problem about wofstream

[mimi]

With a wofstream, we can input wstring into a file. But when i input
some wide character of chinese characters into a file, some error
occur.
With the implementation of STLport 4.62, the wstring of chines
characters were written to the file; But with the implementation of
MS, that is both in VC6.0 and Visual Studio 2005, noting was written
to the file.I don't know why and I wonder to know what the C++
standard said about this? What does the standard guarantees while
dealing with unicode characters?Where could I find such information?
Thanks for any advice.
#include <fstream>

int main()
{
std::wofstream strm2("test.txt");
strm2 << L"ÎÒ°®±±¾©Ìì°²ÃÅ";

return 0;
}

 

[chunhuachunhua1984]

With a wofstream, we can input wstring into a file. But when i input
some wide character of chinese characters into a file, some error
occur.
With the implementation of STLport 4.62, the wstring of chines
characters were written to the file; But with the implementation of
MS, that is both in VC6.0 and Visual Studio 2005, noting was written
to the file.I don't know why and I wonder to know what the C++
standard said about this? What does the standard guarantees while
dealing with unicode characters?Where could I find such information?
Thanks for any advice.
#include <fstream>

int main()
{
std::wofstream strm2("test.txt");
strm2 << L"ÎÒ°®±±¾©Ìì°²ÃÅ";

return 0;

I think u should call strm2.close() before return 0,and i find it can
good work.

 

[Guest]

With a wofstream, we can input wstring into a file. But when i input
some wide character of chinese characters into a file, some error
occur.
With the implementation of STLport 4.62, the wstring of chines
characters were written to the file; But with the implementation of
MS, that is both in VC6.0 and Visual Studio 2005, noting was written
to the file.I don't know why and I wonder to know what the C++
standard said about this? What does the standard guarantees while
dealing with unicode characters?Where could I find such information?
Thanks for any advice.
#include <fstream>

int main()
{
std::wofstream strm2("test.txt");
strm2 << L"我爱北京天安门";

return 0;

}

The wide output streams in MSVC don't act as you would expect. The
streams are not able to handle any characters outside the 0..255
range.

To do better you'll have to get different libraries, or write your own
wrapper to do the translations for the formats you want to use.

You may also find that you have problems with the source code if you
drop those characters in like that too, but I admit to being unsure as
to what text formats the compiler assumes or can read.


K

 

[P.J. Plauger]

With a wofstream, we can input wstring into a file. But when i input
some wide character of chinese characters into a file, some error
occur.
With the implementation of STLport 4.62, the wstring of chines
characters were written to the file; But with the implementation of
MS, that is both in VC6.0 and Visual Studio 2005, noting was written
to the file.I don't know why and I wonder to know what the C++
standard said about this? What does the standard guarantees while
dealing with unicode characters?Where could I find such information?
Thanks for any advice.
#include <fstream>

int main()
{
std::wofstream strm2("test.txt");
strm2 << L"???????";

return 0;

}

The wide output streams in MSVC don't act as you would expect. The
streams are not able to handle any characters outside the 0..255
range.

[pjp] You're talking about the default codecvt facet, which has this
limitation. The wide streams work fine.

To do better you'll have to get different libraries, or write your own
wrapper to do the translations for the formats you want to use.

[pjp] Or get a better codecvt facet. See, for example, our Compleat
Library add-on, which defines dozens.

P.J. Plauger
Dinkumware, Ltd.
http://www.dinkumware.com