emacs and /etc/alternatives
Chip Coldwell
coldwell at redhat.com
Fri Mar 9 14:14:48 UTC 2007
On Fri, 9 Mar 2007, Andrew Haley wrote:
> Matthew Miller writes:
> > On Thu, Mar 08, 2007 at 10:52:33PM +0100, Miloslav Trmac wrote:
> > > > If both are functionally similar, yet the script solution avoids changes
> > > > to the filesystem *and* is much simpler, why not stick to the script
> > > > solution?
> > > If you completely ignore the original purpose of alternatives and focus
> > > only on the mechanism, following a few symlinks set up by alternatives
> > > is actually both more effective and simpler than starting bash to
> > > execute the script.
> >
> > The overhead of bash vs. a symlink is negligible when we're talking about
> > launching *emacs*. The real difference is: one is trivial and
> > self-contained, whereas the other relies on an whole infrastructure.
>
> Exactly. This is a change without a purpose. The existing solution
> works perfectly well: it ain't broken, so don't fix it.
It is broken in the sense that rpmlint pukes on the current emacs spec
file, and a lot of the jiggery-pokery in there is done to support the two
different versions of emacs and the wrapper script+symlinks currently
being used.
Chip
--
Charles M. "Chip" Coldwell
Senior Software Engineer
Red Hat, Inc
978-392-2426
More information about the Fedora-maintainers
mailing list