Hi Justin, On 02/12/2010 7:53 PM, Justin Chase wrote: > I have an implementation for .net also: > http://metasharp.codeplex.com/ > > Which is a lot different but the general goal is the same :)
Jeff Moser also has an implementation in C#: http://ometasharp.codeplex.com/ > In your example rules below, isn't that just indirect left recursion? > If not how is it different? In Warth's paper he explains how to solve > the problem of indirect left recursion and while I was attempting to > implement it I was running into a lot of problematic cases such as > multiple levels of indirect left recursion. For example your C rule > lacks a reference to A, was that intentional? It can get very > complex. Additionally C references B which is another layer of > indirect left recursion. As was discussed on the OMeta list last year (https://lists.csail.mit.edu/pipermail/peg/2009-November/000246.html), Warth's algorithm does not correctly handle this grammar. It gets caught bouncing between the A and B rules, and never sees the C rule. Medeiros's algorithm correctly finds the C rule. > I asked about it on the forums here and the consensus was that while > iLR is possible, it's not completely clear that it's actually > desirable. It may result in an increased complexity of grammars for > those who are authoring the grammars and make it hard to reason about > unexpected results. I would be interested to hear evidence > otherwise. LR is useful for handling operator precedence, for example. I'm afraid that I have no evidence that indirect left-recursion does not result in increased complexity -- being able to handle the increased complexity was the attraction for me in implementing the algorithm. Cheers, Gordon -- Gordon Tisher http://balafon.net _______________________________________________ PEG mailing list PEG@lists.csail.mit.edu https://lists.csail.mit.edu/mailman/listinfo/peg
