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

IRC log for #mojo, 2016-04-18

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

All times shown according to UTC.

Time Nick Message
00:57 jberger Interesting
02:29 mcsnolte joined #mojo
02:37 sri still unsure which stats i want in mainline minion
02:37 noganex_ joined #mojo
02:38 sri size and uptime cost pretty much nothing in postgres, but no clue about other backends
02:38 sri btw. this is nice too pg_size_pretty(pg_total_relation_size('minion_jobs'))
03:59 sri yea, guess i'm removing size and uptime from minion again
04:00 sri one more patch for the fork
04:00 sri ;p
04:24 inokenty-w joined #mojo
06:21 melo joined #mojo
06:33 McA joined #mojo
06:39 McA joined #mojo
06:46 dod joined #mojo
06:51 dod joined #mojo
06:58 odc joined #mojo
07:36 trone joined #mojo
07:39 AndrewIsh joined #mojo
07:42 trone joined #mojo
07:50 punter joined #mojo
08:00 jasanj left #mojo
08:13 Vandal joined #mojo
08:23 vytas joined #mojo
08:35 mtths left #mojo
08:37 osfabibisi joined #mojo
09:05 dod joined #mojo
09:21 meshl joined #mojo
10:18 melo joined #mojo
10:57 Kripton joined #mojo
10:59 dvinciguerra_ joined #mojo
11:00 dod joined #mojo
11:06 pierrick joined #mojo
11:25 tchaves joined #mojo
11:41 marty joined #mojo
11:53 meshl joined #mojo
12:11 kaare_ joined #mojo
12:30 cpan_mojo Mojolicious-Plugin-RoutesAuthDBI-0.410 by MCHE https://metacpan.org/release/MCHE/Mojolicious-Plugin-RoutesAuthDBI-0.410
12:40 perlpilot joined #mojo
12:49 gizmomathboy joined #mojo
12:53 cpan_mojo Mojo-IOLoop-ReadWriteFork-0.20 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Mojo-IOLoop-ReadWriteFork-0.20
12:57 ramortegui joined #mojo
12:58 cpan_mojo Mojolicious-Plugin-RoutesAuthDBI-0.411 by MCHE https://metacpan.org/release/MCHE/Mojolicious-Plugin-RoutesAuthDBI-0.411
13:06 dod joined #mojo
13:48 jacoby joined #mojo
13:51 mcsnolte joined #mojo
13:52 cpan_mojo Mojolicious-Plugin-RoutesAuthDBI-0.412 by MCHE https://metacpan.org/release/MCHE/Mojolicious-Plugin-RoutesAuthDBI-0.412
14:03 pierrick joined #mojo
14:17 dantti_laptop joined #mojo
14:23 dod joined #mojo
14:52 UnLess joined #mojo
14:52 UnLess Hey
14:55 jberger UnLess: o/
15:18 sebbe joined #mojo
15:19 marty joined #mojo
15:24 melo joined #mojo
15:31 kes joined #mojo
15:39 n16gel joined #mojo
15:45 pierrick joined #mojo
15:50 PryMar56 joined #mojo
16:01 stephan48 joined #mojo
16:04 jnbek joined #mojo
16:07 mtths joined #mojo
16:53 n16gel joined #mojo
17:15 n16gel left #mojo
17:16 n16gel joined #mojo
17:18 meshl joined #mojo
17:31 bwf joined #mojo
17:32 trone joined #mojo
17:40 disputin joined #mojo
17:42 disputin joined #mojo
18:23 dod joined #mojo
19:29 disputin joined #mojo
19:48 disputin joined #mojo
20:16 melo joined #mojo
20:19 disputin joined #mojo
20:53 sri slow week
20:54 bpmedley I'm working on apps.. :)
20:55 PopeFelix joined #mojo
20:58 pink_mist the week is just getting started 0_o
20:59 asarch joined #mojo
21:08 sri very slowly
21:15 jberger nothing slow here :D
21:15 jberger got a week's worth of work to do today and tomorrow morning before I head over to QAH
21:15 PopeFelix So I'm trying to clean up the non-blocking tests in Mojo::UserAgent::Mockable. I notice in several places where I make a number of transactions in a loop, I'm calling Mojo::IOLoop->stop inside the callback, and Mojo::IOLoop->start outside the callback but inside the loop.
21:16 PopeFelix And that seems odd to me, but if I take ->stop out of the callback, and just call ->start once, it doesn't work.
21:17 jberger PopeFelix: if you start it you have to stop it to proceed again
21:17 jberger the question is should you be using Delay (answer: probably yes)
21:17 PopeFelix jberger, I think I see.
21:18 PopeFelix Is ->start more of a "tick" than starting the loop? I didn't get that impression from the docs.
21:19 PopeFelix well, there it is: "[start] will block until stop is called".
21:21 PopeFelix jberger, looking at the docs for delay(), if I wanted to run (e.g.) 10 sequential transactions, I'd call "Mojo::IOLoop->delay( sub { $ua->get(...), $ua->get(...) .... )->wait;" ?
21:31 lluad_ joined #mojo
21:32 disputin joined #mojo
21:36 lluad_ joined #mojo
21:41 meshl joined #mojo
22:04 PopeFelix OK, I'm playing with delay a bit, but it's only executing my first query.
22:06 PopeFelix http://paste.scsys.co.uk/510871
22:10 jberger PopeFelix: no that's 10 concurrent
22:11 PopeFelix jberger, Oh, I see. How do I do sequential?
22:13 jberger steps
22:13 jberger sub{ $ua->get(...)}, sub { $ua->get(...)}, ...
22:13 PopeFelix Which is what I'm trying to do.
22:13 PopeFelix https://metacpan.org/pod/Mojo::IOLoop#delay "Callbacks will be passed along to '
22:14 PopeFelix https://metacpan.org/pod/Mojo::IOLoop#delay "Callbacks will be passed along to 'steps' in Mojo::IOLoop::Delay"
22:14 PopeFelix sorry, fat fingered.
22:14 PopeFelix So I've built an array of callbacks. I'm sending that array of callbacks to Mojo::IOLoop->delay(), which I thought would send it to steps().
22:16 jberger it does
22:16 jberger but when you showed the one line (not the paste) above
22:16 PopeFelix Oh, I see.
22:16 gizmomathboy joined #mojo
22:17 jberger now reading through your paste
22:17 PopeFelix Thanks
22:17 jberger you don't call begin
22:17 jberger and you shouldn't pass a callback to the get
22:17 PopeFelix OK.
22:17 jberger you should use the result of $delay->begin
22:18 jberger this is gonna take some reading/thinking on your part
22:18 jberger (that sounded harsher than I meant)
22:18 PopeFelix It's ok.
22:18 sri so, has anyone managed to find out what's so broken about Mojolicious::Types? https://metacpan.org/pod/MojoX::MIME::Types
22:18 jberger this is one of those topics (delay) that takes grokking but earns you a level up once you do
22:18 PopeFelix But it sounds like this is moving away from elegantly testing MUA::Mockable in a non blocking fashion.
22:19 sri or is markov just a jerk?
22:20 PopeFelix sri, I've always found him helpful and polite in my interactions with him.
22:20 sri he criticized Mojolicious::Types in the docs of that module, but never brought it up with us
22:21 PopeFelix sri, I can't speak to that.
22:21 sri don't think i've ever talked to him, so that's all i know
22:22 PopeFelix He's been helpful to me in using XML::Compile::WSDL11.
22:22 jberger sri: no matter how much he argues that ours is broken, I can tell without installing that his is
22:22 jberger the dist metadata (and the makefile) don't specify MIME::Types as a requirement
22:22 sri seems so odd that he would care enough to bundle the module but not talk to us
22:23 Grinnz_ jberger: that is in the MIME::Types distribution
22:23 jberger ah
22:23 jberger hehe
22:23 jberger wow, he bundles that into MIME::Types
22:23 jberger seems like overkill
22:23 sri yea, very puzzling
22:24 jberger sri: I can see the idea of pulling types from a larger types database
22:24 jberger but I don't understand "more correct handling"
22:25 sri sure, i'm all for having a more complete list of mime types on cpan
22:25 sri but i would hope for correctness patches to be pushed upstream
22:26 jberger reading through the "why" section could the issue be text casing?
22:27 sri https://github.com/kraih/mojo/blob/master/t/mojolicious/types.t#L42
22:28 jberger then I'm out of guesses
22:32 sri i mean, it's a very pragmatic little implementation that was only meant to be fast
22:32 sri almost certainly there are bugs
22:33 jberger he is on the network
22:33 jberger should we invite him to #mojo?
22:34 sri sure
22:34 jberger I'm just about to commute home
22:34 jberger either go ahead or else I'll be back on in 30 mins
22:34 sri and i'm not polite enough :)
22:34 jberger you can do it if you'll be polite :-D
22:35 jberger perhaps an optional dependency might make sense here
22:35 genio markov has always been polite and helpful with me.  he tends to not pay so much attention to IRC.  email is better
22:35 sri jberger: on MIME::Types?
22:36 sri doesn't seem like such a good idea, considering the history so far
22:37 sri i'd like to keep new optional deps limited to authors we trust
22:37 sri (mlehmann problem :S)
22:38 genio In my experience, one is helpful, the other...
22:41 sri talking about trusted authors... Oleg still hasn't pushed IO::Socket::Socks and Net::DNS::Native to 1.0 ;p
22:45 PryMar56 joined #mojo
23:00 pierrick hi, i have an issue with my mojolicious app ( hypnotoad behind nginx ), i've got 502 errors like 1 every 5mn ( mainly with googlebot or bingbot )  with no error log in mojolicious : anyone already have this kind of issue ? sorry if it's difficult to understand what i mean i'm french !
23:03 Grinnz_ 502 would be an error from nginx if it cannot proxy to hypnotoad
23:03 Grinnz_ maybe you have requests blocking too long for the number of workers you have?
23:04 pierrick i try to tweak workers and clients in hypnotoad conf like describe in documentation  with no success
23:06 pierrick i see with blocking request it s better to have more workers and less clients but dont know what could be the good number
23:07 pierrick actually i have 16 workers and 40 clients maybe i need to decrease clients
23:32 * sri thinks he will try to use more experimental notes in minion https://github.com/kraih/minion/commit/e8c43597b08893ee99624ffc56ba0aa52dff2e21
23:58 jberger sorry I got a little delayed
23:58 jberger are we going to prefer the email route for markov?
23:59 sri *rimshot*
23:59 jberger :p

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