12 Dec 2025
patrickb that encompasses moving libuv to a temporary fork with the libuv PTY patch. 13:46
When would be the deadline?
thinking about it. it would also be nice to get the windows zef issue solved by then... 13:47
Geth rakudo/lizmat-mixed-int-num-infixes: 1bc6c6b298 | (Elizabeth Mattijsen)++ | src/core.c/Num.rakumod
Add infix + - / * candidates for mixed int / num

Inspired by the benchmark mentioned at
   irclogs.raku.org/raku/2025-12-11.html#22:53
When using native int and num in an expression with these infixes, both the natives where being upgraded, which makes it a *lot* slower. ... (15 more lines)
14:38
rakudo: lizmat++ created pull request #6040:
Add infix + - / * candidates for mixed int / num
melezhik [Coke]: I still need to check I think it’s but a Rakudo star issue , it somehow relates to container environment. I see the same problem with whateverable Rakudo and Rakudo installed from source - if they run on container 15:00
ugexe presumably the issue is exactly where the error message pointed out 15:10
github.com/ugexe/zef/blob/2c6ab9ff...kumod#L922
seems like that would have been the obvious first thing to look at 15:11
lizmat github.com/rakudo/rakudo/issues/6041 15:21
ugexe as far as what zef should do im open to suggestions 15:22
the existing logic aims to open a new shell, but it tries to use the same shell you are already using. this is primarily for windows users who might be launching via the x86 command line tools shell or whatever it was called vs powershell vs command prompt 15:23
i'm not sure what it should do if e.g. ENV<SHELL> isn't defined
ab5tract patrickb: regarding the `anon` scope, there is a method on Package and family that returns the allowed scopes. might be worth adding `anon` there and seeing what happens :)
ugexe maybe just adding // '' would be sufficient
melezhik2 . 16:44
[Coke]: could you please pull the latest commit and start the agent ? I am trying to retest some failed modules and see what it gets 16:46
[Coke] weird, had a reboot, can't run the docker desktop? 17:14
one moment. 17:15
ok, rerunning. 17:17
lizmat m: sub a(--> num) { fail }; a # looks like native return constraints don't pass through Failures 18:25
camelia This type cannot unbox to a native number: P6opaque, Failure
in sub a at <tmp> line 1
in block <unit> at <tmp> line 1
lizmat but this appears fixed in RakuAST :-) 18:26
melezhik [Coke]: when does the next release happen ? I don’t expect any for major changes in brownie , I consider it pretty well tested, it’d interesting what we’ll get with it with the next release … 19:21
Geth rakudo/lizmat-mixed-int-num-infixes: 011c1148d7 | (Elizabeth Mattijsen)++ | src/core.c/Num.rakumod
Some more native infix fixes

  - add implementation-detail method "divzero" that produces a division
   by 0 error for its invocant
  - call this method when attempting to divide by zero for / and %
   hereby reducing bytecode size of methods, therefore making them
   easier to inline.
... (8 more lines)
19:25
japhb lizmat: I commented on 011c114 22:18
patrickb Coke: I fail to reproduce the Windows zef doesn't work bug. Can you provide details on how you set up the VM? Chances should be pretty high to reproduce the issue if I create a VM just like you did. 22:52
[Coke] it's an azure VM; I gave details in the ticket about how I installed rakubrew. 23:05
I can give you the systeminfo if that helps
I can also just give you access to the box. 23:06
gist.github.com/coke/07f2f55fcd569...5e0c582fbd 23:10
patrickb I'm off to bed for today. But system access would be the easiest way forward. How much am I then allowed to mess with the system? 23:11
[Coke] as much as you want, I can always recreate from scratch if needed. 23:12
13 Dec 2025
Geth rakudo/lizmat-mixed-int-num-infixes: 9a17de9efa | (Elizabeth Mattijsen)++ | src/core.c/Num.rakumod
Address japhb++'s thinko catches
00:08
lizmat japhb++ ^^
japhb lizmat++ 00:59