2 ms·
So, if applicable here, what do you parsing experts think of GLL parsing? http://dotat.at/tmp/gll.pdf http://dotat.at/tmp/gll.pdf It mentions retaining easy i
by nickpsecurity 10y ago
So, if applicable here, what do you parsing experts think of GLL parsing?
http://dotat.at/tmp/gll.pdf http://dotat.at/tmp/gll.pdf
It mentions retaining easy implementation and debugging of LL parsers while knocking out limitations with some LR-style stuff. A hybrid of sorts. Result is cubic time.
- PeCaN 10y agoGLL is fantastic! Unlike Earley and GLR, implementing GLL parser combinators is pretty easy. It's pretty fast and has some other nice features (it can be implemented to produce a lazy stream of possible parse sequences). It is, IMO, the current state-of-the-art in parsing general CFGs. Derivative parsing is very interesting and may overtake GLL and related algorithms.
- nickpsecurity 10y agoAppreciate the feedback. My plan was to push some formal methodists or LANGSEC types to try to do a verified GLL if the opportunity arises. So far, there's been SLR [1], LR(1) [2], and PEG [3] verified for correctness. Think verified GLL generator is ideal next target given benefits? [1] http://users.cecs.anu.edu.au/~aditi/esop.pdf http://users.cecs.anu.edu.au/~aditi/esop.pdf [2] http://pauillac.inria.fr/~xleroy/publi/validated-parser.pdf http://pauillac.inria.fr/~xleroy/publi/validated-parser.pdf [3] https://arxiv.org/pdf/1105.2576.pdf https://arxiv.org/pdf/1105.2576.pdf