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

[Benjamin Kosnik]

>Basically, it fixes all the issues we know well (the unbuffered seeks 
>too!, will post later some polished seek testcases split up from the 
>current basic_filebuf/seekoff and seekpos dirs) in a, no so humbly, neat 
>way: just state that 'unbuffered' (27.8.1.4 para 10) means _M_buf_size == 1!

This looks great.

>Using regular buffering (no special cases in _M_underflow) solves in a 
>straightforward way also the seek issue, since we already have the 
>machineery inside seekoff.

Yep. I like this much better.

>Tested x86-linux, seeekoff/* testcases too adapted roughly to the 
>unbuffered case too by adding pubsetbuf(0, 0) before the open.

Outstanding! Thanks Paolo.

-benjamin