Subj : Re: How-to on writing an interpreter?
To   : comp.programming
From : CTips
Date : Sat Jul 16 2005 10:49 am

Phlip wrote:
> C. Rebert wrote:
> 
> 
>>I'm trying to implement an interpreter for a programming language, but
>>so far haven't been able to find any good materials on how to write a
>>simple stack-based interpreter. I would appreciate it if someone could
>>point me to some resources on writing such an interpreter. I'd be open
>>to buying such materials, as long as they aren't ludicrously expensive.
> 
> 
> Industrial-strength parsers come from parser generators like Lex and Yacc.
> Learn them and live them; they have a thriving corpus and community.

I would disagree; "industrial strength parsers" tend *NOT* to be written 
using yacc/bison. There are several reasons, including:
- error handling
- inability to represent many languages (roughly - bison/yacc can handle 
  LALR(1) languages, but most languages are not LALR(1))

flex is a pretty decent scanner generator, but for a variety of reasons 
(including context-sensitive scanning) it doesn't get that much use 
industrihally.

You're better off asking about parsers/scanners as well as interpreters 
in comp.compilers.

If you just want to implement an interpreter, as opposed to an 
interpreter for a specific language, I'd recommend implementing an 
interpreter for Forth. Forth is a stack based language with an 
ultimately simple syntax and grammar (if one can call it that), so you 
can focus on the stack-based interpreter aspect. (It uses RPN so if 
you're an old-time HP calculator user, it will seem pretty obvious).

e.g. look at the following statement in forth

1 3 +

"1": means push 1 onto the stack
	*TOS++ = 1;
"3": means push 3 onto the stack
	*TOS++ = 3;
"+": means take the top two elements on the stack, add them together and 
push back onto the stack
	t0 = *--TOS;
	t1 = *--TOS;
	*TOS++ = t0 + t1;

Forth also has an existing GPL environment (gforth), so you can 
cross-check your results against that.

.