On May 3, 6:13 am, Daniel Krügler <daniel.krueg...@[EMAIL PROTECTED]
>
wrote:
> On 2 Mai, 12:43, "Bo Persson" <b...@[EMAIL PROTECTED]
> wrote:
>
>
>
> > The resolution to issue 581
>
> >http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2008/n2578.html#581
>
> > requires that basic_ostream::flush creates a sentry object to verify
> > the stream state.
>
> > However, for streams with the unit_buf flag set, like std::err, the
> > destructor of the sentry object will again call flush(). This seems to
> > create an infinite recursion for
>
> > std::cerr << std::flush;
>
> > or even
>
> > std::cerr << "Some message" << std::endl;
>
> > Have I missed something here?
The text in [ostream::sentry], p4 should probably be changed
to read something like
If ((os.flags() & ios_base::unitbuf) && !uncaught_exception())
is true, calls os.rdbuf()->pubsync().
to avoid this. Let me fix it along with the similar problem
with the tied stream.
>
> No, I agree with your analysis. Additionally the proposed
> resolution references the wrong section by saying in
> [ostream.unformatted]/7:
>
> "Behaves as an unformatted output function (as described in
> 27.6.2.6.1,
> paragraph 1).[..]"
>
> because 27.6.2.6.1/p.1 describes the semantic of *formatted*
> output functions.
I think the numbers have changed between C++ 98 and the latest
working paper. We should be using section names instead of
numbers, they don't change (i.e., [ostream.unformatted]).
--
[ See http://www.gotw.ca/resources/clcm.htm
for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]
|
