Made file paths and command line arguments UTF-8 on Windows. Fixes #438. (#449)

include/bx/commandline.h

@@ -10,6 +10,7 @@
 
 namespace bx
 {
+	struct AllocatorI;
 	class  Error;
 	struct WriterI;
 
@@ -137,6 +138,34 @@ namespace bx
 		int32_t                  m_numOptions;
 	};
 
+	/// Process arguments as UTF-8.
+	///
+	/// On Windows `main` receives arguments converted to the ANSI code page, which loses every
+	/// character the code page can't represent, so they are taken from the process command line
+	/// instead. On other platforms, or when that fails, `_argv` is used as is.
+	///
+	class CommandLineArgs
+	{
+	public:
+		///
+		CommandLineArgs(AllocatorI* _allocator, int32_t _argc, const char* const* _argv);
+
+		///
+		~CommandLineArgs();
+
+		///
+		int32_t getArgc() const;
+
+		///
+		const char** getArgv() const;
+
+	private:
+		AllocatorI*  m_allocator;
+		void*        m_data;
+		int32_t      m_argc;
+		const char** m_argv;
+	};
+
 } // namespace bx
 
 #endif /// BX_COMMANDLINE_H_HEADER_GUARD

include/bx/filepath.h

@@ -41,6 +41,9 @@ namespace bx
 	///                 ^^^^^^^
 	///                 +-filename
 	///
+	/// Paths are UTF-8 on every platform. On Windows they are converted to UTF-16 before reaching
+	/// the OS, so the process code page has no effect on them.
+	///
 	class FilePath
 	{
 	public:

include/bx/os.h

@@ -128,7 +128,7 @@ namespace bx
 	template<typename ProtoT>
 	ProtoT dlsym(void* _handle, const StringView& _symbol);
 
-	/// Read an environment variable.
+	/// Read an environment variable. Name and value are UTF-8.
 	///
 	/// @param[out]    _out       Buffer that receives the value (may be NULL to query required size).
 	/// @param[in,out] _inOutSize On input the capacity of `_out`, on output the required size including terminator.
@@ -137,7 +137,7 @@ namespace bx
 	///
 	bool getEnv(char* _out, uint32_t* _inOutSize, const StringView& _name);
 
-	/// Set an environment variable.
+	/// Set an environment variable. Name and value are UTF-8.
 	///
 	/// @param[in] _name  Name of the environment variable.
 	/// @param[in] _value Value to assign.
@@ -146,7 +146,7 @@ namespace bx
 
 	/// Change current working directory of the process.
 	///
-	/// @param[in] _path Target directory path.
+	/// @param[in] _path Target directory path, UTF-8.
 	/// @returns 0 on success, non-zero on failure.
 	///
 	int chdir(const char* _path);

include/bx/string.h

@@ -549,6 +549,16 @@ namespace bx
 	/// Converts 64-bit unsigned integer value to string.
 	int32_t toString(char* _out, int32_t _max, uint64_t _value, uint32_t _base = 10, char _separator = '\0');
 
+	/// Converts UTF-8 string to NUL terminated UTF-16 string. Returns number of UTF-16 code units
+	/// written including the terminator, or 0 when the input is not valid UTF-8 or the output doesn't
+	/// fit. With `_max` 0 returns the size required.
+	int32_t toUtf16(uint16_t* _out, int32_t _max, const StringView& _str);
+
+	/// Converts NUL terminated UTF-16 string to NUL terminated UTF-8 string. Returns number of bytes
+	/// written including the terminator, or 0 when the input is not valid UTF-16 or the output
+	/// doesn't fit. With `_max` 0 returns the size required.
+	int32_t fromUtf16(char* _out, int32_t _max, const uint16_t* _str);
+
 	/// Converts string to bool value.
 	bool fromString(bool* _out, const StringView& _str);