japhb patrickb++ 01:03
patrickb: What platforms do you build on? I'm looking at expanding our CI, so wondering what other OS/arch pairs would be useful. 01:04
03:28 vrurg_ left 03:30 vrurg joined 04:56 ShimmerFairy left 04:59 ShimmerFairy joined 06:14 hurufu joined
patrickb japhb: Win x86, arm and x86 mac, x86 Linux. 06:41
tellable6 2026-11-16T06:35:21Z #raku <CIAvash> patrickb your PGP key for Rakudo binaries seems to have been expired
patrickb x86 Linux is using an ancient glibc/distro. 06:42
github.com/rakudo/rakudo/blob/main...s.yml#L880 06:43
that's what's building this (except for arm mac, that's separate) 06:44
CIAvash: Thanks for notifying, 07:02
07:18 hurufu left
patrickb CIAvash: Where did you get the key from that shows the expiry? I rarely (years), but regularly bump the expiry date. Maybe you still have a pre-bump key? 07:58
08:49 hurufu joined 12:28 justanotherjaph joined
Geth rakudo/lizmat-135: 052427c9f5 | (Elizabeth Mattijsen)++ | 13 files
RakuAST: make checks in visit-children use nqp::isconcrete

Turns out that "if nqp::isconrete($foo)" is about 3x as fast as the more general "if $foo" in NQP. So use that at least in the
  "visit-children" methods.
13:30
rakudo: lizmat++ created pull request #6935:
RakuAST: make checks in visit-children use nqp::isconcrete
13:59 shareable6 left 14:01 shareable6 joined
lizmat c: HEAD dd my int %h 15:03
committable6 lizmat, ¦HEAD(95aa98b): «===SORRY!=== Error while compiling /tmp/fFy3LznwzV␤native value types for hashes not yet implemented. Sorry.␤at /tmp/fFy3LznwzV:1␤------> dd my int <HERE>%h␤ «exit code = 1»»
lizmat c: 2026.09 dd my int %h 15:07
committable6 lizmat, ¦2026.09: «int = (my int %)␤»
lizmat bisectable6: 2026.09 dd my int %h
bisectable6 lizmat, Using old=2026.09 new=HEAD in an attempt to do what you mean
lizmat, Bisecting by exit code (old=2026.09 new=95aa98b). Old exit code: 0
lizmat, bisect log: gist.github.com/5428510e7a12df9d66...f0197595d6
lizmat, (2026-10-05) github.com/rakudo/rakudo/commit/07...f5908c9a16
16:35 lizmat left 16:36 lizmat joined
[Coke] patrickb: any idea why on windows, occasionally, doing the rakubrew startup for the shell takes 20s? 17:11
17:38 refactus left 17:40 refactus joined 18:22 hurufu left 18:25 hurufu joined 18:31 iwuvmika left 19:12 hurufu left 19:25 iwuvmika joined
patrickb yeah. Raki tee uses the par packer. It's basically zipping up the app, and on the fly unpacks it into tmp. It ruses the unpacks. but (I don't know exactly what triggers this.) sometimes it re extracts the thing. Then it takes longish to start up. I never investigated in detail though. 19:29
[Coke] ok. it can be very annoying when I'm firing up an unrelated powershell. :) 20:11
(haven't opened a ticket yet because that's on my work machine where I'm typically logged into github as $dayjob 20:12
20:32 hurufu joined 20:36 hurufu left, hurufu joined 21:01 iwuvmika left 21:30 [Coke] left 21:46 iwuvmika joined 22:05 hurufu left 22:44 iwuvmika_ joined, iwuvmika left 22:48 iwuvmika_ is now known as iwuvmika, vrurg left 23:15 vrurg joined 23:26 [Coke] joined 23:32 sjn left 23:33 sjn joined 23:40 sjn left 23:41 sjn joined 23:50 kjp left 23:53 kjp joined