Hello Sérgio,
> We have discussed (and proved) this correspondence here: > https://www.sciencedirect.com/science/article/pii/S0167642314000276 > https://arxiv.org/abs/1304.3177 > Very nice! Thanks! Of course, the theoretical proves are bound to what's possible, but provide little guidance as to how to produce something that is convenient (a PEG grammar with LL(k) lookaheads on each choice is not convenient, for example). > I think it would be valuable if you could present in more detail which are > these problematic patterns, > this would help users of PEG-based tools to avoid them. > I know I've done PEG grammar optimization in the past, but I haven't documented what I've done. I will this time. I know that part of what I do is to manually insert *cut* expressions to avoid choices that will fail (I should look again at the *automatic* cut insertion in Mizushima et. al). -- Juancarlo *Añez*
_______________________________________________ PEG mailing list PEG@lists.csail.mit.edu https://lists.csail.mit.edu/mailman/listinfo/peg
