Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-09-16

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

All times shown according to UTC.

Time Nick Message
00:25 rjbs Just got the best suggestion ever at work.
00:25 rjbs https://metacpan.org/about - disappointed this page isn't https://metacpan.org/meta
00:25 rjbs :-)
00:28 Grinnz so metp even this page
01:25 vanstyn joined #metacpan
04:39 Tempesta joined #metacpan
05:52 Tempesta joined #metacpan
07:07 neilb joined #metacpan
07:29 oiami joined #metacpan
07:57 osfabibisi joined #metacpan
08:12 neilb joined #metacpan
10:01 osfabibisi joined #metacpan
10:15 osfabibisi joined #metacpan
10:23 karjala bug?
10:23 karjala look at this page https://metacpan.org/pod/distribution/perl/pod/perl.pod
10:23 karjala and click on the down-arrow that shows all the versions
10:23 karjala latest version is v2.3 (just released by someone)
10:24 karjala is this a bug?
10:57 osfabibisi joined #metacpan
11:30 mickey seems like a bad upload to pause: https://v1.metacpan.org/release/GRABZIT/perl.2.3.0 , which was later uploaded again with a corrected distribution name https://v1.metacpan.org/release/GrabzItClient
11:37 mickey it's on the list because of the 'perl' distribution name
11:44 neilb joined #metacpan
11:46 osfabibisi joined #metacpan
11:55 neilb_ joined #metacpan
11:55 osfabibi_ joined #metacpan
12:12 neilb joined #metacpan
12:24 neilb joined #metacpan
12:44 neilb joined #metacpan
13:07 oalders i guess we should special case that
13:21 vanstyn joined #metacpan
13:37 osfabibisi joined #metacpan
13:45 neilb joined #metacpan
13:53 vanstyn joined #metacpan
14:11 mickey oalders: i wonder how it got 'authorized' ... i don't see an indication in 06perms for 'perl' as dist name, that may be the reason
14:12 mickey seems like something that should be blocked on PAUSE level
14:12 mst mickey: I don't believe PAUSE will have indexed it
14:12 mickey but we can also add some code to guard it
14:13 mst so it already is as blocked as it's going to get
14:13 mst I'm not sure what you want PAUSE to do that it isn't already?
14:14 mickey well, i'm just surprised it takes perl.X file names in upload
14:14 mickey with no perm. check
14:14 mst huh?
14:15 mst PAUSE allows any file to be uploaded. what it's selective about is indexing.
14:15 mst it *does* check permissions. that's why it didn't index the dist.
14:15 mickey k
14:16 mickey but we did... and i it does pass our permissions check
14:16 mickey which is 06perms
14:16 mickey sec, verifying :)
14:17 mst I think it failed to fail your permissions check because permissions for 'perl' is the pumpking bit system in PAUSE which is separate to the main perms stuff
14:17 BinGOs only these: http://pause.perl.org/pause/query?ACTION=who_pumpkin can upload and have indexed a dist called 'perl'
14:17 BinGOs bazing.
14:19 mickey yeah, we can special-case it the same way i guess
14:38 osfabibisi joined #metacpan
15:20 oiami joined #metacpan
15:24 Grinnz 'perl' being a distribution is a smorgasborg of special cases
15:25 mst Grinnz: s/smorgasborg/hopper.jpg/
16:09 ether if that permissions check was in a CPAN module, then PAUSE and metacpan could share the same code :)
16:09 ether rather than one having to emulate the other
16:09 ether based on smoke signals and half-forgotten memories and old wives' tales
16:09 ether the only other alternative is having a formal specification document, and no one wants that! :D :p
17:13 oiami1 joined #metacpan
17:40 ranguard joined #metacpan
18:11 oalders i'm all for using common code :)
18:13 Grinnz eww common core
18:13 Grinnz oh... wait
18:38 mickey oalders: i see that we do have a special case for it. we skip the permissions check for 'perl' because we trust PAUSE not to 'index' it without permissions... but i think in this case (please do correct me) we are indexing based on file upload and not by PAUSE indexing so we ended up with this case
18:39 oalders mickey: without looking at the code, that sounds right
18:39 oalders we already need to rework how we handle perl indexing *anyway*, so we can add this to the list
18:40 oalders ie we want to stop having to re-index perl manually
18:42 mickey https://github.com/metacpan/metacpan-api/blob/master/lib/MetaCPAN/Document/File.pm#L916
18:42 mickey yes, that's totally an issue to tackle next hackathon :)
18:47 oalders mickey: given today's discussion, i guess that code comment could be amended :)
18:48 oalders also looks like the place we'd special case this particular upload
19:12 neilb joined #metacpan
19:38 oiami joined #metacpan
19:44 ranguard mickey: is api configured to connect to all 3 ES nodes, or just localhost?
19:44 ranguard oalders: ^^ ?
19:45 oalders that should be easy to check. hang on
19:45 ranguard thanks
19:53 oalders ranguard: https://metacpan.org/source/OALDERS/ElasticSearchX-Model-1.0.2/lib/ElasticSearchX/Model/Role.pm#L20
19:55 ranguard oalders: cheers, hmm, wonder what format $ENV{ES} takes!
19:56 oalders ranguard: also https://github.com/metacpan/metacpan-api/blob/master/lib/MetaCPAN/Server/Model/CPAN.pm#L25
19:56 ranguard I'll use Fastly to switch over for now, but really we should connect to all ES nodes from the API, so we don't have to worry about a node being down
19:57 * ranguard starts upgrade of ES.. prepare for wobbles :)
19:57 oalders ranguard: also https://github.com/metacpan/metacpan-api/blob/master/lib/Catalyst/Plugin/Session/Store/ElasticSearch.pm#L10
19:57 oalders AFAICT we're doing this in several places
19:58 ranguard heh - something for the MC::Hack list please :)
19:59 oalders indeed
19:59 oalders re: $ENV, look at https://metacpan.org/pod/Search::Elasticsearch#TEST-SUITE
20:01 ranguard doesn't show how to connect to multiple nodes is what I meant
20:02 ranguard but I don't need this for now :)
20:03 oalders mutli-nodes are in the SYNOPSIS here https://metacpan.org/pod/Search::Elasticsearch
20:03 oalders not sure how to do that via $ENV
20:03 oalders there's now a meta::hack label we can use
20:06 ranguard sweet :)
20:09 ranguard 1 node down and all ok - this cluster thing is good :)
20:17 oalders :)
20:20 ranguard fastly++ # fastapi now using lw-mc-02 (whilst I upgrade 03)
20:20 oalders nice
20:22 sivoais joined #metacpan
20:25 ranguard ES upgraded to 2.4.0 (latest) - and fastapi now back to lw-mc-03
20:25 ranguard 0 downtime :)
20:25 ranguard it's almost like we're a professional setup
20:51 * ranguard adds some disk space to the CPAN volume on lw-mc- .. we were at 90%
21:07 y1mmm joined #metacpan
21:12 oalders ranguard++
21:27 jnbek joined #metacpan
21:31 sivoais joined #metacpan
22:15 jnbek joined #metacpan
22:18 jnbek joined #metacpan
23:06 jnbek joined #metacpan

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