Kristian Larsson

The generated code, whole main init function is this: ```c $R protofuQ_mainD___init__ (protofuQ_main self, $Cont C_cont, B_Env env) { self->f = protofuQ_FooG_new(protofuQ_W_8); ((B_NoneType (*) (protofuQ_Foo, B_Ord, B_int))self->f->$class->insert)(self->f, ((B_Ord)to$int(1))); ////////////////////////// bad...

I'd really prefer if we can explain all of this to someone that doesn't know what boxed values are. I don't think that really changes the proposal in itself but...

Hmm, so I've run into this issue again (#1495 really) and... I'm in quite dire need of a fix. I have to admit that I don't remember exactly what we...

I don't really buy the "the programmers who choose to use these types for some space-saving reason should be trusted to know what they are doing". People make mistakes, even...

Bah, I realize I messed up the CFLAGS when I separated them out for the DB in #1066 - they no longer include -target. Not sure if that actually affects...

I'm thinking maybe that overshadowing should require an extra keyword rather than being the default (depending on context?)? I think some language do this already...

zig doesn't allow shadowing. another point to disallow this... :)

`--trusted`! I think it should be called "trusted modules" and thus `--trusted`

xoshiro256 is already in zigs standard library, so perhaps this can be our first stab at using zig stuff? This way we avoid introducing yet another C library dependency. In...

We now use a particular version of zig so we don't rely on a system CC and thus we can always use --cpedantic. --cpedantic actually broke a bit with the...