[PATCH] Abstract out a display-table-specific API (2)

[Stephen J. Turnbull]

Mike Kupfer writes:

 > >>>>> "ST" == Stephen J Turnbull <stephen at xemacs.org> writes:
 > 
 > ST> This code should be introduced in two places.  First in the
 > ST> mainline, second either in 21.4 core (which will take longer, viz
 > ST> the fact that Vin hasn't weighed in on changing `make-autoload' in
 > ST> 21.4), or in packages.  I think package-future.el is safest.
 > 
 > I was thinking about this some more, and I think there's still a missing
 > piece.  Or maybe it's just my lack of Lisp expertise showing.
 > 
 > Anyway, package-future.el fixes up the compile-time environment, but are
 > those defuns going to be available at runtime on 21.4 systems?

Heh.  No.  Make them defsubsts.  ;-)

I'm not really joking about this case: I think it's a reasonable
answer even though a bit fragile.  However, I acknowledge it's not a
good generic answer.