This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
In its current state the ANTLR module seems to not be buildable or runnable. I have some local changes that seem to (help) fix the problems. I have rewritten the build script to use the <antlr> task. The files generated from .g files that are currently in CVS should be removed from CVS (in src/org/netbeans/modules/antlr/editor/g/: ANTLRLexer.java ANTLRParser.java ANTLRTokenTypes.* Also, I have made some changes to get the module to build against ANTLR 2.7.2a3 instead of whatever version it was previously using. In particular, I have changed the module so that it will not require an ANTLR autoload module (see issue 28061). The upgrade to ANTLR 2.7.2 required some changes to remove dependencies on things in antlr.Tool that were removed in 2.7.2. Everybody that has done any work to get ANTLR to build/run should feel free to attach patches and comments to this issue.
Created attachment 7681 [details] yet another version of the antlr module, which does not work (see thread on nbdev list)
Derrick Oswald is currently listed as a project owner, but it seems clear that others are ready to be more active. Brian, I suggest you request the role - or at least, that of Developer. Holger too. Since the module is currently broken and unmaintained, it could hardly be harmed by people jumping in and working on it.
http://antlr.netbeans.org/servlets/ProjectMemberList by the way.
I requested developer role. I would like Brian to become owner.
I'm now going to append a changelog of all the changes I've made to the antlr sources to make antlr more extendable. The changelog is meant as a starting point for discussions. I can provide a diff, too, but would like to discuss the changes made first. The change made at 2002-10-07 09:10 shows one speciality: it is meant for parsers, which consume tokens from streams which are not antlr lexer in cases where it is preferable to dispense tokens explicitely. My use case is in XML parsing (I mean the layer, which consumes SAX-like events in an application specific manner), where XML element attributes must be copied, when needed at a later point in parsing. To avoid attribute copying I must know, how long a token is actually needed. The change made at 2002-10-02 16:32 extends token name identifiers to a format which betters fits to automatically generated token names having some internal structure. Here `$' is used as separator sign.
Created attachment 7702 [details] changelog of some changes made to antlr sources
Obsolete milestone, please reevaluate