Ralph Jorre writes: > > I don't often use a debugger but I have found JSwat very good and I just > wonder if it wouldn't be more efficient to just include some hooks so as > to be able to incorporate it efficiently. > > Maybe this would be easier to maintain than rewriting the debug. >
I plan to look into this. The problem is that the task of writing an Emacs front-end to JSwat may outway the advantage of replacing JDEbug with JSwat as the backend. In other words, there already exists a Lisp frontend to JDEbug. So it comes down to whether it makes more sense for me to spend my time on creating a frontend for JSwat or fixing the problems with JDEbug. Complicating the decision is the fact that I have created a generalized object-oriented debugger frontend, based on eieio. I decided to do this because jdb and the JDEbug frontends have a lot in common, e.g., essentially the same code for stepping and setting, recording, and displaying breakpoints. The generalized frontend allows both backends to inherit any improvements to the front end. I have already ported jdb to use the generalized frontend and am in the process of porting the JDEbug frontend to the generalized front end. It is conceivable that the generalized JDEE frontend could ultimately support three backends: jdb JDEbug JSwat In fact, I would prefer this as I believe there is value in having a backend that is tuned to the JDEE plus support for alternative backends as backups or to cater to user preferences for those backends. My preference would be for somebody other than myself to take on the task of creating a JDEE frontend to JSwat's console mode. This would free me to concentrate on the generalized frontend and on fixing the JDEbug backend. Developing a JSwat frontend basically entails subclassing the JDEE's generalized debugger frontend (see jde-db-debugger class) to support JSwat's console commands. This should be a fairly straightforward task. I'd be glad to serve as a consultant to anyone who wants to undertake this project. - Paul
