fgetws fails to get the exact wide char string from FILE*

问题

I am using fgetws to get some string line by line from a FILE. The FILE I have is from a popen command. Here is the code snippet:

    FILE* pInstalledApps = popen( command.c_str(), "r" );
    if( NULL != pInstalledApps )
    {
        wchar_t currentAppPath [kMaximumAppPathLength];

        // Reading app paths one line at a time.
        while ( ! feof (pInstalledApps) )
        {
            if ( fgetws  ( currentAppPath, kMaximumAppPathLength, pInstalledApps) == NULL )
            {
                break;
            }
            wchar_t *pCharPos = NULL;
            if ( ( pCharPos = wcschr( currentAppPath, L'\n' ) ) != NULL )
            {
                *pCharPos = L'\0';
            }
            std::wstring appPath( currentAppPath );

                            //Do something with the wstring
        }
        pclose( pInstalledApps );
    }

When the string currentAppPath that i gets has wide char strings, the appPath that I get doesnot have the expected string. For example if the string that I get from the FILE is 10teciêênks my appPath variable will be having 10tecieÌeÌnks.

回答1:

That looks like an encoding (or more specifically, decoding) issue. The wide-character API functions will not automatically detect the character encoding of the data. You need to set this in the application, for example:

#include <locale.h>

setlocale(LC_ALL, "en.UTF-8");

The man page for fgetws states:

The behavior of fgetws() depends on the LC_CTYPE category of the current locale.

so using:

setlocale(LC_CTYPE, "en.UTF-8");

should also work.

NOTE: The above assumes the data is UTF-8 encoded.

UPDATE: It is possible to preserve the current locale by doing something like:

char *prev_locale = strdup(setlocale(LC_CTYPE, NULL));
setlocale(LC_CTYPE, "en.UTF-8");
// ...
setlocale(LC_CTYPE, prev_locale);
free(prev_locale);

来源：https://stackoverflow.com/questions/23936105/fgetws-fails-to-get-the-exact-wide-char-string-from-file