by guenter » 12 Nov 2010, 00:32
Because in POCO Unicode is generally handled by using UTF-8 encoded std::strings. This also applies to paths and other system identifiers. When interfacing the WIN32 API, POCO usually uses the Unicode APIs and transparently handles UTF-8 to UTF-16 conversion and vice versa. This has shown to be the best compromise with regards to cross-platform portability (compared to alternatives like using std::wstring, or, god forbid, yet another custom Unicode string class).