Igor Dejanović

I'm quite sure that it's not theoretically possible :) GLR constructs trees in parallel, not sequential order like some top-down approaches might be doing. All possible sub-trees are constructed as...

That wouldn't be GLR parser anymore. It would be a backtracking parser. The power of GLR comes from its "online" style of work where it takes a token and [do...

In version 0.14.0 GLR parser returns the forest object which is iterable producing lazy trees. It doesn't work as requested here as it wouldn't be possible but from the user...

@hosford42 Thanks much for the contribution! I just skimmed through the implementation and it looks very good. I'll do the full review in the next couple of days.

Maybe you would be interested in the discussion on #78 which covers some of the ideas you presented here (if I understood you correctly). There is a branch with the...

I haven't had a change for a deeper look into it. I'll get to that probably after finishing current rework around error reporting. The thing that I don't like with...

Partly solved on the master branch by persisting parser LR table. See #52

I understand where you are coming from. And your points are valid. I did it in [Arpeggio](http://www.igordejanovic.net/Arpeggio/latest/grammars/) where you can define your grammar either by Python (a.k.a. internal DSL) or...

Thanks for the detailed explanation of the idea. Yes, this approach if implemented would be cool. Dumping grammar to json together with LALR parser table would make possible for parser...

@SupraSummus Just to let you know that I have been working on a redesign lately for the upcoming 0.10 version. There are a lot of changes and one of them...