Split *AndAdjustOffset() functions into their own header, to restore utf_stri...

base/utf_string_conversion_utils.cc

Index: base/utf_string_conversion_utils.cc
===================================================================
--- base/utf_string_conversion_utils.cc	(revision 31511)
+++ base/utf_string_conversion_utils.cc	(working copy)
@@ -2,36 +2,18 @@
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
-#include "base/utf_string_conversions.h"
+#include "base/utf_string_conversion_utils.h"
 
-#include <vector>
-
-#include "base/basictypes.h"
-#include "base/logging.h"
-#include "base/string_util.h"
 #include "base/third_party/icu/icu_utf.h"
 
-namespace {
+namespace base {
 
-inline bool IsValidCodepoint(uint32 code_point) {
-  // Excludes the surrogate code points ([0xD800, 0xDFFF]) and
-  // codepoints larger than 0x10FFFF (the highest codepoint allowed).
-  // Non-characters and unassigned codepoints are allowed.
-  return code_point < 0xD800u ||
-         (code_point >= 0xE000u && code_point <= 0x10FFFFu);
-}
-
 // ReadUnicodeCharacter --------------------------------------------------------
 
-// Reads a UTF-8 stream, placing the next code point into the given output
-// |*code_point|. |src| represents the entire string to read, and |*char_index|
-// is the character offset within the string to start reading at. |*char_index|
-// will be updated to index the last character read, such that incrementing it
-// (as in a for loop) will take the reader to the next character.
-//
-// Returns true on success. On false, |*code_point| will be invalid.
-bool ReadUnicodeCharacter(const char* src, int32 src_len,
-                          int32* char_index, uint32* code_point_out) {
+bool ReadUnicodeCharacter(const char* src,
+                          int32 src_len,
+                          int32* char_index,
+                          uint32* code_point_out) {
   // U8_NEXT expects to be able to use -1 to signal an error, so we must
   // use a signed type for code_point.  But this function returns false
   // on error anyway, so code_point_out is unsigned.
@@ -47,9 +29,10 @@
   return IsValidCodepoint(code_point);
 }
 
-// Reads a UTF-16 character. The usage is the same as the 8-bit version above.
-bool ReadUnicodeCharacter(const char16* src, int32 src_len,
-                          int32* char_index, uint32* code_point) {
+bool ReadUnicodeCharacter(const char16* src,
+                          int32 src_len,
+                          int32* char_index,
+                          uint32* code_point) {
   if (CBU16_IS_SURROGATE(src[*char_index])) {
     if (!CBU16_IS_SURROGATE_LEAD(src[*char_index]) ||
         *char_index + 1 >= src_len ||
@@ -71,9 +54,10 @@
 }
 
 #if defined(WCHAR_T_IS_UTF32)
-// Reads UTF-32 character. The usage is the same as the 8-bit version above.
-bool ReadUnicodeCharacter(const wchar_t* src, int32 src_len,
-                          int32* char_index, uint32* code_point) {
+bool ReadUnicodeCharacter(const wchar_t* src,
+                          int32 src_len,
+                          int32* char_index,
+                          uint32* code_point) {
   // Conversion is easy since the source is 32-bit.
   *code_point = src[*char_index];
 
@@ -84,8 +68,6 @@
 
 // WriteUnicodeCharacter -------------------------------------------------------
 
-// Appends a UTF-8 character to the given 8-bit string.  Returns the number of
-// bytes written.
 size_t WriteUnicodeCharacter(uint32 code_point, std::string* output) {
   if (code_point <= 0x7f) {
     // Fast path the common case of one byte.
@@ -93,6 +75,7 @@
     return 1;
   }
 
+
   // CBU8_APPEND_UNSAFE can append up to 4 bytes.
   size_t char_offset = output->length();
   size_t original_char_offset = char_offset;
@@ -106,8 +89,6 @@
   return char_offset - original_char_offset;
 }
 
-// Appends the given code point as a UTF-16 character to the given 16-bit
-// string.  Returns the number of 16-bit values written.
 size_t WriteUnicodeCharacter(uint32 code_point, string16* output) {
   if (CBU16_LENGTH(code_point) == 1) {
     // Thie code point is in the Basic Multilingual Plane (BMP).
@@ -121,67 +102,8 @@
   return CBU16_MAX_LENGTH;
 }
 
-#if defined(WCHAR_T_IS_UTF32)
-// Appends the given UTF-32 character to the given 32-bit string.  Returns the
-// number of 32-bit values written.
-inline size_t WriteUnicodeCharacter(uint32 code_point, std::wstring* output) {
-  // This is the easy case, just append the character.
-  output->push_back(code_point);
-  return 1;
-}
-#endif  // defined(WCHAR_T_IS_UTF32)
-
 // Generalized Unicode converter -----------------------------------------------
 
-// Converts the given source Unicode character type to the given destination
-// Unicode character type as a STL string. The given input buffer and size
-// determine the source, and the given output STL string will be replaced by
-// the result.
-template<typename SRC_CHAR, typename DEST_STRING>
-bool ConvertUnicode(const SRC_CHAR* src,
-                    size_t src_len,
-                    DEST_STRING* output,
-                    size_t* offset_for_adjustment) {
-  size_t output_offset =
-      (offset_for_adjustment && *offset_for_adjustment < src_len) ?
-          *offset_for_adjustment : DEST_STRING::npos;
-
-  // ICU requires 32-bit numbers.
-  bool success = true;
-  int32 src_len32 = static_cast<int32>(src_len);
-  for (int32 i = 0; i < src_len32; i++) {
-    uint32 code_point;
-    size_t original_i = i;
-    size_t chars_written = 0;
-    if (ReadUnicodeCharacter(src, src_len32, &i, &code_point)) {
-      chars_written = WriteUnicodeCharacter(code_point, output);
-    } else {
-      // TODO(jungshik): consider adding 'Replacement character' (U+FFFD)
-      // in place of an invalid codepoint.
-      success = false;
-    }
-    if ((output_offset != DEST_STRING::npos) &&
-        (*offset_for_adjustment > original_i)) {
-      // NOTE: ReadUnicodeCharacter() adjusts |i| to point _at_ the last
-      // character read, not after it (so that incrementing it in the loop
-      // increment will place it at the right location), so we need to account
-      // for that in determining the amount that was read.
-      if (*offset_for_adjustment <= static_cast<size_t>(i))
-        output_offset = DEST_STRING::npos;
-      else
-        output_offset += chars_written - (i - original_i + 1);
-    }
-  }
-
-  if (offset_for_adjustment)
-    *offset_for_adjustment = output_offset;
-  return success;
-}
-
-// Guesses the length of the output in UTF-8 in bytes, clears that output
-// string, and reserves that amount of space.  We assume that the input
-// character types are unsigned, which will be true for UTF-16 and -32 on our
-// systems.
 template<typename CHAR>
 void PrepareForUTF8Output(const CHAR* src,
                           size_t src_len,
@@ -198,8 +120,10 @@
   }
 }
 
-// Prepares an output buffer (containing either UTF-16 or -32 data) given some
-// UTF-8 input that will be converted to it.  See PrepareForUTF8Output().
+// Instantiate versions we know callers will need.
+template void PrepareForUTF8Output(const wchar_t*, size_t, std::string*);
+template void PrepareForUTF8Output(const char16*, size_t, std::string*);
+
 template<typename STRING>
 void PrepareForUTF16Or32Output(const char* src,
                                size_t src_len,
@@ -217,156 +141,8 @@
   }
 }
 
-}  // namespace
+// Instantiate versions we know callers will need.
+template void PrepareForUTF16Or32Output(const char*, size_t, std::wstring*);
+template void PrepareForUTF16Or32Output(const char*, size_t, string16*);
 
-// UTF-8 <-> Wide --------------------------------------------------------------
-
-bool WideToUTF8(const wchar_t* src, size_t src_len, std::string* output) {
-  PrepareForUTF8Output(src, src_len, output);
-  return ConvertUnicode<wchar_t, std::string>(src, src_len, output, NULL);
-}
-
-std::string WideToUTF8(const std::wstring& wide) {
-  std::string ret;
-  // Ignore the success flag of this call, it will do the best it can for
-  // invalid input, which is what we want here.
-  WideToUTF8(wide.data(), wide.length(), &ret);
-  return ret;
-}
-
-bool UTF8ToWideAndAdjustOffset(const char* src,
-                               size_t src_len,
-                               std::wstring* output,
-                               size_t* offset_for_adjustment) {
-  PrepareForUTF16Or32Output(src, src_len, output);
-  return ConvertUnicode<char, std::wstring>(src, src_len, output,
-                                            offset_for_adjustment);
-}
-
-std::wstring UTF8ToWideAndAdjustOffset(const base::StringPiece& utf8,
-                                       size_t* offset_for_adjustment) {
-  std::wstring ret;
-  UTF8ToWideAndAdjustOffset(utf8.data(), utf8.length(), &ret,
-                            offset_for_adjustment);
-  return ret;
-}
-
-// UTF-16 <-> Wide -------------------------------------------------------------
-
-#if defined(WCHAR_T_IS_UTF16)
-
-// When wide == UTF-16, then conversions are a NOP.
-bool WideToUTF16(const wchar_t* src, size_t src_len, string16* output) {
-  output->assign(src, src_len);
-  return true;
-}
-
-string16 WideToUTF16(const std::wstring& wide) {
-  return wide;
-}
-
-bool UTF16ToWideAndAdjustOffset(const char16* src,
-                                size_t src_len,
-                                std::wstring* output,
-                                size_t* offset_for_adjustment) {
-  output->assign(src, src_len);
-  if (offset_for_adjustment && (*offset_for_adjustment >= src_len))
-    *offset_for_adjustment = std::wstring::npos;
-  return true;
-}
-
-std::wstring UTF16ToWideAndAdjustOffset(const string16& utf16,
-                                        size_t* offset_for_adjustment) {
-  if (offset_for_adjustment && (*offset_for_adjustment >= utf16.length()))
-    *offset_for_adjustment = std::wstring::npos;
-  return utf16;
-}
-
-#elif defined(WCHAR_T_IS_UTF32)
-
-bool WideToUTF16(const wchar_t* src, size_t src_len, string16* output) {
-  output->clear();
-  // Assume that normally we won't have any non-BMP characters so the counts
-  // will be the same.
-  output->reserve(src_len);
-  return ConvertUnicode<wchar_t, string16>(src, src_len, output, NULL);
-}
-
-string16 WideToUTF16(const std::wstring& wide) {
-  string16 ret;
-  WideToUTF16(wide.data(), wide.length(), &ret);
-  return ret;
-}
-
-bool UTF16ToWideAndAdjustOffset(const char16* src,
-                                size_t src_len,
-                                std::wstring* output,
-                                size_t* offset_for_adjustment) {
-  output->clear();
-  // Assume that normally we won't have any non-BMP characters so the counts
-  // will be the same.
-  output->reserve(src_len);
-  return ConvertUnicode<char16, std::wstring>(src, src_len, output,
-                                              offset_for_adjustment);
-}
-
-std::wstring UTF16ToWideAndAdjustOffset(const string16& utf16,
-                                        size_t* offset_for_adjustment) {
-  std::wstring ret;
-  UTF16ToWideAndAdjustOffset(utf16.data(), utf16.length(), &ret,
-                             offset_for_adjustment);
-  return ret;
-}
-
-#endif  // defined(WCHAR_T_IS_UTF32)
-
-// UTF16 <-> UTF8 --------------------------------------------------------------
-
-#if defined(WCHAR_T_IS_UTF32)
-
-bool UTF8ToUTF16(const char* src, size_t src_len, string16* output) {
-  PrepareForUTF16Or32Output(src, src_len, output);
-  return ConvertUnicode<char, string16>(src, src_len, output, NULL);
-}
-
-string16 UTF8ToUTF16(const std::string& utf8) {
-  string16 ret;
-  // Ignore the success flag of this call, it will do the best it can for
-  // invalid input, which is what we want here.
-  UTF8ToUTF16(utf8.data(), utf8.length(), &ret);
-  return ret;
-}
-
-bool UTF16ToUTF8(const char16* src, size_t src_len, std::string* output) {
-  PrepareForUTF8Output(src, src_len, output);
-  return ConvertUnicode<char16, std::string>(src, src_len, output, NULL);
-}
-
-std::string UTF16ToUTF8(const string16& utf16) {
-  std::string ret;
-  // Ignore the success flag of this call, it will do the best it can for
-  // invalid input, which is what we want here.
-  UTF16ToUTF8(utf16.data(), utf16.length(), &ret);
-  return ret;
-}
-
-#elif defined(WCHAR_T_IS_UTF16)
-// Easy case since we can use the "wide" versions we already wrote above.
-
-bool UTF8ToUTF16(const char* src, size_t src_len, string16* output) {
-  return UTF8ToWide(src, src_len, output);
-}
-
-string16 UTF8ToUTF16(const std::string& utf8) {
-  return UTF8ToWide(utf8);
-}
-
-bool UTF16ToUTF8(const char16* src, size_t src_len, std::string* output) {
-  return WideToUTF8(src, src_len, output);
-}
-
-std::string UTF16ToUTF8(const string16& utf16) {
-  return WideToUTF8(utf16);
-}
-
-#endif
+}  // namespace base