|
Fire is step THREE! | github.com/perl6/toolchain-bikeshed | Channel logs: irclog.perlgeek.de/perl6-toolchain/today | useful prior art: metacpan.org/pod/CPAN::Meta::Spec Set by moderator on 7 February 2016. |
|||
|
02:48
ilbot3 joined
|
|||
| moderator | Fire is step THREE! | github.com/perl6/toolchain-bikeshed | Channel logs: irclog.perlgeek.de/perl6-toolchain/today | useful prior art: metacpan.org/pod/CPAN::Meta::Spec | ||
|
02:59
FROGGS_ joined
04:21
MadcapJake joined
04:24
MadcapJake joined
07:11
FROGGS joined
07:33
domidumont joined
07:38
domidumont joined
10:41
cognominal joined
12:35
sufrostico joined
12:43
sufrostico joined
14:10
leont joined
14:11
japhb joined
|
|||
| ugexe | right. i hope we can find a way to do it soon though... the longer no course of action is decided the more modules will rely on auth's one-of-many-possible-values | 15:09 | |
| we have an advantage right now in that no one really uses auth in their code yet | |||
| i like the concept of auth being 2 distinct parts though. it would allow one to request a module from a specific content storage (like :auth('cpan:*')) | 15:14 | ||
| lizmat | authority and storage should not be mixed, feels wrong to me | 15:15 | |
| storage could be a clone / slave / mirror | 15:16 | ||
| ugexe | storage is supposed to be its original location | ||
| (in my concept of it) | |||
| it would not mean "download it from cpan" | 15:17 | ||
| so cpan might host something with auth: "github:ugexe" | 15:19 | ||
| the recomendation manager can then determine any mirrors, wherever they may point | 15:20 | ||
| the key here is no data has to change in the distribution. so the same dist on 2 different services does not require modifying the meta6.json file (aka changing its checksum) | 15:21 | ||
| lizmat | ok, if that's your definition of storage :-) | ||
| ugexe: yes | |||
| ugexe | cpan is a storage yes. the original storage | ||
|
15:25
prammer joined
|
|||
| ugexe | "Please note that this is not an authority, merely an indication of the location where the distribution for that owner was obtained" # this is what leads me to this thinking | 15:26 | |
| maybe i misinterpret it | 15:27 | ||
| lizmat | ugexe: that part of the spec was largely handwavy | 15:29 | |
| ugexe | it makes sense to me though. and it makes a URN easier | 15:31 | |
| think about it from a "i have to type this on the command line to request it in the first place" way | 15:32 | ||
|
15:33
japhb joined
|
|||
| ugexe | email address and urls are nice too, as they are unique. author name not so much ("John Smith", not so unique) | 15:43 | |
| lizmat | the authors section was always intended as additional, non-authoritative information | 15:45 | |
| ugexe | yeah i know... s22 mentions `owner` as filling the role author serves currently | 15:46 | |
| even if changes are decided on any cant be made until another release they still help with the metacpan effort (which is what got me to bring this up in the first place) | 15:56 | ||
| s/any cant/and cant/ | |||
|
16:25
sufrostico joined
16:38
hankache joined
16:45
leont joined
16:54
sufrostico joined
17:24
FROGGS joined
17:45
autarch joined
18:08
japhb joined
|
|||
| jdv79 | i thought auth indicated the actual entity such that cpan:jdv is jdv in the pause db and github:jdv is github user jdv. | 18:34 | |
| not necessarily related to storage though i get why its confusing that is seems to imply such | 18:35 | ||
|
18:39
japhb joined
18:40
kmel joined
|
|||
| jdv79 | for instance, i could just publish the same dist on github, cpan, and some other "content storage" with an auth of cpan:jdv regardless of "original storage", right? I'm not sure what original storage means really. | 18:42 | |
| flussence | .oO( I've half a mind to just start putting an email address there ) |
18:49 | |
| ugexe | that can make it a pita to request from cli | 18:50 | |
|
18:58
japhb joined
19:20
domidumont joined
19:40
cognominal joined
20:07
Kassandry joined
|
|||
| nine | What's hard about email addresses on the cli? | 20:10 | |
|
20:15
hankache joined
20:23
perlpilot joined
|
|||
| ugexe | not that it doesnt apply to other ids, but the allowable characters lend themselves more to needing to quote the string than most user identification | 20:25 | |
| jdv79 | if it was a URN we could allow the email scheme. but on the other hand i don't think we can use the cpan and github scheme officially, right? | 20:29 | |
| oh i meant URI. ugexe: why are you talking about URNs? | 20:31 | ||
| ugexe | because cpan:JDV:Foo-Bar:1.0 is a URN, not a URI | ||
| it could be interpretted as one, but i mean this distribution could be downloaded from various locations | 20:32 | ||
| jdv79 | a URN isa URI though:) | 20:33 | |
| URN feels very 1999 | |||
| but sure | |||
|
20:34
prammer joined
|
|||
| jdv79 | also, iirc, a urn is indicated by a urn scheme so its more like urn:cpan:JDV:Foo-Bar:1.0 | 20:38 | |
| ugexe | sure, but i think we can forgo the urn: | ||
| my only point here is that it points at a resource without implying how to get it (urn) | 20:43 | ||
| so i guess i mean url instead of uri before | |||
| jdv79 | i wonder why there isn't an email URN ns. it seems emails are only officially used an an URI via the mailto scheme which is kinda lame. | 20:46 | |
| then again i'm not sure this level of formalism is worth the effort at this point. | 20:47 | ||
|
20:54
prammer joined
20:57
prammer joined
21:02
prammer joined
21:19
prammer joined
|
|||