01:44 melezhik left 02:22 rakkable left 02:23 rakkable joined 06:51 Voldenet_ joined, Voldenet left, Voldenet_ is now known as Voldenet 07:31 melezhik joined 10:41 finanalyst left 14:38 melezhik_ joined
melezhik_ brownie another round results - brw.sparrowhub.io/report/brw-orch/1 , artifacts, interesting I did not expect so much test failures 14:39
14:44 melezhik_ left
melezhik I added builds link to track previous brownie rounds - brw.sparrowhub.io 14:52
^^ ab5tract 14:53
15:04 guifa joined 15:26 librasteve_ joined 16:11 melezhik_ joined 16:13 melezhik_ left
Geth ¦ rakudo: coke self-assigned bad link in release announcement github.com/rakudo/rakudo/issues/6004 16:24
librasteve_ rakudoweekly.blog/2025/11/03/2025-...drey-tang/ 18:52
lizmat librasteve_++
19:26 guifa left 19:42 guifa joined
timo on the journey towards reproducible rakudo builds again, I think the problem is that we don't use sorted_keys inside RegexCaptures::!from-capnames from the bootstrap.c (and it doesn't seem like we can use it there either, it just gives a VMNull, same for &note and &say, so difficult to diagnose what's going on, too) 20:19
i feel like i've come to this same conclusion before 20:23
oh, two times the same hash of CORE.c.setting.moarvm in a row 20:25
three times 20:26
four times
five times 20:28
i would call this reproducible now 20:30
but i had to copy the implementation of sorted_keys from nqp into rakudo, that's not so great
20:31 melezhik left
lizmat timo: on that note, do you recall the non-nqp way to HLLize an NQPArray ? 20:32
timo hm, not really. doesn't just passing it to a sub or method without being careful cause it to be hllized? 20:35
lizmat ah. yes 20:37
timo but i guess not recursively
lizmat trying to fix github.com/raku-community-modules/.../issues/18 by excluding $!MONITR-lock attribute 20:41
in ^attributes
japhb lizmat: Can a class using Method::Protected inherit from a monitor? 20:54
lizmat Method::Protected uses a different attribute name, so depending on what you want to achieve: yes or no :-) 20:55
popd 20:56
meh
20:57 finanalyst joined
timo we could probably code-gen a method clone that handles the lock specifically and everything else normally 21:15
21:38 finanalyst left
lizmat yeah, that would be my next attempt 23:56
timo it will probably mirror the buildallplan relatively closely 23:57
23:59 rakkable left, rakkable joined