[TestCase(FwNormalizationMode.knmNFKD, 2, 2, "A" + COMBINING_DIAERESIS + COMBINING_MACRON)] // NFKD: no change
public void get_NormalizedForm_NFSC_IsBlockedByRunBoundaries(FwNormalizationMode nm, int expectedRunCount, int expectedFirstRunLim, string expectedText)
{
// The macron will be in a different run; so while the first two compression
// schemes produce a single character, NFSC should produce 2.
string input = "A" + COMBINING_DIAERESIS + COMBINING_MACRON;
var temp = new TsString(input, EnglishWS);
var builder = temp.GetBldr();
// Make two runs, changing the last character
builder.SetIntPropValues(2, 3, (int)FwTextPropType.ktptBold, (int)FwTextPropVar.ktpvEnum, (int)FwTextToggleVal.kttvForceOn);
var tsInput = builder.GetString();
var actual = tsInput.get_NormalizedForm(nm);
Assert.That(actual.Text, Is.EqualTo(expectedText));
Assert.That(actual.RunCount, Is.EqualTo(expectedRunCount));
var normalProps = tsInput.get_Properties(0);
var boldProps = tsInput.get_Properties(1);
Assert.That(actual.get_Properties(0), Is.EqualTo(normalProps));
if (expectedRunCount > 1)
{
Assert.That(actual.get_LimOfRun(0), Is.EqualTo(expectedFirstRunLim));
Assert.That(actual.get_Properties(1), Is.EqualTo(boldProps));
}
}
[TestCase(FwNormalizationMode.knmNFKD, 6, 1, 2, expectedDecomposedTextForReorderingTests)] // NFKD: same as NFD
public void get_NormalizedForm_ReorderingDiacriticsAcrossRuns_ShouldKeepCharacterPropertiesIntact(
FwNormalizationMode nm, int expectedRunCount, int expectedSecondRunMin, int expectedSecondRunLim, string expectedText)
{
string input = "a" + COMBINING_DIAERESIS + COMBINING_DOT_BELOW + a_WITH_DIAERESIS + COMBINING_DOT_BELOW;
var temp = new TsString(input, EnglishWS);
var builder = temp.GetBldr();
// Make three runs, with the underdot having different props from the a and the diaeresis in each case.
builder.SetIntPropValues(2, 4, (int)FwTextPropType.ktptBold, (int)FwTextPropVar.ktpvEnum, (int)FwTextToggleVal.kttvForceOn);
var tsInput = builder.GetString();
Assert.That(tsInput.get_IsNormalizedForm(nm), Is.False);
var actual = tsInput.get_NormalizedForm(nm);
Assert.That(actual.Text, Is.EqualTo(expectedText));
Assert.That(actual.RunCount, Is.EqualTo(expectedRunCount));
// Second run is a good way to test that all the expected runs are in the right place
Assert.That(actual.get_MinOfRun(1), Is.EqualTo(expectedSecondRunMin));
Assert.That(actual.get_LimOfRun(1), Is.EqualTo(expectedSecondRunLim));
// For all cases, first run of output should be normal, just like first run of input was
// For all cases, second run of output should be bold, just like second run of input was
var normalProps = tsInput.get_Properties(0);
var boldProps = tsInput.get_Properties(1);
Assert.That(actual.get_Properties(0), Is.EqualTo(normalProps));
Assert.That(actual.get_Properties(1), Is.EqualTo(boldProps));
Assert.That(actual.get_PropertiesAt(0), Is.EqualTo(normalProps));
Assert.That(actual.get_PropertiesAt(expectedSecondRunMin), Is.EqualTo(boldProps));
Assert.That(actual.get_PropertiesAt(expectedSecondRunLim - 1), Is.EqualTo(boldProps));
if (expectedRunCount > 2)
{
Assert.That(actual.get_PropertiesAt(expectedSecondRunLim), Is.EqualTo(normalProps));
}
}
public void get_NormalizedForm_SimpleCases_WorkAsExpected(FwNormalizationMode nm, string expectedNormalizedText, int expectedStartIndexOfSecondRun)
{
TsString oneRunInput = new TsString(inputForBasicNormalizationTests, EnglishWS);
ITsString actual = oneRunInput.get_NormalizedForm(nm);
Assert.That(actual.Text, Is.EqualTo(expectedNormalizedText));
Assert.That(((TsString)actual).IsAlreadyNormalized(nm), Is.True);
Assert.That(actual.get_IsNormalizedForm(nm), Is.True);
// Make second run starting at the capital C
ITsStrBldr builder = oneRunInput.GetBldr();
builder.SetIntPropValues(7, oneRunInput.Length, (int)FwTextPropType.ktptBold, (int)FwTextPropVar.ktpvEnum, (int)FwTextToggleVal.kttvForceOn);
ITsString twoRunInput = builder.GetString();
actual = twoRunInput.get_NormalizedForm(nm);
Assert.That(actual.get_MinOfRun(1), Is.EqualTo(expectedStartIndexOfSecondRun));
// This run does not split any diacritic clusters, so it should not change any normalization results (even NFSC)
Assert.That(actual.Text, Is.EqualTo(expectedNormalizedText));
}
[TestCase(FwNormalizationMode.knmNFKD, 8, 1, 2, expectedDecomposedTextForUnderdotTests)] // NFKD: same as NFD
public void get_NormalizedForm_MultipleInterveningRuns_CombinesOrReordersCorrectly(
FwNormalizationMode nm, int expectedRunCount, int expectedSecondRunMin, int expectedSecondRunLim, string expectedText)
{
string input = "a" + COMBINING_DIAERESIS + COMBINING_DIAERESIS + COMBINING_DIAERESIS + COMBINING_DOT_BELOW + a_WITH_DIAERESIS + COMBINING_DOT_BELOW;
var temp = new TsString(input, EnglishWS);
var builder = temp.GetBldr();
// Make five total runs in the builder, so that the first underdot will have to reorder "past" multiple intervening runs
builder.SetIntPropValues(1, 2, (int)FwTextPropType.ktptForeColor, (int)FwTextPropVar.ktpvDefault, (int)FwTextColor.kclrRed);
builder.SetIntPropValues(2, 3, (int)FwTextPropType.ktptForeColor, (int)FwTextPropVar.ktpvDefault, (int)FwTextColor.kclrGreen);
builder.SetIntPropValues(3, 4, (int)FwTextPropType.ktptForeColor, (int)FwTextPropVar.ktpvDefault, (int)FwTextColor.kclrBlue);
builder.SetIntPropValues(4, 6, (int)FwTextPropType.ktptBold, (int)FwTextPropVar.ktpvEnum, (int)FwTextToggleVal.kttvForceOn);
var tsInput = builder.GetString();
Assert.That(tsInput.get_IsNormalizedForm(nm), Is.False);
var actual = tsInput.get_NormalizedForm(nm);
Assert.That(actual.Text, Is.EqualTo(expectedText));
Assert.That(actual.RunCount, Is.EqualTo(expectedRunCount));
// Second run is a good way to test that all the expected runs are in the right place
Assert.That(actual.get_MinOfRun(1), Is.EqualTo(expectedSecondRunMin));
Assert.That(actual.get_LimOfRun(1), Is.EqualTo(expectedSecondRunLim));
// For all cases, output should have runs in order normal, bold
// For all cases except NFC and NFKC, the red, green and blue runs should be preserved
var normalProps = tsInput.get_Properties(0);
var redProps = tsInput.get_Properties(1);
var greenProps = tsInput.get_Properties(2);
var blueProps = tsInput.get_Properties(3);
var boldProps = tsInput.get_Properties(4);
Assert.That(actual.get_PropertiesAt(0), Is.EqualTo(normalProps));
Assert.That(actual.get_Properties(0), Is.EqualTo(normalProps));
Assert.That(actual.get_Properties(1), Is.EqualTo(boldProps));
if (expectedRunCount > 2) // NFC and NFKC will remove the red/green/blue runs
{
Assert.That(actual.get_Properties(2), Is.EqualTo(redProps));
Assert.That(actual.get_Properties(3), Is.EqualTo(greenProps));
Assert.That(actual.get_Properties(4), Is.EqualTo(blueProps));
}
}
public void NfdAndFixOffsets_WithStackedDiacriticsThatDoNotOverlapRunBoundaries_FixesOffsetsCorrectly()
{
// This test is copied over verbatim from the old C++ unit tests
string input = "Stacked diacritics: We" + // 0..21
COMBINING_DOUBLE_ACUTE_ACCENT + COMBINING_RING_BELOW + COMBINING_GRAVE_ACCENT_BELOW + // 22..24
"lc" + COMBINING_LEFT_TACK_BELOW + COMBINING_MINUS_SIGN_BELOW + // 25..28
o_WITH_CIRCUMFLEX + "o" + // 29..30 (+1)
"m" + COMBINING_SEAGULL_BELOW + COMBINING_GRAVE_ACCENT + COMBINING_DIAERESIS + //31..34
COMBINING_MACRON + //35
e_WITH_GRAVE + //36 (+1)
" to" + COMBINING_DIAERESIS + COMBINING_CIRCUMFLEX_ACCENT + //37..41
" Wo" + COMBINING_DOT_ABOVE + COMBINING_INVERTED_BREVE + //42..46
"r" + COMBINING_SQUARE_BELOW + //47..48
"l" + COMBINING_TILDE + //49..50
"d" + COMBINING_DOWN_TACK_BELOW + COMBINING_TILDE_BELOW + //51..53
"a" + COMBINING_DIAERESIS + COMBINING_DOT_BELOW + //54..56
a_WITH_DIAERESIS + COMBINING_DOT_BELOW + //57..58 (+1)
"d" + COMBINING_ACUTE_ACCENT_BELOW + //59..60
"!"; //61
var temp = new TsString(input, EnglishWS);
var builder = temp.GetBldr();
// Make two runs, to make things interesting
builder.SetIntPropValues(0, 30, (int)FwTextPropType.ktptFontSize, (int)FwTextPropVar.ktpvMilliPoint, 20000);
builder.SetIntPropValues(30, input.Length, (int)FwTextPropType.ktptForeColor, (int)FwTextPropVar.ktpvDefault, (int)FwTextColor.kclrGreen);
var tsInput = (TsString)builder.GetString();
int[] origOffsets = { 30, 33, 49, 57, 0, 21, 29, input.Length - 1, 36, 37, 54, 55, 56, 58, 62 };
int[] expectedOffsets = { 31, 34, 51, 59, 0, 21, 29, input.Length - 1 + 3, 37, 39, 56, 58, 57, 60, 65 };
ITsString nfd;
int[] actualOffsets;
using (ArrayPtr nativeOffsets = ConvertToNativeArray(origOffsets))
{
tsInput.NfdAndFixOffsets(out nfd, nativeOffsets, origOffsets.Length);
actualOffsets = ConvertToManagedArray(nativeOffsets, origOffsets.Length);
}
CollectionAssert.AreEqual(expectedOffsets, actualOffsets);
// Verify that the characters at the new offsets are still the "same" characters that were at the old offsets
// "Same" because some offsets pointed to composed characters, and we expect to find the base character at the new offset
for (int i = 0; i < origOffsets.Length; i++)
{
if (origOffsets[i] == input.Length)
{
continue;
}
int oldOffset = origOffsets[i];
int newOffset = actualOffsets[i];
char oldCh = input[oldOffset];
char newCh = nfd.Text[newOffset];
char expectedCh = oldCh;
// Three offsets pointed to composed characters, which should be matched to the base character in the output
if (oldCh == '\u00e4') // 'ä'
{
expectedCh = 'a';
}
if (oldCh == '\u00f4') // 'ô'
{
expectedCh = 'o';
}
if (oldCh == '\u00e8') // 'è'
{
expectedCh = 'e';
}
Assert.That(newCh, Is.EqualTo(expectedCh),
String.Format("Old char '{0}' (U+{1:X4}) at {2} should match new char '{3}' (U+{4:X4}) at {5}, but didn't match.",
oldCh, (int)oldCh, oldOffset, newCh, newOffset, (int)newCh));
}
}
public void GetBldr_TwoRuns_ReturnsBldrWithCorrectData()
{
TsString tss = CreateTwoRunString();
Assert.That(tss.GetBldr().GetString(), Is.EqualTo(tss));
}
