xemacs21 and unbundled lisp (was [XEmacs] New Maintainer, soonish.)
karlheg@cathcart.sysc.pdx.edu (Karl M. Hegbloom) writes:
> James LewisMoss has offered me the maintainership of the Debian
> XEmacs packages, and I have accepted.
Good luck!
> - The XEmacs lisp has been unbundled (upstream) in this version, and
> so I'll be creating a large set of individual library packages for
> it. Users can choose which ones to install and which ones to leave
> behind as never-gonna-use-it-anyways. I've only begun to have
> vague notions of a need to make some sort of plan before I finally
> get on with actually tackling this one... I know it'll require a
> makefile. 8-p That subproject is on next week's agenda... or maybe
> over this weekend. It's raining anyway. (darn)
Regarding unbundled lisp, there would seem to a few different cases:
If some elisp package (say, gnus) has been unbundled from xemacs, and
the vanilla version and the xemacs21 version are identical, then the
existing vanilla package (say, gnus) just needs to be taught to
install for xemacs21. Does someone need to file a buncha bugs about
elisp packages that don't know about xemacs21 yet?
If the upstream version and the xemacs21 version diverge, then you can
either create a new package (say, psgml-xemacs21) which can live
alongside the vanilla elisp package (say, psgml), or else you can try
to get the xemacs21 changes merged into the vanilla package. I would
prefer trying to merge the two whenever possible -- that way, we don't
have to worry about propogating fixes between two packages (say,
psgml-xemacs21 and psgml).
Could you perhaps supply a list of the current potato elisp packages
which need to be taught about XEmacs21?
> -- I will name them prefixed by "xemacs" so they sort to the right
> place in `dselect'.
Well, see above. First off, you should name it 'xemacs21-*'.
Secondly, I would hope that we can avoid all this double packaging of
elisp packages.
> - I sent two wishlist bugs in against xemacs21...
>
> -- I'd like to have a Debian BTS minor-mode for Gnus. I guess it
> ought to work for either emacsen. It could use a set of nnml
> folders, and work with the email commanded bts server somehow to
> make it simpler to work with. I think it's got more potential
> than a CGI interface. We'll see... it's a wishlist item for sure
> right now. AFAIC, any serious Linux admin uses XEmacs anyway.
>
> -- I'd like to write or help write an `adduser' interface for XEmacs.
> I suppose it ought to work in either emacsen also... <sigh> You
> can read what I wrote in the wishlist bug report if you like; just
> pull up the ones for xemacs21 and look for it.
These should definately just be elisp so that FSF Emacs users can also
benefit.
--
.....Adam Di Carlo....adam@onShore.com.....<URL:http://www.onShore.com/>
Reply to: