Re: [boost] [locale] Strange(?) result in example code

Subject: Re: [boost] [locale] Strange(?) result in example code
From: Paul A. Bristow (pbristow_at_[hidden])
Date: 2011-04-18 07:30:13

> -----Original Message-----
> From: boost-bounces_at_[hidden] [mailto:boost-bounces_at_[hidden]]
> On Behalf Of Artyom
> Sent: Saturday, April 16, 2011 2:32 PM
> To: boost_at_[hidden]
> Subject: Re: [boost] [locale] Strange(?) result in example code
> 
> > >
> > > >  std::cout.imbue(std::locale()); // Now global locale imbued to
> > > > stream  as
> > well.
> > >
> > > LOL, DUH!  OK.  Thanks.
> >
> > Well join  the club - I've fallen into this pit recently :-(
> >
> > Am I correct in  thinking that the reason is that std::cout is
> > constructed
> > *before* the new  locale is  'globalled'?
> >
> 
> Yes.
> 
> > And so std::cerr needs imbuing too but that  fstream and stringstreams
> > constructed after the  std::locale::global("en_US.UTF-8") statement
> > will get the new  locale?
> 
> Yes, basically if you want to do it fully for standard library and boost:
> 
> std::locale::global(loc);
> std::cout.imbue(loc);
> std::cin.imbue(loc);
> std::cerr.imbue(loc);
> std::clog.imbue(loc);
> 
> std::wcout.imbue(loc);
> std::wcin.imbue(loc);
> std::wcerr.imbue(loc);
> std::wclog.imbue(loc);
> 
> // this is for boost filesystem
> boost::filesystem::path::imbue(loc);

Thanks for this confirmation.

This seems such an obviously attractive pit, I am sure Noah and I won't be
the last to fall in it ;-)

I think Boost.Locale docs would be a good place to remind users of locale of
its existence, including a reference to 27.4.2.3/4 of C++03.

Would you like to add it as a [warning ....] while you are crossing the 't's
and dotting the 'i's ?

Thanks

Paul

---
Paul A. Bristow,
Prizet Farmhouse, Kendal LA8 8AB  UK
[phone]  07714330204
pbristow_at_[hidden]