This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: return-value policy question
On Tue, Apr 29, 2003 at 02:07:38PM -0400, Phil Edwards wrote:
> On Mon, Apr 28, 2003 at 10:30:12PM -0700, Nathan Myers wrote:
> > or the even simpler and more traditional
> >
> > return (_M_out_cur < _M_out_end) ?
> > traits_type::to_int_type(*_M_out_cur++ = __c) :
> > this->overflow(traits_type::to_int_type(__c));
>
> Normally I'm not a style bigot, but Sweet Lord, that's hideously ugly.
"Traditional" is the polite way to express it. (Although of course it's
really the "traits_type::to_int_type" call that makes it ugly. :-)
I included it as an extremum, not a proposal. However, if on some
platform it led to a 10% speed difference in streambuf performance,
I would barely hesitate to code it that way. (I.e. I probably
would try variations to see if something clearer produced the same
performance.)
Nathan Myers
ncm-nospam@cantrip.org