[CEDET-devel] More Analyzer

[Eric M. L. <er...@si...>]

Hi all,

  I've been tweaking the grammar mode analyzer support, and it
resulted in more changes in the Emacs Lisp support for the analyzer.
There is a new concept in semantic-ctxt.el, and the function is:
`semantic-ctxt-current-class-list'.  The idea is that at a given
cursor position, you can predict the class of the tokens you can put
there.  For example, in Emacs lisp, things after a ( is almost always
a function.  I think different languages will grow the  functionality
here as completion gets more polish.

  Anyway, I've added rough support here for Emacs Lisp as an
experiment at it works pretty well.  It required a batch of changes
in different places.

  Next up is (probably) more grammar support.

  David, in the semanic-grammar file, there is both grammar rules and
Emacs Lisp code.  This is handled in most cases by code like this:

  (if (in-lisp)
     (emacs-lisp-do-something)
   (more code for grammars))

  For overload methods, it would be nifty to have a method like this:

  (if (in-lisp)
      (semantic-with-mode-variables 'emacs-lisp-mode
         (do stuff with Emacs Lisp))

    (do grammar stuff.))

  to execute code as if we were in an Emacs Lisp buffer.  Do you have
a notion for doing this?

  I know there is mmm-mode or some such that does something similar.
In this case I'm mostly worried about the overload methods which are
in some sort of hash table.

  A cool additional feature would be that semanticdb could offer to
execute overload methods from the originating buffer of the tags.

Thanks
Eric

-- 
          Eric Ludlam:                 za...@gn..., er...@si...
   Home: http://www.ludlam.net            Siege: www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: www.gnu.org