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 15 July 2016.
ugexe nine: ignoring complexity, is there any reason why the repo-chain can't be a repo-tree (possibly as a circular doubly linked list)? 01:41
as a means of either not having to disable precomp when the repo-chain changes, or possibly optimizing the best branch of dependencies to precompile if it has multiple branches but can only precomp one 01:46
01:48 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
jdv79 sounds confusing 01:55
ugexe it will fit right in then 01:57
02:43 Madcap^Jake^ joined 06:30 domidumont joined 06:35 domidumont joined 07:21 domidumont joined 08:19 FROGGS joined
nine ugexe: how would we choose a branch if more than one are available? 08:44
15:12 camelia joined 19:06 Madcap^Jake^ joined
ugexe No idea. A naive approach might be to chose the branch that has the most constraints/dependencies already precompiled. A real solution probably involves a type of running backtracking tree decomposition that would make any compsci guru blush 19:48
mst I think that would be unfortunately complicated
I'm unsure why changing the chain needs to disable precomp though 19:49
surely a precomp store is a wrapper around a complete chain
and you'd just need a new store
I mean, a CUR doesn't own its precomp data, the precomp data owns a CUR chain, no?
nine mst: use Test; use lib 't/spec/packages'; use Test::Util; # which uses Test again 20:15
mst: the first Test is from lib/ while Test::Util was compiled against another precomped Test
mst: but I could fix that specific case :)
mst mmm 20:16
I thought you could load more than one version of Test inside different things? (not that that would likely be DWIm in this case)
nine more versions of a module but not more different precompiled files of the same version 20:21
mst oh, of course, Test::Util was precomped against the Test in its normal repo chain, and now resolves the one from lib/ 20:22
yeah, I see
nine: sorry, my brain had inverted the problem 20:31
21:01 sjn_ joined 23:52 b2gills joined