Re: [s-d] dice server enhancements, part 2

[Antonio Dolcetta <antonio.dolcetta@xxxxxxxxx>]

On 21 Dec 2006, at 23:12, Joel Uckelman wrote:

> So, I've encountered a difficulty that maybe one of you here is  
> familiar
> with: I made the observation just now that my original rolling  
> language
> is easy to evaluate because you can parse it into a tree. The whole
> program is an expression, possibly composed of subexpressions, etc.
> But once you have things like function calls and jumps, it's not so
> obvious (to me) that programs in such a language can or should be
> represented as trees. I think this is something I'd know about if  
> I'd ever
> had a compilers course, but alas, I've not.
>
> Thoughts?

I've never had courses about compliers either, but here's my 2 cents:
i think that the implementation of the language should remain the  
same as it was before, an expression evaluator. The function calls,  
jumps etc. should be implemented via a wrapper that calls the   
evaluator when needed.


_______________________________________________
spoon-discuss mailing list
spoon-discuss@xxxxxxxxx
http://lists.ellipsis.cx/mailman/listinfo/spoon-discuss