Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2015-12-17

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

All times shown according to UTC.

Time Nick Message
00:00 jdv79 ranguard: yes, META.list is used to generate the eco repo list
00:04 jdv79 that might have just been to flag some temporary erroneous dups in the eco at the time
00:11 jdv79 there isn't great (any?) error detection or validation of that data so every once in a while it gets mucked up
00:13 jdv79 anyone know if pause would be ok the same distvname (I think its called) for multiple users?
00:13 jdv79 s//with/
02:11 punter joined #metacpan
05:37 oiami joined #metacpan
07:53 oiami joined #metacpan
07:55 neilb joined #metacpan
08:04 neilb_ joined #metacpan
08:28 ribasushi joined #metacpan
09:01 Relequestual joined #metacpan
09:10 melo joined #metacpan
09:16 neilb joined #metacpan
09:30 Ralesk joined #metacpan
10:34 oiami joined #metacpan
10:40 Peter joined #metacpan
15:38 oalders jdv79: PAUSE cares about filenames. there's nothing stopping multiple users from uploading things with the same names
15:51 mst jdv79: uniqueness on PAUSE is at the AUTHOR/File.tar.gz level
15:51 mst jdv79: AUTHOR1/SomeFileName.tar.gz and AUTHOR2/SomeFIleName.tar.gz are totally different things
16:05 oalders there we go.  that was expressed much more eloquently
16:38 jdv79 that's what i thought from looking at the pause code.  i've just never seen it concretely so wasn't 100% sure.
16:40 jdv79 oalders: well, i'm not sure thats entirely correct since the first table in pause has a unique key on user and filename
16:40 jdv79 which again is what mst said.  thanks
17:07 neilb joined #metacpan
17:47 ranguard jdv79: any reason you added 'v' onto version numbers?
17:47 ranguard evening btw :)
17:57 ranguard jdv79: do you also have a feel for if people actually increase their version numbers in git repo?
18:08 oiami joined #metacpan
18:24 jdv79 i believe the v is spec (s22). i dont remember offhand if its optional.
18:25 jdv79 versioning is largely not done atm
18:25 jdv79 we only recently (weeks) gained ver support in use...
18:26 jdv79 evening:)
18:27 jdv79 right now installing from gh uses head
18:27 jdv79 the installers dont support versioning last i knew
18:28 jdv79 well, master i mean.
18:30 jdv79 maybe use the date like dns for better artificial vers?
19:03 ranguard jdv79: so v0.0.YYYYMMDDhhmmss - then if someone does want to release it them selves we've not released a version higher than theirs?
19:05 ranguard mst: ^^ is that sane?
19:21 ranguard jdv79: do you have a running mockup of mc6?
19:36 y1mmm joined #metacpan
19:48 neilb_ joined #metacpan
20:15 jdv79 ranguard: idk.  i was thinking about YYYY.MM.DD.HH.mm.  didn't think about it too hard yet.
20:17 jdv79 yeah, maybe yours is better.  the author is part of the auth which is part of the longname of the ns so conflicts are less of a concern
20:18 jdv79 though it may become an issue depending on how the installers resolve ambiguities
20:19 jdv79 http://hack.p6c.org:5001 - doesn't have all te changes but most iirc
20:20 jdv79 http://hack.p6c.org:5001/author/JDV/releases for instance
20:20 ranguard jdv79: cool, just want to mention it in docs
20:21 jdv79 do you have a pause id yet?
20:21 ranguard https://metacpan.org/author/PSIXDISTS :)
20:21 jdv79 i'd be interested in seeing how same ver and/or same dist work out.  haven't tried yet
20:22 jdv79 oh. local?
20:23 * ranguard is getting there, got the .tar.gz's being created
20:23 ranguard local?
20:23 jdv79 oh, i guess cause you haven't uploaded anything there is no cpan dir yet
20:23 jdv79 nice
20:25 ranguard jdv79: https://github.com/perl6modules/perl6-module-uploader/blob/master/publish_to_cpan.pl
20:25 * ranguard stops for dinner
20:26 jdv79 cool!  yeah, late to the pub myself.
20:26 jdv79 til tomorrow
21:32 ranguard great, so http://design.perl6.org/S22.html#META6.json doesn't specify how to set a repo URL
21:43 ranguard bah, seriously szabgab - https://github.com/szabgab/Perl6-Maven/blob/main/ <- main? - why not `master`
21:57 Khisanth joined #metacpan
21:59 Khisanth joined #metacpan
22:02 jdv79 ranguard++
22:03 jdv79 ranguard: please explain wrt repo url
22:10 ranguard jdv79: some of the META.info files don't have a source-url field at all
22:12 ranguard and S22 spec doesn't even mention how one should do it
22:18 grantm joined #metacpan
22:27 ether that document doesn't read like a spec; it's loose and informal enough that it's more of a guideline.
22:27 ether a spec would clearly say what things are strings, what things are listrefs, what restrictions exist on the string contents..
22:27 * ranguard agrees
22:27 ether and it seems to have forgotten some of the lessons we've learned in perl5's meta spec
22:29 ranguard jdv79: Please review: https://github.com/perl6modules/perl6-module-uploader/blob/master/publish_to_cpan.pl
22:29 ranguard I think it's done, but need to do more testing before running in anger :)
22:31 * ranguard heads to bed
22:33 ether anger++
22:33 vroom joined #metacpan
22:46 jdv79 so:)  meta.infi is not spec.  meta6.json is.
22:47 jdv79 meta.info is what panda understands "off spec"
22:47 jdv79 ether: the p6 "specs" are more like guidelines...
22:47 jdv79 aka design docs
22:48 jdv79 in short they will be firmed up after we implement something well enoygh
22:50 jdv79 ranguard: its a differnt field for meta6.json.  its a support.source.
22:51 jdv79 this shenanigans is implied/dealt with in my hacky script
22:52 jdv79 anyway, morrow
23:11 neilb joined #metacpan
23:58 Khisanth joined #metacpan

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