[00:22] <japhb> [Coke]: https://gist.github.com/japhb/09b54273146ecc5a7863854dfd8b0677

[01:42] *** hulk joined
[01:43] *** kylese left
[01:44] <[Coke]> japhb: looks great. want to submit a PR to put in in util/ in raku/blin so you get the commit credit?

[02:10] <japhb> Oh sure, give me a couple

[02:15] *** hulk left
[02:15] *** kylese joined
[02:19] <japhb> [Coke]: https://github.com/Raku/Blin/pull/57

[02:23] *** rnddim joined
[02:23] *** rnddim left
[02:24] *** rnddim joined
[02:24] *** ShimmerFairy left
[02:31] <[Coke]> japhb++

[02:47] *** rnddim is now known as ShimmerFairy

[03:10] <[Coke]> japhb: using on the current blin run, thanks!

[03:16] <japhb> <3

[03:33] *** ShimmerFairy left
[03:33] *** ShimmerFairy joined
[03:37] *** ShimmerFairy left
[03:38] *** ShimmerFairy joined
[04:38] *** vrurg_ joined
[04:39] *** vrurg left
[05:23] *** Sgeo left
[06:36] *** camelia joined
[07:59] *** dakkar joined
[08:10] *** jcallen left
[08:11] *** jcallen joined
[09:02] *** bazzrrr0 joined
[09:10] <bazzrrr0> Does anyone know whether El_Che intends to add a sister service providing pre-built rakupp native package binaries, to go alongside the current very useful rakudo-pkg service?

[10:04] *** hurufu joined
[10:28] *** hulk joined
[10:29] *** kylese left
[11:11] *** bazzrrr0 left
[11:13] *** xinming joined
[11:52] *** hurufu left
[11:52] *** hurufu joined
[12:43] <tbrowder> bazzrrr0: it seems github is causing problems with rakudo-pkg generation. i'm in process of using another method for my systems. i hate to do it, but i like to upgrade rakudo asap.

[12:43] <tellable6> tbrowder, I'll pass your message to bazzrrr0

[12:44] <El_Che> the packages are up

[12:44] <tbrowder> ah, just in the nick of time, thanks!

[12:49] <tbrowder> painless, as usual, very cool!!

[13:09] <tbrowder> bazzrrr0: ^^

[13:09] <tellable6> tbrowder, I'll pass your message to bazzrrr0

[13:17] *** hurufu left
[13:32] *** hurufu joined
[13:47] *** hurufu left
[13:47] *** Sgeo joined
[13:50] *** hurufu joined
[13:54] *** jjido joined
[14:21] *** jjido left
[14:37] *** jjido joined
[15:00] *** jjido left
[16:11] *** hurufu left
[16:43] <[Coke]> japhb: on windows, the refresh on the script is still showing up (lots of: counts{'Fail'} of type Any I can see in the once-a-second flash)

[16:44] <[Coke]> sorry: running on linux under tmux, then ssh'd in from windows.

[16:44] <[Coke]> ssh'd in from *mac*, no flashes. (So maybe not the Terminal module's problem)

[16:45] <japhb> That's really odd ... it shouldn't have any difference in behavior between the OSen, I tried to be completely minimal.

[16:45] <japhb> I can imagine that a terminal that didn't do at least 256 colors would be confused, but even that is a positively ancient requirement.

[16:45] <japhb> What terminal are you running on Windows?

[16:45] <[Coke]> powershell

[16:46] <japhb> Is that using Windows Terminal or Windows Console?

[16:46] <[Coke]> an older one, probably.

[16:46] <[Coke]> no idea?

[16:46] <japhb> Ah, yeah, old versions of Windows Console were ... LTA.

[16:47] <japhb> The 'type Any' bit is weird to me too, because I thought I caught all of those cases with my final version.

[16:48] <[Coke]> well, I did commit a change. :)

[16:48] <japhb> if you pipe stderr to a file, can you see which line the errors are happening on?

[16:48] <japhb> :-D

[16:48] <[Coke]> I think this is windows terminal (has tabs across the top)

[16:48] *** dakkar left
[16:49] <[Coke]> hilariously, it is me, I think.

[16:50] <[Coke]> ok, it *IS* me, but I swear the flashing didn't happen when starting on mac. :)

[16:51] <[Coke]> https://github.com/Raku/Blin/commit/b305cd5a37e2ca658d03512ad207462410103f19

[16:51] <[Coke]> japhb++

[16:53] <[Coke]> Thanks again. This is so much nicer than the simple one I had.

[16:53] <japhb> Ah, I was wondering if you wanted a running total.  :-)

[16:53] <japhb> You're quite welcome!  'Twas fun.  :-)

[16:55] <[Coke]> speaking of blin, the pancake emoji in windows font looks very different. :)

[16:59] *** human-blip left
[17:01] *** human-blip joined
[17:55] <[Coke]> japhb: huh: Died because of the exception:  Malformed UTF-8

[17:57] <[Coke]> japhb: weirdly, I restarted it, and it's running again? (wonder where it could have seen bad data during the middle of a run)

[17:58] <japhb> Maybe it caught the overview being partially written?

[17:58] <[Coke]> ah, maybe.

[17:58] <[Coke]> Should probably have it just give up that refresh and try again in 1s.

[18:00] <japhb> Yeah, a catch around the parsing contents would help.  Might need to write the overview to a temp file and atomic-rename it into place to avoid that happening in the first place.

[18:00] <japhb> Of course "Both is good."

[18:04] <[Coke]> yah, blin had no reason to think anyone else was reading it when it was dumping it

[18:04] <japhb> nodnod

[18:06] <[Coke]> of course, I won't be able to easily test that the fix fixes it. :)

[18:32] <[Coke]> aaand, there's a bug in output/overview handling in blin. One of the records is this:

[18:32] <[Coke]> Foo::

[18:32] <[Coke]> Foo – AlwaysFail

[18:32] <El_Che> hi

[18:35] *** belluzj joined
[18:51] <disbot> <antononcube> Trying to install one of my NativeCall packages, "Math::SparseMatrix::Native", with "moar-2026.09" fails with: >  This representation (CStruct) cannot unbox to a native int (for type Math::SparseMatrix::Native::CSRStruct)

[18:51] <disbot> <antononcube> Is there a documented way to figure out how to fix this?

[18:52] <disbot> <antononcube> I assume, I can see a NativeCall package that is installable with "moar-2026.09".

[19:03] *** bazzrrr0 joined
[19:13] <ugexe> there is an error in your module that is just exposed now

[19:13] <ugexe> https://github.com/antononcube/Raku-Math-SparseMatrix-Native/blob/af05ae9f0726434abe90d7a8e15d4a7190705e1a/lib/Math/SparseMatrix/Native.rakumod#L478

[19:13] <ugexe> that returns `self`, but the return type is `int32` https://github.com/antononcube/Raku-Math-SparseMatrix-Native/blob/af05ae9f0726434abe90d7a8e15d4a7190705e1a/lib/Math/SparseMatrix/Native.rakumod#L459

[19:15] *** belluzj left
[19:30] *** bazzrrr0 left
[19:41] <disbot> <antononcube> Thank you!

[20:00] <ugexe> error message could certainly be improved

[20:13] <disbot> <antononcube> I was thinking my mistake is much more fundamental.

[20:14] *** jdv_ is now known as jdv

[20:15] <ugexe> m: sub foo(--> int32) { "x" }; say foo()

[20:15] <camelia> rakudo-moar e17a1d43b: OUTPUT: «x␤»

[20:15] <ugexe> basically that is fixed now

[20:15] <[Coke]> ... why is it not fixed here?

[20:15] <[Coke]> c: HEAD sub foo(--> int32) { "x" }; say foo()

[20:15] <committable6> [Coke], ¦HEAD(5136581): «This type cannot unbox to a native integer: P6opaque, Str␤  in sub foo at /tmp/8MCvfYhLoH line 1␤  in block <unit> at /tmp/8MCvfYhLoH line 1␤␤ «exit code = 1»»

[20:16] <[Coke]> guessing m needs an update to at least 2026.09?

[20:16] <ugexe> m: say $*RAKU.compiler.version

[20:16] <camelia> rakudo-moar e17a1d43b: OUTPUT: «v2026.08.407.ge.17.a.1.d.43.b␤»

[20:17] <[Coke]> that'd do it.... I bet it died when we made RAKUAST the default.

[20:34] <disbot> <antononcube> 🙂

[21:19] *** swaggboi left
[21:29] *** swaggboi joined
[21:42] *** johnjay left
[21:45] *** snonux left
[21:47] *** snonux joined
[21:51] *** lizmat left