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 28 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
06:50 ugexe joined 07:04 domidumont joined 07:09 domidumont joined 08:14 domidumont joined 08:48 domidumont joined 14:02 perlpilot joined 17:24 sufrostico joined 17:28 sufrosti1o joined 17:36 TimToady joined
ugexe the one part of Distributions i do not have a plan for is `files`. Specifically those that are not listed in the META6, such as bin/ and hooks/ (if/when) 17:44
all i can think of is `files` as an optional field. it would act as it does now when generated for an install; that is, if its not filled out then its reasonably to be filled in automatically 17:46
but if your META6 is pulled from github.com you may not have access to the actual files it provides, or a way to find them besides downloaded the entire thing
so i propose that a user be allowed to fill these in if they wish. a role with default behavior would probbly be added to each distribution though, so paths can do a recursive ls, git can do git ls-files, etc 17:47
in this way, a user is essentially allowed to explicitly create their MANIFEST 17:48
this goes well with: when installed META6 (whats in dists/) have info added, they use a key/value instead of a string for the leaf. ::FileSystem could then be made to understand the normal META6 json format, whereas ::Installation takes the same format except leafs like { "MyModule" : {"lib/MyModule.pm6", "<sha1>"} } instead of the current method of inverting the structure like 17:52
.<pm><MyModule>[0]<file>
as `files` could act in the same way for ::FileSystem vs ::Installation 17:53
i started thinking about this when llfourn discovered `%?META` in one of the design docs. i think that would be a cool thing to have 17:58
18:33 Kassandry joined 18:55 domidumont joined
mst ugexe: perhaps the answer for .perl is to ensure that each repo object has a unique name, and have a %repositories somewhere, and .perl can just reference %repositories{'My name is bob'} or whatever 19:25
ugexe mst: irclog.perlgeek.de/perl6-toolchain/...i_12113587 we discussed that too :) 19:27
mst ugexe: erm, yes, I was riffing on top of that discussion, having read it already 19:29
ugexe so you can already do that, there just isnt a mechanism to get a list of the available names
mst also site/vendor/perl are standard names
my point was that when you're stacking random things on top with -I etc. those should have unique names *too*
ugexe their spec: inst#xxx, file#xxx, theoretically github#xxx 19:30
mst that's not sufficient though
X atop Y atop Z is not the same thing as X atop Z
and the foo#bar syntax only specifies X as yet 19:31
ugexe a name for the entire REPO chain used to build? is a naive example like a name for "foo#x, foo#y, foo#z" would have a different name fhan "foo#x, foo#z"? 19:37
19:44 sufrostico joined, sufrosti1o joined
nine_ Please all have a look at github.com/rakudo/rakudo/commit/12...e3b7d6de11 22:11
mst: almost there ^^^
lizmat ugexe: where you still working on the Signature / punning madness ? 22:13
mst nine_: YEEEEEESSSSSSSS 22:18
nine_: damn, I remember going over that with you and assuming I'd probably have to write it due to it being batshit insane
nine_: I am extremely glad to discover that you're also sufficiently crazy, given how useless I've been 22:19
nine++ for 1..Inf;
22:47 xdg joined
ugexe lizmat: from yesterday re: bufs? after your PRs yesterday the test suite went from 19s to 16.5s (2015.12 runs it in 12s) 22:51
its a lot of IO::Socket.recv(1, :bin)
lizmat oops, I guess I meant skids: gist.github.com/skids/b7db42fc7d029e5c0f9e
23:20 sufrostico joined, sufrosti1o joined