Hi Steve, Am Montag, den 11.12.2006, 19:32 +0000 schrieb Steve Brewin:
> You might want to look at the jSieve code to see how something similar to > your objectives is achieved using JavaCC rather than ANTLR. This is a > different, not necessarily better, approach. Okay, I looked at the jSieve code, googled and read a bit docs. ANTLR was my first experience with parsers and use of ABNF like notations apart from a few exercises at university. >From a first look its syntax seems to be comparable. Some notations are very similar. The steep learning curve for non compiler builders might be the same. :-) > It would be interesting to compare and contrast the pro's and con's of each > approach. My hazy memory says I looked at both approaches and plumped for > JavaCC's as being easier to work with (to me). Well, for an objective judgment I would have to spend another day with JavaCC, for sure. :-) I'm not going to do that now, but maybe I get some more insight to JavaCC when I'm more familiar with ANTLR. But yes, it could be quite interesting to discuss the differences we assume. With a lot of IMOs and AFAIKs, pros for ANTLR: - well suited for unit tests because parsing could start at any rule - full control over tree-structure (which should be a node? Which should be a List of arguments?) without writing java code - presentation logic for testing/debugging like a text or a gui representation - more (probably not better) documentation - wider use in similar projects (?) - inheritance for lexer and parser > Of course the JSieve code is parsing Sieve commands rather than IMAP's. This could maybe make a difference for the way ANTLR and JavaCC build trees. Joachim --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]