private static void AssertIsInvalidFourByteSequence(byte[] invalidSequence)
{
Assert.Equal(4, invalidSequence.Length);
byte[] knownGoodBytes = Utf8Tests.DecodeHex(GRINNING_FACE);
byte[] toTest = invalidSequence.Concat(invalidSequence).Concat(knownGoodBytes).ToArray();
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 0, 0, 0);
toTest = knownGoodBytes.Concat(invalidSequence).Concat(knownGoodBytes).ToArray();
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 4, 1, 1);
}
private static void GetIndexOfFirstInvalidUtf8Sequence_Test_Core(string inputHex, int expectedRetVal, int expectedRuneCount, int expectedSurrogatePairCount)
{
byte[] inputBytes = Utf8Tests.DecodeHex(inputHex);
// Run the test normally
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(inputBytes, expectedRetVal, expectedRuneCount, expectedSurrogatePairCount);
// Then run the test with a bunch of ASCII data at the beginning (to exercise the vectorized code paths)
inputBytes = Enumerable.Repeat((byte)'x', 128).Concat(inputBytes).ToArray();
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(inputBytes, (expectedRetVal < 0) ? expectedRetVal : (expectedRetVal + 128), expectedRuneCount + 128, expectedSurrogatePairCount);
// Then put a few more ASCII bytes at the beginning (to test that offsets are properly handled)
inputBytes = Enumerable.Repeat((byte)'x', 7).Concat(inputBytes).ToArray();
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(inputBytes, (expectedRetVal < 0) ? expectedRetVal : (expectedRetVal + 135), expectedRuneCount + 135, expectedSurrogatePairCount);
}
private static void AssertIsInvalidThreeByteSequence(byte[] invalidSequence)
{
Assert.Equal(3, invalidSequence.Length);
byte[] knownGoodBytes = Utf8Tests.DecodeHex(EURO_SYMBOL);
byte[] toTest = invalidSequence.Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // at start of first DWORD
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 0, 0, 0);
// Run the same tests but with extra data at the beginning so that we're inside one of
// the 3-byte processing "hot loop" code paths.
toTest = knownGoodBytes.Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // straddling first and second DWORDs
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 3, 1, 0);
toTest = knownGoodBytes.Concat(knownGoodBytes).Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // straddling second and third DWORDs
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 6, 2, 0);
toTest = knownGoodBytes.Concat(knownGoodBytes).Concat(knownGoodBytes).Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // at end of third DWORD
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 9, 3, 0);
}
private static void AssertIsInvalidTwoByteSequence(byte[] invalidSequence)
{
Assert.Equal(2, invalidSequence.Length);
byte[] knownGoodBytes = Utf8Tests.DecodeHex(E_ACUTE);
byte[] toTest = invalidSequence.Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // at start of first DWORD
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 0, 0, 0);
toTest = knownGoodBytes.Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // at end of first DWORD
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 2, 1, 0);
// Run the same tests but with extra data at the beginning so that we're inside one of
// the 2-byte processing "hot loop" code paths.
toTest = knownGoodBytes.Concat(knownGoodBytes).Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // at start of next DWORD
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 4, 2, 0);
toTest = knownGoodBytes.Concat(knownGoodBytes).Concat(knownGoodBytes).Concat(invalidSequence).Concat(knownGoodBytes).ToArray(); // at end of next DWORD
GetIndexOfFirstInvalidUtf8Sequence_Test_Core(toTest, 6, 3, 0);
}