🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
[Coke] release this saturday - any help on cleaning up roast appreciated: github.com/rakudo/rakudo/issues/5841 00:39
The changelog is up to date as of right now.
another blin run in progress now. 00:44
⏳ 1659 out of 2303 modules processed (72.03%) 00:45
ugexe i just added 'use v6.d' to all zef source files since modern versions of zef dont work on rakudos so old they dont support v6.d. maybe the non-blocking await will give a boost to performance 00:53
i bumped the version to v1.0.0 and also made the first release on the zef ecosystem 00:54
[Coke] O_o 01:00
I will probably do another blin run mid week then to test out the updated zef. 01:16
just in case.
01:18 kjp left 01:19 kjp joined 01:20 kjp left, kjp joined 03:30 sourceable6 left, bloatable6 left, nativecallable6 left, committable6 left, linkable6 left, notable6 left, bisectable6 left, coverable6 left 03:31 unicodable6 left, unicodable6 joined, notable6 joined, coverable6 joined, bisectable6 joined, sourceable6 joined, linkable6 joined 03:32 bloatable6 joined, nativecallable6 joined 03:39 unicodable6 left, committable6 joined 03:40 bloatable6 left, nativecallable6 left, notable6 left, sourceable6 left, notable6 joined, nativecallable6 joined, sourceable6 joined, bloatable6 joined, unicodable6 joined
Geth rakudo/main: 25117e3001 | (Stefan Seifert)++ | src/Raku/ast/package.rakumod
RakuAST: fix accidental sharing of meta object in nested declarations

Fixes: module Foo { class Foo { has $.foo } }
07:24
nine This gets you to the redeclaration error with module MoarVM::Profiler { class MoarVM::Profiler { } } 07:25
MasterDuke: yes, I've seen those messages. Tried to debug them and failed. They are about a mismatch between $*PACKAGE and nqp::getattr($!braid, Braid, '$!package') but I can't for the life of mine find where its coming from. Also haven't found any negative effects. 07:26
tellable6 nine, I'll pass your message to MasterDuke
nine Wity more debug info: 07:28
Out-of-sync package detected in LANG1 at se::L10N::$slang;
RakuAST::Dep (value in braid 1665795248: RakuAST::Class Map, value in $*PACKAGE: RakuAST::Class RakuAST::Deparse) at NQP::src/core/IO.nqp:297 (/home/nine/rakudo/install/share/nqp/lib/NQPCORE.setting.moarvm:)
Seems to help with the EVAL problem: 08:10
my $context := nqp::defined($ctx) ?? $ctx !! CALLER::LEXICAL::;
+ unless $context<$_>:exists { # Probably called from NQP where there's no implicit $_
+ $context := { LEXICAL:: }();
+ }
08:10 finanalyst joined 09:53 lizmat left 11:12 lizmat joined 11:22 lizmat left 11:27 liztormato joined
liztormato PSA: back from travelling, now waiting for my new MBP to restore a full backup from my old MBP 11:28
I do not expect to be able to do a weekly today 11:29
11:33 liztormato left
[Coke] github.com/rakudo/rakudo/issues/5849 12:23
blin run had only 1 failure which I'm guessing is a false positive.
esp. as the most recent failing output says: 12:24
[Text::Markdown::Discount] Failed to get passing tests, but continuing with --force-test
... but there are no failing tests in the output
Geth setup-raku: skaji++ created pull request #40:
update package.json
13:10
13:15 JimmyZhuo joined
Geth setup-raku: 0f5fe214b0 | (Shoichi Kaji)++ | CHANGELOG.md
v1.9.0
13:15
finanalyst [Coke] ping 13:23
13:47 JimmyZhuo left
finanalyst .tell [Coke] i just sent you an email 13:48
tellable6 finanalyst, I'll pass your message to [Coke]
finanalyst .tell lizmat pls check email 13:49
tellable6 finanalyst, I'll pass your message to lizmat
13:53 finanalyst left 16:49 finanalyst joined
finanalyst [Coke] ping 17:06
[Coke] pong 17:09
finanalyst [Coke] I was just testing on my mobile 17:22
It seems to be a CSS problem. I was being too clever and styling based on the value of a data-attribute 17:23
I will rewrite to include an invisible checkbox. This sort of html/css styling is more widespread. Its the way I handle the REPL bar 17:24
18:31 lizmat joined, lizmat left, lizmat joined
lizmat is somewhat back 18:32
tellable6 2025-04-14T13:49:04Z #raku-dev <finanalyst> lizmat pls check email
finanalyst lizmat Hi somewhat back :) 18:33
lizmat ok, looks like I can't compile Rakudo on a M4Max CPU 18:52
actually, can't build NQP 18:53
perl Configure.pl --optimize --prefix=/Users/liz/Github/nqp/install --make-install 18:54
Can't exec "perl": Bad CPU type in executable at /Users/liz/Github/nqp/3rdparty/nqp-configure/lib/NQP/Config.pm line 1502.
releasable6 Next release in ≈4 days and ≈23 hours. 2 blockers. Please log your changes in the ChangeLog: github.com/rakudo/rakudo/wiki/ChangeLog-Draft 19:00
[Coke] weird. I don't see "Bad" in 3rdparty/nqp-configure that would match that 19:06
timo that's not something nqp-configure says 19:07
[Coke] ah. 32 bit perl maybe?
timo that's another part of the OS that says that
lizmat yeah, it's the OS that says that :-(
timo yeah, check what the perl it's trying to execute actually is
i don't know how you work with rosetta 19:08
[Coke] maybe change the sub around line 1502 to diag the system args first?
lizmat tried that, nothing wrong to see there :-( 19:09
timo the args are probably correct
[Coke] I assume new mac, you're forced to be on 15.x?
I'm on 14.7.5 here.
timo please also check the environment variables, like PATH
or alternatively 19:10
[Coke] could also be intel vs arm64?
lizmat It came with 15.x something, on 15.4 now
timo system a /bin/sh so you can type commands yourself and inspect from "the inside"
[Coke] just using sys perl?
lizmat hmmm.... double checking
timo it could be it's executing like an arm64 bash, then that tries to exec "perl" and finds an x86_64 one, and whatever is necessary for cross-architecture execing isn't properly set up in that particular case? 19:11
lizmat hmmm... could be a perl a built long ago
yeah, removing that fixed the issue it seems 19:12
gist.github.com/lizmat/708e871e2b9...d0badf71cc 19:14
now Moar doesn't want to build 19:15
timo which moar do you have there? we had a patch for that months ago i thought
lizmat heh good question 19:16
2022.06-9-gced366862
where did that come from?
timo good question :)
lizmat ah, it's still on "master"
timo that'd do it
lizmat I guess I'll do a fresh clone
timo in theory we can push a completely empty state to the master branch with like a readme 19:17
lizmat yeah, that might be a good idea 19:18
timo but i'm not sure how many machines are still out there that have a moar checkout that's from like mid 2022 19:24
lizmat In my case it was part of a Time Machine backup I used to set up my new MBP 19:30
hnmmm... looks like NQP's "make test" insists on a /usr/local/bin/perl 19:48
prove -j0 -r --exec '/Users/liz/Github/nqp/nqp-m' t/nqp t/hll t/qregex t/p5regex t/qast t/moar t/serialization t/nativecall t/concurrency
sh: /usr/local/bin/prove: /usr/local/bin/perl: bad interpreter: No such file or directory
timo sounds like ti has a hashbang 19:49
lizmat hmmm.... 19:50
timo please check if your nqp-m is up to date, too? it's supposed to be a binary nowadays, but it might be a text file
oh 19:51
actually, that sounds more like your prove is the thing with the hashbang?
lizmat aha... yes 19:52
there was a /usr/local/bin/prove
timo looks like your whole system is just full of land-mines just waiting to be stepped on one by one
lizmat now nuked
yeah 19:53
timo would be kinda nice to have some mac-native containers-like thing to have some amount of isolation and reproducibility 19:59
[Coke] wonder if minikube would help 20:03
timo that'll give you a linux
much like docker for windows will also give you linux containers 20:04
[Coke] I am seeing "mac os workloads" but no deets.
timo i can look into that
do you have a source where exactly you saw that? 20:05
[Coke] some google result ai hint, but that did lead to medium.com/agoda-engineering/how-w...0includes:
heh. step 1: start with a hundred or so mac minis. :) 20:07
timo ah, ok 20:09
no that post is mostly about making macos virtual machines manageable with your existing kubernetes-centric workloads
so it's not as containers-like as i'd like
[Coke] Yup, unsuprisingly, the ai snippet wasn't as helpful as it thought it was. :) 20:12
timo we can of course look into making liz's raku development more VM-focused, but stuff based on containers have certain benefits. a bit easier to run a tool like rak that actually lives inside a container to search through files on your host machine? possible with VMs of course, but probably a bit more fiddly? 20:13
ugexe i wonder how you end up with perl/prove in /usr/local/bin. afaik macs come with those in /usr/bin 20:24
if those aren't actually default macos binaries i would guess they are built for the wrong arch 20:45
timo that's my guess as well, yeah
they were perhaps compiled manually from source at some point and then put in the system with "make install"?
ugexe run `file /usr/local/bin/perl` and see what it says 20:46
could be, or something like homebrew
timo and when the backup put them back in place, they were of course not replaced by system-provided binaries with the correct arch, because the system doesn't have anything there
but they come earlier in the PATH, so they are preferred
even though they don't actually work - the PATH doesn't care, the program looking to invoke something doesn't get some kind of "retry with next candidate" semantics for free
ugexe file /usr/bin/perl
/usr/bin/perl: Mach-O universal binary with 2 architectures: [x86_64:Mach-O 64-bit executable x86_64] [arm64e:Mach-O 64-bit executable arm64e]
/usr/bin/perl (for architecture x86_64):Mach-O 64-bit executable x86_64
/usr/bin/perl (for architecture arm64e):Mach-O 64-bit executable arm64e 20:47
timo congrats on the 1.0.0 btw, that's a big symbolic step :) 20:49
wondering if Nil.iterator should be implemented to give an Iterator::Empty 22:03
m: my @foo; Nil.iterator.push-all(@foo); dd @foo
camelia [Any]
timo m: my $foo := IterationBuffer.new; Nil.iterator.push-all($foo); dd $foo
camelia (Nil,).IterationBuffer
timo so those would both give no elements
it's currently specced to give OneValue(Nil) 22:05
i will try a spectest to see what the fallout is 22:06
would be lovely if we had a cancellable "read bytes from stdin" 22:49
22:54 finanalyst left