Re: kill ring menu

[Colin Walters <walters@debian.org>]

On Mon, 2002-05-06 at 02:39, Miles Bader wrote:

> For my part, I've never seen any complaints on any of the emacs
> lists/newsgroups from people that dislike fontification in non-font-lock
> modes (this is the important case, because most package currently seem
> to either use font-lock, or if not, to default to always having
> fontification).

It is true that this does change the default; it removes fontification
in an *Occur* buffer where there was fontification before.  But in the
end, the user gets more control, in a more consistent fashion, and I
think this win outweighs the change of default.

Using font-lock from the start is a win for other reasons, too.  Suppose
that Occur mode was changed so you could actually edit the buffer, and
changes would be reflected in the original buffer.  If the user typed
additional matches for the search string into the *Occur* buffer, we
would expect them to be made bold, too.  This is possible in a clean way
if we use a dynamic font-lock mechanism.

> I think it is true that using font-lock is wasteful if all it does is
> examine text-properties which could have just as easily been face
> properties.

It's only "wasteful" for those users who liked fontification in special
modes like *Occur*, but not elsewhere.  I really do think that those
users are a relatively rare breed.

And finally, this is a situation where we have a user-interface win
(consistency of interface to enabling/disabling fontification) over
resource consumption (loading of font-lock.el).  I think that as a
general principle, the former should win over the latter.