----- Original Message ----- From: "Jason van Zyl" <[EMAIL PROTECTED]> [...] > Adding prelimary support for JavaDoc processing. By preliminary I mean > that there is no formal support to process what is found but the JavaDoc > ar are being place in the AST instead of being ignored which is what the > default grammar did. Jude here we come!
kewl ! > Stephane, I also removed the process() node which was something I put in > the velocity grammar as part of making a reusable parser. ok. FWIW I'm investigating metrics (org.apache.maven.metrics) I'm doing experiments since I would probably need to modify a couple of nodes and add helper methods (ie adding getMethods, getFields to Class and things like this). I would like to add a getter to simple node getId() which would get the id of the node so that it is easier to identify them than using some costly instanceof. Question: For now I assume that the methods and fields of a class are referenced (ie cached) in a lazy way (that is, it will iterate at the first request to find them and store the references to avoid unecessary iterations again). I'm not that easy with jjtree, but do you think it would make sense to link the node at parsing time ? (ie once for the method production it will iterate til the class parent and add itself to the class). I'm not so sure because of cluttering the parser grammar but just thinking loud. thoughts ? -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
