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

IRC log for #mojo, 2016-01-30

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

All times shown according to UTC.

Time Nick Message
00:35 jontaylor joined #mojo
01:04 bjoernfan joined #mojo
01:25 gtodd joined #mojo
02:25 romel joined #mojo
02:25 sugar joined #mojo
03:11 romel joined #mojo
03:13 romel joined #mojo
03:17 marty joined #mojo
03:20 marty__ joined #mojo
03:36 noganex joined #mojo
03:39 kaare joined #mojo
04:48 c--_ joined #mojo
06:08 davido joined #mojo
07:17 dod joined #mojo
08:17 Vandal joined #mojo
08:20 marty joined #mojo
08:31 dod joined #mojo
10:14 kes joined #mojo
11:00 lluad_ joined #mojo
11:08 jontaylor joined #mojo
14:44 jberger I still haven't gotten my stickers :'(
14:46 sri :(
14:52 jberger It says they were shipped so I don't think it's stickermule's fault
15:02 disputin joined #mojo
15:02 disputin joined #mojo
15:27 asarch joined #mojo
15:38 dod joined #mojo
15:42 good_news_everyon joined #mojo
15:42 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vzhaO
15:42 good_news_everyon mojo/master cbf534d Sebastian Riedel: better content_for tests
15:42 good_news_everyon left #mojo
16:43 romel joined #mojo
17:58 PryMar56 joined #mojo
18:10 oliver hi folks. Is there a way to know when a group of Minion jobs have completed, and react to that?
18:11 sri no, you have to make your own notifications
18:13 oliver sri: okay, thanks
18:14 oliver I'm using Minion::Notifier for reacting to individual jobs, but I now have a case of wanting to do something after a set has finished. Perhaps I'll try to come at it from another angle.
18:17 sri oh, i know nothing about Minion::Notifier
18:19 oliver sri: heh, okay. perhaps the author would be interested in your opinion. I certainly find it useful in my app.
18:22 oliver sri: by the way, thanks for Mojolicious and Minion, they're ace :)
18:32 jberger sri: notifier was extracted out of my "project codenamed banana"
18:33 oliver jberger: ah, it's yours.. thank you also
18:34 oliver I'm registering a once() handler against "job:$id" after enqueue
18:35 jberger job:$id still scares me since I still worry about a race condition on very fast jobs
18:36 oliver hmm yes
18:45 jberger Current status: http://imgur.com/gjL7vc4
18:51 disputin joined #mojo
19:10 bpmedley I ordered a coffee mug.. :-)
19:21 sri \o/
19:21 sri need more pictures on twitter
19:32 good_news_everyon joined #mojo
19:32 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vzhQx
19:32 good_news_everyon mojo/master d5000f4 Sebastian Riedel: fix typo in Mojo::WebSocket
19:32 good_news_everyon left #mojo
19:37 jberger I wish I could post some stickers
19:37 * jberger weeps
19:37 * jberger drowns sorrows http://imgur.com/2j1yxGF
19:52 * sri is tempted to consider this issue failed https://github.com/kraih/minion/issues/24
20:01 sri there is clearly interest, but i don't see any hope for a decent solution
20:02 sri nobody is going to invest like two weeks to design a nice distributed scheduler
20:18 chansen Designing a reliable distributed scheduler requires more than "two weeks" ;)
20:18 sri i think it's doable in this case, since you can cheat a little
20:19 chansen cheat?
20:19 sri even if two workers end up being the scheduler for some time the shared locks will prevent duplicated jobs, it would just be a liitle more load on the database
20:20 chansen Ok, how would the "shared lock" be persisted?
20:21 sri there was one proposal already where it was simply a field in the postgres table
20:21 sri https://github.com/kraih/minion/pull/22
20:22 * chansen reads
20:23 sri super simple
20:23 sri (the patch has flaws, but the idea is sensible imo)
20:25 sri sidekiq uses a key like $queue:$job_type:$args in redis with a time based expiration
20:29 chansen I hope his servers is time synchronized to sub second! Otherwise he is in a deep shit of problems ;)
20:31 sri https://github.com/mperham/sidekiq/wiki/Ent-Periodic-Jobs#limitations
20:32 sri i have yet to find an implementation without countless limitations
20:32 sri bad accuracy seems to be fine for most use cases though
20:33 sri once a day, once a week...
20:34 sri i even keep hearing that some folks don't care about scheduler failover at all :)
20:35 sri actually the main reason why i keep thinking that maybe cron+unique jobs is enough after all
20:47 chansen bad accuracy can be acceptable depending on application, but designing/providing a scheduler that doesn't meet the developer/user's expectation, probably causes more harm than good.
20:51 chansen Personally I would not rely on cron.
21:04 Adura joined #mojo
21:15 * jberger current status: http://imgur.com/3gXdTi7
21:26 * sri changes the channel topic to "jberger's instagram"
21:30 sri hmmmm... i guess the fact that none of us actually have an instagram means we are getting old :(
21:36 mcsnolte joined #mojo
21:40 jberger My wife does, does that count?
22:29 sugar joined #mojo
23:44 sugar_ joined #mojo
23:44 genio How old am I if I don't use Twitter?
23:49 sugar__ joined #mojo
23:49 jberger GET OFF genio's LAWN!!!

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