David Kastrup <dak@gnu.org> writes:

> If you thing it reasonable that key-binding should also react to
> bona-fide events _not_ wrapped in an array, I can do this easily
> enough in the change I am working on.

Ok, here is my current patch (incidentally: why doesn't it work to use
PCL-CVS to do a multi-file diff? It ignores the marked files, just
working on the current line).  I have added an additional argument to
key-binding and documented it (that required changes in keymap.h and
callers of key-binding from C in keyboard.c).

As far as I can tell, this works with all key sequences.  If people
agree that this is reasonable, I would install it.  I'd be glad if
someone looked it over with regard to the SPECDCL and garbage
collection protection: I don't feel comfortable about those.  I also
copied code from keyboard.c that aborts if the current frame is not
live and can't be returned to.  No idea whether this is required (I
assume it was put there to avoid a busy lockup when Emacs got its
frame killed or something) or even a good idea.

Anybody have an opinion about that?

If people think this is the right way to go, I can install it.  I have
verified it with quite a few tries.  However, I have not yet tested
the LOCATION argument.  It should come in handy for mouse.el and
follow-link.