[04:06] *** Geth joined
[04:06] *** lizmat joined
[04:08] <Geth> ¦ docker: 3366cd371d | AntonOks++ | 6 files

[04:08] <Geth> ¦ docker: Bump to 2026.04 [skip workflow]

[04:08] <Geth> ¦ docker: review: https://github.com/rakudo/docker/commit/3366cd371d

[04:33] *** mehbark left
[04:34] *** mehbark joined
[05:01] *** mehbark left
[05:01] *** sibl left
[05:02] *** sibl joined
[05:02] *** mehbark joined
[05:08] *** sibl left
[05:09] *** sibl joined
[05:12] *** sibl left
[05:14] *** mehbark left
[05:15] *** mehbark joined
[05:15] *** sibl joined
[05:18] *** sibl left
[05:19] *** atcroft joined
[05:20] *** sibl joined
[05:30] *** sibl left
[05:34] *** sibl joined
[05:36] *** sibl left
[05:36] *** sibl joined
[05:39] *** sibl left
[05:39] *** sibl joined
[05:43] *** ShimmerFairy left
[05:43] *** sibl left
[05:44] *** ShimmerFairy joined
[05:52] *** Aedil joined
[05:56] *** abraxxa joined
[06:01] *** abraxxa left
[06:03] *** abraxxa joined
[06:11] *** sibl joined
[06:49] *** peder joined
[06:57] *** Aedil left
[07:01] *** Sgeo left
[07:35] *** sibl left
[07:36] *** sibl joined
[07:40] *** sibl left
[07:41] *** sibl joined
[08:10] *** sibl left
[08:11] *** sibl joined
[08:21] *** sibl left
[08:27] *** sibl joined
[10:06] <apogee_ntv> All of my regularly updated packages CI is now free of node 20.

[10:06] <disbot2> <librasteve> \o/

[10:06] <apogee_ntv> If anyone uses msvc-dev-cmd in CI feel free to use my fork: https://github.com/m-doughty/msvc-dev-cmd

[10:08] <apogee_ntv> I had to bump some deps to make it pass npm audit.

[10:10] *** sibl left
[10:11] *** abraxxa left
[10:14] *** abraxxa joined
[10:15] *** ShimmerFairy left
[10:15] *** ShimmerFairy joined
[10:16] <disbot2> <librasteve> npm 🤮 🤮 🤮

[10:19] <apogee_ntv> Yes, hate node but most of the actions use it :(

[10:19] *** abraxxa left
[10:19] <apogee_ntv> Almost as bad as python

[10:20] <apogee_ntv> I would rather write C++ than either of them

[10:21] <apogee_ntv> with cmake

[10:22] *** abraxxa joined
[10:22] *** librasteve_ joined
[10:23] <apogee_ntv> At least the C++ will probably still work in 5yrs.

[10:23] <librasteve_> i have made a toy DSL - can use to test Selkie::UI … https://github.com/librasteve/Actionable/blob/main/invoice_dsl.raku

[10:24] <librasteve_> the output is text - like:-

[10:24] <librasteve_> https://www.irccloud.com/pastebin/SiJW4V5P

[10:24] <disbot2> <librasteve> Smokemachine: ^^

[10:25] <disbot2> <librasteve> so no need for HTML / PDF rendering in the view (at least not for MLP)

[10:25] <disbot2> <librasteve> would be good to have a view on Grammar::Tracer

[10:27] <apogee_ntv> I'm currently building a bundler that pulls a bunch of stuff for Selkie apps to be 1 click install across platforms.

[10:27] <apogee_ntv> The pain is real

[10:27] <apogee_ntv> Why cant all platforms just use flatpak like civilized OS's?

[10:28] <apogee_ntv> Apple has flatpak but you pay $99/yr for it

[10:28] <apogee_ntv> Windows has hey you should write stuff everywhere and btw uninstallers are really fun to write.

[10:33] *** annamalai left
[10:35] *** annamalai joined
[10:40] *** annamalai left
[10:40] *** annamalai joined
[10:44] <apogee_ntv> I'm running into an intermittent spesh bug I think on 2026.03 -- it's hard to nail a repro but it's trying to cast a P6bigint into int and crashing.

[10:45] <apogee_ntv> I got around it by adding a nqp::isbig_I check and no-oping on 1 but not sure why the misspec is being hit, I need to deep dive into spesh code at some point.

[10:47] <apogee_ntv> (the bigint seems to be confined to 1 hot pane on terminal resize where it's probably the OS doing something stupid and reporting a massive coord)

[10:49] *** annamalai left
[10:49] *** annamalai joined
[11:04] <apogee_ntv> Might be a missing deopt guard

[11:05] <apogee_ntv> (I am guessing, I haven't explored yet)

[11:05] <lizmat> if you're running on Intel hardware, it might be a JIT issue as well

[11:06] <lizmat> MVM_JIT_DISABLE=1 would disable that

[11:13] <apogee_ntv> I'm on M4 Max.

[11:14] <apogee_ntv> Apple Silicon. I do have an Intel machine I can test on.

[11:14] <apogee_ntv> Xeon though

[11:14] <apogee_ntv> Not sure if that matters

[11:15] <lizmat> if it happens on Apple silicon, it's not the JIT  :-)

[11:15] <lizmat> as ARM processors don't hae JITv

[11:15] <apogee_ntv> I need to learn moarvm code I think :D

[11:15] <apogee_ntv> Then I can be helpful :D

[11:16] <apogee_ntv> Any good reading or just read the source?

[11:17] <lizmat> 6guts.wordpress.com  has a lot of instructive materials

[11:18] <lizmat> remastered versions at: https://raku.github.io/CCR/Remaster/Jonathan%20Worthington/

[11:26] *** masayoshi joined
[11:27] <m6locks> did jnthn quit

[11:33] <lizmat> I like to think of him being on an extended undetermined length leave

[11:34] <apogee_ntv> Thanks, will make sure to read through.

[11:34] <apogee_ntv> Happy to help with VM stuff where I can.

[11:36] <lizmat> ++apogee_ntv 

[11:39] <lizmat> any MoarVM specific questions should probably be asked on #moarvm  :-)

[11:39] *** masayoshi left
[11:51] <apogee_ntv> Yeah I am in there :D

[11:53] *** oodani left
[11:57] *** oodani joined
[11:57] *** annamalai left
[11:58] *** annamalai joined
[11:58] *** annamalai left
[11:58] *** annamalai joined
[12:05] *** annamalai left
[12:06] *** annamalai joined
[12:10] *** annamalai left
[12:10] <disbot2> <librasteve> massive kudos to jnthn - he did an enormous amount of work and had the vision for many aspects of where we are now - Cro, AST to mention but two

[12:11] *** annamalai joined
[12:18] *** annamalai left
[12:23] <apogee_ntv> K trying to do something ambitious today: mouse support in Selkie

[12:23] <apogee_ntv> fml why do I torture myself

[12:26] *** annamalai joined
[13:01] *** annamalai left
[13:02] *** annamalai joined
[13:56] *** kurahaupo left
[13:56] *** kurahaupo joined
[14:51] *** Aedil joined
[15:20] *** apogee_ntv left
[15:21] *** apogee_ntv joined
[16:04] <apogee_ntv> Pushing Selkie 0.6.0 with mouse support.

[16:09] *** abraxxa left
[16:30] <librasteve_> apogee_ntv ++

[16:33] *** justache left
[16:35] *** justache joined
[17:44] <lizmat> apogee_ntv++

[17:55] *** human-blip left
[17:56] *** human-blip joined
[18:00] <apogee_ntv> https://raku.land/zef:apogee/Noise::Simplex::Native ~2900x faster than Noise::Simplex

[18:01] <timo> i bet a mature C compiler can do a lot with SIMD in a simplex noise implementation

[18:08] <timo> apogee_ntv: i'd be interested in details about the spesh bug you've spotted in 2026.03

[18:12] <timo> if it reliably crashes and otherwise behaves deterministically, you can use the spesh bisect tool to find the first specialization that is created in the crashing case

[18:12] <timo> if the crash is hard to provoke reliably, a recording with `rr` can be extremely useful. `rr pack` makes it possible to share with other computers, though I'm not sure about different architectures

[18:13] <lizmat> timo: afaik, apogee_ntv is on Apple silicon

[18:14] <timo> ios on apple silicon or something like asahi linux?

[18:14] <timo> sorry, macos obviously

[18:14] <lizmat> yeah

[18:15] <timo> then I don't think rr would run

[18:30] <timo> but hopefully it reproduces cleanly on linux and not just on macos

[19:55] *** abraxxa joined
[20:09] *** corwin left
[20:35] *** corwin joined
[20:37] *** abraxxa left
[20:44] *** corwin left
[20:51] *** corwin joined
[20:52] *** eseyman left
[20:52] *** abraxxa joined
[21:14] *** manu_ joined
[21:26] *** abraxxa left
[22:02] *** cm_ joined
[22:03] *** cm left
[22:03] *** cm_ is now known as cm

[22:04] *** apogee_ntv left
[22:05] *** apogee_ntv joined
[22:08] *** Sgeo joined
[22:27] *** dustinm` left
[22:41] *** dustinm` joined