[4.0.x] may reach end warning in system headers

[Mike Stump]

On Friday, July 1, 2005, at 12:00 PM, Gabriel Dos Reis wrote:
> That is a view but I do not share it.  Geoff's local change sounds
> much more sensible to me, and is less invasive.

I like Geoff approach first and foremost.  Beyond that, I think we 
should welcome any cleanups to the code that are appropriate.  I think 
the libstdc++ people can decide that, but they should be free to rely 
on Geoff cnange (or what ever fix makes it not warn) if they want.