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

IRC log for #mojo, 2016-08-31

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

All times shown according to UTC.

Time Nick Message
01:32 disputin joined #mojo
02:41 noganex joined #mojo
02:45 itaipu joined #mojo
02:54 Bender joined #mojo
02:58 jehdin joined #mojo
03:32 itaipu joined #mojo
04:02 asarch joined #mojo
04:10 alilles joined #mojo
04:11 itaipu joined #mojo
04:19 jberger https://www.aidanwoods.com/blog/faulty-login-pages
04:20 jberger I especially love the brotastic way they shrug him off too
05:00 itaipu joined #mojo
05:44 Guest-questX joined #mojo
05:53 inokenty-w joined #mojo
06:10 mbudde joined #mojo
06:33 Atog joined #mojo
06:36 dod joined #mojo
06:41 dod joined #mojo
07:08 coolo joined #mojo
07:08 Lee joined #mojo
07:20 alilles joined #mojo
07:27 Vandal joined #mojo
07:31 itaipu joined #mojo
07:32 AndrewIsh joined #mojo
07:48 itaipu joined #mojo
07:59 osfabibisi joined #mojo
08:52 margeas joined #mojo
09:13 dod joined #mojo
09:30 stephan48 joined #mojo
09:39 mpapec1 thumbs up for Merkel, http://sputniknews.com/europe/20160831/1044802816/frankfurt-airport-Evacuation.html
09:47 mpapec1 http://deutsche-wirtschafts-nachrichten.de/2016/08/31/frankfurter-flughafen-wird-teilweise-evakuiert/
09:47 nic wrong channel
09:48 pink_mist unsure what merkel has to do with anything
09:49 kes joined #mojo
09:53 mpapec1 she welcomed the refugees?
09:53 nic definitely the wrong channel
09:54 mpapec1 tnx nic for the notice ;)
09:56 pink_mist mpapec1: huh? what does merkel welcoming refugees have to do with frankfurt airport being evacuated?
09:58 pink_mist but sure, thumbs up for merkel for that ...
10:12 cuechan joined #mojo
10:22 meshl joined #mojo
10:25 cuechan joined #mojo
10:28 itaipu joined #mojo
10:41 cuechan_ joined #mojo
10:45 cuechan joined #mojo
10:47 cuechan_ joined #mojo
10:57 cuechan joined #mojo
11:03 cuechan_ joined #mojo
11:06 cuechan joined #mojo
11:11 tchaves joined #mojo
11:22 tchaves joined #mojo
11:25 tchaves joined #mojo
11:51 jberger Not to get to political
11:53 jberger But I'd say this channel leans to the "don't blame immigrants for things that aren't the immigrants fault; oh dear gods you know you're talking about Germany right?!" side of the spectrum
11:53 jberger Or maybe it's just me
11:57 cpan_mojo Mojolicious-Plugin-OAuth2-Server-0.27 by LEEJO https://metacpan.org/release/LEEJO/Mojolicious-Plugin-OAuth2-Server-0.27
11:57 Lee now with implicit grant flow!
12:04 cuechan_ joined #mojo
12:23 jberger Lee: can you describe that? I don't know what that means
12:23 Lee allows javascript calls to get access tokens
12:24 Lee not recommended unless scopes are restricted
12:24 Lee (it's basically the auth code flow minus the auth code)
12:24 Lee client -> "gimme an access token"
12:24 Lee server -> "ok, your redirect_uri and scopes look good so here you go"
12:37 cpan_mojo Mojolicious-Plugin-OAuth2-Server-0.28 by LEEJO https://metacpan.org/release/LEEJO/Mojolicious-Plugin-OAuth2-Server-0.28
12:37 Lee and it's a mindfield :/
12:38 sri that was a cool the prodigy song
12:39 Lee also a minefield :)
12:41 sri in case someone wants to work on minion https://github.com/kraih/minion/issues/39
12:41 sri (my motivation is still pretty low after yesterday)
12:48 gizmomathboy joined #mojo
13:07 dod joined #mojo
13:26 cuechan joined #mojo
13:30 zivester joined #mojo
13:48 jberger I've been thinking some more about that, especially falling asleep last night
13:48 blonewolfs joined #mojo
13:48 jberger and while I guess you can say there is some demotivation about minion (I don't agree, but I can understand it)
13:49 Dandre joined #mojo
13:49 jberger I think it says more about the excitement in the Mojo community
13:49 jberger people are building real things and solving real problems using Mojo
13:49 jberger and that's awesome
13:49 jberger and maybe Minion isn't right for everyone
13:49 jberger it is great for what it is great for
13:50 jberger like say for MetaCPAN, indexing new dists as they come in, but then sometimes doing a full reindex when they need to change something
13:50 jberger minion is really helping them
13:51 jberger they've had to do full reindexes of the cpan several times in this effort to move to the new version of elasticsearch
13:51 jberger and they picked minion because a reindex was taking days sequentially (even with some forking for parallelism)
13:52 jberger and inevitably something would break on some random dist that does something crappy
13:52 jberger and so minion has made it so much easier, they have 3 worker boxes
13:52 jberger each dist is its own job
13:52 jberger the few that fail don't stop all the rest, big win
13:54 * jberger rests his case
13:54 sri it's not about minion being right for everyone, it's about competing with you
13:55 jberger I don't need to open mine, I've said that already
13:55 sri even if i win, it won't really be winning :(
13:55 jberger also this isn't zero sum
13:55 jberger batman and I didn't compete over forkcall vs readwritefork
13:56 jberger there were cases where I would recommend rwf and he uses forkcall in the new convos
13:56 sri thing is you'll keep working on the job queue for servercentral, and the better it gets the more you will feel an itch to release it
13:56 jberger (though he will probably use the subprocess)
13:56 ramortegui joined #mojo
13:56 sri that's now a damocles sword hanging over minion
13:57 jberger I think that's a pretty pessimistic view
13:57 sri and being a servercentral backed project, you'd have resources minion doesn't have
13:58 jberger hehe
13:58 jberger you know my team is like 3 people right?
13:58 jberger of which I'm not even the lead
13:58 sri still more than minion has
13:59 jberger you know that my interest is in making the community the best it can be, I don't want to cause strife, indeed I don't like even being near strife
14:00 jberger but I also can't nuke my project either, I need it for $work, at least until we could backport things into minion
14:00 jberger so I don't know what to say next, other than I hope you'll trust that I won't hurt this community
14:01 sri i don't think there is anything that could be done differently, and i'm not criticizing
14:02 perlpilot joined #mojo
14:02 jberger I think the thing I regret most is pushing for multiple backends for minion
14:03 jberger I think that having file backed minion helps push adoption, since it is easy to try (the same reason the new convos' first backend it a file, redis was a barrier)
14:03 jberger but now most of the things I'd like to add to minion really would be pg specific
14:05 sri like?
14:06 jberger sequential queues are so easy with pg array types
14:06 sri what is a sequential queue?
14:06 jberger fifo queue
14:07 sri minion queues are already fifo
14:07 jberger sorry, where you can't pick the next one while the current one is running
14:07 jberger blocking fifo queue? I'm not sure what to call that
14:08 sri how is that problem not solved by job dependencies?
14:09 jberger in mine I have two concepts batches which are like minion job dependencies which fail if any one fails
14:09 jberger and the queue of batches which are sequential and blocking and the next one does not fail if the previous one did
14:10 dod joined #mojo
14:10 jberger I made a rule that all jobs must be in a batch even if there is only one
14:10 jberger and it made the design relatively simple (ish)
14:10 jberger the batch also declares the locking tags etc
14:11 jberger that was mostly convenience, the jobs could also have done that
14:11 jberger but in practice it seems like a batch of jobs mostly needs the same things locked
14:12 jberger (don't reset the switch while I'm installing a server OS etc)
14:13 jberger installing the server OS might be a few different jobs, but I don't want to allow the switch lock to go away while they progress through
14:17 sri why would you want to continue in a dependency chain if one of the jobs failed?
14:19 jberger that was discussed, but it was decided that they are different batches and they should continue
14:20 jberger things like polling for system status
14:20 jberger batches need to encompass all the things that would need to fail if one thing fails
14:22 jberger indeed the first job in a batch might be to verify that the system is in a correct state to perform the rest of the batch
14:26 perlpilot joined #mojo
14:26 asarch joined #mojo
14:26 sri so a simple parent_may_fail flag in minion would have worked
14:32 jberger can we please not discuss in terms of what could have or might have worked? I needed to get something working in a month and the number modifications that I would have needed to make to minion would have been impossible to plan and execute safely for other users
14:32 jberger I'm happy to mine it for minion ideas now
14:32 jberger and if I can move back then great
14:33 sri see, this is what i mean, feature ideas are not pushed back upstream anymore :(
14:33 sri i was interested in it as a minion feature
14:33 jberger I think that's unfair, it still IS interesting as a minion feature
14:34 jberger so lets design it right and put it in
14:34 jberger would you really have let me bash in all the things I needed for minion that quickly?
14:35 sri we'll never know now
14:35 jberger I didn't have time to discuss them with myself most times :-P
14:35 sri what's the performance of yours like anyway?
14:36 jberger I haven't tried it at any kind of scale
14:36 jberger and we have a 7s limit between jobs, so, bad
14:36 jberger :D
14:37 jberger the backside controllers on these servers are surprisingly underpowered for what they cost to be included
14:37 jberger given them too many requests (and I mean like, a request every 3s is pushing it) and they flip out
14:38 jberger also I'm testing against physical servers (many of the tasks I'm doing have no analog for vm s) so I can't just pretend like I have 1000 servers, I literally have one in my dev env and I think 8 in the qa env
14:39 jberger I suppose I could make a test app that defines some other unrelated tasks but I really haven't had time
14:45 marty_ joined #mojo
14:48 cuechan joined #mojo
14:54 dod joined #mojo
15:05 mcsnolte joined #mojo
15:10 itaipu joined #mojo
15:10 umask001 joined #mojo
15:22 zivester joined #mojo
15:49 romel joined #mojo
15:49 mpapec joined #mojo
15:58 disputin joined #mojo
16:07 cuechan joined #mojo
16:11 irqq joined #mojo
16:13 sri mishanti1: so, i guess we missed the opportunity to get your designer to make something for minion?
16:31 perlpilot joined #mojo
17:03 cpan_mojo Mojolicious-Plugin-OAuth2-Server-0.29 by LEEJO https://metacpan.org/release/LEEJO/Mojolicious-Plugin-OAuth2-Server-0.29
17:21 sri hmm, i don't like sequences in postgres
17:21 sri if you do "select last_value from foo_id_seq" after creating a table it returns 1
17:22 sri if you do it after inserting the first row, it still returns 1
17:23 jberger that's weird
17:24 sri after inserting the second row, it returns 2
17:24 jberger I would have expected that to be null
17:24 jberger yeah it is the boundary behavior
17:24 PryMar56 joined #mojo
17:25 jberger is there some way to test that the sequence is "fresh"
17:25 jberger fresh is a bad word, new is probably better
17:25 sri the sequence value would almost work for getting the number of jobs ever enqueued
17:25 sri except for the starting value
17:25 Grinnz_ " It is an error to call lastval if nextval has not yet been called in the current session." according to the docs
17:25 sri Grinnz_: i don't care about sessions state
17:25 Grinnz_ i guess this isn't lastval() though
17:27 sri lastval(), currval()... are for the current session
17:27 sri what i want is the last number generated by the sequence across all sessions
17:28 sri and i would like it to say 0 or null if no number has been generated yet
17:35 haarg check is_called maybe?
17:37 Grinnz_ i'd be wary of that because sequences can be incremented in a transaction that's never committed
17:37 Grinnz_ that = using last_value as the total number
17:38 Grinnz_ i have the same thing happen with AUTO_INCREMENT in mysql
17:39 sri haarg: that did the trick! \o/
17:39 itaipu joined #mojo
17:39 sri Grinnz_: i know, and i'm ok with that
17:40 sri i'm calling it a rough estimate of how many jobs have ever been enqueued
17:40 Grinnz_ ok
17:45 sri and the official backend methods should not be doing anything that messes with sequences
17:45 sri so by default it should be rather accurate
17:45 Grinnz_ i have no idea if the sqlite rowid will be usable in the same way but i'm sure i can finagle some number out of it :P
17:46 sri :)
17:47 sri https://github.com/kraih/minion/commit/58912ac33beb70a65960a6018307a14ef23587ef
17:47 sri wonder what other stats one can squeeze out of postgres
17:53 batman joined #mojo
17:54 sri disk usage might be interesting
18:12 irqq joined #mojo
18:13 bpmedley joined #mojo
18:21 cuechan joined #mojo
18:47 dave joined #mojo
18:48 dod joined #mojo
18:52 sri anyway, overwatch time
19:05 mishanti1 sri: Not at all. I offered our help and stand by that. It is simply a matter of planning on our end.
19:11 mishanti1 joined #mojo
19:24 dave you ranked yet sri?
19:32 mishanti1 joined #mojo
19:33 Bean joined #mojo
19:34 mishanti1 joined #mojo
19:38 sri mishanti1: oh, that's great, maybe i'll make a basic admin ui in the next few days then
19:38 sri dave: the season ended right as i started playing ;p
19:39 mishanti1 sri: Would be if you could make something like that yes. Also, if the designer could check something out that did something (put tasks in queues) so he could get a feel of how it would be in real life use then that would be even better.
19:40 mishanti1 sri: As mentioned we are a little short on developer hours, so the less devs need to be involved the better.
19:41 sri makes sense, maybe we can make a little sqlite example
19:41 mishanti1 Somthing like that yes.
19:41 Grinnz_ season 2 next tuesday!
19:42 dave after the recent nerfs, some friends of mine have stopped playing (one was a genji player)
19:42 mishanti1 sri: Feel free to alert me whenever you have something and I'll check it out myself and see what it would take to get him started.
19:42 sri mishanti1: will do
19:42 Grinnz_ dave: heh, people love to overreact to nerfs and even buffs
19:42 Grinnz_ pros too
19:42 dave oh yeah
19:42 sri if anyone wants to team up, my battletag is still kraih#2445 :)
19:42 dave I get popcorn ;)
19:42 Grinnz_ watch the pros playing comp on PTR right now and genji is still perfectly fine
19:43 Grinnz_ his ult nerf was definitely needed
19:43 Grinnz_ the other two changes dont really make him weaker
19:43 perlpilot joined #mojo
19:44 dave I haven't played in a while myself, had to finish DOOM
19:44 sri i'm not even that bad anymore ;p
19:44 dave compared to those pros, I suck
19:44 Grinnz_ watching twitch streams and youtube leaves me little time to play too :P
19:45 dave well, 5 programming projects, 3 bands, and 1 girlfriend ... I'm surprised I had time to finish DOOM
19:45 dave I dont watch twitch much
19:45 Grinnz_ i still haven't started that, nor witcher 3
19:45 sri it's ridiculous how much more fun overwatch is in groups, even 2-3 people
19:45 dave oh I had to finish DOOM
19:45 Grinnz_ i heard it was pretty good
19:45 dave sri; indeed...I was just a pair of people
19:45 sri not having 5 pants on head types makes all the difference
19:46 dave DOOM is great :)
19:46 dave it's decently difficult and the demons are modern looking
19:46 dave plus the idea of meleeing to get more ammo drops is awesome, makes you feel like a demon killer
19:48 * sri has to hit level 100 now
19:48 dave pssh, I'm not even much over 60
19:48 Grinnz_ i wanted to hit 200 during the summer event to get lootboxes but i wasnt even close
19:48 Grinnz_ still havent hit it
19:49 Grinnz_ now im wondering if i should wait until the halloween event :P
19:49 dave that will take you a while
19:51 dave I still play tribes: ascend to keep in FPS shape so I should be gtg with overwatch
20:03 disputin joined #mojo
20:03 y1mmm joined #mojo
20:24 anton joined #mojo
20:28 disputin joined #mojo
20:35 margeas joined #mojo
20:41 itaipu joined #mojo
20:54 tchaves1 joined #mojo
22:42 meshl joined #mojo
23:23 cuechan joined #mojo

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