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

IRC log for #mojo, 2016-04-10

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

All times shown according to UTC.

Time Nick Message
00:09 woz joined #mojo
01:25 asarch joined #mojo
02:16 woz joined #mojo
02:16 noganex joined #mojo
02:36 asarch joined #mojo
03:17 woz joined #mojo
04:19 woz joined #mojo
04:24 kaare joined #mojo
05:51 elrod_ joined #mojo
06:00 woz joined #mojo
06:37 woz joined #mojo
06:54 woz joined #mojo
07:28 Vandal joined #mojo
08:09 woz joined #mojo
08:51 mdom jberger: Have you thought about releasing your Mojo::URLQueue?
08:52 mdom jberger: And i would be interested why you directly access the $self hash for concurrent and delay and don't add accessors for those two?
08:53 dod joined #mojo
09:09 woz joined #mojo
09:14 dotan joined #mojo
09:17 woz joined #mojo
09:30 woz joined #mojo
09:49 anon3252523 joined #mojo
10:03 woz joined #mojo
11:15 woz joined #mojo
11:38 woz joined #mojo
12:04 meshl joined #mojo
12:18 cpan_mojo Paws-0.24 by JLMARTIN https://metacpan.org/release/JLMARTIN/Paws-0.24
12:50 denny joined #mojo
13:00 jberger mdom: I've considered it but mostly decided against it
13:00 jberger It is a decent demonstration of the technical parts of a spider but none of the other necessary but
13:01 jberger s/but/bits/
13:01 jberger Like rate limiting, robots.txt, etc
13:02 jberger And as for accessing the private data, i do it because its private data that i don't want there to be an accessor for
13:10 asarch joined #mojo
13:21 tchaves joined #mojo
13:26 denny joined #mojo
13:42 woz joined #mojo
14:38 tchaves joined #mojo
14:42 meshl joined #mojo
15:25 PryMar56 joined #mojo
15:29 Adura joined #mojo
15:34 woz joined #mojo
15:50 elrod_ joined #mojo
15:58 sri this needs feedback https://github.com/kraih/minion/pull/31
15:59 tchaves joined #mojo
16:00 disputin joined #mojo
16:01 jberger sri: do you have any initial comment?
16:07 disputin joined #mojo
16:07 sri no
16:09 jberger I could expand on my use-case
16:09 disputin joined #mojo
16:09 jberger we are building a tool which we use for administering servers in a data center
16:10 jberger most of the manipulations of the servers are done via the minion tasks/jobs
16:11 jberger currently I'm using the ajax call which enqueue's the job to start front-end display of the job being in progress and using Minion::Notifier I wait for it to complete and respond to the ajax call to display the job's completion
16:12 jberger recently though I have also added a websocket stream which can update other information like when a new server is added it gets push notified to all clients
16:12 jberger it now makes me want to do the same for the minion jobs
16:13 sri it's the first event that would be emitted from the "client" right?
16:13 jberger yes
16:13 sri think that what makes me cautious
16:13 jberger I struggled a little on the documentation for that
16:13 jberger I settled on "Emitted in the application process after a job has been enqueued."
16:14 sri yea, the docs are bad
16:14 sri should ay something along the lines of emitted in the process that enqueued the job
16:14 jberger I would use the event in conjunction with Minion::Notifier (which I would add handlers to) to notify other processes about the enqueuing
16:15 sri application process is not a thing
16:15 jberger I'm perfectly happy to change the documentation
16:15 sri guess it's also questionable what information should be passed to the event
16:16 jberger I think I also want to move the test to a few lines lower (after confirming there is a job enqueued)
16:16 jberger sri: meaning a job object vs an id?
16:16 sri i don't know
16:16 jberger I went with id so that there wasn't a performance impact on users' that didn't use that event
16:17 jberger subscribers can lookup the job by id (which is the same thing we would have to do)
16:17 jberger pay for what you need
16:17 jberger I guess it could build the object if there are subscribers, like the json decode in websocket events
16:18 sri i guess it depends on the use cases
16:19 sri so far there is only notifications?
16:21 zivester joined #mojo
16:21 jberger "Emitted in the enqueuing process after a job has been enqueued." ?
16:22 sri https://github.com/kraih/minion/blob/master/lib/Minion/Job.pm#L83
16:22 sri some examples
16:23 jberger ah good point
16:24 sri Emitted after a job has been enqueued, in the process that enqueued it.
16:25 jberger oh I like that
16:26 sri the test would fit in here, no? https://github.com/kraih/minion/blob/master/t/pg.t#L358
16:27 jberger ah, yeah
16:27 jberger seemed appropriate under the general enqueue heading, but I didn't look further to see that
16:28 sri think i liked keeping the events section together because it's irrelevant for backend authors
16:31 jberger sure
16:31 jberger with the worker event you can use once to clean up
16:32 jberger in the structure of this test I could either check that I get two enqueue events and have to unsubscribe at the end, or else I could once on it and only check that it got called once
16:32 sri once should be good enough
16:32 sri basically the test you have already
16:34 jberger oh you want me to test the id and not just like $enqueue++ ?
16:34 jberger I was trying to mimic the style around the other tests
16:35 sri they all actually test the argument i hope
16:35 jberger I don't think it is doing that
16:36 jberger nor is it testing the job ids
16:36 jberger oh I guess it does
16:36 jberger there are other on()s being called later
16:37 sri yea, just merge your test and it will be fine
16:37 jberger k
16:42 dvinciguerra_ joined #mojo
16:46 csson joined #mojo
16:47 jberger updated
16:51 sri merged
16:51 jberger woo hoo
16:52 sri you didn't perltidy
16:52 * jberger flails
16:53 jberger hmmmm, does minion not bundle the .perltidy from mojo?
16:53 jberger how do I use .perltidyrc in another directory?
16:54 * sri wonders if you intentionally leave in little flaws so he can feel useful :)
16:54 jberger sure, let's say it is that
16:55 jberger perltidy's interface is so hard to understand I just use the .vimrc command I added ages ago
16:55 jberger I don't know how else to use it
16:55 preaction robert heinlein has something on that topic ;) http://www.goodreads.com/quotes/801627-you-have-to-give-an-editor-something-to-change-or
16:55 jberger hahahahah
16:57 jberger argh, seriously
16:57 jberger perltidy why you so difficult?
17:01 woz joined #mojo
17:03 jberger perltidy -b -nst -pro=../mojo/.perltidyrc lib/Minion.pm
17:03 jberger geesh
17:03 cpan_mojo Minion-5.03 by SRI https://metacpan.org/release/SRI/Minion-5.03
17:06 jberger anyway, thanks
17:06 dvinciguerra_ joined #mojo
17:06 jberger time to start hacking on Minion::Notifier
17:09 good_news_everyon joined #mojo
17:09 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vV9F3
17:09 good_news_everyon mojo/master 544595a Sebastian Riedel: remove deprecated build_frame and parse_frame methods from Mojo::Transaction::WebSocket
17:09 good_news_everyon left #mojo
17:10 good_news_everyon joined #mojo
17:10 good_news_everyon [mojo] kraih tagged v6.58 at 78bf551: https://git.io/vV9FZ
17:10 good_news_everyon left #mojo
17:12 good_news_everyon joined #mojo
17:12 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vV9F4
17:12 good_news_everyon mojo/master 147bf30 Sebastian Riedel: bump version
17:12 good_news_everyon left #mojo
17:28 sri jberger++
17:48 woz joined #mojo
17:54 mdom jberger: Would you be fine with using the code in a project on cpan under a different name?
17:54 jberger I'd be ok with the project on CPAN with the modifications I mentioned, under any name
17:55 jberger without them I think it encourages bad spidering
17:55 jberger and I only haven't deleted it from gist because I think it can be useful for internal use by site authors
17:56 mdom jberger: I'm not actually using it for spidering. My users subscribe to a list of urls they want to check every time they call my programm
17:56 jberger and what kinds of urls are those?
17:57 jberger are they on your own site? do they conform to robots.txt? do they conform to terms of use?
17:58 mdom It's for my twtxt client txtnix ... it's more like a browser than anything else
17:58 jberger I don't know what twtxt is
17:58 mdom https://github.com/buckket/twtxt
17:59 mdom Or better http://twtxt.readthedocs.org/en/latest/user/usage.html
17:59 jberger anyway, I support using that gist and I especially like the example of using it as a "condvar" pattern on the delay object
18:00 jberger but I'm afraid to put it on cpan without safeguards that prevent it being misused
18:00 mdom sure, i can understand that
18:01 mdom But thanks for allowing to use the code!
18:01 jberger sure
18:01 jberger just use it responsibly
18:01 jberger and if you do want to try to clean it up and add some safeguards, at least by default, I wouldn't refuse it going on cpan
18:04 zivester joined #mojo
18:04 jberger mdom: twtxt seems interesting
18:25 mcsnolte joined #mojo
18:49 mdom Yeah, it's fun, but it's a very small project. The main client has a lot of stars on github, but that the result on being on hackernews. I would be surprised if we have more than 15 active user ... :)
18:56 elrod_ Is there a comparison somewhere between Gearman and Minion?
18:57 elrod_ Are they solving the same problem?
19:08 woz joined #mojo
19:09 asarch joined #mojo
19:16 tchaves joined #mojo
19:25 skinch wts ladder
19:45 jberger hmmmm, tests for Minion::Notifier have gotten a bit trickier
19:47 jberger I might have to ditch perform_jobs
19:51 woz joined #mojo
20:00 woz joined #mojo
20:01 jberger yeah, manually registering the worker and performing the job is much more reliable
20:13 cpan_mojo Minion-Notifier-0.05 by JBERGER https://metacpan.org/release/JBERGER/Minion-Notifier-0.05
20:14 asarch joined #mojo
20:17 sebbe joined #mojo
21:12 bpmedley Should there be something like Mojolicious::CLI?  This could have support some of the DSL provided by Mojolicious::Lite (plugin, app, IOLoop support, and probably others).
21:15 dvinciguerra_ joined #mojo
21:39 cpan_mojo WWW-EZTV-0.07 by DIEGOK https://metacpan.org/release/DIEGOK/WWW-EZTV-0.07
21:41 sri preaction: oh man, i think the final boss of the claptrap dlc just ruined the game for me
21:42 sri a giant bullet sponge that takes like an hour to kill...
21:42 * sri rage quits
21:42 bpmedley What's a bullet sponge?
21:44 hesco joined #mojo
21:45 Adura Borderlands games have enemies that do not die in a reasonable amount of shooting.
21:45 bpmedley Egads!! And that is paid for content?!
21:49 hesco In browser I see the redirect, but the setting for immediate expiration does not seem to be respected.  When I browse next to an end point which ought to be protected by my authentication, I get a 200 and access.  Are there other steps required to completely purge my session?  Should mention that $self in this code isa Mojolicious::Controller, what I access as $t->app->session in a Test::Mojo::More test
21:50 hesco script.  https://gist.github.com/hesco/759ef510a5638d65ccf3c7d2922afba0
22:00 woz joined #mojo
22:17 sri bpmedley: yea, bit of a bummer
22:17 sri apparently the dlc was rushed and thanks to an unfixed glitch the boss is almost impossible to beat solo
22:17 sri argh!
22:19 sri preaction: poke me when you get there and we'll kill him coop :)
22:21 preaction alright. might be awhile. my game time is severely limited
22:21 preaction also, at least it isn't depressingly easy like the Borderlands 2 end boss?
22:22 sri the normal end boss is not bad, 4 phases
22:24 sri took me 5 tries i think, but i looked up the strategy
22:49 bpmedley hesco: Would you create a minimal Lite app for debugging?
22:51 jantore joined #mojo
23:01 woz joined #mojo
23:26 hesco bpmedley: will give that a try I guess.
23:36 bpmedley https://github.com/brianmed/AtStartup <-- Anyone have time that can run once.pl in the "eg" directory?
23:42 meshl joined #mojo
23:46 jberger Sorry, /me is mobile atm
23:56 tchaves joined #mojo

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