Pressing RET in woman-mode calls man

[Sven Joachim]

Pressing RET in woman-mode when point is on a "reference word" (e.g. in the
"SEE ALSO" section of the man page) invokes man, not woman, on the manpage.
The same applies to mouse-1 and mouse-2. Pressing C-h k in woman-mode gives
the following information:

<mouse-1> at that spot is remapped to <mouse-2>
  which runs the command push-button
   which is an interactive compiled Lisp function in `button'.
It is bound to <mouse-2>, RET.
(push-button &optional pos use-mouse-action)

If I want to invoke woman instead of man, I have to press M-mouse-2, which
is cumbersome:

woman-follow-word is on <M-mouse-2>.

I used "M-x debug-on entry man-follow" to find out that push-button in turn
calls man-follow, as seen in the backtrace:

Debugger entered--entering a function:
* man-follow("dpatch")
  #[(button) "ÁÂ!!‡" [button man-follow button-label] 3](#<marker (moves after 
insertion) at 11300 in *WoMan 1 dpatch*>)
  push-button(11300)
  call-interactively(push-button)

Probably push-button should call woman-follow-word rather than man-follow
in woman-mode.

I noticed this in

GNU Emacs 22.0.50.1 (i486-pc-linux-gnu, X toolkit, Xaw3d scroll bars)
 of 2005-09-23 on debian, modified by Debian
X server distributor `The X.Org Foundation', version 11.0.60802000 .