const TextRun& getTextRun(const TextRun& originalRun)
{
// Convert the |originalRun| to NFC normalized form if combining diacritical marks
// (U+0300..) are used in the run. This conversion is necessary since most OpenType
// fonts (e.g., Arial) don't have substitution rules for the diacritical marks in
// their GSUB tables.
//
// Note that we don't use the icu::Normalizer::isNormalized(UNORM_NFC) API here since
// the API returns FALSE (= not normalized) for complex runs that don't require NFC
// normalization (e.g., Arabic text). Unless the run contains the diacritical marks,
// Harfbuzz will do the same thing for us using the GSUB table.
for (unsigned i = 0; i < originalRun.length(); ++i) {
UBlockCode block = ::ublock_getCode(originalRun[i]);
if (block == UBLOCK_COMBINING_DIACRITICAL_MARKS) {
return getNormalizedTextRun(originalRun);
}
}
return originalRun;
}
const TextRun& getTextRun(const TextRun& originalRun)
{
// Normalize the text run in two ways:
// 1) Convert the |originalRun| to NFC normalized form if combining diacritical marks
// (U+0300..) are used in the run. This conversion is necessary since most OpenType
// fonts (e.g., Arial) don't have substitution rules for the diacritical marks in
// their GSUB tables.
//
// Note that we don't use the icu::Normalizer::isNormalized(UNORM_NFC) API here since
// the API returns FALSE (= not normalized) for complex runs that don't require NFC
// normalization (e.g., Arabic text). Unless the run contains the diacritical marks,
// Harfbuzz will do the same thing for us using the GSUB table.
// 2) Convert spacing characters into plain spaces, as some fonts will provide glyphs
// for characters like '\n' otherwise.
for (unsigned i = 0; i < originalRun.length(); ++i) {
UChar ch = originalRun[i];
UBlockCode block = ::ublock_getCode(ch);
if (block == UBLOCK_COMBINING_DIACRITICAL_MARKS || (Font::treatAsSpace(ch) && ch != ' ')) {
return getNormalizedTextRun(originalRun);
}
}
return originalRun;
}
