更新时间:2023-02-26 12:56:47
The problem with wofstream is that it accepts the wide string for the open() method but does not actually write wide characters to the file. You have to be explicit about that and imbue() it with a locale that has a codecvt with the encoding you want. Implementation of such a codecvt that produces a UTF encoding is still spotty, here's an example that uses Boost.