On Jan 17, 2010, at 8:57 PM, Sam Harwell wrote: > The only thing *really* causing headaches for the Visual Studio > integration is the association of error handling to threads instead of > to the grammar generator when it's actually a property of the latter.
Oh right. Well, that would be more proper, but what situation does this cause problems? Surely you can use the same thread to compile to grammars at once > Before rewriting the code, I think the following major issues affecting > syntax should be covered so you know the new design will meet the needs: > > * Syntax for expression parsing you bet! > * Support for ambiguous tree nodes (similar to backtracking=true, except > it keeps all viable alts instead of just the first one) I do like that idea; seems like we would have to invent a construct that said match all ambiguous alternatives rather than picking one. added to list. > The design of the profiler and debugger grammar configurations concerns > me because grammars that support debugging have to be derived from a > different base class than regular grammars. does this cause a problem in Dflat er...C#? ;) Ter _______________________________________________ antlr-dev mailing list [email protected] http://www.antlr.org/mailman/listinfo/antlr-dev
