Update on some PRs I'm working on

[Paolo Carlini]

Paolo Carlini wrote:

> - 9423
> This is apparently the easiest one by far: we have just to
> destroy the internal buffer when is_open() is true. However,
> regtesting revealed a surprise: filebuf_virtuals.cc, test5
> calls setbuf(0, 0) _after_ open(), leading to undefined
> behaviour.

Another error, sorry, not just after open(), that's ok, after some I/O
operations!

Paolo.