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

IRC log for #mojo, 2015-10-28

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

All times shown according to UTC.

Time Nick Message
00:11 good_news_everyon joined #mojo
00:11 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vW9Um
00:11 good_news_everyon mojo/master 0914cab Sebastian Riedel: slightly better encapsulation
00:11 good_news_everyon left #mojo
00:11 sri i guess the fact that you can call ->warmup again to update the index might be a bonus
00:17 sri perhaps a better name than ->warmup would be good
00:18 sri can't think of anything though
00:24 good_news_everyon joined #mojo
00:24 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vW9ID
00:24 good_news_everyon mojo/master 40f7458 Sebastian Riedel: no need for two entries
00:24 good_news_everyon left #mojo
00:27 tencendur joined #mojo
00:29 good_news_everyon joined #mojo
00:29 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vW9tJ
00:29 good_news_everyon mojo/master 96c4350 Sebastian Riedel: mention that it happens automatically
00:29 good_news_everyon left #mojo
00:50 aborazmeh joined #mojo
01:09 PopeFelix joined #mojo
01:19 aborazmeh joined #mojo
01:40 Zoffix joined #mojo
01:49 davido_ joined #mojo
01:59 arpadszasz joined #mojo
02:00 dysfun joined #mojo
02:00 crab joined #mojo
02:00 esh joined #mojo
02:00 mikegrb joined #mojo
02:00 dabudabu joined #mojo
02:00 worklapperl joined #mojo
02:00 stephan48 joined #mojo
02:00 Repaster joined #mojo
02:00 KindOne joined #mojo
02:00 skaji joined #mojo
02:00 gabiruh joined #mojo
02:00 y1mmm joined #mojo
02:00 Nemix joined #mojo
02:00 Adura joined #mojo
02:00 Ralesk joined #mojo
02:00 bzero joined #mojo
02:00 Andreas joined #mojo
02:00 jwang joined #mojo
02:00 glauber joined #mojo
02:00 bowtie joined #mojo
02:00 fvox joined #mojo
02:00 allison joined #mojo
02:00 jnbek joined #mojo
02:00 oalders joined #mojo
02:02 lestrrat joined #mojo
02:04 Grinnz joined #mojo
02:06 renormalist joined #mojo
02:07 bc547 joined #mojo
02:12 rwp joined #mojo
02:13 jb360 joined #mojo
02:14 caeles joined #mojo
02:18 cfedde joined #mojo
02:25 zivester joined #mojo
02:55 arpadszasz joined #mojo
02:59 davido_ joined #mojo
03:16 mattastrophe joined #mojo
03:20 kaare_ joined #mojo
03:25 good_news_everyon joined #mojo
03:25 good_news_everyon [mojo] kraih pushed 2 new commits to master: http://git.io/vW9iu
03:25 good_news_everyon mojo/master 4f93f47 Stefan Adams: use only one space character...
03:25 good_news_everyon mojo/master ac83c8e Sebastian Riedel: Merge pull request #860 from s1037989/patch-4...
03:25 good_news_everyon left #mojo
04:00 nnutter joined #mojo
04:42 initialo joined #mojo
04:43 nnutter joined #mojo
04:47 cfedde joined #mojo
05:19 * sri decides to delete his facebook account
05:23 irqq joined #mojo
05:28 absolut_todd joined #mojo
05:37 inokenty-w joined #mojo
06:09 davido_ joined #mojo
07:38 dod joined #mojo
07:40 marcus I'm mostly using Facebook for groups and im these days.
07:53 dod joined #mojo
07:56 ashimema joined #mojo
07:58 Vandal joined #mojo
08:01 trone joined #mojo
08:11 meshl joined #mojo
08:16 eseyman joined #mojo
08:18 cheesekun joined #mojo
08:45 ashimema joined #mojo
09:01 bc547 joined #mojo
09:06 vanHoesel joined #mojo
09:07 miquelet joined #mojo
10:01 irqq joined #mojo
10:03 dvinciguerra joined #mojo
10:15 HtbaaPi when constructing a URL with Mojo::URL, using query(). How can I make sure query(foo => [1,2,3]) results in &foo[]=1&foo[]=2&foo[]=3 ?
10:17 mattastrophe joined #mojo
10:22 jontaylor joined #mojo
10:23 pink_mist what does it result in now? and in what way is that inadequate for your task?
10:23 HtbaaPi I need to talk to an API that expects this form
10:24 pink_mist which property of the form are you not getting?
10:24 HtbaaPi now it simply does foo=1&foo=2&foo=3
10:24 pink_mist so, did you try ->query('foo[]' => [1,2,3])?
10:24 Lee can you not do  query("foo[]" => [1,2,3]) ?
10:25 Lee jinx!
10:25 HtbaaPi Lee: no that doesn't work. It encodes the [ and ]
10:25 Lee ah!
10:25 pink_mist well that's annoying :P
10:26 HtbaaPi yes :P
10:28 * nic points to the sign on the wall that says "RFC"
10:29 * Lee had to deal with this for gocardless: https://metacpan.org/source/LEEJO/Business-GoCardless-0.12/lib/Business/GoCardless/Utils.pm#L127
10:29 Lee not using mojo though
10:30 HtbaaPi I don't know if it's part of the RFC but I know PHP does handle multiple values with the same name that way
10:30 HtbaaPi so I guess the API I'm talking to is written in PHP
10:32 HtbaaPi I guess for now I'll construct the URL without Mojo::URL
10:33 vanHoesel1 joined #mojo
10:34 Zoffix If anyone's close to Toronto, join Toronto Perl Mongers today at 7 for an "Introduction to Mojolicious" series of talks: http://www.meetup.com/Toronto-Perl-Mongers/events/225031099/
10:34 Lee why should it matter that the [ and ] are encoded? shitty parser on the other end?
10:36 Zoffix HtbaaPi, by the time it gets to PHP it should still become foo[] even if encoded.
10:44 HtbaaPi Zoffix: maybe so. But the API I'm talking to doesn't recognize it
10:44 HtbaaPi so yeah, I guess it's a shitty parser on the other end :)
10:44 CIAvash joined #mojo
10:47 Lee is this a public API?
10:50 HtbaaPi no
10:52 Lee aw ;)
10:58 vanHoesel joined #mojo
11:09 neilhwatson joined #mojo
11:20 asarch joined #mojo
11:21 hernan605 joined #mojo
11:42 cpan_mojo Toadfarm-0.69 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Toadfarm-0.69
12:12 vanHoesel joined #mojo
12:14 tencendur joined #mojo
12:18 kivilahtio joined #mojo
12:54 jberger I rarely use my facebook
12:54 jberger if no one messaged me about anything I'm sure I'd almost never use it
12:56 * nic detects too much denial
13:13 CandyAngel jberger: Set your status to something like "No-one messages me on facebook. No-one likes me :( I'm ugly"
13:14 CandyAngel Or does that only work for girls?.. I think it only works for girls, ignore that.
13:15 CandyAngel I'd message you, but I don't use facebook at all :P
13:17 * Grinnz has had his facebook deactivated for a couple years now
13:19 Grinnz HtbaaPi, if it's a PHP API then youshould send it foo=1&foo=2&foo=3 and it will decode it as foo[] itself
13:20 Grinnz i could be mistaken, i don't use PHP
13:21 ajr_ joined #mojo
13:25 Grinnz at work the javascript is always wanting to send us those stupid x[] params, so i just wrote a function which retrieves a combined array of 'x' and 'x[]' params
13:26 Grinnz whoever started doing that in URL parameters should be shot
13:29 gryphon joined #mojo
13:59 Vandal joined #mojo
14:03 jberger CandyAngel: hehe
14:03 tencendur joined #mojo
14:15 PopeF joined #mojo
14:21 doby joined #mojo
14:38 doby joined #mojo
14:47 disputin joined #mojo
15:04 vanHoesel joined #mojo
15:07 CIAvash joined #mojo
15:14 sri HtbaaPi: for the record, chrome percent encodes foo[] too
15:19 vanHoesel joined #mojo
15:23 Grinnz_ as does URI::QueryParam
15:24 meshl joined #mojo
15:38 PryMar56 joined #mojo
15:39 meshl joined #mojo
15:42 sri hmm
15:42 sri the minion file backend is pretty much unusable at 5000 jobs
15:51 vanHoesel joined #mojo
15:53 jontaylor_ joined #mojo
15:54 sh4 joined #mojo
16:01 sri think the file backend needs to be completely redesigned for better performance
16:02 sri have a few ideas, but no time
16:09 sri tbh. doubt i'll ever have time for that ;p
16:09 hernan606 joined #mojo
16:09 jberger imo the file backend was never really designed for large-scale usage
16:09 sri so, if anyone else wants a faster file backend, patches welcome ;)
16:09 jberger it is for prototyping and tests
16:10 jberger demos etc
16:10 disputin joined #mojo
16:10 sri well, it could at least be faster
16:10 sri by storing jobs more DBM::Deep friendly
16:10 sri right now it's just $jobs->{$id} = {...}
16:10 sri you can change that to $jobs->{$state}{$id} = {...} and the like
16:11 sri to reduce the number of jobs minion has to check
16:11 jberger the info calls would get harder, but that's probably a fair tradeoff
16:12 jberger what about two files?
16:12 sri all possible
16:12 jberger "completed" jobs go in the second file
16:12 jberger they can always move back
16:13 sri sorting is also expensive, you could make your own fake index to avoid resorting for every dequeue
16:14 sri sorting by created/priority
16:14 sri lots of optimizations possible
16:14 asarch joined #mojo
16:16 sri right now it's literally unusable at 5000 jobs
16:16 sri takes 20+ seconds to dequeue a job
16:17 sri and reads pretty much the entire file
16:18 sri and this is an ssd
16:18 eseyman joined #mojo
16:22 sri i'd say 500 jobs is the max for keeping the file backend usable
16:44 ajr_ joined #mojo
17:01 vanHoesel joined #mojo
17:10 csson joined #mojo
17:21 ajr_ joined #mojo
17:32 anon joined #mojo
17:37 Grinnz SQLite shouldnt have those problems... :)
17:44 sri nope, indexes help *a lot*
17:44 CIAvash joined #mojo
17:45 sri now i'm a little curious how the old file backend would perform at 5000 jobs
17:46 sri it used Storable+gzip and rewrote the entire file every time
17:46 sri have to admit, DBM::Deep is a lot weaker than i expected
17:46 jb360 joined #mojo
17:47 sri 5000 smal hashes on disk shouldn't be such a big problem
17:47 dvinciguerra joined #mojo
18:12 Trelane When should I use the MojoX namespace?
18:12 sri there are no rules, i don't use the namespace at all personally
18:13 Trelane So, is there any reason to use MojoX:: vs Mojo:: vs Mojolicious:: in a CPAN module?
18:13 Kogurr joined #mojo
18:14 * pink_mist tends to use Mojolicious:: when it's specifically related to the HTTP backend stuff, such as when writing a plugin for example
18:14 sri thing about using Mojo:: and Mojolicious:: is that you shouldn't use names that might clash with future core modules
18:14 pink_mist otherwise I use Mojo::
18:16 pink_mist or as the case may be, something completely different. I don't see the use for MojoX personally
18:16 sri Mojolicious is the framework, stuff we mostly made up ourselves, and Mojo the infrastructure used to build frameworks, often based on RFCs
18:17 sri sooo, DBM::Deep uses 10mb for 500 jobs, while Storable uses 188kb
18:18 pink_mist heh
18:18 sri wow, Storable wins by a lot when processing jobs
18:18 pink_mist what does SQLite use?
18:19 pink_mist (if you've checked?)
18:19 sri 0.5s per job vs 20s per job at 5000 jobs in the queue
18:19 pink_mist wow
18:19 sri there is no SQLite backend yet i believe
18:20 pink_mist oh, pity :P
18:20 sri and Storable actually has to rewrite the entire file every time a job gets enqueued or dequeued
18:22 trone joined #mojo
18:22 cpan_mojo Mojolicious-Plugin-DbicSchemaViewer-0.0100 by CSSON https://metacpan.org/release/CSSON/Mojolicious-Plugin-DbicSchemaViewer-0.0100
18:23 sri Storable+gzip is terrible in theory.... but unlike DBM::Deep actually still usable in practice :S
18:23 sri this is annoying!
18:25 sri jberger: you were so keen on using DBM::Deep, any thoughts?
18:27 avenj if I had to pick a word to describe DBM::Deep it would be 'glacial'
18:28 tencendur joined #mojo
18:29 sri i expected it to not be very fast, but the read performance is just abysmal
18:30 sri 20s to filter through 5000 little hashes
18:30 * sri shrugs
18:34 sri starting to wonder again if minion should at all ship a file backend in core
18:35 sri DBM::Deep is nice in theory, but doesn't work if you have more than a few hundred jobs
18:36 sri Storable+gzip works up to a few thousand, but wears out your ssd with big writes
18:37 Zoffix Oh
18:38 Zoffix sri, do you have a larger Mojolicious logo? I'd like to include it on the title slide of my presentation and the largest one I found is tiny: http://i.imgur.com/CA5Hag0.png
18:38 sri Zoffix: no, that's a pixel image
18:38 Zoffix :( Too bad
18:39 * Zoffix is tempted to use the rainbow-puking rapto-unicorn, but is afraid most people in the audience won't get the reference.
18:40 pink_mist I don't think I get the reference tbh 0_o I just thought it a weird quirk
18:40 Zoffix Well, not a "reference" but that it's something you'd see when working with Mojolicious. It's shown on the default production 500 page
18:42 pink_mist ah, right
18:42 sri Zoffix: i do have an unpolished vector version, it's not as nice though
18:43 Zoffix sri, I think it should work :)
18:43 sri https://gist.github.com/anonymous/ac8b5f57b13914359858
18:43 sri never finished it though, since the core logo needs to be pixel optimized anyway
18:44 pink_mist why is that not part of the Mojolicious codebase anyway?
18:44 sri the logos in core are too small
18:44 pink_mist yes, that's why an svg would be nice to have
18:44 sri you can't just scale down svg images that small
18:44 pink_mist oh you can't?
18:45 sri no, they look odd
18:45 sri small images you always want to hand optimize
18:46 pink_mist alright sure, but I don't quite understand why this would prevent the svg from being in the git?
18:46 sri why would it be there if we don't use it?
18:46 jberger sri: I don't have strong opinions on DBM::Deep vs other file backing
18:46 pink_mist so that /people/ can use it
18:47 jberger I like DBM::Deep, but I won't argue that it is performant
18:47 sri pink_mist: there is no canonical mojolicious svg logo
18:47 jberger I think having a file backed db in the core is really useful though
18:47 sri the one i gave Zoffix i would never ship myself
18:47 sri it's unfinished
18:47 pink_mist :/
18:47 jberger to images, it might be nice to have an official repo of images (with license)
18:48 jberger kraih/mojo-images
18:48 sri i don't have polished images
18:48 Zoffix sri, thanks. I can see it's unfinished, but it should be fine for my purpose :) Thanks.
18:48 sri besides the pixel ones in core
18:49 sri maybe some day i'll get around to putting something together, but polishing svg images takes time
18:50 genio graphics are cumbersome for me (I'm terrible at it)
18:53 sri releasing the raptor was not such a big success either https://github.com/kraih/perl-raptor
18:54 genio What do you use to make your vector graphics?
18:55 sri Sketch
18:57 sri what i do have in vector format is the frontpage images (including the glossy menubar logo) http://mojolicio.us/
18:58 sri they are not very reusable though
19:00 sri jberger: re file backends, problem is they all have problems
19:01 sri DBM::Deep is a dependency that doesn't even work very well
19:01 sri Storable+gzip wears out your ssd
19:09 cpan_mojo Mojolicious-Plugin-DbicSchemaViewer-0.0101 by CSSON https://metacpan.org/release/CSSON/Mojolicious-Plugin-DbicSchemaViewer-0.0101
19:30 dvinciguerra joined #mojo
19:38 zoo-zed joined #mojo
19:50 zoo-zed joined #mojo
19:51 zoo-zed joined #mojo
19:52 mudler joined #mojo
19:55 tencendur joined #mojo
19:59 jberger sri: its ok for a job to call $job->finish($result), correct?
19:59 jberger if so, is that documented somewhere?
19:59 jberger Minion::Job::finish doesn't make it clear that that is from inside the job body
20:09 sri jberger: where should it be documented?
20:16 sri anyway, already spent too much time on minion today
20:20 lluad joined #mojo
20:25 sri odds are the file backend will be going away in 3.0, if only to make my life easier, since i don't actually use it myself
20:26 sri and nobody else has invested anything in it
20:27 sri so, if someone here wants the file backend to stay, i expect you to step up and make it work reasonably well at a few thousand jobs
20:32 jberger the convenience of just being able to spin up a minion for testing isn't worth anything to anyone else?
20:32 jberger I'll try to take a look
20:32 jberger I never use it for 1000s of jobs, I use it for maybe dozens
20:33 sri well, the bad performance already caused complaints
20:33 Grinnz_ i would say it's useful to leave as testing, but to recommend against for production
20:33 jberger complaints?
20:33 sri last time fro HtbaaPi
20:34 jberger doesn't that just garner a "don't use morbo as your production server" type response?
20:34 sri and honestly, 5000 jobs is not an outrageous number
20:34 sri the file backend makes minion look bad as it stands
20:34 Grinnz_ it's more "don't use for anything except test files" tbh
20:34 jberger sri: define "look bad", you mean performance-wise or code-wise?
20:35 sri performance wise
20:36 sri just sitting there, eating cpu, keeping disks spinning, 20+ seconds to find the next job
20:37 sri the pg backend on the other hand is so nice...
20:37 sri with pub/sub notifications and everything, keeping latency super low and resource usage at a minimum
20:37 jberger just like back in the early days of minion, I think without an easy-to-try setup, minion adoption will be much slower
20:38 jberger do you think anyone would try my send email script from my perltricks article if they had to setup a postgres server first?
20:38 sri the file backend scares people away
20:38 sri that's how i see it
20:39 sri there actually is people complaining
20:39 sri you can have a file backend in a 3rd party cpan module
20:40 sri or even different file backends
20:40 sri have a DBM::Deep one and a Storable one
20:40 firnsy people complaining indicates they're actually using it, "no complaints" should not be your goal
20:41 jberger ^^ ++
20:41 sri still bugs me, the fila backend is not up to my standards
20:41 sri s/s/e/
20:42 jberger I bet a directory tree backend beats most of them
20:42 nic That one complaint was from someone who hadn't adjusted their thresholds to match their wants
20:42 sri jberger: that's an argument for removal ;p
20:42 sri to encourage more file backends
20:42 PryMar56 joined #mojo
20:43 jberger sri: I know how your logic works, I'm already trying to plan my future course of action
20:43 jberger I use the file backend quite a bit
20:44 * nic <3 File
20:44 sri i'd imagine an SQLite backend will take its place
20:44 jberger does one exist?
20:44 sri best way to get performance and ease of setup
20:45 * Grinnz_ will work on that as soon as he is not at work :P
20:45 sri not yet
20:45 jberger that would probably be the best of both worlds
20:45 jberger especially if it can work :memory: for in-process testing (ie ->perform_jobs)
20:46 sri does that work with forked processes?
20:46 Grinnz_ no
20:47 Grinnz_ Mojo::SQLite will emulate it with tempfiles though
20:47 sri then it would break a lot
20:47 sri perform_jobs still forks processes, and you need to be able to access the backend from both processes
20:47 jberger ah, right
20:47 sri access and modify
20:48 jberger oh well, file backed is ok, its the same as what I have now
20:48 Grinnz_ also, every ->db from Mojo::SQLite could be a different database if you use :memory:
20:48 Grinnz_ thats why it defaults to the tempfile
20:49 Grinnz_ since it sets WAL mode, access to the tempfile from forked processes is no problem
20:49 sri honestly, i'm very disappointed with DBM::Deep
20:49 sri with that read performance i don't see it being usable for much
20:50 osfameron sri: really?  surely it's *obviously* a terrible idea? ;-)
20:50 osfameron or at least, I seem to remember being aware it had its limitations
20:50 trone_ joined #mojo
20:59 sri are there any other job queues that support a file backend?
21:03 dvinciguerra joined #mojo
21:10 sri for the record, i did like having an easy o set up backend in minion
21:11 sri but for all use cases i've seen so far, even Storable would be better
21:12 sri "i'll never use more than a few dozen jobs" <- then Storable is so much better for you!
21:18 jberger I'm not married to an implementation
21:19 jberger with a file-per-job you can do job-level locking
21:22 dvinciguerra joined #mojo
21:23 meshl joined #mojo
21:27 lb speaking of storable - a billion years ago chansen wrote a script to compare the performance of serializers, that might be helpful; it's gone from its official url now, but i uploaded my slightly modified version to https://gist.github.com/lbalker/79f3dbf315248ed3c257
21:27 lb also, chansen is pretty old
21:34 sawtooth joined #mojo
21:44 Grinnz_ huh, count(distinct foo) works in sqlite but i can't find the documentation that says it works :P
21:44 Grinnz_ oh, there it is
21:44 Grinnz_ at the top of the page...
21:44 Grinnz_ "In any aggregate function that takes a single argument, that argument can be preceded by the keyword DISTINCT."
21:51 sri and Minion::Backend::File is gone from master for now https://github.com/kraih/minion/commit/be5b07e842f96ae55e8dc24f8e3561bd40ad1b58
21:52 jberger sri: does that mean that the next minion is going to be 3.0?
21:52 sri yea
21:53 jberger is someone going to adopt it on cpan, perhaps under a a different name?
21:53 sri was about to ask ;p
21:54 sri if someone wants to adopt the DBM::Deep backend now is a good time
21:54 sri and yes, with a name change
21:54 sri so we can keep Minion::Backend::File open in case a great one pops up that is core worthy
21:55 sri in fact, the Storable backend is also up for adoption (needs to be updated for the 2.0 changes though)
21:57 jberger if no one else wants it, I can take it, though I can't say I'm probably the best one for it
21:57 jberger if someone else wants it, please speak up
21:58 sri best maintained 3rd party backend currently is Minion::Backend::mysql
21:58 sri the mongodb ones didn't get the versioning changes
22:00 sri really too bad we couldn't get the list_jobs thing settled yet, sucks to have so many backends on cpan knowing they will all break again
22:00 jberger are any other breaking changes expected at 3.0, ie, will I (or the new maintainer) need to change anything right away?
22:00 jberger list_jobs ?
22:00 sri i have no plans, but expect requests from ui builders
22:01 sri more stats, and ways to filter jobs for listing
22:01 jberger so far bpmedley and I are them :{
22:01 sri like, you can't filter by queue yet
22:01 jberger :P
22:02 sri yes, so i guess... jberger... do *you* have plans to make me break minion again? ;p
22:02 jberger you saw the amount of time I've had to do the ui :s
22:02 jberger I had a few hours to restructure bpmedley's attempt
22:03 bpmedley I've been busy the last few days, so haven't fully looked at those changes.  Hopefuly soon.
22:04 sri lol, i found a new concurrency issue in Mojo::Pg
22:04 sri you can isolate different tests with a schema, but notifications are of course still shared
22:04 sri https://travis-ci.org/kraih/minion/builds/87995935
22:05 jberger http://www.sadtrombone.com/
22:05 jberger actually: http://www.sadtrombone.com/?autoplay=true
22:11 sri :(
22:12 jberger can you namespace the notifications?
22:12 sri don't think so
22:12 jberger I guess that's ugly
22:12 jberger I mean manually
22:12 sri there's a trigger function sending it
22:12 jberger oh
22:13 sri it's an insert trigger
22:13 sri anyway, this is quite interesting
22:13 sri otherwise namespaces work great
22:14 sri i've converted the file app test to postgres https://github.com/kraih/minion/commit/fc6a90f745d603c1cb7b27a3cabaf144a0fb8031
22:14 sri and it can run concurrently with the other tests
22:14 sri even though there's automatic migrations, and resetting involved
22:21 disputin joined #mojo
22:23 ajr_ joined #mojo
22:33 sri this is the benchmark i currently use for the postgres backend btw. https://gist.github.com/anonymous/213b15792a98cf06fef6
22:35 sri and the results on my little macbook air https://gist.github.com/anonymous/d2919c05ef949eb76a0c
22:46 irqq joined #mojo
22:57 zivester joined #mojo
22:59 sawtooth hello friends, should flash(message => 'blah'); redirect_to('somewhereelse'); etc.. work when the 'somewhereelse' controller uses $c->render_later?
23:17 d4rkie joined #mojo
23:21 sri jberger: btw. there is an example http://mojolicio.us/perldoc/Minion#add_task
23:26 dvinciguerra joined #mojo
23:28 sri looking through existing job queues with multiple backends
23:29 sri i couldn't find any using a file backend, closest would be backends that just perform jobs right away and keep all data in memory
23:34 sri guess i'll add a recommendation for the sqlite backend to Changes, should Grinnz get everything working
23:36 Grinnz sri, to go by polling instead of using pubsub (which isnt really pubsub in sqlite) i just have to make dequeue() like the File backend right?
23:36 sri right

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