Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2014-12-19

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

All times shown according to UTC.

Time Nick Message
01:31 alnewkirk joined #metacpan
02:06 klapperl_ joined #metacpan
03:07 riveter rwstauner, around?
03:44 tomboh joined #metacpan
04:02 punter joined #metacpan
04:29 rwstauner riveter: what's up?
04:29 rwstauner oh, i see a pr just came in
04:43 * rwstauner &
05:03 jwang joined #metacpan
08:23 neilb joined #metacpan
09:06 neilb joined #metacpan
09:23 neilb joined #metacpan
11:33 kentnl joined #metacpan
12:39 andreeap joined #metacpan
13:31 andreeap joined #metacpan
13:34 andreeap joined #metacpan
14:01 neilb joined #metacpan
14:16 punter joined #metacpan
15:03 ranguard oalders++ # focus on the ES upgrades
15:04 vroom joined #metacpan
15:08 oalders ranguard: i learned that from you ;)
15:13 ranguard me++ :)
15:13 riveter watch out for overflows, ranguard
15:18 Ptolemarch joined #metacpan
15:19 riveter oalders, so this machine is completely separate from prod and I can do horrible unspeakable things to it without anyone screaming, yes?
15:21 oalders riveter: it's yours to abuse
15:21 riveter awesome
15:21 riveter I'm probably only getting a couple hours in today
15:22 ranguard riveter: it is part of a ES cluster - but that's not in production
15:23 ranguard the lw-mc-* machine's arn't in use yet, but will become production when we can create a new api using ES 1.4
15:23 oalders but if you can make it sing, it'll one day go into production
15:23 ranguard exactly - so we want to see how the cluster stuff actually operates and configure anything with that in mind
15:25 riveter so, es doesn't seem to be listening on 9200 on this machine
15:25 ranguard riveter: will be on the public IP, not localhost (because of clustering)
15:25 riveter oh right
15:25 riveter doi
15:26 ranguard but the firewall prevents public access
15:26 ranguard so you need to ssh tunnel in if you want to see stuff...
15:27 ranguard riveter: ssh -L localhost:9091:lw-mc-01:9200 lw-mc-01.metacpan.org
15:27 ranguard oops
15:27 ranguard make that lw-mc-03
15:28 ranguard then on your desktop http://localhost:9091/_plugin/kopf/ or http://localhost:9091/_plugin/marvel/
15:28 ranguard you'll see some logstash indexes I've been playing with
15:30 ranguard oh, or just http://localhost:9091/ if you want to make `curl` requests against it from your desktop
15:30 riveter hoop, there it is
15:30 riveter yeah, I know what port forwarding is :)
15:30 ranguard :)
15:31 riveter hm
15:31 riveter so, I was just trying to get the code working with 1.0
15:31 riveter there are more breaking changes in 1.4
15:32 riveter and esx::model, which I think we use somewhat heavily, does not even have 1.0 support released yet
15:32 ranguard we should just go straight to latest
15:32 riveter ok
15:32 riveter so this job might involve forking esx::model and updating it
15:33 * ranguard should actually upgrade to 1.4.2
15:33 riveter if there's breaking changes between 1.4.0 and 1.4.2 I'm going to take up scuba diving instead
15:34 ranguard if you fork https://github.com/monken/p5-elasticsearch-model I'm sure we can then get mo to do a release once ready
15:34 dipsy [ monken/p5-elasticsearch-model · GitHub ]
15:34 riveter mateu was saying they use something different at his work
15:35 riveter good morning dipsy
15:35 * ranguard steps back and points to the sysadmin hat on his head, not the coder one :)
15:35 mateu Elastic::Model, but only lightly, mostly just Search::Elasticsearch directly (formerly Elastic(s|S)earch)
15:35 riveter :P
15:36 ranguard we're using S::Es directly at work as well
15:36 riveter oalders, any feedback on whether I should look at forking ESX::Model or look at migrating to Elastic::Model
15:37 ranguard right, time to pickup $family and go for dinner - enjoy :)
15:37 riveter I was kind of optimistic that the unreleased 0.1.8 would work as is with 1.0
15:37 riveter but it certainly won't with 1.4
15:37 riveter ranguard, bon appetit :)
15:37 mateu last time I spoke w/ Gormely the future of Elastic::Model is uncertain
15:37 riveter great
15:38 mateu the changes required on the perl side from 1.0 to 1.4 are not large iirc
15:38 riveter so, we might as well stick with esx::model
15:38 mateu i think so
15:39 oalders we do depend a fair bit on esx::model
15:39 riveter ok, so maybe I should actually leave the api code alone for now
15:39 riveter and see what needs to be done on esx::model to get it up to speed
15:39 oalders that's a good plan.  if that turns out to be a huge problem, we can re-evaluate
15:40 riveter ok, I'll take a quick look today
15:40 riveter I'm probably only doing a couple hours, dunno if you saw my note
15:40 oalders i did see your note. it's not a problem :)
15:41 oalders i haven't looked at Elastic::Model, so i don't know how it compares
15:41 riveter cool, I am keeping track & will make up the time in jan
15:41 oalders :)
15:41 riveter well, if they're both unmaintained there's not much point in switching
15:42 riveter so mateu, probably not much progress until monday
15:42 riveter ok, lemme fork esx::model & see what I can see
15:42 mateu riveter: no worries
16:34 * gtodd throws out a rhetorical question for discussion ...
16:35 gtodd I wonder if github issues or RT bug reporting could be gamified and "reputationally vetted" whether the quality of bug reports and general S/N would improve ... and specific to #metacpan ... could metacpan have an even better set of "reputation" features
16:44 Seveas left #metacpan
16:52 gtodd if it were possible to grab data from other sources, that would broaden the set of factors going in to the calculation of reputation (make it more "meta") which might in turn make the ratings given by high reputation users/devs  of various modules, features, pragmas, (other devs ?!)  ... useful
16:53 gtodd not sure what "metric" would come out at the end or if it would be helpful in any way :-|   it's probably been discussed
16:53 gtodd already
17:08 vroom joined #metacpan
17:38 gtodd on the "supply side", each distribution's "home page" on metacpan already has a host of those kind of features (which you can sort of see at a glance) so it's more a question of if/how a similar concept could apply to the "user masses" to "engage and encourage"  and/or/if that would be useful
17:40 neilb joined #metacpan
18:31 Ptolemarch joined #metacpan
18:42 ranguard gtodd: "reputation" features? - for example?
18:47 oalders gtodd: is this a wishlist item or are you looking for a quality way to spend your spare time?
18:48 oalders i've done a bit of work on something vaguely related https://github.com/oalders/github-mergevelocity
18:48 dipsy [ oalders/github-mergevelocity · GitHub ]
19:25 Ptolemarch joined #metacpan
19:59 Ptolemarch joined #metacpan
20:02 ranguard riveter: https://github.com/monken/p5-elasticsearch-model/commit/c39c4765c1dc45771c5790298569ff631cc3989d - looks like mo has already dropped ES < 1 support and switched to the recommended  Search::Elasticsearch - he tested with ES 1.1.1
20:02 dipsy [ require 1.0.0 for testing · c39c476 · monken/p5-elasticsearch-model · GitHub ]
20:04 ranguard oh, maybe Search::Elasticsearch 1.11 it's not clear :)
20:09 Ptolemarch joined #metacpan
20:23 root joined #metacpan
22:12 Ptolemarch joined #metacpan
22:16 gtodd oalders: not sure I'd be much help but I could always try :)  merge-velocity is pretty neat  ... the advent.pod describes it action?
22:17 oalders gtodd: it was supposed to be an advent article, but i got frustrated with it.  but that's should mostly be correct
22:17 oalders *that
22:17 oalders gtodd: if you're interested in working on this kind of thing, it's not too hard to get started
22:18 oalders i never got around to creating a proper algorithm
22:18 oalders neilb has some good ideas
22:18 gtodd great
22:19 neilb it’s a fun idea — happy to bounce round more thoughts
22:20 oalders :)
22:22 gtodd ranguard: "reputation" might be the wrong word for it but e.g. it's not necessary to have "points and badges" for perl distributions since it's easy to glean from metacpan that developer or distribution is experienced, diligent (bug fixes and releases), their code is used in other projects (dependences) etc. ... with a little more work you can see if a dist/module figures in one of the curated collections (Tasks? some
22:24 gtodd and after all the real work happens (the mere matter of implementing all the modules) there's all this great automatically generated useful information
22:26 mst bah
22:26 mst now I want to provide a Task::MSTPAN or something
22:26 mst except that makes no sense because it's all varying choices
22:26 mst don't mind me
22:26 gtodd heheh
22:29 gtodd of course if you write something useful and good people will never let your contribution  go unpunished  :)
22:39 mattisbusycom joined #metacpan

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