Ondřej Bílka Your comment:
> Longest match just tries to hide problems under the rug. > Instead (A | A B) you have problem with (A B | A) B not matching AB True, it's the same problem, but I don't think there is anything hidden under the rug... That's the price for a committed choice grammar such as PEG. As Roman says, it would be nice to have some way to structure how we express things... but in practice its does not seem to be a big issue. I found I made more mistakes with priority choice (wrong order, or a first choice that could match empty), so I much prefer the longest choice. I still don't think there is a significant speed problem: thanks Alex, your experience sounds similar to mine: but Lowell seems to have had less luck.. The parser does need to fail fast, and commit late, and memos sometimes help, and of course character sets need to be compiled down to eliminate explicit choice operators anyway, regardless of the type of choice...... Cheers, Peter
_______________________________________________ PEG mailing list PEG@lists.csail.mit.edu https://lists.csail.mit.edu/mailman/listinfo/peg
