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

IRC log for #mojo, 2016-04-16

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

All times shown according to UTC.

Time Nick Message
00:01 gizmomathboy joined #mojo
00:06 asarch joined #mojo
00:22 cpan_mojo Ado-0.930 by BEROV https://metacpan.org/release/BEROV/Ado-0.930
00:22 berov joined #mojo
00:53 berov left #mojo
01:03 woz joined #mojo
01:44 Adura joined #mojo
02:29 _dave_ joined #mojo
02:39 noganex_ joined #mojo
03:05 woz joined #mojo
04:52 Kripton joined #mojo
05:03 irqq joined #mojo
05:07 woz joined #mojo
05:14 Kripton joined #mojo
05:29 Kripton joined #mojo
05:47 Kripton joined #mojo
06:02 woz joined #mojo
07:24 Kripton joined #mojo
07:29 woz joined #mojo
08:02 woz joined #mojo
08:19 Vandal joined #mojo
08:41 umask001 joined #mojo
09:01 woz joined #mojo
09:19 meshl joined #mojo
09:41 woz joined #mojo
09:53 woz joined #mojo
10:09 woz joined #mojo
11:12 stuart joined #mojo
11:38 pink_mist jberger: what's the status of Mojo::ACME? will it automagically renew yet? :P
11:40 pink_mist jberger: "The plugin itself reads configuration out of the application's config method. This can be set directly in the application or loaded from a file via say Mojolicious::Plugin::Config in the usual way. It looks for a config key acme containing a hash of configuration options. Those options can be seen in the Mojolicious::Plugin::ACME documentation." <-- so I go to check the Mojolicious::Plugin::ACME
11:40 pink_mist documentation, and it's completely empty :P
11:52 pink_mist jberger: https://github.com/jberger/Mojo-ACME/blob/master/lib/Mojolicious/Plugin/ACME.pm#L48 <-- this looks like you're conditionally declaring $auth and $hmac_res ... that seems like a bad idea 0_o
11:53 woz joined #mojo
12:25 jberger No automagic yet
12:25 jberger Can you explain your fears about those lines?
12:26 jberger pink_mist: ^^
12:28 pink_mist jberger: having a conditional 'my' is a good way to get perl to segfault
12:29 pink_mist I mean, it used to be one of the tricks used to get state variables, but it was always prone to get the perl interpreter into an undefined state
12:30 pink_mist I've no idea if you're saved by the return statement returning if any of the statements are falsy, but I just always react when I see something like that
12:35 pink_mist I'd expect Grinnz_ could help throw some light on the issue
12:41 woz joined #mojo
12:43 jberger The return statements are very intentional
12:44 pink_mist as I said: I don't know if they save you or not.
12:44 pink_mist either way, it's still bad style to have a conditional 'my' declaration
12:45 jberger Respectfully disagree
12:47 jberger And I'm pretty sure that kind of construct (return unless my) is in the mini codebase too
12:48 pink_mist return unless my isn't bad
12:48 pink_mist return unless $foo && my is
12:48 jberger But it returns if it doesn't declare the variable
12:49 pink_mist no, return unless my always declares the variable
12:49 jberger my $× = 3 if $y; is the one that always scares me
12:50 pink_mist that's the SAME thing as $foo && my
12:50 jberger No it isn't
12:50 pink_mist yes it is
12:50 jberger Because the subroutine doesn't end
12:52 woz joined #mojo
13:11 haarg while it's goofy behavior you shouldn't rely on, it's not something i'd call dangerous
13:12 haarg and since you are returning, you aren't relying on the behavior of it
13:18 haarg and by 'it' i mean a conditional my that wasn't executed
14:22 woz joined #mojo
14:35 woz joined #mojo
14:41 Kripton joined #mojo
14:47 marty joined #mojo
15:01 woz joined #mojo
15:18 irqq joined #mojo
15:23 woz joined #mojo
16:10 woz joined #mojo
16:17 marty joined #mojo
16:26 PryMar56 joined #mojo
16:38 dvinciguerra joined #mojo
16:40 stryx` joined #mojo
16:57 lluad joined #mojo
16:59 sri any thoughts on this minion patch? https://gist.github.com/anonymous/ffbb9c0da1b5f2bac5cad0d94e6ba361
17:01 sri performace cost doesn't seem so bad
17:01 sri for minion_bench.pl i get 200 stats calls per second instead of 245
17:05 sri hmm, maybe also time i put minion_bench.pl into the minion dist
17:07 sri https://github.com/kraih/minion/blob/master/examples/minion_bench.pl
17:13 sri knowing how many jobs are actually ready for processing and how many are just delayed seems like valuable information
17:14 sri if there wasn't a performance cost i guess i wouldn't even hesitate
17:38 bpmedley sri: That could allow introspection for a queue, is that correct?  For example, feedback to a user could be given if there are too many jobs in a transcoding queue, right?
17:38 sri no
17:39 sri that's exactly what it doesn't do
17:43 bpmedley Odd, when I conceptualize a lot of delayed jobs it's that there are a number of them that will be ready in the future.  Is that correct, at lease?
17:53 sri not sure what you're talking about
17:53 sri sounds odd
17:54 bpmedley Fair enough, sorry for the ignorance on my part..
17:58 dod joined #mojo
17:58 sri adding queue specific stats was actually still an open question
17:59 bpmedley So, the added delayed stat is for the entire system?
17:59 sri all stats are
17:59 bpmedley Doh! Let me think..
18:05 bpmedley Users and developers love stats, so the more the merrier.. And, I wonder if something like a Minion::Queue would be useful for retrieving stats.
18:06 bpmedley Said another way, in my opinion a Minion::Queue::info that returned delayed, inactive, and active jobs would be sweet.  Lots of possibilities for what to retrieve exist.
18:13 sri the high level api is pretty irrelevant
18:13 sri it has to fit into Minion::Backend
18:18 dod joined #mojo
18:23 sri honestly, this is rather frustrating
18:23 bpmedley https://gist.github.com/brianmed/a594ef61e7308955ca84bc422bf001bf <-- My initial thought is that something as simple as copying and modifying job_info could be used for Queue specific stats.
18:23 sri for ages i've been asking for input on how to integrate named queues into stats and job lists
18:23 sri and here you are assuming it's already there
18:24 bpmedley My apologies. I should have done due diligence and read the latest documentation, code, and back log.
18:24 bpmedley However, I did just create a smallish (untested) patch..
18:26 bpmedley Honestly, without more context; I'm not sure how useful the number of delayed jobs is for the system as a whole.
18:36 sri ok, i'll remove it again
18:38 sri figured it would be useful for planning the worker pool
18:38 bpmedley Gaw; what days of the back log should I read to get more context.. I apologize, I was away for about a week.
18:39 sri like, there are a million inactive jobs in the queue, but you see 999k of them are delayed, so you know you don't have to add more workers
18:39 bpmedley So, you wanna dynamically add / subtract workers?  That would be sweet..
18:39 sri no idea, i bring the topic up every few weeks
18:40 sri personally, i've lost interest again, maybe we'll try again in a few weeks ;p
18:40 bpmedley I personally believe that introspection at the queue level would provide for a better minion experience.
18:41 bpmedley That's understandable; I hope I'll be around for the banter.  I love learning.
18:45 sri https://github.com/kraih/minion/issues/26#issuecomment-210872633
18:46 sri so much that could be added to minion
18:46 jberger sri: At $work we are on the verge of really tackling job dependencies
18:47 jberger I hope that might shake out some ideas
18:47 bpmedley We're glad you a fantastich gatekeeper of features.. Don't want more than necessary.. :-0
18:47 sri i'm not actually trying to be a gatekeeper ;p
18:48 sri if i had the time i'd add all of those features :D
18:48 bpmedley Wowzers, then I'm always in a perpetual state of confusion.. :-)
18:49 sri pretty admin ui and really good stats are a nobrainer imo
18:49 sri but getting them right requires you to be lucky or have a lot of time
18:50 jberger Some of that might come out of our efforts too
18:51 jberger I'll certainly be developing with an eye towards cpan release at least for parts
18:51 jberger Minion::Notifier being the first bit
19:14 sri named queue specific stats are really hard, not sure we'll ever have a good proposal for that
19:14 sri so far we don't even have a list of all named queues in use
19:19 cpan_mojo MojoX-Transaction-WebSocket76-0.05 by DIONYS https://metacpan.org/release/DIONYS/MojoX-Transaction-WebSocket76-0.05
19:30 batman sri: why doesn't Mojolicious::Types include eot and ttf?
19:31 sri think i based it on what formats are mentioned in important specs
19:31 batman ok
19:32 batman would you mind including eot and ttf? i think they are very common when serving font types...
19:32 sri i don't even know what eot means
19:33 batman "Embedded OpenType (EOT) fonts are a compact form of OpenType fonts designed by Microsoft for use as embedded fonts on web pages"
19:33 sri so, -1 i guess
19:34 sri although, the minion thing got me into a bad mood, so i'm -1 on pretty much everything today ;p
19:35 batman sorry to hear that :/
19:37 kaare joined #mojo
20:04 meshl joined #mojo
20:17 dvinciguerra joined #mojo
20:42 tchaves joined #mojo
20:48 irqq_ joined #mojo
20:53 new_student joined #mojo
20:56 csson joined #mojo
22:58 woz joined #mojo

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