This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: std::num_put<char, char*>


> Table 51 only mandates that has_facet<num_put<char> >(loc) shall be
> true, for any loc either constructed or returned by locale::classic().
> Nothing is said about, legal, additional, facets identified by a given
> locale. In principle, that is an unbounded set, one for each possible
> output iterator type...

Hmm, I must admit that that reply was a little beyond me. Thinking about
it more after reading your reply, is the issue that num_put is only
implemented as a specialization? Or is it implemented as a generic
template that will function with any output iterator type but only
instanced with ostreambuf_iterator? Am I on the right track? Long
question short, is there any way I can get code asking for num_put<char,
char*> to work with gcc without writing a specialization for it? I know
I could go through the code and convert all the stacked char arrays to
stringstreams, but that seems like an unnecessary de-optimization.

Thanks.
--
	Eddie


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]