[antlr-interest] Possible incremental parse approach?

[Rick Mann]

On Oct 14, 2007, at 2:34 PM, Austin Hastings wrote:

> That sounds workable. But you should give some consideration to how  
> the user is going to "escape" when she gets into a parse context  
> and wants to reset back to just evaluating expressions.
>
> Some of what you're doing is basically a line editor. But the need  
> to be able to abort the parse (or suspend/resume it) and just do  
> commands is going to make life really complex.

Yeah, I think you're right about that. I may have to settle for "user- 
directed incremental parsing." Basically, you enter a number of lines  
of text, then hit a different key (on this Mac, <enter>) to signal  
that a complete translation unit has been typed.

Boo.

-- 
Rick