I’m not sure if taking the longest match is necessarily right. Imagine a (nonsensical) rule like: LeftRec <= LeftRec ‘+’ Term / “ok” /

I’m not sure if taking the longest match is necessarily right. Imagine a (nonsensical) rule like: LeftRec <= LeftRec ‘+’ Term / “ok” / [a-z]*[0-9] According to PEG, “ok3” should consume only “ok” here, and leave the 3.