[Patch] Clean up sbumpc, move back _M_buf_size, do not special case unbuffered _M_underflow...

[Paolo Carlini]

Benjamin Kosnik wrote:

>No, this sounds like a great plan. I just didn't want you to lose test
>cases.
>
Ok, I have just committed the patch as posted.

In the next days we can consider tweaking it here and there: e.g., 
probably Nathan is right that setbuf should have on top 'else if (__s && 
__n >= 1)'; also _M_destroy_pback probably now can be simplified, 
removing the 'if (_M_pback_cur_save)', since can't happen anymore that 
the get area pointers are null.

Many thanks to everyone for the attention,
Paolo.