QTextEngine - treat a fullstop (0x2E) as the same script as the preceeding text when dividing up strings

Many languages use a fullstop to indicate an abbreviation, making the
fullstop part of the word.  For languages like thai, it is required to
pass the fullstop along for correct word breaking.

Change-Id: I5ad0ddbc66ea96e08913446dad8fd3c5d5dd0905
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
This commit is contained in:
John Tapsell 2012-02-08 10:12:16 +00:00 committed by Qt by Nokia
parent 785e95ef0a
commit 2b23d7214f
2 changed files with 20 additions and 7 deletions

View File

@ -115,7 +115,20 @@ private:
return;
const int end = start + length;
for (int i = start + 1; i < end; ++i) {
if ((m_analysis[i] == m_analysis[start])
// According to the unicode spec we should be treating characters in the Common script
// (punctuation, spaces, etc) as being the same script as the surrounding text for the
// purpose of splitting up text. This is important because, for example, a fullstop
// (0x2E) can be used to indicate an abbreviation and so must be treated as part of a
// word. Thus it must be passed along with the word in languages that have to calculate
// word breaks. For example the thai word "ครม." has no word breaks but the word "ครม"
// does.
// Unfortuntely because we split up the strings for both wordwrapping and for setting
// the font and because Japanese and Chinese are also aliases of the script "Common",
// doing this would break too many things. So instead we only pass the full stop
// along, and nothing else.
if (m_analysis[i].bidiLevel == m_analysis[start].bidiLevel
&& m_analysis[i].flags == m_analysis[start].flags
&& (m_analysis[i].script == m_analysis[start].script || m_string[i] == QLatin1Char('.'))
&& m_analysis[i].flags < QScriptAnalysis::SpaceTabOrObject
&& i - start < MaxItemLength)
continue;

View File

@ -1283,7 +1283,7 @@ void tst_QTextScriptEngine::thaiIsolatedSaraAm()
void tst_QTextScriptEngine::thaiWithZWJ()
{
QString s(QString::fromUtf8("ร‍ร‌ร“ร…ร”ร\xA0ร本ร") + QChar(0x0363)/*superscript 'a', for testing Inherited class*/);
QString s(QString::fromUtf8("ร‍ร‌..“ร…ร”ร\xA0ร本ร") + QChar(0x0363)/*superscript 'a', for testing Inherited class*/);
QTextLayout layout(s);
layout.beginLayout();
layout.createLine();
@ -1296,7 +1296,7 @@ void tst_QTextScriptEngine::thaiWithZWJ()
// The current implementation hides them, so we test for that.
// But make sure that we don't hide anything else
QCOMPARE(e->layoutData->items.size(), 11);
QCOMPARE(e->layoutData->items[0].num_glyphs, ushort(5)); // Thai: The ZWJ and ZWNJ characters are inherited, so should be part of the thai script
QCOMPARE(e->layoutData->items[0].num_glyphs, ushort(7)); // Thai: The ZWJ and ZWNJ characters are inherited, so should be part of the thai script
QCOMPARE(e->layoutData->items[1].num_glyphs, ushort(1)); // Common: The smart quotes cannot be handled by thai, so should be a seperate item
QCOMPARE(e->layoutData->items[2].num_glyphs, ushort(1)); // Thai: Thai character
QCOMPARE(e->layoutData->items[3].num_glyphs, ushort(1)); // Common: Ellipsis
@ -1310,15 +1310,15 @@ void tst_QTextScriptEngine::thaiWithZWJ()
//A quick sanity check - check all the characters are individual clusters
unsigned short *logClusters = e->layoutData->logClustersPtr;
for (int i = 0; i < 5; i++)
for (int i = 0; i < 7; i++)
QCOMPARE(logClusters[i], ushort(i));
for (int i = 0; i < 10; i++)
QCOMPARE(logClusters[i+5], ushort(0));
QCOMPARE(logClusters[15], ushort(1));
QCOMPARE(logClusters[i+7], ushort(0));
QCOMPARE(logClusters[17], ushort(1));
// The only characters that we should be hiding are the ZWJ and ZWNJ characters in position 1
// and 3.
for (int i = 0; i < 16; i++)
for (int i = 0; i < 18; i++)
QCOMPARE((bool)e->layoutData->glyphLayout.attributes[i].dontPrint, (i == 1 || i == 3));
}