00:23 [Coke] left 00:24 [Coke] joined
timo the bane of my existence: "value has been optimized out" 00:47
gist.github.com/timo/340a02f49638b...aedaf98cc2 random output from "moar timeline 1000" and a `moar bt` with details (arguments and such) from a random spot inside there 01:17
japhb Woah. That's pretty cool! 03:06
Voldenet that's neat, very detailed 03:16
08:44 nine joined 08:45 nine left 09:22 finanalyst joined 10:00 sp1983 joined 10:04 sp1983 left 10:05 Sp1983_ joined 10:07 Sp1983_ left 10:11 El_Che joined
El_Che a problem building moarvm: 10:11
git checkout --recurse-submodules 2026.07 ; fatal: not a git repository: ../../.git/modules/3rdparty/ryu ; fatal: could not reset submodule index
it's a fresh download/build in a container, no trailing files 10:12
10:41 finanalyst left 10:48 nine joined 11:17 nine left 11:58 nine joined
[Coke] That shouldn't have landed until post release, no? 12:35
was this resolved in #raku? 12:42
El_Che I did a workaround 12:43
the submodule is dirty because probably a sunmodule was deleted between the release tag en head
[Coke] OK. If I did something wrong and we can figure it out, we'll get it into the release guide for moar. 12:45
I do have a step where I clean up the submodules explicitly because I have had a dirty checkout in the past.
(but that's to avoid committing any changes on the release branch) 12:46
(changes to submodules, I mean)
El_Che I don't think it's wrong, per se. It that my way cloning+checkout-to-tag strubbles with it when the submodules are "dirty" 12:47
just changed it a little to deal with that
ab5tract m: Q|sub infix:<darling>($x,$y --> Bool) { return $x and $y }; 1 !darling 0 ==> say()|.AST.EVAL 12:50
camelia ===SORRY!=== Error while compiling EVAL_0
Cannot negate darling because default-infix operators are not iffy enough
at EVAL_0:1
ab5tract I think we can pretty easily make that work via RakuAST, right? I guess it doesn't matter to fix unless iffy and diffy can coexist, though. 12:51
m: sub infix:<darling>($x,$y) is equiv(&infix:<eq>) { return $x and $y }; 1 !darling 0 ==> say() 12:52
camelia False
ab5tract but shouldn't that be True? (and please ignore my extraneous return statements .. 😅 12:53
[Coke] github.com/MoarVM/MoarVM/blob/main...lain=1#L33 13:00
(... wow, the clone of moarvm on my irc box was on master still)
ab5tract the past is ever with us :) 13:03
m: sub infix:<darling>($x,$y) is equiv(&infix:<eq>) { $x and $y }; 1 !darling 0 ==> say(); [&&] True Zdarling False ==> say(); [darling] (True, False) xx 2 ==> say()
camelia ===SORRY!=== Error while compiling <tmp>
Cannot reduce with darling because chaining operators are diffy and not chaining
at <tmp>:1
------> ); [&&] True Zdarling False ==> say(); [<HERE>darling] (True, False) xx 2 ==> say()
expecting any …
13:22 nine left 13:33 nine joined 13:37 nine left, nine joined 13:46 thundergnat joined 14:07 thundergnat left
[Coke] whoops, left the blin/release box on post-release. 14:26
15:36 El_Che left 15:37 apogee_ntv left 15:38 apogee_ntv joined
[Coke] ... ok with multiple heavy months of blin usage, releases, etc., time to ditch azure and get a standalone linux box for this. :| 16:46
19:24 finanalyst joined
[Coke] Ordered. Blah. 20:22
japhb Specs, out of curiosity? 20:31
[Coke] 32G mem, 1TB disk, mini form factor, intel core i9/14 20:42
dual boot linux/win
Can throw it under my monitor stand and remote into it. 20:43
20:56 Geth left 20:57 Geth joined
Geth rakudo/main: 96c7630c32 | (Elizabeth Mattijsen)++ | src/core.c/RakuAST/LegacyPodify.rakumod
RakuAST: allow for empty =comment/code/data/input/output

Spotted by Marcel Timmerman, raised by Richard Hainsworth
20:57
[Coke] lizmat: is github.com/Raku/geth/issues the current repo for Geth? 21:17
looks like it had a commit a month ago, so *guessing* yes
question: if our default branch is main in a repo, can/should we remove the old master? 21:29
I would vote should. (e.g. on raku/nqp)
22:10 finanalyst left