Deleting skeleton-proxy

Deleting skeleton-proxy

[Stefan Monnier]

While looking at abbrevs and friends, I bumped into the skeleton-proxy horror.
Trying to understand what happened, I finally discovered that
skeleton used to use something like

	(defmacro define-skeleton (name &rest skel)
          `(progn (defvar ,name ',skel)
	          (defalias ',name 'skeleton-proxy)))

and so skeleton-proxy looked up the stack to find the name of the
alias used, then looked up the var-binding of that name to find
the actual skeleton.

Someone (apparently Richard) saw how disgusting it was and fixed it
to do the right thing.  This was way back in 96.

	revision 1.15
	date: 1996/06/28 08:11:18;  author: rms;  state: Exp;  lines: +53 -3
	(skeleton-proxy-new): New function.
	(define-skeleton): Define the skeleton as a defun
	using skeleton-proxy-new.

Now, skeleton-proxy was kept at that time because it was still needed
by code byte-compiled with the old macro.  I don't know if the code
still works (its stack-walking thing is something that I'd expect to
break at least somewhat in the course of 7 years), but it is still
autoloaded.

Any objection to getting rid of this thing ?


	Stefan

Re: Deleting skeleton-proxy

[Richard Stallman]

    Now, skeleton-proxy was kept at that time because it was still needed
    by code byte-compiled with the old macro.  I don't know if the code
    still works (its stack-walking thing is something that I'd expect to
    break at least somewhat in the course of 7 years), but it is still
    autoloaded.

    Any objection to getting rid of this thing ?

I don't think those old byte-compiled files would even work now, if
there are any.  So please get rid of it.