[comp.emacs] buffer-local bindings w/o copying keymaps

gaynor@pilot.njin.net (Silver) (09/04/88)

How do *you* locally bind characters without copymaps?

I figured out a half-baked scheme that works for sparse keymaps using shared
lists.  (Mail me for the fugly details, if you're curious.)  But, using the
existing utilities, I can't think of how to do it properly for non-sparse
keymaps.

Is there some application of the (keymap . char) definition type that I've
overlooked that *properly* does the trick?  If not, what approach would you
suggest for a source-level implementation?  Hmm, using the (variable) value of
a symbol if it doesn't have a function value is kind of easy, and almost
backwards-compatable.  Or would it be better to add a buffer-local-map to
buffers?

Comment?

Regards, [Ag] gaynor@rutgers.edu
	      ...!rutgers!gaynor

jbw@bucsb.UUCP (Joe Wells) (09/04/88)

In article <Sep.4.02.25.11.1988.15619@pilot.njin.net> gaynor@rutgers.edu writes:
>How do *you* locally bind characters without copymaps?
[deleted stuff]
>Regards, [Ag] gaynor@rutgers.edu
>	      ...!rutgers!gaynor

Personally, I was disappointed when I discovered that copy-keymap was
inadequate for this purpose.  I think that copy-keymap *ought* to
recursively copy both types of keymaps, instead of just copying the
a-lists of a sparse keymap.  When I get around to it, I'll write one
that does that ...

For all of you reading comp.emacs or gnu.emacs, I have a question.
The article to which I am responding went to both comp.emacs and
gnu.emacs.  Rn filled in the followup field for me with comp.emacs.
Where would *you* rather have followups to this kind of article go?
gnu.emacs seems more appropriate, but a lot of places don't get that
group.

Joe Wells
UUCP:  ..!harvard!bu-cs!bucsb!jbw
ARPANET: jbw@bucsb.bu.edu