00:02 Altai-man_ joined 00:05 sena_kun left 02:03 sena_kun joined 02:05 Altai-man_ left 03:09 Xliff left 03:54 HarmtH joined 04:02 Altai-man_ joined 04:05 sena_kun left 06:04 sena_kun joined 06:05 Altai-man_ left 07:53 cognominal joined 07:55 cognominal left 08:02 Altai-man_ joined 08:05 sena_kun left 08:35 patrickb joined 08:45 patrickb left
lizmat Files=1307, Tests=113021, 217 wallclock secs (28.89 usr 8.37 sys + 3029.08 cusr 289.21 csys = 3355.55 CPU) 09:09
09:29 leont_ joined 10:03 sena_kun joined 10:06 Altai-man_ left 12:00 vrurg left, vrurg joined 12:03 Altai-man_ joined 12:05 sena_kun left, vrurg left 12:33 softmoth joined 12:34 vrurg joined 12:35 vrurg left 12:36 vrurg joined 13:15 lucasb joined 13:19 softmoth left 14:03 sena_kun joined 14:05 Altai-man_ left
nine Is there a reliable way during compilation to get a package's full name, i.e. Cool::Cat::Goes instead of just Goes? 14:24
duh....it's the one argument I didn't look closer at, the to be installed object itself which in the case of install_package_symbol_unchecked seems to be a package anyway and gets its name via $longname.fully_qualified_with($target_package) 14:37
14:44 dogbert17 left 16:02 Altai-man_ joined 16:05 sena_kun left
ShimmerFairy m: say $*VM.version 16:25
camelia v2020.06.14.g.5832.b.2628
ShimmerFairy I was just thinking, it'd be nice if, in a future language version, Version could possess a trailing "uncomparable" component for things like git hashes, where Raku can't be expect to know how to compare that component in two version numbers, but some other software does know. 16:26
(for cmp I'd say it pretends the uncomparable component doesn't exist, with the justification that if that component is the only thing differing, the two things being described are most likely very similar) 16:28
nine Oh boy...this complicates things: require Cool::Cat; require Cool::Cat::Goes::Splat; During compilation of the test script we create the full package chain for Cool::Cat::Goes::Splat. Then we precompile Cool::Cat during which it picks up Goes from the pre-existing stub. Thus we end up with a reference to the scripts SC in Cool::Cat 16:42
ShimmerFairy I suppose moving package precomp up to the point of a "require" or similar statement is out of the question? 16:49
nine Yes, as especially for a require we don't even know which file to precompile at the point we compile the require statement. The repository chain could be changed between compilation and running that statement. 16:51
Taking about 17 steps back I wonder why during precompilation it even sees the existing Cool::Cat package when it has its own empty GLOBAL package 17:03
I mean the whole reason for having nested packages and eschewing the flat global view of Perl 5's our scope is to be able to have independent roots for the package tree after all. 17:06
If we have to deal with all the compliations this brings, shouldn't we actually get the benefits as well?
ShimmerFairy Yeah, I can't think of any good reason why Cool::Cat would even want to know about a separated Cool::Cat::Goes::Splat (or else it would've brought that package in itself) 17:07
And considering all you'd be able to know is that there is a thing with that name in the world, it's not like you'd even get _that much_ info out of the reference. 17:08
18:03 sena_kun joined 18:05 Altai-man_ left 19:14 rypervenche left 19:17 rypervenche joined, softmoth joined
nine Ah, of course. This example is so insidious because the Cool in Cool::Cat::Goes::Splat comes from the setting. And the setting is a very global thing as well. 19:42
19:46 patrickb joined 20:03 Altai-man_ joined 20:05 sena_kun left 20:16 patrickb left 20:44 patrickb joined
Altai-man_ releasable6, status 20:46
releasable6 Altai-man_, Next release in ≈6 days and ≈22 hours. There are no known blockers. Changelog for this release was not started yet
Altai-man_, Details: gist.github.com/00687d20eba35387da...02588fd6a7
22:04 sena_kun joined 22:05 Altai-man_ left 22:47 patrickb left 23:52 dogbert17 joined