Interesting new language

[Jarrett Billingsley]

"Bruno Medeiros" <brunodomedeiros+spam at com.gmail> wrote in message 
news:f3p16p$2bfd$1 at digitalmars.com...
> LOOOOOOOOOOOOOOOOOOOOOOOOOL AWESOME. :D
> A friend of mine had mentioned this already, and I too was about to post 
> it here in the NG, despite flagrantly OT.
> More seriously though, like anderson mentioned, it would be really cool 
> (and maybe some good D publicity too) to implement a D lolcode compiler 
> with mixin text macros. Anyone up to the challenge? (I don't have time for 
> that unfortunately :S)

One problem with compiling the Lolcode into D code is that Lolcode seems to 
be dynamically typed.  We'd have to implement a variadic type, and things 
get even trickier if Lolcode functions, when a syntax is decided for them, 
are first-class and with variable numbers of parameters.

On the other hand, we already have the MiniD VM at our disposal, a 
dynamically-typed language VM with first-class functions with variable 
numbers of parameters ;)  The only thing that would have to be changed would 
be the lexer and parser; virtually all the Lolcode constructs have analogues 
in MiniD code, and can be represented by a MiniD AST.  This is all done at 
runtime, though.

If we still want to use compile-time constructs for compiling the Lolcode 
source, it might be better to compile to MiniD source, and then compile that 
at run-time.  Targeting the MiniD bytecode directly using compile-time code 
would probably be pretty awful :S