[Coke] one failure so far on blin: github.com/coke/raku-blin-release-.../2026.07.1 01:00
github.com/coke/raku-blin-release-...3AZH_0.0.3
⏳ 1760 out of 2482 modules processed (70.91%)
that one bisects to github.com/rakudo/rakudo/commit/db...0abcc7b094 which is an AST one? 01:10
so that seems suspect.
(the failure, not the commit)
if you're looking at these before the run is done, the "bisected to" is one of the lines in the overview, btw.
ugexe the l10n stuff uses rakuast, although i'm not sure we can actually count breaking those as regressions 01:58
yeah i think that failure is more of an issue with L10N itself 02:01
which has other issues that i'm not sure can be fixed 02:02
for example i tried to fix github.com/rakudo/rakudo/issues/6046 but my conclusion at the time was the l10n stuff wasn't viable the way it is implemented 02:03
If something is wrong with that commit though i expect more stuff should show up so i suppose we should just wait and see 02:12
The weird part is I would expect that commit to actually fix l10n modules 02:20
releasable6 Next release in ≈6 days and ≈15 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 03:00
ugexe "l10n stuff wasn't viable the way it is implemented" i meant viable as something to use with blin 03:05
Geth rakudo: ugexe++ created pull request #6462:
RakuAST: make the require adverb constraint tests self-contained		 04:10
ugexe github.com/Raku-L10N/ZH/pull/2 04:58
Geth rakudo/main: e4ef743148 | (Nick Logan)++ | t/12-rakuast/xx-fixed-in-rakuast.rakutest
Limit the attribute-signature-default todo to the legacy frontend

The default-from-signature attribute test passes under the RakuAST frontend, so the todo now applies only where the dispatch error remains.		 05:10
rakudo/main: d3b85b99ac | (Nick Logan)++ | 3 files
Test require constraints against a fixture distribution

Previously the require adverb tests constrained requires of the installed Test module. What enforces a constraint depends on which repository resolves the name and what its distribution declares: the test harness sets RAKULIB=./lib, where rakudo's own lib/Test.rakumod resolves as a bare source that satisfies any constraint, so the ... (6 more lines)
rakudo/main: bcd2c0a5ba | (Nick Logan)++ (committed using GitHub Web editor) | 3 files
Merge pull request #6462 from ugexe/ugexe/require-constraint-test-fixes

RakuAST: make the require adverb constraint tests self-contained
06:12 coleman left 10:14 finanalyst joined 11:08 finanalyst left
[Coke] blin done. just one more failure: cro 15:42
ugexe that one is certainly more concerning :) 15:54
[Coke]: might be flakey / racey, could you try running that one again? 16:03
Geth rakudo/main: 5 commits pushed by (Nick Logan)++ 16:05
rakudo: ugexe++ created pull request #6463:
RakuAST: skip unused implicits, flatten more statement forms, and inline compile-time decided dispatches		 16:07
ugexe i think thats all the optimizations from Optimizer.nqp that can be ported 16:08
the last of them rather
16:22 hurufu joined
ugexe bah, i guess there are a couple more still 16:35
hllbool stripping and some smartmatch stuff
➜ rakudo git:(ugexe/rakuast-lex2local-riders) ./install/bin/raku -e 'my $t = now; my int $i; my int $n; while $i < 10_000_000 { $n = $n + $i; $i++ }; printf "%.3f\n", now - $t' 16:39
0.144
➜ rakudo git:(ugexe/rakuast-lex2local-riders) RAKUDO_RAKUAST=1 ./install/bin/raku -e 'my $t = now; my int $i; my int $n; while $i < 10_000_000 { $n = $n + $i; $i++ }; printf "%.3f\n", now - $t'
0.139
➜ rakudo git:(ugexe/rakuast-lex2local-riders) ./install/bin/raku -e 'my $t = now; my $x = 0; my $i = 0; while $i < 1_000_000 { { $x = $x + 1 }; $i = $i + 1 }; printf "%.3f\n", now - $t'
0.325
➜ rakudo git:(ugexe/rakuast-lex2local-riders) RAKUDO_RAKUAST=1 ./install/bin/raku -e 'my $t = now; my $x = 0; my $i = 0; while $i < 1_000_000 { { $x = $x + 1 }; $i = $i + 1 }; printf "%.3f\n", now - $t'
0.217
so rakuast is actually faster right now for native int while loops and bare blocks 16:40
apparently legacy cant flatten the bare block case at all 16:41
legacy closure clones and invokes the inner { $x = $x + 1 } every iteration because its arity check trips on the block optional topic parameter 16:42
18:00 camelia left 18:02 MasterDuke joined 18:04 camelia joined
MasterDuke ugexe: i don't know how much it would trigger, but something the static optimizer doesn't really do right now is dead-code-elimination. so `raku -e 'if 2 > 4 { say "dead" } else { say "alive!" }'` at every `--target=` level (both legacy and RakuAST), shows the `if` being evaluated. obviously it would make the most sense to run after inlining 18:06
[Coke] ugexe. sure. 18:11
Rerunning just Cro::HTTP against the new HEAD. 18:14
releasable6: next 18:22
releasable6 [Coke], Next release in ≈6 days and ≈0 hours. There are no known blockers. 216 out of 224 commits logged
[Coke], Details: gist.github.com/feadd6dfcad54996ac...440b1fab83
[Coke] are those 8 commits all AST?
ugexe yeah
[Coke] 👍 18:23
yup, was fine on a rerun 18:29
18:39 camelia left 19:32 MasterDuke left 19:35 camelia joined
nine m: say "alive?" 19:35
camelia alive?
20:36 finanalyst joined 20:59 finanalyst left 21:10 finanalyst joined 21:18 finanalyst left 21:36 hurufu left, hurufu joined
ugexe MasterDuke: Thanks, I have a branch adding some dead code elimination now but it’s behind like 4 other branches I have yet to review and push 21:39
tellable6 ugexe, I'll pass your message to MasterDuke
patrickb nine: Hi there! I'm off to bed, but still nice seeing you around! 21:44
22:21 hurufu left
releasable6 Next release in ≈5 days and ≈19 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 23:00

Powered by Cro and the Raku® Programming Language.

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

Thank you!