create an elisp portgroup to simplify installing modules for emacs/emacs-app

[drkp (Dan Ports)]

We've long had an issue with ports that install emacs modules, because they need to have a dependency on emacs, and usually use the emacs binary to byte-compile elisp code. The problem is doing this in a way that supports emacs-app as well. (Path dependencies aren't useful, because emacs and emacs-app intentionally don't install files into the same location.)

Some ports have been dealing with this using a +emacs_app variant, which isn't an ideal solution. Several others don't have a solution at all, which is even worse.

The gnuplot port checks at installtime which version is installed, and uses that. The attached portgroup tries to factor out the common code for that test so that we can use it in other ports. The assumption here is that all of these emacsen install into the same site-lisp directory (which is true of our emacs and emacs-app ports) and that they are the same major version so that bytecode will be compatible, so it doesn't really matter which one we build with.

The portgroup just sets up some variables for use by portfiles. I didn't try to set up any of the build environment directly in the portgroup; the build systems for these modules are so different that it's not clear there's any common settings. Also attached an example of how to modify the org-mode portfile to use the portgroup.

Any comments are appreciated!

[drkp (Dan Ports)]

proposed portgroup

[drkp (Dan Ports)]

example of usage

[cooljeanius (Eric Gallager)]

Cc Me!

[mojca (Mojca Miklavec)]

Thank you very much for starting this (I have no idea how portgroup files are created). One more thing that would be potentially nice for gnuplot is the lispdir variable used in --with-lispdir=${prefix}/share/emacs/site-lisp

I don't have much to add (I'm not competent enough to swim in the sea of zillion emacs ports), but now that elisp is there: what do you think about creating a wxwidgets portgroup? ;)

[drkp (Dan Ports)]

Added the portgroup in r106035. Any additional comments are still appreciated, of course.