michael@0: // Copyright (c) 2010 The Chromium Authors. All rights reserved.
michael@0: // Use of this source code is governed by a BSD-style license that can be
michael@0: // found in the LICENSE file.
michael@0: 
michael@0: #include "base/strings/utf_string_conversions.h"
michael@0: 
michael@0: #include "base/strings/string_piece.h"
michael@0: #include "base/strings/string_util.h"
michael@0: #include "base/strings/utf_string_conversion_utils.h"
michael@0: 
michael@0: namespace base {
michael@0: 
michael@0: namespace {
michael@0: 
michael@0: // Generalized Unicode converter -----------------------------------------------
michael@0: 
michael@0: // Converts the given source Unicode character type to the given destination
michael@0: // Unicode character type as a STL string. The given input buffer and size
michael@0: // determine the source, and the given output STL string will be replaced by
michael@0: // the result.
michael@0: template<typename SRC_CHAR, typename DEST_STRING>
michael@0: bool ConvertUnicode(const SRC_CHAR* src,
michael@0:                     size_t src_len,
michael@0:                     DEST_STRING* output) {
michael@0:   // ICU requires 32-bit numbers.
michael@0:   bool success = true;
michael@0:   int32 src_len32 = static_cast<int32>(src_len);
michael@0:   for (int32 i = 0; i < src_len32; i++) {
michael@0:     uint32 code_point;
michael@0:     if (ReadUnicodeCharacter(src, src_len32, &i, &code_point)) {
michael@0:       WriteUnicodeCharacter(code_point, output);
michael@0:     } else {
michael@0:       WriteUnicodeCharacter(0xFFFD, output);
michael@0:       success = false;
michael@0:     }
michael@0:   }
michael@0: 
michael@0:   return success;
michael@0: }
michael@0: 
michael@0: }  // namespace
michael@0: 
michael@0: // UTF-8 <-> Wide --------------------------------------------------------------
michael@0: 
michael@0: bool WideToUTF8(const wchar_t* src, size_t src_len, std::string* output) {
michael@0:   PrepareForUTF8Output(src, src_len, output);
michael@0:   return ConvertUnicode(src, src_len, output);
michael@0: }
michael@0: 
michael@0: std::string WideToUTF8(const std::wstring& wide) {
michael@0:   std::string ret;
michael@0:   // Ignore the success flag of this call, it will do the best it can for
michael@0:   // invalid input, which is what we want here.
michael@0:   WideToUTF8(wide.data(), wide.length(), &ret);
michael@0:   return ret;
michael@0: }
michael@0: 
michael@0: bool UTF8ToWide(const char* src, size_t src_len, std::wstring* output) {
michael@0:   PrepareForUTF16Or32Output(src, src_len, output);
michael@0:   return ConvertUnicode(src, src_len, output);
michael@0: }
michael@0: 
michael@0: std::wstring UTF8ToWide(const StringPiece& utf8) {
michael@0:   std::wstring ret;
michael@0:   UTF8ToWide(utf8.data(), utf8.length(), &ret);
michael@0:   return ret;
michael@0: }
michael@0: 
michael@0: // UTF-16 <-> Wide -------------------------------------------------------------
michael@0: 
michael@0: #if defined(WCHAR_T_IS_UTF16)
michael@0: 
michael@0: // When wide == UTF-16, then conversions are a NOP.
michael@0: bool WideToUTF16(const wchar_t* src, size_t src_len, string16* output) {
michael@0:   output->assign(src, src_len);
michael@0:   return true;
michael@0: }
michael@0: 
michael@0: string16 WideToUTF16(const std::wstring& wide) {
michael@0:   return wide;
michael@0: }
michael@0: 
michael@0: bool UTF16ToWide(const char16* src, size_t src_len, std::wstring* output) {
michael@0:   output->assign(src, src_len);
michael@0:   return true;
michael@0: }
michael@0: 
michael@0: std::wstring UTF16ToWide(const string16& utf16) {
michael@0:   return utf16;
michael@0: }
michael@0: 
michael@0: #elif defined(WCHAR_T_IS_UTF32)
michael@0: 
michael@0: bool WideToUTF16(const wchar_t* src, size_t src_len, string16* output) {
michael@0:   output->clear();
michael@0:   // Assume that normally we won't have any non-BMP characters so the counts
michael@0:   // will be the same.
michael@0:   output->reserve(src_len);
michael@0:   return ConvertUnicode(src, src_len, output);
michael@0: }
michael@0: 
michael@0: string16 WideToUTF16(const std::wstring& wide) {
michael@0:   string16 ret;
michael@0:   WideToUTF16(wide.data(), wide.length(), &ret);
michael@0:   return ret;
michael@0: }
michael@0: 
michael@0: bool UTF16ToWide(const char16* src, size_t src_len, std::wstring* output) {
michael@0:   output->clear();
michael@0:   // Assume that normally we won't have any non-BMP characters so the counts
michael@0:   // will be the same.
michael@0:   output->reserve(src_len);
michael@0:   return ConvertUnicode(src, src_len, output);
michael@0: }
michael@0: 
michael@0: std::wstring UTF16ToWide(const string16& utf16) {
michael@0:   std::wstring ret;
michael@0:   UTF16ToWide(utf16.data(), utf16.length(), &ret);
michael@0:   return ret;
michael@0: }
michael@0: 
michael@0: #endif  // defined(WCHAR_T_IS_UTF32)
michael@0: 
michael@0: // UTF16 <-> UTF8 --------------------------------------------------------------
michael@0: 
michael@0: #if defined(WCHAR_T_IS_UTF32)
michael@0: 
michael@0: bool UTF8ToUTF16(const char* src, size_t src_len, string16* output) {
michael@0:   PrepareForUTF16Or32Output(src, src_len, output);
michael@0:   return ConvertUnicode(src, src_len, output);
michael@0: }
michael@0: 
michael@0: string16 UTF8ToUTF16(const StringPiece& utf8) {
michael@0:   string16 ret;
michael@0:   // Ignore the success flag of this call, it will do the best it can for
michael@0:   // invalid input, which is what we want here.
michael@0:   UTF8ToUTF16(utf8.data(), utf8.length(), &ret);
michael@0:   return ret;
michael@0: }
michael@0: 
michael@0: bool UTF16ToUTF8(const char16* src, size_t src_len, std::string* output) {
michael@0:   PrepareForUTF8Output(src, src_len, output);
michael@0:   return ConvertUnicode(src, src_len, output);
michael@0: }
michael@0: 
michael@0: std::string UTF16ToUTF8(const string16& utf16) {
michael@0:   std::string ret;
michael@0:   // Ignore the success flag of this call, it will do the best it can for
michael@0:   // invalid input, which is what we want here.
michael@0:   UTF16ToUTF8(utf16.data(), utf16.length(), &ret);
michael@0:   return ret;
michael@0: }
michael@0: 
michael@0: #elif defined(WCHAR_T_IS_UTF16)
michael@0: // Easy case since we can use the "wide" versions we already wrote above.
michael@0: 
michael@0: bool UTF8ToUTF16(const char* src, size_t src_len, string16* output) {
michael@0:   return UTF8ToWide(src, src_len, output);
michael@0: }
michael@0: 
michael@0: string16 UTF8ToUTF16(const StringPiece& utf8) {
michael@0:   return UTF8ToWide(utf8);
michael@0: }
michael@0: 
michael@0: bool UTF16ToUTF8(const char16* src, size_t src_len, std::string* output) {
michael@0:   return WideToUTF8(src, src_len, output);
michael@0: }
michael@0: 
michael@0: std::string UTF16ToUTF8(const string16& utf16) {
michael@0:   return WideToUTF8(utf16);
michael@0: }
michael@0: 
michael@0: #endif
michael@0: 
michael@0: std::wstring ASCIIToWide(const StringPiece& ascii) {
michael@0:   DCHECK(IsStringASCII(ascii)) << ascii;
michael@0:   return std::wstring(ascii.begin(), ascii.end());
michael@0: }
michael@0: 
michael@0: string16 ASCIIToUTF16(const StringPiece& ascii) {
michael@0:   DCHECK(IsStringASCII(ascii)) << ascii;
michael@0:   return string16(ascii.begin(), ascii.end());
michael@0: }
michael@0: 
michael@0: }  // namespace base