On Jan 17, 2010, at 11:17 PM, Gerald Rosenberg wrote: >> Are there changes I could make > I would be very interested in understanding more what you are planning. > Improvements in the tree walker/tree rewrite capabilities would be highly > desirable.
Pass 'em along so we can discuss and I will add to the new v4 planning page. > Completely hiding "hoisting" from the user's perspective is another (Antlr, > as a language, should not have unexpected side-effects). that would be really great; should be easy to fix even if people want to use local variables; I can use a stack that also. >> that would help other tool developers like Gerald's eclipse plug-in? I'd >> like to make it as easy as possible for people to integrate ANTLR stuff. >> > The tools will adapt, as tools should. By design, AntlrDT uses (and only > depends on) an unmodified ANTLRv3.g, so updating to v4 should be relatively > simple. kewl. > That said, there are a number of minor things that would be helpful: > > 1) error reporting could be more modular/have a defined API. That would make > it easier for a parser wrapper to intercept and redirect errors on the fly. > Having to tee stdout and stderr is not desirable. can you be more specific? Right now it just calls the listener, right? > 2) having to decode the different error messages and line/col number based on > error type is a bit complex (they can be unexpectedly null in some places and > buried in different locations in others). Would be nice to be a bit more > well-defined. > 3) reduce the a number of errors that are reported simply as occurring at 0:0. Yeah, I will make a concerted effort to get good error messages and line numbers. I was waiting until the rewrite of the front and grammar to get to this. I've done a pretty good job with STv4 in this respect so far, though I notice a couple of places where a screwup still. Ter _______________________________________________ antlr-dev mailing list [email protected] http://www.antlr.org/mailman/listinfo/antlr-dev
