This is probably the worst patch you've ever seen - not sure whether the 
"advice" I added is the optimal one to add, it just seems to solve the 
problem for me.  Adding unused "&rest args" to a function so that it can 
be used as advice is probably incorrect as well.

Problem was, with global-whitespace-mode enabled, when I switched to a 
new buffer (say C-x b asdf <RET>), whitespace was not visible.  The 
`whitespacesturn-on-if-enabled' function is already added to 
after-change-major-mode-hook, but apparently that isn't triggered when a 
new buffer is created in fundamental-mode.  If I turn 
global-whitespace-mode off and back on again, or if I type "M-x 
fundamental-mode", then whitespace is visible.

The fix I came up with in my init file was to add the code:

     (defun whitespace-turn-on-if-enabled-2 (&rest args)
       (whitespace-turn-on-if-enabled))
     (advice-add 'switch-to-buffer :after 'whitespace-turn-on-if-enabled-2)

But I figured I would create a patch and let you guys look at it and 
decide what should be done.  I have attached it.  It works for me, but 
might not fix a wider root cause, and is almost certainly not the 
optimal elisp code in any case.


Jefferson Carpenter

P.S. Even with the above fix (either init file or whitespace.el patch), 
when I switch to a new buffer, whitespace is shown but in the wrong 
color.  It remains the wrong color if I turn global-whitespace-mode off 
and on again, but uses the "whitespace-space" face, etc. as it should, 
when I run "M-x fundamental-mode" (the buffer already being in 
fundamental-mode).