|
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 5 February 2016. |
|||
|
07:17
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 | ||
| nine | ugexe: I actually think we should just check for existence and die loudly (yet with an explanation) if the META6.json is invalid, i.e. is a directory, has no contents or just has invalid contents. | 09:13 | |
| When something called META6.json is there, we have to assume that the user meant for it to be used, so informing her of a problem would be the prudent thing to do. | 09:23 | ||
| And as we've seen, we're not good at anticipating all the valid ways of providing a META6.json. For example it could even be a named pipe that doesn't have a size, but will supply us with valid meta data. | 09:24 | ||
| Sounds a bit contrived, but who knows what the crazy people in the Perl community come up with? Might just be some test setup. Even tests for CURI itself. | 09:27 | ||
|
09:41
cognominal joined
|
|||
| ugexe | sure, but this issue is related specifically to symlinks (which serve the same purpose, and work fine on linux and osx in this regard) | 16:16 | |
| point being: it should act the same on every OS due to the "automatically loads your META6.json" nature of ::FileSystem | 16:17 | ||
| on windows a symlink named META6.json pointing to a valid META.info that is cloned from git will be a valid file existing as META6.json but only contain the string "META.info" | 16:18 | ||
| dying loudly is fine (what happens currently). differing behavior by OS is not | 16:20 | ||
| let me be clear though: i am all for ramping up the strictness on all this stuff | 16:32 | ||
| ignore meta.info all together, make required meta fields actually required, etc | 16:33 | ||
| and make auth a specific required field to make it easier to search for (instead of having to search auth, author, and authority to find which one it came from) | 16:37 | ||
| for the sake of parsing a URN (like cpan:AUTHOR:Module-XXX:1.0) it also helps if auth requires a delimiter (':'), or instead have 2 required fields (content storage: cpan, owner: AUTHOR) | 16:41 | ||
| this all boils down to being able to know what meta fields to search for given only a modules full identity | 16:47 | ||
|
16:48
sufrostico joined
17:02
Cheery joined
|
|||
| nine | I think we share a pretty similar idea of how the goal looks like :) It's just the way there that's foggy for me | 17:13 | |
| And by that I mean a deployable way mostly. The coding seems pretty straight forward | 17:14 | ||
|
17:17
FROGGS joined
17:43
sevvie joined
19:40
hankache joined
|
|||