PATCH for save_tree_status

[Jeffrey A Law]

  In message <199902242141.NAA16222@adsl-206-170-148-33.dsl.pacbell.net>you wri
  > 
  > The save_tree_status function mishandles the case where the context
  > of the new function is NULL, but we're at global scope.  It would be
  > useful to handle this case, so that we can use save_tree_status during
  > template instantiation in C++ front-end.  
  > 
  > This change has been bootstrapped/checked on x86-linux-gnu.
  > 
  > Would someone care to approve this patch?
  > 
  > Thanks,
  > 
  > -- 
  > Mark Mitchell 			mark@markmitchell.com
  > Mark Mitchell Consulting	http://www.markmitchell.com
  > 
  > Wed Feb 24 13:38:39 1999  Mark Mitchell  <mark@markmitchell.com>
  > 
  > 	* tree.c (save_tree_status): Don't treat functions with no context
  > 	as nested.
This doesn't really seem right to me, at least not without further comments
and/or explanation.

save_tree_status says explicitly that it's used before starting a nested
function and that CONTEXT is the context of the nested function (which I
don't think can ever be NULL).

If we're using it for some other purpose, then we'll probably want to make
the change.  We'll definitely want to update the comments about when 
save_tree_status is used.

It may also be the case that the comment is simply out of date and just
needs to be updated to match current reality.

Do we have any corresponding changes to make to restore_tree_status?

jeff