Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
Nicholas good *, * 07:34
nine Good *, indeed 08:19
Geth MoarVM/master: 7 commits pushed by (Stefan Seifert)++ 15:01
lizmat meh, looks like the timeout logic of IRC::Client still needs some tuning 15:06
MasterDuke could developers.redhat.com/blog/2020/06...un-speeds# be relevant for moarvm? 15:37
github.com/rakudo/rakudo/issues/4729 ugh, bigint ops also need the same fixing as nativecall 15:56
another reason to switch to gmp, you can pass it allocation functions to use 15:57
lizmat does it? the version used there misses at least 23 commits ?
aaah,. that's related to mimalloc, not the uint merge 15:58
lizmat goes silent again
MasterDuke Nicholas++ has been on a roll fixing/finishing my incorrect/incomplete PRs, maybe i can trick^h^h^h^h^h convince him to look into the gmp one after i rebase it... 16:00
Nicholas and the reward is cake? 16:01
(I am not familiar with the game, but I am aware of it)
MasterDuke virtual cake now, physical beer if we ever meet in-person 16:02
lizmat nine: P5getnetbyname got borked 16:09
===SORRY!===
Calling _getnetbyaddr(uint, int) will never work with declared signature (int32, uint32 --> P5getnetbyname::NetStruct)
I guess that's an error that previously went unnoticed
nine Yeah, definitely looks wrong in the module. Previously uint was essentially the same as int, so we didn't mind 16:12
We can be more lenient with this. Even C compilers wouldn't complain about this case. But I wonder where the line is between the compiler detecting this being helpful and it being just noise. 16:13
lizmat no, it's fine as far as I am concerned :-) 16:14
lizmat nine: looks like IO::Socket::Async::SSL got borked 16:43
This container does not reference a native unsigned integer
in block at /Users/liz/Github/rakudo.moar/install/share/perl6/site/sources/B354B3F623BE4824406EEA4C8A5D0C9910B12079 (IO::Socket::Async::SSL) line 768
which in turn breaks Cro::TLS
nine steps to reproduce? 16:47
tests seem to run fine here
lizmat running my REA update script 16:50
hmmm... I was already considering not using Cro::HTTP::Client, but instead of just shelling out to curl 16:52
nine What does --ll-exception point at? 16:53
lizmat This container does not reference a native unsigned integer
at SETTING::src/core.c/Buf.pm6:187 (/Users/liz/Github/rakudo.moar/install/share/perl6/runtime/CORE.c.setting.moarvm:)
nine Are you on 2021.12-141-g3823f88ce?
lizmat aaah... need to pull rakudo :-) 16:54
am now :-)
MasterDuke actually, now i'm not so sure the bigint functions do need anything. we don't directly alloc/free their internals, libtommath does that and it's just using regular libc calls. not exactly sure what the problem is 17:08
japhb Welp, with nine++'s latest commits, it appears Cairo can build/test cleanly now. However NativeHelpers::Array is failing. :-/ 21:37

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!