|
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 2 August 2017. |
|||
|
01:07
http_GK1wmSU joined
01:08
http_GK1wmSU left
01:52
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 | ||
|
03:28
camelia joined
|
|||
| ugexe | nine: i'm getting the feeling that everywhere $dist<source> (.meta<source> in my pr) could instead be referring to a CompUnit attribute. For instance, consider gist.github.com/ugexe/c31932c2c4b2...239f28d63c | 03:59 | |
| if we do this, that means .candidates could return CompUnit instead of Distribution. and this is actually good because it means CompUnit's attributes (name/auth/ver) keep from reaching into its :$distribution (which would parse json) *and* that $dist<source> becomes more like $comp-unit<source> which actually makes sense | 04:02 | ||
| the only problem then is you could not query directly for a distribution by name that did not have a provides by the same name | 04:06 | ||
| will have to try and see how it feels i suppose | 04:11 | ||
| nine: what do you think about a CompUnit $.version alias to $.ver? (otherwise we could change all the Distribution stuff to use version instead of ver) | 04:14 | ||
|
06:09
camelia joined
|
|||
| nine | ugexe: the not being able to query for a distribution by name bothers me a bit. Maybe we'd need an additional interface for querying distro info? Have a clean separation between those two. | 07:11 | |
| I imagine zef will be much more interested in distros than in compunits | 07:12 | ||
| ugexe: I'm not big on aliases unless they are required for backwards compatibility. | |||
| ugexe | well, you'd still be able to query for a distribution. my gripe is that it would be like it is *currently*, where `use DistName-Thats-Not-Module-Name` tries to load something instead of failing to find such a module | 15:10 | |
| a separate interface would work, but also seems like overkill | 15:11 | ||
| nine | Yep that's quite unpretty | ||
| ugexe | the non-clean part here is that in the case of searching for a *distribution* there is no single file to reference (like with .script referencing bin/foo, or .resolve referencing Some::Module) | 15:12 | |
| nine | But there's the meta data file? | 15:13 | |
| ugexe | well, thats true. but thats available through $dist.meta - an actual file need not exist | 15:14 | |
| nine | I mean when you are looking for a distribution, you'll be looking for its name | ||
| and expect to be able to get ethe meta data | |||
| Anyway I think slow progress on this is a good thing. We already have a rushed API. Let's get it right on the second try | 15:16 | ||
| ugexe | a problem with that is that every day in #perl6 there are reports of new perl6 installs not allowing people to update their zef due to .script always choosing the first repo | 15:18 | |
| mspo | zef is bundled with star, I thought | 15:21 | |
| ugexe | yea, but that doesn't mean its the first one that gets seen when you do `zef` command | 15:22 | |
| before .06 or .07 the bin wrapper would sort based on version. after it started resolving like a module (highest version *inside first repo that has any version*) | 15:23 | ||
| so if you have an old zef in ~/.perl6 you forgot about it'll probably still get chosen | 15:24 | ||
| mspo | yeah the whole loading and precomp and repo path thing is crazy complex | 15:26 | |
| ugexe | its not really so much complex as its hard to build a mental model in your head when everything is Jj8u98u92fji2fu38li2u8fi | 15:27 | |
| mspo | overall perl6 has been incredibly ... nitpicky? pedantic semantics? I don't know it's hard to describe | ||
| mst | I'm not sure it could be made any simpler without losing at least one of the important features it provides over simple approaches | 15:28 | |
|
15:31
perlpilot_ joined
|
|||
| mspo | I'm not necessarily complaining about anything | 15:31 | |
| just sharing some impressions | |||
| mst | oh, yes, the thing is, I agree, but what's really needed is to make it easier to get your head around how it works, I think | 15:32 | |
| ugexe | if we switched to puny encoding instead of sha1 it would be easier to reason with. we just have to con someone into doing it | ||
| jdv79 | ugexe: do you have nightmares about random strings? | 15:33 | |
| mspo | like I'm starting to put together packages and am not even sure if I should include the source (.pl/.pm) files somewhere! | ||
| or just the precomp/ dirs | |||
| and, of course, precomp/ dirs aren't deterministic so that's another complexity :) | 15:34 | ||
| which, thankfully, you guys are changing | 15:35 | ||
| there probably needs be to a Moar Runtime From The Ground Up paper or something | |||
|
20:57
perlpilot_ joined
|
|||