Re: CEDET Merge

[Edward John Steere <edward.steere@gmail.com>]

> CEDET's Java support works pretty well actually. I use it every day for
> my day job. It'd be great if it could be included (eventually if not
> part of this) in Emacs-proper.
>
> Of course it could be improved since it's a little out of date, but
> it remains useful.

I can report the same thing: CEDET's Java support although dated in some
places is very good.  I've submitted patches which prevent the parser
from bombing out on diamond syntax and annotations as well as fixing
starry imports, but there are other areas where it's lacking,
e.g. support for static imports.

If we were to bring it up to standard and fix some debilitating
performance issues which crop up on larger projects then it would
compete with many of the projects which try to bring semantic Java
editing to Emacs with server programs.  (You can also work around the
performance problems with some hacks which limit search scope
aggressively but that's obviously not ideal.)

I don't understand what's preventing us from porting this and other
aspects of CEDET right away -- other than scope creep perhaps?  I would
advise against using upstream as a fallback though because:
 a) it's likely to receive even less attention after the merge;
 b) most of it doesn't work with the latest Emacs because of the changes
    to EIEIO in core;