🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | log inspection situation still under development | For MoarVM see #moarvm
Set by lizmat on 22 May 2021.
japhb ugexe: FWIW, as someone who has *NOT* used them heavily, it feels like CU* provide APIs that are powerful but require deep understanding to use properly. Kindof plumbing versus porcelain, or Lock/Thread/Scheduler versus Promise/Supply/Channel. 04:29
Each time someone asks about how to do something with those APIs, we either get XY questions (leading to "Why would you want to do that?" answers) or "Just use this long one-liner that you would have no hope of knowing, but really isn't any more complex than implementing monads in SQL". 04:31
I myself want (and I think I see people asking) for more convenience, more procelain, more DWIM. 04:32
Hell, I'd be in favor of someone making a CompUnit::Utils dist with just a bajillion examples and convenience routines and such. And I can't do it myself, because I lack the mental model (so far) to be able to produce them. 04:34
japhb Possible Rakudo incompatibility canary? github.com/raku-community-modules/URI/issues/48 06:48
nine lizmat: CompUnit::Repository::Staging is NOT just a part of our build chain. It's explicitly meant to be used by module installers and github.com/rakudo/rakudo/blob/mast...ll-dist.p6 is a) an example for how to do this and b) itself a tool that's used for packaging modules: build.opensuse.org/package/view_fi...c?expand=1 07:41
nine ugexe: AFAIR CompUnit::Repository::Staging could easily live in the core. It just seemed to be something with a very narrow use case and as such better suited as an external module. But it has to be shipped with rakudo as our installation process needs it. It was also simply a test bed for custom CUR implementations 07:42
lizmat m: CompUnit::RepositoryRegistry.repository-for-name("core").installed.map(*.meta).elems.say # sanity check 09:29
camelia 1
lizmat locally I got 2, most likely because of github.com/rakudo/rakudo/pull/4613
nine yes, that will be the reason 09:32
ugexe japhb: most of what i post is long because the names of things and methods are long 12:25
otherwise pretty much everything is done by chaining 3 steps together $*REPO.repo-chain to iterate over all the repo chains, a .grep to remove the ones you dont want, and then a .map(.candidates(...) | .resolve(...) | .?installed) 12:28
iterate over all the repos rather^ 12:35
lizmat anyways: raku.land/zef:lizmat/Rakudo::CORE::META 12:51
lizmat hmmm... interesting: some distros refer to "Pod::To::HTML:auth<github:Raku>" as a dependency 13:12
sena_kun lizmat, mm? 13:21
or maybe "what's curious about that"? 13:22
lizmat argh, confused it with Pod::To::Text 13:23
sena_kun ah, then yes, there are no other alternatives
lizmat and Pod::To::Text is core indeed 13:24
ugexe nine: i guess i was thinking of github.com/rakudo/rakudo/commit/ac...t-18128990 but i suppose that could have been caused by something else 14:51
nine ugexe: yes, that's exactly why I thought keeping Staging out of core was a good test. Was quite tricky to get it all working 15:24
Geth roast: 90eba8098b | (Elizabeth Mattijsen)++ | S32-list/sort.t
Remove tests from a PR that will not be merged

  github.com/rakudo/rakudo/pull/3849 to be precise		 20:25
rakudo: 6e5c17f1ab | (Elizabeth Mattijsen)++ | src/core.c/Exception.pm6
Remove Exception class that was needed for a PR

But the PR never made it. See github.com/rakudo/rakudo/pull/3849		 20:27
melezhik Uploaded file: uploads.kiwiirc.com/files/5d4834ab.../image.png
MasterDuke lizmat: re that pr and roast test, i think it does make sense to disallow cmp'ing with a junction. adding `multi sub infix:<cmp>(\a, Junction:D $b) { die "Can't cmp a Junction" }; multi sub infix:<cmp>(Junction:D $a, \b) { 20:55
die "Can't cmp a Junction" };` to Order.pm6 passes a spectest
m: say (9, 3, 1, any(4, 5, 6, 7, 500), "a").sort # and this now dies with that message 20:56
camelia (a any(4, 5, 6, 7, 500) 1 3 9)
lizmat MasterDuke: fwiw I'm looking at github.com/Raku/problem-solving/issues/297 20:57
MasterDuke cool
lizmat if you want to create a PR for junctioning cmp, well... I was pretty burned out on that one
Geth rakudo: 3ef09a72e2 | (Elizabeth Mattijsen)++ | src/core.c/StrDistance.pm6
Make StrDistance.ACCEPS also a multi

Just like all other .ACCEPTS methods		 20:58
MasterDuke i can do that, what about adding back your exception in my pr?
lizmat well, I felt that the exception in that form would be pretty useless for handling #297 20:59
lizmat so I thought I'd remove it to have a clean slate 20:59
nine Could not find symbol '&Junction' in 'X::Cannot' in block <unit> at t/spec/S32-list/sort.rakudo.moar line 266 21:35
lizmat ddi you update roast?
nine Ah, local change prevented the automatic pull in roast
lizmat *phew* 21:36
MasterDuke i get that on my branch where i reverted removing the exception and throw it in Order.pm6, when i try to test 21:40
dogbert11 'zef install URI' has started to fail with the following error message:
[URI] # Failed test 'right exception type (X::Assignment::RO)'
[URI] # at SETTING::src/core.c/Exception.pm6 line 62
[URI] # Expected: X::Assignment::RO
[URI] # Got: X::Attribute::Required
[URI] # Exception message: The attribute '$!value' is required, but you did not provide a value for it.
nine Now I get: # Failed test 'found original file name in the backtrace' 21:41
dogbert11 must be a relatively recent regression
nine # at t/spec/integration/error-reporting.t line 127
MasterDuke i just had a clean spectest on my branch (which is at HEAD + revert removing the exception + throw it in cmp) 21:42
nine Huh...that spectest passes after removing .precomp directories. 21:44
lizmat dogbert11: URI errors confirmed
nine Apparently we still found the same precomp file after the file got renamed. And the test explicitly checked the compiled file name. Weird coincidence :) 21:45
MasterDuke the URI error is probably github.com/rakudo/rakudo/pull/4594 right? 21:47
but unrelated to URI, why can't rakudo find X::Cannot::Junction? 21:48
hm. `raku -e 'X::Cannot::Junction.new(junction=>"a|b", for=>"for nothing").throw'` works fine 21:50
lizmat because I just removed it ? 21:53
MasterDuke this is on my branch which added it back in 21:54
Geth rakudo: MasterDuke17++ created pull request #4618:
Disallow cmp with a Junction		 22:04

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!