Hi, I think what you're looking for is Sylvain Schmidt's paper entitled "Modular syntax demands verification". IIRC in general the problem is undecidable, but Sylvain develops a conservative NFA approach and demonstrates it on a tricky ML grammar snippet. It's far from trivial though.
Cheers, Colin On 27 May 2010 19:57, Panciz <dav.pole...@gmail.com> wrote: > > Hi, > > I'm wrinting a tool to that collect sparse single grammr productions > and combine them together in order to generate the complete grammar > definition. > > In particular the colleted rules are transformed in Rats! modules in > order to generate a Java Packrat Parser. > > My problem is that I need to give to the productions that define the > same non terminal the correct priority (that in > Rats! means the definition order in the module) in order > to ensure that a shorter rules does not prevent a > longer one to be choosen. > > Is there an efficient strategy to automatically check this constrain?? > It's trivial in simple grammars like > > NonTerm1 <-- NonTerm3 / NonTerm2 > NonTerm2 <-- NonTerm3 "sym" > > find out that NonTerm3 must have low priority than NonTerm2. > but when I have something like this > > NonTerm1 <-- NonTerm2 / NonTerm3 > NonTerm2 <-- NonTerm2A / NonTerm2B / NonTerm2C > NonTerm3 <-- NonTerm3A / NonTerm3B > etc..... > > In order to compare NonTerm3 with NonTerm2 I need to check for their > definition > and then recursevly check in their definition until I found > out if NonTerm2 prevents or not the choice NonTerm3. > > Thanks > Davide > > > > > _______________________________________________ > PEG mailing list > PEG@lists.csail.mit.edu > https://lists.csail.mit.edu/mailman/listinfo/peg > _______________________________________________ PEG mailing list PEG@lists.csail.mit.edu https://lists.csail.mit.edu/mailman/listinfo/peg
