Geth rakudo: ugexe++ created pull request #6706:
Make frontend detection and switching in tests less fragile		 00:32
ugexe you can use that if you want, or close it if you have something else. it makes the tests DTRT regarding rakuast detection and enabling/disabling assuming RAKUDO_RAKUAST is the switch 00:33
it can be merged by itself as well 00:36
Geth rakudo: ugexe++ created pull request #6707:
Run the CI frontend variant as legacy rather than RakuAST		 00:39
ugexe and you can use that one^ for the CI although it shouldn't be merged on its own and instead should be cherry-picked into whatever PR enables rakuast by default 00:40
01:03 [Coke] left 01:23 ShimmerFairy left 01:26 ShimmerFairy joined
Geth roast: 1f521d7986 | (Nick Logan)++ (committed using GitHub Web editor) | S28-named-variables/slangs.t
Remove Perl5 regex leftover
 01:28
rakudo: ugexe++ created pull request #6708:
RakuAST: reject a Perl 5 regex adverb like any other unknown adverb		 01:33
02:59 [Coke] joined
Geth rakudo/main: 1cdf64d61b | (Nick Logan)++ | 2 files
RakuAST: reject a Perl 5 regex adverb as an unknown adverb

Support for Perl 5 regexes was removed, but the :P5 and :Perl5 adverbs were still recognized as compilation adverbs, so a construct such as m:P5/o+/ compiled and then never matched. Dropping them from the adverb tables leaves them unrecognized, so they now get the same compile time
  "Adverb P5 not allowed on m" error as any other adverb a construct does
not take.		 08:09
rakudo/main: 0dd07126b4 | (Nick Logan)++ | 3 files
RakuAST: remove the Perl 5 regex slang

Nothing selects the P5Regex slang now that the adverb naming it is an error, so its grammar, its actions and the flag that chose between it and the Raku regex slang go away.
rakudo/main: 8f6418198a | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 4 files
Merge pull request #6708 from ugexe/ugexe/rakuast-p5-adverb-unknown

RakuAST: reject a Perl 5 regex adverb like any other unknown adverb
rakudo/main: 87d3151487 | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
Assign to a new Array or Hash element from its own callsite

Previously ASSIGN-POS and ASSIGN-KEY ran one p6assign for both a new element and an existing one, with the fresh container or the stored one as its target. MoarVM specializes that assignment on the type of the value the target holds. A fresh container holds nothing, so code that first fills an array or hash and then overwrites its elements got a ... (18 more lines)		 08:12
rakudo/main: 57ba63f2a6 | (Nick Logan)++ | 16 files
Test the compiler frontend instead of RAKUDO_RAKUAST

Tests that branch on which frontend compiled them read the RAKUDO_RAKUAST environment variable. That reads the knob rather than the result: the variable says what was asked for, while the COMPILER-FRONTEND hllsym says what the compiler became, and only the hllsym stays correct if the default ever changes.
Reads of the variable now go through nqp::gethllsym('Raku',
  'COMPILER-FRONTEND'). Files that set it for a subprocess keep doing so,
since that is how a child process picks its frontend.		 08:14
rakudo/main: 0a9e73be11 | (Nick Logan)++ | 2 files
Pin the legacy side of a frontend matrix with RAKUDO_RAKUAST=0

Tests that drive both frontends through subprocesses select the legacy frontend by removing RAKUDO_RAKUAST from the child environment, which selects whatever the default is rather than legacy. Setting it to 0 names the frontend the assertions expect.
rakudo/main: 4152f0e83e | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 18 files
Merge pull request #6706 from ugexe/ugexe/tests-compiler-frontend-check

Make frontend detection and switching in tests less fragile
nqp/main: d8ec7766b8 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM for zmij fix on older GCC's patrickb++
 08:30
rakudo/main: 4f26b03bb1 | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP for zmij fix on older GCC's patrickb++
 08:44
lizmat notable6: weekly 09:47
notable6 lizmat, 1 note: 2026-09-16T22:48:28Z <lizmat>: andrewshitov.com/2026/09/15/raku-l...-italiano/
lizmat notable6: weekly reset 09:48
notable6 lizmat, Moved existing notes to “weekly_2026-09-21T09:48:00Z”
09:50 finanalyst joined 10:37 finanalyst left
lizmat [Coke]: re fixing GLib, forked into community modules, installed the deps, tried to run tests 11:10
An exception X::AdHoc occurred while evaluating a constant:
Do not know how to handle the darwin OS. Please contact the project maintainer!
meh
so I won't be doing that 11:11
14:15 sjn left 14:22 sjn joined
[Coke] I'm fine with creating a ticket on the root offending module and then linking to that from blin and then skipping those modules for now 14:40

Powered by Cro and the Raku® Programming Language.

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

Thank you!