Move utf_string_conversions to strings/ and add namespace.

This keeps a forwarding header in the old location and adds "using" statements to avoid changing all callers.

BUG=None
R=tfarina@chromium.org

Review URL: https://codereview.chromium.org/14126006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@198542 0039d316-1c4b-4281-b951-d872f2087c98

3 files changed, 468 insertions, 0 deletions

diff --git a/base/strings/utf_string_conversions.cc b/base/strings/utf_string_conversions.cc
new file mode 100644
index 0000000..b75ed0c
--- /dev/null
+++ b/base/strings/utf_string_conversions.cc
@@ -0,0 +1,185 @@
+// Copyright (c) 2010 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/strings/utf_string_conversions.h"
+
+#include "base/string_util.h"
+#include "base/strings/string_piece.h"
+#include "base/strings/utf_string_conversion_utils.h"
+
+namespace base {
+
+namespace {
+
+// 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) {
+  // 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;
+    if (ReadUnicodeCharacter(src, src_len32, &i, &code_point)) {
+      WriteUnicodeCharacter(code_point, output);
+    } else {
+      WriteUnicodeCharacter(0xFFFD, output);
+      success = false;
+    }
+  }
+
+  return success;
+}
+
+}  // namespace
+
+// UTF-8 <-> Wide --------------------------------------------------------------
+
+bool WideToUTF8(const wchar_t* src, size_t src_len, std::string* output) {
+  PrepareForUTF8Output(src, src_len, output);
+  return ConvertUnicode(src, src_len, output);
+}
+
+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 UTF8ToWide(const char* src, size_t src_len, std::wstring* output) {
+  PrepareForUTF16Or32Output(src, src_len, output);
+  return ConvertUnicode(src, src_len, output);
+}
+
+std::wstring UTF8ToWide(const StringPiece& utf8) {
+  std::wstring ret;
+  UTF8ToWide(utf8.data(), utf8.length(), &ret);
+  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 UTF16ToWide(const char16* src, size_t src_len, std::wstring* output) {
+  output->assign(src, src_len);
+  return true;
+}
+
+std::wstring UTF16ToWide(const string16& utf16) {
+  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(src, src_len, output);
+}
+
+string16 WideToUTF16(const std::wstring& wide) {
+  string16 ret;
+  WideToUTF16(wide.data(), wide.length(), &ret);
+  return ret;
+}
+
+bool UTF16ToWide(const char16* src, size_t src_len, std::wstring* 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(src, src_len, output);
+}
+
+std::wstring UTF16ToWide(const string16& utf16) {
+  std::wstring ret;
+  UTF16ToWide(utf16.data(), utf16.length(), &ret);
+  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(src, src_len, output);
+}
+
+string16 UTF8ToUTF16(const StringPiece& 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(src, src_len, output);
+}
+
+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 StringPiece& 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
+
+std::wstring ASCIIToWide(const StringPiece& ascii) {
+  DCHECK(IsStringASCII(ascii)) << ascii;
+  return std::wstring(ascii.begin(), ascii.end());
+}
+
+string16 ASCIIToUTF16(const StringPiece& ascii) {
+  DCHECK(IsStringASCII(ascii)) << ascii;
+  return string16(ascii.begin(), ascii.end());
+}
+
+}  // namespace base
diff --git a/base/strings/utf_string_conversions.h b/base/strings/utf_string_conversions.h
new file mode 100644
index 0000000..ec6a87f
--- /dev/null
+++ b/base/strings/utf_string_conversions.h
@@ -0,0 +1,72 @@
+// Copyright (c) 2011 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef BASE_STRINGS_UTF_STRING_CONVERSIONS_H_
+#define BASE_STRINGS_UTF_STRING_CONVERSIONS_H_
+
+#include <string>
+
+#include "base/base_export.h"
+#include "base/string16.h"
+#include "base/strings/string_piece.h"
+
+namespace base {
+
+// These convert between UTF-8, -16, and -32 strings. They are potentially slow,
+// so avoid unnecessary conversions. The low-level versions return a boolean
+// indicating whether the conversion was 100% valid. In this case, it will still
+// do the best it can and put the result in the output buffer. The versions that
+// return strings ignore this error and just return the best conversion
+// possible.
+BASE_EXPORT bool WideToUTF8(const wchar_t* src, size_t src_len,
+                            std::string* output);
+BASE_EXPORT std::string WideToUTF8(const std::wstring& wide);
+BASE_EXPORT bool UTF8ToWide(const char* src, size_t src_len,
+                            std::wstring* output);
+BASE_EXPORT std::wstring UTF8ToWide(const StringPiece& utf8);
+
+BASE_EXPORT bool WideToUTF16(const wchar_t* src, size_t src_len,
+                             string16* output);
+BASE_EXPORT string16 WideToUTF16(const std::wstring& wide);
+BASE_EXPORT bool UTF16ToWide(const char16* src, size_t src_len,
+                             std::wstring* output);
+BASE_EXPORT std::wstring UTF16ToWide(const string16& utf16);
+
+BASE_EXPORT bool UTF8ToUTF16(const char* src, size_t src_len, string16* output);
+BASE_EXPORT string16 UTF8ToUTF16(const StringPiece& utf8);
+BASE_EXPORT bool UTF16ToUTF8(const char16* src, size_t src_len,
+                             std::string* output);
+BASE_EXPORT std::string UTF16ToUTF8(const string16& utf16);
+
+// We are trying to get rid of wstring as much as possible, but it's too big
+// a mess to do it all at once.  These conversions should be used when we
+// really should just be passing a string16 around, but we haven't finished
+// porting whatever module uses wstring and the conversion is being used as a
+// stopcock.  This makes it easy to grep for the ones that should be removed.
+#if defined(OS_WIN)
+# define WideToUTF16Hack
+# define UTF16ToWideHack
+#else
+# define WideToUTF16Hack WideToUTF16
+# define UTF16ToWideHack UTF16ToWide
+#endif
+
+// These convert an ASCII string, typically a hardcoded constant, to a
+// UTF16/Wide string.
+BASE_EXPORT std::wstring ASCIIToWide(const StringPiece& ascii);
+BASE_EXPORT string16 ASCIIToUTF16(const StringPiece& ascii);
+
+}  // namespace base
+
+// TODO(brettw) remove these when callers are fixed up.
+using base::WideToUTF8;
+using base::UTF8ToWide;
+using base::WideToUTF16;
+using base::UTF16ToWide;
+using base::UTF8ToUTF16;
+using base::UTF16ToUTF8;
+using base::ASCIIToWide;
+using base::ASCIIToUTF16;
+
+#endif  // BASE_STRINGS_UTF_STRING_CONVERSIONS_H_
diff --git a/base/strings/utf_string_conversions_unittest.cc b/base/strings/utf_string_conversions_unittest.cc
new file mode 100644
index 0000000..974184e
--- /dev/null
+++ b/base/strings/utf_string_conversions_unittest.cc
@@ -0,0 +1,211 @@
+// Copyright (c) 2010 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/basictypes.h"
+#include "base/logging.h"
+#include "base/string_util.h"
+#include "base/strings/string_piece.h"
+#include "base/strings/utf_string_conversions.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace base {
+
+namespace {
+
+const wchar_t* const kConvertRoundtripCases[] = {
+  L"Google Video",
+  // "网页 图片 资讯更多 »"
+  L"\x7f51\x9875\x0020\x56fe\x7247\x0020\x8d44\x8baf\x66f4\x591a\x0020\x00bb",
+  //  "Παγκόσμιος Ιστός"
+  L"\x03a0\x03b1\x03b3\x03ba\x03cc\x03c3\x03bc\x03b9"
+  L"\x03bf\x03c2\x0020\x0399\x03c3\x03c4\x03cc\x03c2",
+  // "Поиск страниц на русском"
+  L"\x041f\x043e\x0438\x0441\x043a\x0020\x0441\x0442"
+  L"\x0440\x0430\x043d\x0438\x0446\x0020\x043d\x0430"
+  L"\x0020\x0440\x0443\x0441\x0441\x043a\x043e\x043c",
+  // "전체서비스"
+  L"\xc804\xccb4\xc11c\xbe44\xc2a4",
+
+  // Test characters that take more than 16 bits. This will depend on whether
+  // wchar_t is 16 or 32 bits.
+#if defined(WCHAR_T_IS_UTF16)
+  L"\xd800\xdf00",
+  // ?????  (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 : A,B,C,D,E)
+  L"\xd807\xdd40\xd807\xdd41\xd807\xdd42\xd807\xdd43\xd807\xdd44",
+#elif defined(WCHAR_T_IS_UTF32)
+  L"\x10300",
+  // ?????  (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 : A,B,C,D,E)
+  L"\x11d40\x11d41\x11d42\x11d43\x11d44",
+#endif
+};
+
+}  // namespace
+
+TEST(UTFStringConversionsTest, ConvertUTF8AndWide) {
+  // we round-trip all the wide strings through UTF-8 to make sure everything
+  // agrees on the conversion. This uses the stream operators to test them
+  // simultaneously.
+  for (size_t i = 0; i < arraysize(kConvertRoundtripCases); ++i) {
+    std::ostringstream utf8;
+    utf8 << WideToUTF8(kConvertRoundtripCases[i]);
+    std::wostringstream wide;
+    wide << UTF8ToWide(utf8.str());
+
+    EXPECT_EQ(kConvertRoundtripCases[i], wide.str());
+  }
+}
+
+TEST(UTFStringConversionsTest, ConvertUTF8AndWideEmptyString) {
+  // An empty std::wstring should be converted to an empty std::string,
+  // and vice versa.
+  std::wstring wempty;
+  std::string empty;
+  EXPECT_EQ(empty, WideToUTF8(wempty));
+  EXPECT_EQ(wempty, UTF8ToWide(empty));
+}
+
+TEST(UTFStringConversionsTest, ConvertUTF8ToWide) {
+  struct UTF8ToWideCase {
+    const char* utf8;
+    const wchar_t* wide;
+    bool success;
+  } convert_cases[] = {
+    // Regular UTF-8 input.
+    {"\xe4\xbd\xa0\xe5\xa5\xbd", L"\x4f60\x597d", true},
+    // Non-character is passed through.
+    {"\xef\xbf\xbfHello", L"\xffffHello", true},
+    // Truncated UTF-8 sequence.
+    {"\xe4\xa0\xe5\xa5\xbd", L"\xfffd\x597d", false},
+    // Truncated off the end.
+    {"\xe5\xa5\xbd\xe4\xa0", L"\x597d\xfffd", false},
+    // Non-shortest-form UTF-8.
+    {"\xf0\x84\xbd\xa0\xe5\xa5\xbd", L"\xfffd\x597d", false},
+    // This UTF-8 character decodes to a UTF-16 surrogate, which is illegal.
+    {"\xed\xb0\x80", L"\xfffd", false},
+    // Non-BMP characters. The second is a non-character regarded as valid.
+    // The result will either be in UTF-16 or UTF-32.
+#if defined(WCHAR_T_IS_UTF16)
+    {"A\xF0\x90\x8C\x80z", L"A\xd800\xdf00z", true},
+    {"A\xF4\x8F\xBF\xBEz", L"A\xdbff\xdffez", true},
+#elif defined(WCHAR_T_IS_UTF32)
+    {"A\xF0\x90\x8C\x80z", L"A\x10300z", true},
+    {"A\xF4\x8F\xBF\xBEz", L"A\x10fffez", true},
+#endif
+  };
+
+  for (size_t i = 0; i < ARRAYSIZE_UNSAFE(convert_cases); i++) {
+    std::wstring converted;
+    EXPECT_EQ(convert_cases[i].success,
+              UTF8ToWide(convert_cases[i].utf8,
+                         strlen(convert_cases[i].utf8),
+                         &converted));
+    std::wstring expected(convert_cases[i].wide);
+    EXPECT_EQ(expected, converted);
+  }
+
+  // Manually test an embedded NULL.
+  std::wstring converted;
+  EXPECT_TRUE(UTF8ToWide("\00Z\t", 3, &converted));
+  ASSERT_EQ(3U, converted.length());
+  EXPECT_EQ(static_cast<wchar_t>(0), converted[0]);
+  EXPECT_EQ('Z', converted[1]);
+  EXPECT_EQ('\t', converted[2]);
+
+  // Make sure that conversion replaces, not appends.
+  EXPECT_TRUE(UTF8ToWide("B", 1, &converted));
+  ASSERT_EQ(1U, converted.length());
+  EXPECT_EQ('B', converted[0]);
+}
+
+#if defined(WCHAR_T_IS_UTF16)
+// This test is only valid when wchar_t == UTF-16.
+TEST(UTFStringConversionsTest, ConvertUTF16ToUTF8) {
+  struct WideToUTF8Case {
+    const wchar_t* utf16;
+    const char* utf8;
+    bool success;
+  } convert_cases[] = {
+    // Regular UTF-16 input.
+    {L"\x4f60\x597d", "\xe4\xbd\xa0\xe5\xa5\xbd", true},
+    // Test a non-BMP character.
+    {L"\xd800\xdf00", "\xF0\x90\x8C\x80", true},
+    // Non-characters are passed through.
+    {L"\xffffHello", "\xEF\xBF\xBFHello", true},
+    {L"\xdbff\xdffeHello", "\xF4\x8F\xBF\xBEHello", true},
+    // The first character is a truncated UTF-16 character.
+    {L"\xd800\x597d", "\xef\xbf\xbd\xe5\xa5\xbd", false},
+    // Truncated at the end.
+    {L"\x597d\xd800", "\xe5\xa5\xbd\xef\xbf\xbd", false},
+  };
+
+  for (int i = 0; i < arraysize(convert_cases); i++) {
+    std::string converted;
+    EXPECT_EQ(convert_cases[i].success,
+              WideToUTF8(convert_cases[i].utf16,
+                         wcslen(convert_cases[i].utf16),
+                         &converted));
+    std::string expected(convert_cases[i].utf8);
+    EXPECT_EQ(expected, converted);
+  }
+}
+
+#elif defined(WCHAR_T_IS_UTF32)
+// This test is only valid when wchar_t == UTF-32.
+TEST(UTFStringConversionsTest, ConvertUTF32ToUTF8) {
+  struct WideToUTF8Case {
+    const wchar_t* utf32;
+    const char* utf8;
+    bool success;
+  } convert_cases[] = {
+    // Regular 16-bit input.
+    {L"\x4f60\x597d", "\xe4\xbd\xa0\xe5\xa5\xbd", true},
+    // Test a non-BMP character.
+    {L"A\x10300z", "A\xF0\x90\x8C\x80z", true},
+    // Non-characters are passed through.
+    {L"\xffffHello", "\xEF\xBF\xBFHello", true},
+    {L"\x10fffeHello", "\xF4\x8F\xBF\xBEHello", true},
+    // Invalid Unicode code points.
+    {L"\xfffffffHello", "\xEF\xBF\xBDHello", false},
+    // The first character is a truncated UTF-16 character.
+    {L"\xd800\x597d", "\xef\xbf\xbd\xe5\xa5\xbd", false},
+    {L"\xdc01Hello", "\xef\xbf\xbdHello", false},
+  };
+
+  for (size_t i = 0; i < ARRAYSIZE_UNSAFE(convert_cases); i++) {
+    std::string converted;
+    EXPECT_EQ(convert_cases[i].success,
+              WideToUTF8(convert_cases[i].utf32,
+                         wcslen(convert_cases[i].utf32),
+                         &converted));
+    std::string expected(convert_cases[i].utf8);
+    EXPECT_EQ(expected, converted);
+  }
+}
+#endif  // defined(WCHAR_T_IS_UTF32)
+
+TEST(UTFStringConversionsTest, ConvertMultiString) {
+  static wchar_t wmulti[] = {
+    L'f', L'o', L'o', L'\0',
+    L'b', L'a', L'r', L'\0',
+    L'b', L'a', L'z', L'\0',
+    L'\0'
+  };
+  static char multi[] = {
+    'f', 'o', 'o', '\0',
+    'b', 'a', 'r', '\0',
+    'b', 'a', 'z', '\0',
+    '\0'
+  };
+  std::wstring wmultistring;
+  memcpy(WriteInto(&wmultistring, arraysize(wmulti)), wmulti, sizeof(wmulti));
+  EXPECT_EQ(arraysize(wmulti) - 1, wmultistring.length());
+  std::string expected;
+  memcpy(WriteInto(&expected, arraysize(multi)), multi, sizeof(multi));
+  EXPECT_EQ(arraysize(multi) - 1, expected.length());
+  const std::string& converted = WideToUTF8(wmultistring);
+  EXPECT_EQ(arraysize(multi) - 1, converted.length());
+  EXPECT_EQ(expected, converted);
+}
+
+}  // base