[RFC] A doubt about our istream::sentry::sentry

[Paolo Carlini]

Martin Sebor wrote:

> I think this is the same as issue 419:
> http://www.open-std.org/jtc1/sc22/wg21/prot/14882fdis/n1635.html#419

By the way, I agree 100% with you about this issue, in particular I 
agree with
your reply to Jerry Schwarz. Interestingly, if I'm not mistaken, v3 to 
be safe ;)
probably sets the failbit *2* times, both in the sentry, and then again, 
in the
extractor!

Paolo.