22 bjoernfan And I'm guessing charliebrady deserves a thank you for the test case maybe? Just to keep people wanting to contribute. I'd say it but it doesn't really feel like my place to do so.
18:22 bjoernfan Thoughts?
18:22 purl Moonlight shines through the dark night / clouds move overhead, casting shadows / dancing in the firelight
18:22 bjoernfan :)
18:27 sri of course you can thank him
18:34 sri but you could also criticize that his test case is not straight to the point, the extra level of indirection seems unnecesary
18:37 sri someone could also say that the suggestion in the oetiker issue breaks features, or that we have a gazillion environment variables that affect tests in the HeikoSchlittermann issue
18:37 sri (just some pointers)
18:40 sri i'm really staying out of issues now, either the community gets more involved or open issues become a more common sight... which might not be a bad thing either
18:47 dwierenga joined #mojo
18:48 bjoernfan Those are the kind of things I'd like to see as comments to the issues, but I totally understand if you want to stay away from issues for a while, and if you put those comments there you're going to get more involved than you'd like... so... yeah...
18:50 bjoernfan But sure, I can do something at least. Think I found the commit that broke the localization thing.
18:51 sri if i feel strongly about an issue of there's a pull request to vote on i will comment... so it's safe to assume i'm indifferent on anything without comments
18:53 bjoernfan So I can trust you'll rip my pull requests apart if I submit anything stupid? :)
18:54 sri sure :)
18:54 bjoernfan yay
18:56 sri the _ thing for example depends a lot on opinions... you need to have a discussion about it... possibly hunt down core devs to vote... that's the stuff i'd really like volunteers for
18:59 sri guess i'll write to the list about it as well
19:02 bjoernfan Is it safe to assume all core devs read all the comments on issues?
19:05 sri i can only speak for myself, but i read most
19:25 tianon joined #mojo
19:39 maxhq joined #mojo
19:58 sri this is one of the few jokes today i appreciate :) http://drafts.csswg.org/css-scoping/#shadow-dom
20:12 asarch joined #mojo
20:21 genio dear gods, that site hurts my eyes
20:21 genio the '90s were a terrible, terrible time for web development
20:22 sri you mean fabulous time
20:25 genio fabulously terrible?
20:36 cfedde do I need to do anythign special to make a mojolite app work under fcgi?
20:41 stephan48 i got this test chain http://paste.stejau.de/?486&download but text_is reports the attrib behind the selector as empty. i checked the transactions and it's body contains the ul with the mentioned id... any idea what the problem could be?
20:45 punter joined #mojo
21:03 rem_lex joined #mojo
21:08 rwstauner joined #mojo
21:17 rwstauner are the default values for optional placeholders required to be true values?
21:20 rwstauner that wasn't the case in the past, but appears to be now.  i'm trying to write up a test to confirm with git bisect
21:27 espent joined #mojo
21:34 batman cfedde: not really, but why do you want to use fcgi?
21:43 bluescreen_ joined #mojo
21:59 rwstauner awesome.  i can't reproduce it in a clean app, so there's something in my existing app that is confusing things. sigh.
22:27 mire joined #mojo
22:31 jberger_ joined #mojo
22:32 jberger_ As to the env var issue, yes there are many that would affect the test, but is there any reason not to fix this one?
22:35 sri "fix" implies that it's a bug
22:35 sri if one env var is a bug, all are
22:35 sri if all are bugs, then there needs to a consistent fix
22:36 jberger_ The test is meant to test the default for the attribute which read the variable
22:37 jberger_ Is it not correct then to put a local $ENV{...} before the first test in the block
22:37 jberger_ And the block above it for that matter?
22:37 sri i'm certain there's a dozen more like it
22:38 jberger_ If you would like, when I get home, I will attempt to find them
22:38 sri all i want is consistency
22:39 jberger_ Ok then, I will look at it when I get off the train
22:39 sri jberger++
22:40 jberger_ That just leaves the build_controller thingy
22:41 jberger_ Not sure what to tell him
22:48 * jberger_ starts the conversation
22:48 * sri doesn't like oetikers passive aggressive response and comments too
22:50 al joined #mojo
22:52 jberger_ I suspected hypnotoad was the feature that would be broken, but then I'm not really the expert there
22:52 cfedde batman: my need is to fit some mojo apps into an existing apache environment that does lots of ssl/ldap stuff.  the easy rooute seemed to be fcgi or mod_perl.
22:53 jberger_ I would need to see how it would break things
22:53 jberger_ At which point I realized that that was my out, close as a no code provided feature request :-P
22:54 jberger_ I couldn't see what would break because I had no code to run
22:54 sri morbo too, the manager claims the listen socket and holds it open, the worker starts a whole new daemon and reclaims the inherited file descriptor https://github.com/kraih/mojo/blob/master/lib/Mojo/Server/Morbo.pm#L37
22:55 sri it's one of those insanely clever little tricks you're extremely proud of but that nobody ever notices :)
22:55 sri *mad scientist laugh*
22:56 sri of course the real magic happens here https://github.com/kraih/mojo/blob/master/lib/Mojo/IOLoop/Server.pm#L52
23:12 * jberger hands sri a brush for that mad scientist hair
23:14 d4rkie joined #mojo
23:15 jberger ENOGITHUB!
23:18 punter joined #mojo
23:33 d4rkie joined #mojo
23:34 aggrolite joined #mojo
23:47 good_news_everyone joined #mojo
23:47 good_news_everyone [mojo] jberger created test_env (+1 new commit): http://git.io/ihNAgw
23:47 good_news_everyone mojo/test_env b6b59c0 Joel Berger: protect several tests from the local environment
23:47 good_news_everyone left #mojo
23:48 jberger these seemed like the most obvious ones
23:48 jberger cases where clearly the environment was being tested
23:53 jberger these aren't sufficient, more coming :-)
23:56 sri i guess this serves the purpose for now https://github.com/kraih/minion/commit/b7cdc17d5021e2e336505f4b16d930abf3273326
23:56 sri not sure where to go with the job management command yet

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