…e FsLexYacc project instead of the old F# PowerPack.
… dependency for the next Facio package. FsLexYacc.Runtime is going to have a breaking namespace change in version 7.0.0 which we don't (yet) want to pick up.
…he command line when called from xbuild/Mono.
… be moved to Graham.Tests later.
…y (the dominator relation was reversed, causing it not to be calculated correctly). - Renamed some of the types (and fixed the comments) to just "State Position Graph" (SPG); the name "Parser State Position Graph" (PSPG) wasn't accurate -- in Purdom and Brown's paper, PSPG is "Partial State Position Graph".
…was only computing a partial result.
- Removed a bit of unused code from the LeftCornerParserAction.ToString() method.
… correctly). - Fixed the 'open' references to namespaces in Graham. - Removed unnecessary references to System.Core and System.Numerics assemblies.
…e old unit tests over from the FSharpYacc.Tests project; however, the unit tests don't currently build -- they need to be modified a bit before they'll compile.
…ser tables into a new library called "Graham". This will allow Graham to be expanded and used for other purposes, and for FSharpYacc to stay focused on what it needs to do (e.g., generate code, emit diagnostic messages, etc.)
…ementation; the code needs a bit of clean-up but overall it appears to be working correctly.
…rser table to SLR(1).
- Created the Relation and PartialFunction type aliases to simplify the parameters of 'traverse' and 'digraph'. - Added some documentation comments.
…algorithm instead; the LA sets are now computed correctly.
…ation -- our second implementation appears to be working correctly now.
…nello algorithm for computing LALR(1) look-ahead sets.