🦋 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: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
02:02 Voldenet left 02:36 Voldenet joined
Geth nqp/main: 6881b371f1 | (Daniel Green)++ | tools/templates/Makefile-backend-common.in
Don't (re)compile P5QREGEX every time test are run
nqp/main: 22af4c8b02 | MasterDuke17++ (committed using GitHub Web editor) | tools/templates/Makefile-backend-common.in
Merge pull request #815 from MasterDuke17/second_attempt_at_not_compiling_P5QREGEX_every_time_tests_are_run

Don't (re)compile P5QREGEX every time test are run
rakudo/main: 555186626e | (Daniel Green)++ | src/Perl6/Metamodel/MethodContainer.nqp
Unbreak the JVM build

Not sure *why* this is required since cf474c9 is fine for the MoarVM backend...
lizmat cf474c9 08:20
.tell MasterDuke I guess JVM's nqp::istype doesn't like nqp::null as one of the arguments 08:21
tellable6 lizmat, I'll pass your message to MasterDuke48
Geth rakudo/main: b82fa83db0 | (Elizabeth Mattijsen)++ | 2 files
Treat =place as a directive

Not as a block
rakudo/main: f997c36cc0 | (Elizabeth Mattijsen)++ | src/Raku/Actions.nqp
RakuAST: ignore =finish if there is no code

The =finish construct only makes sense in code. If it is being used in a RakuDoc context only, it tried to attach the text to the CompUnit when there was none, because there was no code.
This now just ignores the =finish text if there is no CompUnit to attach to. Potentially, it could also be converted to a RakuDoc, but this does not actually appear to make sense (yet).
09:33 sena_kun joined
Geth rakudo/main: c49e05d848 | (Elizabeth Mattijsen)++ | 6 files
Streamline Metamodel::RoleContainer

  - make adding a role threadsafe: this however broke several spectests
   because apparently during role composition, in some cases roles were
   being added. Which was fine in the non-threadsafe handling.
  - add a threadsafe "pop_role_to_compose" method which will either
   produce a role to be composed, or nqp::null (in a threadsafe way).
  - adapt clients of this role to the new way of handling role composition
  - remove unneeded positional arguments where possible
This appears to have a little bit of a negative effect on performance
japhb lizmat: After all parts of c49e05, are the spectests still broken, or did the later parts fix that? 14:21
lizmat the later parts fixed that, otherwise I wouldn't have pushed that to main 19:00
japhb lizmat: Gotcha, it just wasn't clear to me. Thanks for clarifying. :-) 21:13