[00:11] *** grondilu joined
[00:34] *** kurahaupo left
[00:41] *** tjr left
[00:44] <wayland76> So I know we have/had backends for moarvm, jvm, and javascript.  Have we ever considered partial backends, ie. whether parts of Raku could be compiled to SQL?  

[00:45] *** tjr joined
[01:22] *** hulk joined
[01:23] *** kylese left
[02:09] *** kst joined
[02:15] *** hulk left
[02:15] *** kylese joined
[02:21] *** wayland76 left
[03:22] *** lichtkind_ joined
[03:25] *** lichtkind left
[04:39] *** wayland76 joined
[04:46] *** vrurg joined
[04:48] *** vrurg_ left
[04:52] *** kurahaupo joined
[05:02] *** wayland76 left
[06:05] *** kurahaupo left
[06:05] *** kurahaupo1 joined
[06:12] *** Sgeo left
[07:17] *** kurahaupo1 is now known as kurahaupo

[07:20] *** rnddim left
[07:21] *** ShimmerFairy joined
[07:25] *** belluzj joined
[07:26] *** belluzj left
[07:53] *** belluzj joined
[08:08] <Voldenet> I wouldn't say that compiling ast to sql is a backend, more like a translation lib

[08:10] <Voldenet> it's a good idea to have stuff like this overall, but it shouldn't built into compilers imo

[09:11] *** polettix joined
[09:17] *** tjr left
[09:28] *** polettix left
[09:28] *** polettix joined
[09:57] <apogee_ntv> mi6 question, is there a RunBeforeBuild that just runs *once* on mi6 release?

[09:58] <apogee_ntv> I have Selkie docs gen in RunBeforeBuild but it runs multiple times during a release and makes it tediously slow

[10:30] *** wayland76 joined
[10:32] <wayland76> Voldenet: Agreed.  Well, I think we should (eventually) have a standard querying lanugage built into Raku, but the compile-to-SQL option should be an addon.  

[10:51] <disbot> <melezhik.> Voldenet: would you like to try out my very alpha version of dsci git server ?

[10:51] <disbot> <melezhik.> I already have something to show

[12:02] *** wayland joined
[12:02] *** wayland76 left
[12:07] <apogee_ntv> melezhik. Just saw your issue on Notcurses::Native and replied, I think you are on the most recent unsupported rakudo for NativeCall varargs syntax :)

[12:07] <apogee_ntv> Looks like it was added in the changelog for 2025.12

[12:08] <apogee_ntv> Your runner is on 2025.11

[12:11] *** belluzj left
[12:23] *** belluzj joined
[12:41] *** SmokeMachine left
[12:42] *** xinming left
[12:43] *** swaggboi left
[12:43] *** summerisle left
[12:43] *** ugexe left
[12:43] *** summerisle joined
[12:44] *** xinming joined
[12:55] <tbrowder> apogee_ntv: i'm looking at your run-before-build line. what is the %x supposed to do?

[12:56] *** SmokeMachine joined
[12:57] *** swaggboi joined
[12:57] *** ugexe joined
[13:02] <apogee_ntv> tbrowder: It's expanded to $*EXECUTABLE, in this case 'raku'

[13:04] <apogee_ntv> Thats whats in the github docs so I copied it assuming its a preferred convention instead of hardcoding 'raku'

[13:24] <disbot> <melezhik.> Sure, will try with fresh Rakudo and let you know

[13:25] <apogee_ntv> ty

[13:46] <apogee_ntv> You did help me find a different bug with the precompiled binaries

[13:46] <apogee_ntv> It shouldn't have tried to fall back to native compilation for the shim

[13:48] <tbrowder> got it. no clue as to 3x build, but nice output anyway

[14:01] <[Coke]> does class X { has $no-twigil } ... do anything useful? Should we make that a warning?

[14:02] <[Coke]> Just spent 10m trying to figure out why my attribute wasn't getting set on instantiation. :)

[14:10] <tbrowder> apogee_ntv: have you tried having chapgpt analyzing that code for some kind of circular path?

[14:11] <tbrowder> *ChatGPT

[14:11] <apogee_ntv> tbrowder: Claude thinks the issue is in mi6: "Easy, and the mi6 bug is structural: release calls build, and the RegenerateFiles + DistTest steps each call build again, so RunBeforeBuild fires 3×."

[14:14] *** belluzj left
[14:17] <tbrowder> yeah, mi6 does some complex stuff, @skaji might take the challenge if you file an issue.

[14:18] *** belluzj joined
[14:41] <apogee_ntv> I've been trying to look at release trains tbh, some of my packages are... a pain to release to say the least lol. The delay on things turning up in zef means the dependent package has to wait to run CI until it exists. I have packages with 3-4 levels of dependencies and sometimes I touch all of them.

[14:42] <apogee_ntv> Notcurses::Native -> Selkie -> Any Selkie app for example, or LLM::Chat -> LLM::Data::Inference -> Consumer

[15:21] <apogee_ntv> Speaking of Selkie, pushing 0.13.0 with typed store.

[15:54] *** belluzj left
[17:25] *** sp1983 joined
[17:26] <sp1983> weekly: http://dsci.sparrowhub.io is alive - minimalistic Self Hosted Git Server + CI with general languages ( including Raku ), use this link as demo server

[17:26] <notable6> sp1983, Noted! (weekly)

[17:26] <sp1983> ^^ Voldenet

[17:26] <sp1983> everyone indeed is welcome to test it out - documentation is here - http://deadsimpleci.sparrowhub.io/

[17:29] <sp1983> so it can be installed in any Linux VM, as binary

[17:30] <sp1983> to run CI all you need is docker or podman

[17:30] <sp1983> one gets self hosted git server plus CI ( dsci )

[17:30] <sp1983> as a result

[17:32] *** sp1983 left
[17:53] *** wayland left
[17:53] *** wayland joined
[19:05] <disbot> <melezhik.> Example of git repo - http://dsci.sparrowhub.io/repo/demo-raku-cro.git

[19:52] *** jdv joined
[21:45] <apogee_ntv> melezhik any luck with Notcurses?

[22:16] *** sftp left
[22:54] *** sftp joined
[22:54] *** sftp left
[22:54] *** sftp joined
[23:13] *** sftp left
[23:26] *** sftp joined
[23:26] *** sftp left
[23:26] *** sftp joined
[23:27] *** Sgeo joined