| 14 Oct 2025 | |||
| ab5tract | what does 'specialize' mean in terms of roles? | 15:35 | |
| lizmat | create something that can be mixed in with the given parameterization? | 15:39 | |
| ab5tract | I wish I understood that sentence :/ | 15:50 | |
| R => R[Int] == speciailization? | |||
| Basically I'm trying to understand how CurriedRoleHOW interacts with ParametricRoleGroupHOW | 15:51 | ||
| as there doesn't seem to be much of a connection. but also CurriedRoleHOW is supposedly an intermediate metamodel object but I still see plenty of them in regular usage | 15:53 | ||
| m: Positional[Int].HOW.say | |||
| camelia | Perl6::Metamodel::CurriedRoleHOW.new | ||
| ab5tract | the metamodel code is mind bending | 15:54 | |
| m: dd [1,2] ~~ Positional[Int] | 16:14 | ||
| camelia | Bool::False | ||
| ab5tract | ^^ wish I hadn't been using this as a sanity check | ||
| lizmat | well, [] is basically Positional[Mu] | 16:21 | |
| and roles are not bytecode, not until they are consumed | |||
| ab5tract | m: dd Positional.HOW.raku; dd Positional[Mu].HOW.raku | ||
| camelia | "Perl6::Metamodel::ParametricRoleGroupHOW.new" "Perl6::Metamodel::CurriedRoleHOW.new" |
||
| ab5tract | what does "consumed" mean here? | 16:22 | |
| btw, I think I may have cracked the generics thing | |||
| lizmat | role B { }; class A does B { } # class A is consuming role B | 16:23 | |
| my $a = 42 but False # enums provide a role with a method with the name of the enum, so .Bool will produce False in this case | 16:24 | ||
| you could say that the but created a new Int type that consumed that special role | |||
| ab5tract | Ok, so I got t/spec/02-types/generics.t passing, but it feels like way fewer TODOs are passing overall in make spectest :( | 16:27 | |
| Geth | rakudo/r5970-fix-generics: 88db3dd4ee | ab5tract++ | src/Raku/ast/type.rakumod RakuAST - Address some issues with generics Previously the following code: role R[::T] { has @.a is Array[T] } dd R[Int].new: :a(1,2,3) ... (33 more lines) |
16:34 | |
| ab5tract | nine: I think this one is a lot less gross, and (almost) entirely works (small regression, noted in commit message) | 16:35 | |
| lizmat: thank you for the help | 16:38 | ||
| lizmat | jdv [Coke] I believe all blin issues are addressed | 17:57 | |
| Math::Sequences I cannot reproduce, so assume a flapper | 17:58 | ||
| PDF::Content / SBOM::CycloneDX have been fixed with github.com/rakudo/rakudo/commit/1ac895f152 | |||
| all other modules have just been updated and uploaded to zef | |||
| Geth | rakudo/main: c9714bf939 | (Elizabeth Mattijsen)++ | src/core.c/RakuAST/Raku.rakumod RakuAST: fix gisting of &?ROUTINE Spotted by David Warring, fixes #5977 |
19:50 | |
| [Coke] | lizmat++ | 23:04 | |
| 15 Oct 2025 | |||
| japhb | lizmat: Why were some modules worthy of change? Where they doing something previously only working by luck? | 00:50 | |
| lizmat | CALLER::CALLER::<::?PACKAGE> | 06:56 | |
| in 6.e | |||
| releasable6 | Next release in ≈3 days and ≈7 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft | 11:00 | |
| [Coke] | get-command raku on powershell on windows shows version "0.0.0.0" for raku 2025.03 | 13:28 | |
| Need to consider pulling META6 into raku-community-modules. | 14:41 | ||
| timo | i thought i had written code for the moarvm/HLL/Backend.nqp to change how all the sql insert statements are written out to use a common function that handles splitting the statements up when they get too long, maybe that's still hidden in a branch somewhere | 16:20 | |
| re: github.com/rakudo/rakudo/issues/5972 | |||
| Geth | ¦ rakudo: lizmat self-assigned RakuAST - incorrect DEPARSE of a rule with leading declarator github.com/rakudo/rakudo/issues/5978 | 18:25 | |