Implement fast path in UTF8ToUTF16 for pure ASCII strings

base/strings/string_util_unittest.cc

 // Copyright 2013 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/string_util.h"
 
 #include <math.h>
 #include <stdarg.h>
 
 #include <algorithm>
@@ skipping 368 matching lines @@
   // Check that we support Embedded Nulls. The first uses the canonical UTF-8
   // representation, and the second uses a 2-byte sequence. The second version
   // is invalid UTF-8 since UTF-8 states that the shortest encoding for a
   // given codepoint must be used.
   static const char kEmbeddedNull[] = "embedded\0null";
   EXPECT_TRUE(IsStringUTF8(
       std::string(kEmbeddedNull, sizeof(kEmbeddedNull))));
   EXPECT_FALSE(IsStringUTF8("embedded\xc0\x80U+0000"));
 }
 
+TEST(StringUtilTest, IsStringASCII) {
+  static char char_ascii[] =
+      "0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF";
+  static char16 char16_ascii[] = {
+      '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', '0', 'A',
+      'B', 'C', 'D', 'E', 'F', '0', '1', '2', '3', '4', '5', '6',
+      '7', '8', '9', '0', 'A', 'B', 'C', 'D', 'E', 'F', 0 };
+
+  // Test a variety of the fragment start positions and lengths in order to make
+  // sure that bit masking in IsStringASCII works correctly.
+  // Also, test that a non-ASCII character will be detected regardless of its
+  // position inside the string.
+  {
+    const size_t string_length = arraysize(char_ascii) - 1;
+    for (size_t offset = 0; offset < 8; ++offset) {
+      for (size_t len = 0, max_len = string_length - offset; len < max_len;
+           ++len) {
+        EXPECT_TRUE(IsStringASCII(StringPiece(char_ascii + offset, len)));
+        for (size_t char_pos = offset; char_pos < len; ++char_pos) {
+          char_ascii[char_pos] |= '\x80';
+          EXPECT_FALSE(IsStringASCII(StringPiece(char_ascii + offset, len)));
+          char_ascii[char_pos] &= ~'\x80';
+        }
+      }
+    }
+  }
+
+  {
+    const size_t string_length = arraysize(char16_ascii) - 1;
+    for (size_t offset = 0; offset < 4; ++offset) {
+      for (size_t len = 0, max_len = string_length - offset; len < max_len;
+           ++len) {
+        EXPECT_TRUE(IsStringASCII(StringPiece16(char16_ascii + offset, len)));
+        for (size_t char_pos = offset; char_pos < len; ++char_pos) {
+          char16_ascii[char_pos] |= 0x80;
+          EXPECT_FALSE(
+              IsStringASCII(StringPiece16(char16_ascii + offset, len)));
+          char16_ascii[char_pos] &= ~0x80;
+          // Also test when the upper half is non-zero.
+          char16_ascii[char_pos] |= 0x100;
+          EXPECT_FALSE(
+              IsStringASCII(StringPiece16(char16_ascii + offset, len)));
+          char16_ascii[char_pos] &= ~0x100;
+        }
+      }
+    }
+  }
+}
+
 TEST(StringUtilTest, ConvertASCII) {
   static const char* char_cases[] = {
     "Google Video",
     "Hello, world\n",
     "0123ABCDwxyz \a\b\t\r\n!+,.~"
   };
 
   static const wchar_t* const wchar_cases[] = {
     L"Google Video",
     L"Hello, world\n",
@@ skipping 786 matching lines @@
   const std::string live = kLive;
   std::string dead = live;
   strncpy(WriteInto(&dead, 5), kDead, 4);
   EXPECT_EQ(kDead, dead);
   EXPECT_EQ(4u, dead.size());
   EXPECT_EQ(kLive, live);
   EXPECT_EQ(4u, live.size());
 }
 
 }  // namespace base