Lars Ingebrigtsen <larsi@gnus.org> writes:

> I've now taken another whack at this (but storing the actual symbol), and
> `C-h S' now works better than before, and no tests are failing.  I'm not
> quite sure whether the change here will spill over into Emacs Lisp mode,
> but some testing seems to say that it won't...

What I liked about the 'display approach is that it didn't seem to
require changes in the describe-* commands to work...

So I thought we could perhaps limit the places where we propertize the
string, and I came up with the attached patch.  But if you still think
it is the wrong way to do it, or if you like your changes better, I
promise I won't insist.

Note that with my patch, we'd still need the info-lookup patch from
yesterday.