>>>>> "Paul" == Paul Landes <lan...@mailc.net> writes:

    Paul> Again, I have thought about forking/canabalizing etc.  The
    Paul> primary motivation is that the project is large and very
    Paul> dependent on CEDET.  I have nothing bad to say about CEDET but
    Paul> it's current integration with Emacs post 23 has been
    Paul> non-compat and created a lot of issues for installation.

One thing that working on JDEE has forced me to do is learn more about
CEDET, specifically its inclusion into Emacs 23+. The conclusion I've
drawn is this: JDEE should work with GNU Emacs CEDET version alone. It's
the right thing to do for *users* and it resolves the compatibility
issues by making the choice of platform. I would welcome feedback on
this (in particular I'm still unclear about the relationship between the
CEDET and Emacs code bases, though it seems friendly enough).

    Paul> ABCL (common lisp implementation) would replace the now dead
    Paul> Beanshell language, and frankly, seems like a better fit since
    Paul> its lisp and not Java.

Has anyone looked at implementing the beanshell functionality in elisp?
While I *love* coding in CL, I'm not sure I've talked myself to where
I'm sold on switching beanshell out with still another languge. The
truth is, I actually like elisp, and it will always be a well supported
language in Emacs :-)

Getting rid of, or improving on, beanshell is probably #2 on the list of
things I need though (mostly to get Java 5+ language support).

    Paul> Another idea I had is to move everything to maven so there
    Paul> would be maven integration (bridged through ABCL) for
    Paul> compilations.  If others still want to use Ant or something
    Paul> else we can talk about some abstraction layer.

Maven is #1 on my list. I'd really like to fold Maven support into
JDEE. There have been so many efforts out there to do it in the past
that I feel it just a matter of picking and choosing appropriately
licensed code and adding it to JDEE. I personally need this big time, so
from a motivation perspective I will work on it personally.

    Paul> IMPORTANT: I am only one person with a day job.  The only way
    Paul> we get a solid Java IDE environment is for everyone to pool
    Paul> their time and resources to create this all together!

This is key. I use JDEE when my day job asks me to write Java. So I'm
very focused on always having a working, maintainable, and useful
JDEE. This leads to the following "roadmap" in my head:

   1. Get JDEE to work with Emacs CEDET (I believe this is now done on
   trunk, at least I am using it daily)

   2. Get the code base out (which I'm proposing we do now with a minor
   release, and I will also work on getting this version back into
   Debian).

   3. Now clean out all the cruft to support old versions of Emacs,
   CEDET, and other libraries. Build for GNU Emacs with it's included
   CEDET. Period. No third party dependencies other than beanshell (and
   Ant to build). This cleans up build and start up code, the CEDET
   integration, sub-process management, and many other minor things that
   are annoyances when making changes/improvements to JDEE (to me
   anyway, because I find myself testing to many configurations since it
   is so hard to tell which ones to worry about).

   4. Documentation updates, and another minor release

I can get (3) and (4) done relatively quickly after the release. It will
mean trunk will target only GNU Emacs going forward.

With that out of the way my list continues to:

   5. Add Maven support (at a minimum like jde-ant today, but I'd like
   better EDE/prj.el integration)

   6. Do a release with working Maven support for Emacs 24+?

   7. Replace Beanshell (probably along the lines that Paul is
   suggesting)

   8. Major release

Step (7) is where a fork is a realistic thing to do. But I think we
should get this done in this code base. I'm not sure there is a
community large enough to support a fork, and without some basic
housekeeping the existing community, as it were, will fade away. 

Cheers!
Shyamal

------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
_______________________________________________
jdee-devel mailing list
jdee-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jdee-devel

Reply via email to