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

IRC log for #mojo, 2017-11-19

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

All times shown according to UTC.

Time Nick Message
00:27 mohawk Grinnz, did you get benchmark numbers?
00:27 mohawk i'm very interested in how this kind of task performs with nosql
00:30 karjala_ joined #mojo
00:30 sri mohawk: redis is made for this and an optimized implementation will outperform any rdbms
00:31 sri the minion api does have some requirements that make it tricky to optimize for redis
00:31 sri since it was made with rdbms in mind
00:31 mohawk fair enough, thanks
01:00 mohawk well, that seems to be schema-complete for cpantester's OpenAPI -> GraphQL
01:05 CandyAngel With regards to Minion jobs, why would you use notes instead of passing as arguments?
01:05 sri you make notes while the job is running
01:06 sri $job->note(progress => '10%');
01:07 sri or keep job meta data such as memory/cpu usage
01:07 CandyAngel Just possible through enqueue for completeness?
01:08 sri or to set defaults, like progress => '0%'
01:08 CandyAngel Ah, makes sense. Thankies :)
01:09 sri it's also preparation for possible future minion extensions
01:10 sri ->enqueue(foo => [] => {notes => {memory_limit => '10MB'}})
01:10 sri we don't have a formal extension mechanism yet, but you could do that already with events
01:11 CandyAngel Ah coolies
01:47 aborazmeh joined #mojo
01:52 kaare_ joined #mojo
02:55 ilbot2 joined #mojo
02:55 Topic for #mojo is now 🍩 nom nom | http://mojolicious.org | http://irclog.mojolicious.org | http://code-of-conduct.mojolicious.org
03:09 noganex joined #mojo
03:09 kaare joined #mojo
04:44 jamesaxl joined #mojo
04:52 itaipu joined #mojo
05:04 dboehmer_ joined #mojo
05:09 preaction okay, so the django admin console is pretty slick, but it's basically what i'm going for. totally stealing some of the UX though
05:10 mohawk preaction, tomorrow could i ask for some help testing the gql -> cpantesters thing?
05:10 preaction maybe? ether found a problem in the code i released today that i have to fix (reports can't be viewed anymore, which is great)
05:10 mohawk i'm a bit overawed by how to hook up gql -> mojo::ua/test::mojo -> cpantesters::api::base test
05:11 mohawk uh oh
05:11 mohawk well, that'll be tomorrow anyway :-)
05:11 mohawk good luck with this fire
06:02 Vandal joined #mojo
08:32 ccakes joined #mojo
08:37 trone joined #mojo
09:12 geospeck joined #mojo
09:51 dod joined #mojo
09:56 dod joined #mojo
10:16 ghenry joined #mojo
10:32 [vlad] joined #mojo
11:34 itaipu joined #mojo
11:49 pheleas_frog joined #mojo
12:27 sh14 joined #mojo
12:52 sri popover form for creating jobs looks kinda good https://i.imgur.com/yKtOYgw.png
12:53 sri downside is that validation errors can't re-render the form for correction
12:54 sri would have to show a red notification at the top with a sentence listing the fields that failed validation
13:02 sri well, i'll keep these experiments as branches for later https://github.com/kraih/minion/compare/admin_create_job_popover
13:08 good_news_everyon joined #mojo
13:08 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vFHMn
13:08 good_news_everyon mojo/master 35fc9d6 Sebastian Riedel: silence a warning on older versions of Perl
13:08 good_news_everyon left #mojo
14:17 ghenry joined #mojo
15:16 geospeck joined #mojo
15:18 sh14|2 joined #mojo
15:23 karjala_ joined #mojo
15:26 mib_91butb joined #mojo
16:13 sri i've added notifications though https://i.imgur.com/q5pTAUD.png
16:35 dod joined #mojo
16:38 pheleas_frog does ojo since v7.51 enable strict on purpose?
16:39 jberger ojo has always enabled strict
16:39 pheleas_frog really?
16:40 sri jberger: actually it was borked before
16:40 pheleas_frog with v7.50, perl -Mojo -E '$s = "hello"; say $s' was fine
16:40 jberger oh, hmmmmm
16:40 jberger I stand corrected then
16:40 jberger it was intended to enable strict I believe
16:40 sri it was always meant to though
16:40 pheleas_frog even for one-liners?
16:41 jberger I mean, what it does it create an inline Mojolicious::Lite app
16:41 jberger so it seems to me that that is consistent
16:41 jberger but I don't have a strong opinion I suppose
16:47 pink_mist I think not enabling strict for ojo makes sense
16:48 pink_mist but I don't have a strong opinion either
16:59 mohawk i think one-liners are the one place where no strict makes sense
17:07 sh14 joined #mojo
17:23 mohawk_pts joined #mojo
17:23 mohawk joined #mojo
17:24 preaction if a minion job fails with "Worker went away", is it supposed to be re-attempted? it doesn't seem to be retried, even though i set attempts => 10
17:26 sri which backend?
17:26 purl which backend are you using?
17:27 sri i'll assume your backend is missing the feature
17:27 mohawk purl, forget which backend
17:27 purl mohawk: I forgot which backend
17:27 preaction i mean, i thought i implemented it. i could have implemented it wrongly though
17:28 Grinnz preaction: that's one of the steps in repair
17:29 preaction ahh. yeah, so, no, i failed to implement that correctly
17:29 sri as far as i can see there's even tests
17:29 sri you really need to go over the whole backend again method by method
17:30 sri if you started from the old implementation and just added stuff
17:30 sri it was never any good
17:30 preaction that is, in fact, what i did: hammer things until it worked on Minion 7
17:32 mohawk_pts joined #mojo
17:32 Grinnz for the redis backend i had the source for the pg and sqlite backends open side by side, sometimes one or the other is easier to read, maybe that will help
17:33 sri Grinnz: did you ever find anything not covered in the tests?
17:34 sri i would hope just passing all the tests would be enough
17:34 Grinnz not so far
17:34 Grinnz so make a copy of the current tests first
17:34 preaction and then s/pg/mysql/gi
17:34 Grinnz you'll have to edit some backend queries too probably but yeah
17:35 sri yea, no way around that
17:35 preaction yeah
17:35 mohawk_pts sorry for another dumb question[tm] - is dbix::class not suitable for this?
17:35 sri hell no, way slow
17:36 mohawk_pts fair enough
17:36 sri every single query in the pg backend is hand optimized with pg-isms
17:36 Grinnz sqlite has to use a lot of interesting joins as well, that dbix::class would not be good at replicating
17:37 karjala_ joined #mojo
17:37 mohawk_pts interesting, thanks
17:37 preaction yeah. i'll end up replicating those joins in mysql, since mysql is closer to sqlite than it is pg
17:37 Grinnz right, no arrays is a big difference
17:37 preaction and i really want to require mysql 5.6 for its JSON support to get some of the pg-isms (i can make arrays via the json, probably)
17:38 preaction (i practically have required it since 5.5->5.6 had a lot of changes to the TIMESTAMP columns that are not backwards-compatible)
17:38 Grinnz oh yeah a lot of things are way easier with at least the ability to query json
17:38 Grinnz i use json for a couple of the arrays in sqlite
17:39 Grinnz the docs do say "This backend requires at least v5.6.5 of MySQL."
17:39 preaction ... i don't remember doing that. maybe it was the guy who's been cleaning things up a few weeks ago
17:40 Grinnz maybe the constructor can check for it like pg does
17:45 preaction yeah, probably
17:48 sri almost wonder if there should be a Minion::Lint to check backends :)
17:55 preaction i mean, i copy/pasted the test just now, and there were only like 5 Pg-isms i had to fix
17:56 ChmEarl joined #mojo
18:21 karjala_ joined #mojo
19:42 kaare joined #mojo
22:00 mohawk sri, have you considered getting Mojo's promises listed here? https://promisesaplus.com/implementations
22:01 sri mohawk: no
22:01 mohawk no you haven't or no you're not interested?
22:04 mattp joined #mojo
22:07 sri haven't, and don't really care
22:08 sri feel free if you think it's worth it
22:18 mohawk https://github.com/promises-aplus/promises-spec/pull/261
22:20 zivester joined #mojo
22:31 sri mohawk++
22:31 good_news_everyon joined #mojo
22:31 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vFQts
22:31 good_news_everyon mojo/master 149fd9e Sebastian Riedel: mention Promises/A+
22:31 good_news_everyon left #mojo
22:34 mohawk when they merge it i'll let you know
22:34 mohawk they encourage those who conform to put a logo on website, what do you think?
22:35 mohawk (dear god, how do i test promises without having to choose a promises framework to go with)
22:35 sri only when it's advantageous to the design
22:35 mohawk i'll let you know when merged and then you can figure whether to do anything ;-)
22:38 mohawk it's enough to make you want to use Future, i tells ya
22:39 mohawk sri, am i right in assuming that returning a Future in a Mojo callback will not lead to good things?
22:47 itaipu joined #mojo
23:54 disputin joined #mojo

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