ab5tract camelia: 04:54
Whoops
Geth rakudo/fix-negated-junctions: 4 commits pushed by ab5tract++ 07:21
rakudo/fix-negated-junctions: 4 commits pushed by ab5tract++
Geth rakudo/main: f1980a8cfc | (Stefan Seifert)++ | 2 files
RakuAST: prohibit placeholders with same name as variables already used in scope
 08:29
nine 1336 (t/spec/S06-signature/positional-placeholders.t fixed)
nine lizmat: sure. I just haven't come across a spectest demanding it. 08:51
Geth rakudo/main: 25c5350f82 | (Stefan Seifert)++ | src/Raku/ast/type.rakumod
RakuAST: fix enum values missing from surrounding package stash
 09:05
nine Fixes the enum problem in t/spec/S06-other/main.t
Geth rakudo/main: 550ad11d50 | (Stefan Seifert)++ | src/Raku/ast/signature.rakumod
RakuAST: fix copy pasta in check for rw params of MAIN subs

Also tightens the check to only cover subs, so we don't accidentally warn about an unrelated method MAIN		 09:14
nine 1337! We shall never fix another spec test :D
patrickb But reducing the number of total spec tests without increasing the number of succeeding ones is allowed. 09:29
nine Absolutely! Best case we end up with 100 % passing without changing that number ;) 09:33
Geth rakudo/main: 46f933a756 | (Stefan Seifert)++ | 2 files
RakuAST: complain at compile time about assigning to a r/o native parameters
 11:22
nine Quick! Someone delete some spectest file!
nine Oh how I love spectests prohibiting us from detecting an error at compile time... 11:56
JimmyZhuo RAKUDO_RAKUAST=1 ./rakudo-m --ll-exception --target=ast -e 'while "D" { last }' 12:00
This type cannot unbox to a native integer: P6opaque, Failure
nine: for '$!condition.maybe-compile-time-value == $while', is it ok change to '$!condition.maybe-compile-time-value.Bool == $while' ? 12:01
or maybe a better one 12:02
nine .Bool may cause bootstrap issues but it's worth a try 12:03
Maybe nqp::istrue($!condition.maybe-compile-time-value) == $while 12:04
JimmyZhuo thanks 12:17
patrickb JimmyZhuo: Did you manage to validate the Windows native runner PR? 12:21
[Tux] Rakudo v2025.03-174-g46f933a75 (v6.d) on MoarVM 2025.03-11-g778b5d309
csv-ip5xs0.264 - 0.267
csv-ip5xs-201.092 - 1.106
csv-parser1.107 - 1.177
csv-test-xs-200.138 - 0.138
test1.852 - 1.858
test-t0.410 - 0.411
test-t --race0.270 - 0.273
test-t-204.975 - 5.100
test-t-20 --race1.234 - 1.255
tux.nl/Talks/CSV6/speed4-20.html / tux.nl/Talks/CSV6/speed4.html tux.nl/Talks/CSV6/speed.log
JimmyZhuo patrickb: git clone github is very unstable and slow here 12:24
can't even do a fresh clone rakudo 12:25
[Coke] O_o; 12:29
Geth rakudo: zhuomingliang++ created pull request #5840:
RakuAST: Fix while "D" { last }		 12:32
JimmyZhuo patrickb: your branch builds ok, but it's hard to test by me. 12:40
patrickb JimmyZhuo: Ok, cool. Thanks for all the testing and feedback so far! 12:42
[Coke] patrickb: need something tested on windows? I can try something if needed. 12:44
finanalyst [Coke] pls test new tweak new-raku.finanalyst.org/en/type/AST 14:32
[Coke] that works! 14:36
patrickb [Coke]: We've been talking about github.com/rakudo/rakudo/pull/5725 I've originally developed that PR on Windows, but I welcome all testing. 15:06
[Coke]: After all JimmyZ uncovered something like 4 separate issues in the build of that PR by trying it out on their Windows machine.... 15:08
Geth rakudo/main: 28b4b3377c | (Stefan Seifert)++ | src/Raku/ast/variable-declaration.rakumod
RakuAST: fix overzealous compile time attr default value check

If the attribute type has a definedness constraint, a mere type won't be able to match it. So check a possible compile time value against the full type and compare mere types with the attribute base type.		 20:49
rakudo/main: 7f1cf26242 | (Stefan Seifert)++ | 2 files
RakuAST: return-type for Type::Simple
rakudo/main: 3a8a337c89 | (Stefan Seifert)++ | src/Raku/ast/variable-declaration.rakumod
RakuAST: use private name when reporting attribute default type errors
roast: 64359c41a7 | (Stefan Seifert)++ | S12-attributes/smiley.t
Change a bunch of attribute default tests to expect compile time errors

These tests as written originally expect us to throw a runtime error for certain situations where a default value does not pass an attribute's type constraint. This in turn means that catching such errors at compile time is actually prohibited (as that would make the whole test file fail to compile). It's clearly better to detect problems at compile time, so this changes those tests and marks them as todo. This way the RakuAST frontend can exceed the old compiler's capabilities without disrupting the build checks.

Powered by Cro and the Raku® Programming Language.

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

Thank you!