On Sat, Mar 23, 2002 at 08:39:38AM +0000, Vaclav Barta wrote:
> > Nascif Abousalh-Neto wrote:
> > (http://transmogrify.sourceforge.net/). I used the approach suggested
> > by Philip, implementing a layer on top of gnuclient to make calls from
> > Java to Emacs (to do things like obtain cursor position, etc.). Using
> > gnuclient still gives me the feeling of a hacked up solution, and some
> > designers had problems with it; I would like to evolve it further and
>
> I would be one of those designers... :-) I'm using Emacs to edit 
> the source code: isn't it unnatural to use some other GUI for
> refactorings, which are just source code transformations? Why don't you
> integrate the refactoring functionality into Emacs (like JDE)?

that is what I am planning to do.  I think the solution is as follows:

For my tool to query emacs I will use the gnuclient interface.  This is
definitely feasible as I have written routines to get the current line,
column and file from emacs via a socket.

For emacs (ie JDE) to query my tool, e.g. to initiate a particular
refactoring, I will probably use the jde-jeval functionality to call
static methods on my code.  I can therefore keep an instance of my tool
in a static member somewhere and just call methods on that.

This way all the interface will be done in emacs, not with a separate
java gui as things like Jacob seem to use.

Any thoughts?

-- 
Graham Bennett
[EMAIL PROTECTED]  ICQ: 17566658

Reply via email to