timotimo nekovm.org/ - i wonder if this has anything interesting in it? 17:41
nekovm.org/doc/nxml - they have their own way of writing asts directly in their code, like our nqp::stmts and such nodes 17:47
their FAQ contrasts neko vs parrot 17:50
Neko is using the Boehm GC, which is a conservative multithreaded mark and sweep collector. However, since all calls to the GC are wrapped by the Neko API in vm/alloc.c, it might be possible to easily switch to another garbage collector in the future. 17:51
that's cute :P
jnthn I'd guess a bunch of options will run into needing write barriers. 18:35
"a single VM shouldn't be used to execute code on several threads at the same time."
Guess we won't be borrowing any neat threading ideas :) 18:36
brrt interesting thing 19:59
japhb What's the MoarVM idiom for writing out debugging info, akin to fprintf(stderr, ...) but safe WRT threads, GC, libuv, etc.? 23:22
I'm trying to figure out the problem with getting Malformed UTF-8 on binary stdout from an async proc, and I can't see where the binaryness is getting lost, so now I'm going to try debug output ... 23:25
geekosaur had been wondering... was it stdout, or stderr? 23:35
japhb stdout
Thankfully, all the IO happens on a single thread, so I won't get a complete mess in the debug output, but it is callback-and-bespoke-structure hell because libuv, so it will be "interesting" to figure out. 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!