01:36 sjn left 02:22 sjn joined 05:43 djinni` left 05:48 djinni` joined 09:00 librasteve_ left 09:44 librasteve_ joined 11:55 ShimmerFairy left, rnddim joined 11:57 ab5tract left 11:58 ab5tract joined, librasteve_ left 12:14 rnddim is now known as ShimmerFairy
[Coke] blin run in progress - looks like it's hanging on Subsets::IO trying to bisect it. 12:44
(it's still processing other stuff while that's hanging) 12:46
⏳ 2324 out of 2443 modules processed (95.12%) 14:12
weird, it stopped complaining about versions of rakudo being locked during the run. 14:13
was getting e.g. /tmp/whateverable/rakudo-moar/d8aeb7eeb8a342e10d19c1faf4025713d6d8ac2e is locked. Waiting…
it would complain about a version being locked a few dozen times, then another... then it stopped. we'll see what the final output says
14:20 donaldh left
[Coke] Bisecting Calendar, Pod::To::PDF::Lite 14:28
16:25 librasteve_ joined 16:27 donaldh joined
[Coke] ⏳ 2442 out of 2443 modules processed (99.95%) (left: PDF::Document) 17:43
looks like 85 failures. 17:59
lizmat meh
[Coke] github.com/rakudo/rakudo/commit/dc...fa1cba4108
that's 81 of them. 18:00
three here: github.com/rakudo/rakudo/commit/ab...0e3c9e9943
lizmat ok, I guess we'll have to revert the first
[Coke] 1 here: github.com/rakudo/rakudo/commit/4e...174e0d9750
I would verify that it also fails for someone outside of this blin machine, the bisects were making a LOT of noise.
I may have to rerun it if you can't verify one. 18:01
lizmat that last one is basically the same as the first
so it feels we're going to need a language level check for those
[Coke] * [ ] [Physics::Unit](raku.land/zef:librasteve/Physics::Unit) – Fail, Bisected: [dc5e945](github.com/rakudo/rakudo/commit/dc...a1cba4108)
gist.githubusercontent.com/coke/0a...rtlist.txt 18:06
that's just the bulleted list with no error output
lizmat confirmed the issue with ab6efcab1a90d74c15af46cc1bde890e3c9e9943 18:11
linkable6 (2026-04-14) github.com/rakudo/rakudo/commit/ab6efcab1a Properly support coercive types on parameterized Set(Hash)
lizmat fixing that now
[Coke]: it appears fixed in a more massive overhaul that I worked on today 18:12
yeah, all 3 modules affected by that, install fine 18:13
I was originally planning on merging that work after the release, but it appears it could be done now already :-) 18:14
[Coke] ok. happy to kick off another blin run when you do 18:19
Geth rakudo/main: 5bceb7cada | (Elizabeth Mattijsen)++ | 23 files
Massive overhaul of QuantHash infrastructure

This basically in preparation to fully supporting coercive type parameterization.
The story: so far, an empty QuantHash would *not* have an instantiated Rakudo::Internals::IterationSet (a HLL version of an nqp::hash) in it. ... (26 more lines)
18:40
lizmat [Coke]: that should fix the 3 modules attributed to ab6efcab1a90d74c15af46cc1bde890e3c9e9943
linkable6 (2026-04-14) github.com/rakudo/rakudo/commit/ab6efcab1a Properly support coercive types on parameterized Set(Hash)
[Coke] cool 18:42
!
releasable6 Next release in ≈4 days and ≈23 hours. There are no known blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
[Coke] timo: can you provide changelog notes for: 19:02
+ Bump NQP for bytecode validation error fix, timo++ [8051244f]
+ Bump NQP for utf8-c8 fix, timo++ [9dcbc23c]
20:04 librasteve_ left
timo - Fix crash when reporting errors of broken bytecode files [8051244f] 20:11
- Fix memory corruption when using utf8-c8 codec in a streaming decoder [9dcbc23c]
[Coke] thank you 21:50
lizmat [Coke]: OpenSSL 0.2.7 with ugexe's fix is on its way to the ecosystem 21:54
[Coke] ... and might that resolve all the downstream issues we saw?
I'll kick off a blin run when it hits (or tomorrow morning at the latest) 21:55
21:55 rakkable left, rakkable joined
lizmat unsure how much that will fix, but OpenSSL is pretty upstream, so there might be a lot of modules fixed 21:56
[Coke] ok, cool. worth a reblin
0.2.7 landed. 22:56
kicked off run #2 23:11
23:38 [Tux] left 23:53 [Tux] joined