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 *, * 05:20
nine Good summer! 07:32
Nicholas "while stocks last" 07:56
lizmat nine: re github.com/rakudo/rakudo/commit/d8ff3ade1e is that something that should be fixed on master as well ? 16:39
nine Unless someone intends to give precompiled scripts another try, I don't see a need to rush. 16:41
If you are curious, you could test if we actually need that GLOBAL preserving in Perl6::ModuleLoader at all anymore. Since that only loads NQP modules, I don't see how Raku's GLOBAL could get overwritten 16:42
lizmat will check 17:02
nine Oooooh! find-lexical on a scope caches the negative. So my check for conflicts on IMPORT actually prevents those imported symbols from getting found 18:48
Well, it doesn't actually cache a negative. It's more that it generates the list of generated lexical declarations just the once and caches the result, preventing future additions from getting found. 18:51
lizmat I think I just had a record in the Rakudo code base: fixed a problem introduced in 2012, and *nobody* has noticed that until now 19:04
nine Could be. My ModuleLoader bug was from 2013 19:06
400! 19:16
The 4 additional spec tests failed before because they `use Test;` twice: github.com/Raku/roast/blob/master/...-0.t#L2306 19:18
MasterDuke os just updated to gcc 12 and there's a new warning 22:54
compiling src/jit/tile.o
src/jit/tile.c: In function ‘MVM_jit_tile_make’:
src/jit/tile.c:51:25: warning: writing 1 byte into a region of size 0 [-Wstringop-overflow=]
   51 |         tile->values[i] = (MVMint8)va_arg(arglist, MVMint32);
      |                         ^
In file included from src/moar.h:286,
                 from src/jit/tile.c:1:
src/jit/tile.h:21:14: note: at offset 4 into destination object ‘values’ of size 4
   21 |     MVMuint8 values[4];
      |              ^~~~~~
drakonis sounds like gcc added some new checkers 23:19
hot take: how long would it take to rewrite moarvm with rust
Voldenet probably depends on whether you're willing to start today ( ¬‿¬) 23:23
drakonis its wayyyy out of my depth 23:24
japhb drakonis: A long time. 23:29
tellable6 2022-05-10T20:29:09Z #raku <tonyo> japhb i understand how macros work, referring mostly to the trend of code generation vs just using macros (whether it's syntactic or semantic). go's use of code generation, in practice, is abhorrent 23:30
drakonis it does that because it used to lack generics 23:33
it offers it now
still no macros
Voldenet golang does still lack proper generics 23:34
or rather the strategy of executing generics is peculiar
drakonis it seems entirely in line with golang's modus operandi 23:35
Voldenet and regarding rust, it probably isn't good language for this 23:37

Powered by Cro and the Raku® Programming Language.

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

Thank you!