Fix various problems with inline autocomplete and URLs that change length dur...

base/i18n/icu_string_conversions_unittest.cc

-// Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
+// Copyright (c) 2009 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 <math.h>
 #include <stdarg.h>
 
 #include <limits>
 #include <sstream>
 
 #include "base/basictypes.h"
+#include "base/i18n/icu_string_conversions.h"
 #include "base/logging.h"
 #include "base/utf_string_conversions.h"
-#include "base/i18n/icu_string_conversions.h"
 #include "testing/gtest/include/gtest/gtest.h"
 
 namespace base {
 
 namespace {
 
 // Given a null-terminated string of wchar_t with each wchar_t representing
 // a UTF-16 code unit, returns a string16 made up of wchar_t's in the input.
 // Each wchar_t should be <= 0xFFFF and a non-BMP character (> U+FFFF)
 // should be represented as a surrogate pair (two UTF-16 units)
 // *even* where wchar_t is 32-bit (Linux and Mac).
 //
 // This is to help write tests for functions with string16 params until
 // the C++ 0x UTF-16 literal is well-supported by compilers.
 string16 BuildString16(const wchar_t* s) {
 #if defined(WCHAR_T_IS_UTF16)
   return string16(s);
 #elif defined(WCHAR_T_IS_UTF32)
   string16 u16;
   while (*s != 0) {
     DCHECK(static_cast<unsigned int>(*s) <= 0xFFFFu);
     u16.push_back(*s++);
   }
   return u16;
 #endif
 }
 
-static const wchar_t* const kConvertRoundtripCases[] = {
+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(StringUtilTest, ConvertCodepageUTF8) {
+TEST(ICUStringConversionsTest, ConvertCodepageUTF8) {
   // Make sure WideToCodepage works like WideToUTF8.
   for (size_t i = 0; i < arraysize(kConvertRoundtripCases); ++i) {
     std::string expected(WideToUTF8(kConvertRoundtripCases[i]));
     std::string utf8;
     EXPECT_TRUE(WideToCodepage(kConvertRoundtripCases[i], kCodepageUTF8,
                                OnStringConversionError::SKIP, &utf8));
     EXPECT_EQ(expected, utf8);
   }
 }
 
@@ skipping 67 matching lines @@
    true,
    L"\x00A5\x00A8",
    NULL},
   // Chinese (GB18030) : A 4 byte sequence mapped to plane 2 (U+20000)
   {"gb18030",
    "\x95\x32\x82\x36\xD2\xBB",
    OnStringConversionError::FAIL,
    true,
 #if defined(WCHAR_T_IS_UTF16)
    L"\xD840\xDC00\x4E00",
-#else
+#elif defined(WCHAR_T_IS_UTF32)
    L"\x20000\x4E00",
 #endif
    L"\xD840\xDC00\x4E00"},
   {"big5",
    "\xA7\x41\xA6\x6E",
    OnStringConversionError::FAIL,
    true,
    L"\x4F60\x597D",
    NULL},
   // Greek (ISO-8859)
@@ skipping 57 matching lines @@
   // Thai (windows-874)
   {"windows-874",
    "\xCA\xC7\xD1\xCA\xB4\xD5" "\xA4\xC3\xD1\xBA",
    OnStringConversionError::FAIL,
    true,
    L"\x0E2A\x0E27\x0E31\x0E2A\x0E14\x0E35"
    L"\x0E04\x0E23\x0e31\x0E1A",
    NULL},
 };
 
-TEST(StringUtilTest, ConvertBetweenCodepageAndWide) {
+TEST(ICUStringConversionsTest, ConvertBetweenCodepageAndWide) {
   for (size_t i = 0; i < ARRAYSIZE_UNSAFE(kConvertCodepageCases); ++i) {
     std::wstring wide;
     bool success = CodepageToWide(kConvertCodepageCases[i].encoded,
                                   kConvertCodepageCases[i].codepage_name,
                                   kConvertCodepageCases[i].on_error,
                                   &wide);
     EXPECT_EQ(kConvertCodepageCases[i].success, success);
     EXPECT_EQ(kConvertCodepageCases[i].wide, wide);
 
     // When decoding was successful and nothing was skipped, we also check the
@@ skipping 41 matching lines @@
   // Invalid characters should fail.
   EXPECT_TRUE(WideToCodepage(L"a\xffffz", "iso-8859-1",
                              OnStringConversionError::SKIP, &encoded));
   EXPECT_STREQ("az", encoded.c_str());
 
   // Invalid codepages should fail.
   EXPECT_FALSE(WideToCodepage(L"Hello, world", "awesome-8571-2",
                               OnStringConversionError::SKIP, &encoded));
 }
 
-TEST(StringUtilTest, ConvertBetweenCodepageAndUTF16) {
+TEST(ICUStringConversionsTest, ConvertBetweenCodepageAndUTF16) {
   for (size_t i = 0; i < ARRAYSIZE_UNSAFE(kConvertCodepageCases); ++i) {
     string16 utf16;
     bool success = CodepageToUTF16(kConvertCodepageCases[i].encoded,
                                    kConvertCodepageCases[i].codepage_name,
                                    kConvertCodepageCases[i].on_error,
                                    &utf16);
     string16 utf16_expected;
     if (kConvertCodepageCases[i].u16_wide == NULL)
       utf16_expected = BuildString16(kConvertCodepageCases[i].wide);
     else
       utf16_expected = BuildString16(kConvertCodepageCases[i].u16_wide);
     EXPECT_EQ(kConvertCodepageCases[i].success, success);
     EXPECT_EQ(utf16_expected, utf16);
 
     // When decoding was successful and nothing was skipped, we also check the
     // reverse conversion. See also the corresponding comment in
     // ConvertBetweenCodepageAndWide.
     if (success &&
         kConvertCodepageCases[i].on_error == OnStringConversionError::FAIL) {
       std::string encoded;
       success = UTF16ToCodepage(utf16, kConvertCodepageCases[i].codepage_name,
                                 kConvertCodepageCases[i].on_error, &encoded);
       EXPECT_EQ(kConvertCodepageCases[i].success, success);
       EXPECT_EQ(kConvertCodepageCases[i].encoded, encoded);
     }
   }
 }
 
+static const struct {
+  const char* codepage_name;
+  const char* encoded;
+  size_t input_offset;
+  size_t u16_output_offset;
+  size_t wide_output_offset;
+} kAdjustOffsetCases[] = {
+  {"gb2312", "", 0, string16::npos, std::wstring::npos},
+  {"gb2312", "\xC4\xE3\xBA\xC3", 0, 0, 0},
+  {"gb2312", "\xC4\xE3\xBA\xC3", 2, 1, 1},
+  {"gb2312", "\xC4\xE3\xBA\xC3", 4, string16::npos, std::wstring::npos},
+  {"gb2312", "\xC4\xE3\xBA\xC3", 1, string16::npos, std::wstring::npos},
+  {"gb2312", "\xC4\xE3\xBA\xC3", std::string::npos, string16::npos,
+   std::wstring::npos},
+  {"gb18030", "\x95\x32\x82\x36\xD2\xBB", 2, string16::npos,
+   std::wstring::npos},
+  {"gb18030", "\x95\x32\x82\x36\xD2\xBB", 4, 2, 1},
+};
+
+TEST(ICUStringConversionsTest, AdjustOffset) {
+  for (size_t i = 0; i < ARRAYSIZE_UNSAFE(kAdjustOffsetCases); ++i) {
+    string16 utf16;
+    size_t offset = kAdjustOffsetCases[i].input_offset;
+    EXPECT_TRUE(CodepageToUTF16AndAdjustOffset(kAdjustOffsetCases[i].encoded,
+        kAdjustOffsetCases[i].codepage_name,
+        OnStringConversionError::FAIL, &utf16, &offset));
+    EXPECT_EQ(kAdjustOffsetCases[i].u16_output_offset, offset);
+
+    std::wstring wide;
+    offset = kAdjustOffsetCases[i].input_offset;
+    CodepageToWideAndAdjustOffset(kAdjustOffsetCases[i].encoded,
+        kAdjustOffsetCases[i].codepage_name,
+        OnStringConversionError::FAIL, &wide, &offset);
+#if defined(WCHAR_T_IS_UTF16)
+    EXPECT_EQ(kAdjustOffsetCases[i].u16_output_offset, offset);
+#elif defined(WCHAR_T_IS_UTF32)
+    EXPECT_EQ(kAdjustOffsetCases[i].wide_output_offset, offset);
+#endif
+  }
+}
+
 }  // namespace base