"Davide G. M. Salvetti" wrote:

> >>>>> "PSG" == Peter S Galbraith <[EMAIL PROTECTED]> writes:
> 
> PSG> James LewisMoss wrote:
> >> >>>>> On Mon, 15 Oct 2001 14:30:09 -0400, Peter S Galbraith <[EMAIL 
> >> >>>>> PROTECTED]
> o-m
> po.gc.ca> said:
> >> 
> Peter> I just noticed the use of the debian-emacs-flavor symbol in
> Peter> auctex's /usr/share/emacs/site-lisp/tex-site.el file.
> 
> Peter> I'm wondering if debian-emacs-policy should forbid the use of
> Peter> debian-emacs-flavor outside of the /etc/emacs* directories, or
> Peter> rather inside of directories usually visible by unpackaged
> Peter> Emacsen (i.e setup whatever variable you want using
> Peter> debian-emacs-flavor in /etc/emacs/site-start.d/50somefile.el,
> Peter> and use that variable later).
> >> 
> >> I believe it should forbid its usage outside the /etc/(x)emacs*
> >> files. 
> 
> On the contrary, I think we should not restrict its use.
> 
> Rationale: Emacs it's a big beast; user who install their own, not
> prepackaged, version should be expected to know how to cope with
> integration with the rest of the Debian system.

I don't really see why.  I know elisp and it took me until now to
figure out how to do that, after months of using emacs21 without
full integration (and I understand how Debian emacs packaging is
basically done).

But I may have overstated the problem.  While the directory
/usr/share/emacs/site-lisp/ is outside the scope of
/etc/(x)emacs, it's not in the load-path of a self-installed
emacs21 either (which will rather look under
/usr/local/share/emacs/site-lisp/).
 
> OTOH, if people will ask for it, as the auctex maintainer, I may add a
> couple of lines to tex-site.el checking for debian-emacs-flavor
> existence and recovering with some default if it isn't.  Please, file a
> wishlist bug against auctex if this is the case.

I was just picking on your package because what's where I noticed
it.  I'd rather we adopt a general solution: Documenting how to
make a /usr/local installed emacs work with Debian elisp add-ons
packages would be better.  I assume this problem will go away
when emacs21 is released anyway.

Peter


Reply via email to