[OT] Partial-reparsing thory?

[Nick Sabalausky]

Is anyone aware of any (formal or informal) 
theory/information/research/articles/etc. that's already been done on the 
idea of reparsing *part* of an already-parsed source? Either lexing, parsing 
or both together. Preferably LR, but LL might be helpful, too. (I'd imagine 
it may be notably more difficult with LR. May need to keep track of all the 
parsing steps taken the first time around, and may need to continue 
re-parsing to the end of the source.)

I know code editors deal with that sort of thing a lot, but my understanding 
is they often accept a certain amount of inaccuracy (for the sake of being 
more line-oriented) and also tend to be more lexing-oriented and very light 
on parsing.