Update QString::isRightToLeft() to take directional hints into account

Make sure we properly handle both directional embeddings as well as
directional isolates to determine the direction of the string.

According to the latest version of the Unicode bidi algorithm,
parts of the string contained inside an directional isolate is
to be ignored when determining the paragraph direction. Embedding
markers themselves are to be ignored as well, but not the characters
inside an explicit directional embedding or override.

This is also some required pre-work to get our BiDi algorithm
updated to the latest version of the standard.

Move the implementation to QStringView and implement the methods
in QString and QStringRef through that implementation.

Task-number: QTBUG-57743
Change-Id: I7f24e09198e22d6359c6534c9ae40a904e94c46e
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>

1 files changed, 19 insertions, 0 deletions

diff --git a/tests/auto/corelib/tools/qstring/tst_qstring.cpp b/tests/auto/corelib/tools/qstring/tst_qstring.cpp
index f6840e60a8..98af7748c5 100644
--- a/tests/auto/corelib/tools/qstring/tst_qstring.cpp
+++ b/tests/auto/corelib/tools/qstring/tst_qstring.cpp
@@ -6901,6 +6901,25 @@ void tst_QString::isRightToLeft_data()
 
     static const ushort unicode3[] = { QChar::highSurrogate(0x10800u), QChar::lowSurrogate(0x10800u), QChar::highSurrogate(0x10805u), QChar::lowSurrogate(0x10805u) };
     QTest::newRow("surrogates-cypriot") << QString::fromUtf16(unicode3, 4) << true;
+
+    QTest::newRow("lre") << (QString("12345") + QChar(0x202a) + QString("9") + QChar(0x202c)) << false;
+    QTest::newRow("rle") << (QString("12345") + QChar(0x202b) + QString("9") + QChar(0x202c)) << false;
+    QTest::newRow("r in lre") << (QString("12345") + QChar(0x202a) + QString::fromUtf16(unicode1, 2) + QChar(0x202c) + QString("a")) << true;
+    QTest::newRow("l in lre") << (QString("12345") + QChar(0x202a) + QString("a") + QChar(0x202c) + QString::fromUtf16(unicode1, 2)) << false;
+    QTest::newRow("r in rle") << (QString("12345") + QChar(0x202b) + QString::fromUtf16(unicode1, 2) + QChar(0x202c) + QString("a")) << true;
+    QTest::newRow("l in rle") << (QString("12345") + QChar(0x202b) + QString("a") + QChar(0x202c) + QString::fromUtf16(unicode1, 2)) << false;
+
+    QTest::newRow("lro") << (QString("12345") + QChar(0x202d) + QString("9") + QChar(0x202c)) << false;
+    QTest::newRow("rlo") << (QString("12345") + QChar(0x202e) + QString("9") + QChar(0x202c)) << false;
+    QTest::newRow("r in lro") << (QString("12345") + QChar(0x202d) + QString::fromUtf16(unicode1, 2) + QChar(0x202c) + QString("a")) << true;
+    QTest::newRow("l in lro") << (QString("12345") + QChar(0x202d) + QString("a") + QChar(0x202c) + QString::fromUtf16(unicode1, 2)) << false;
+    QTest::newRow("r in rlo") << (QString("12345") + QChar(0x202e) + QString::fromUtf16(unicode1, 2) + QChar(0x202c) + QString("a")) << true;
+    QTest::newRow("l in rlo") << (QString("12345") + QChar(0x202e) + QString("a") + QChar(0x202c) + QString::fromUtf16(unicode1, 2)) << false;
+
+    QTest::newRow("lri") << (QString("12345") + QChar(0x2066) + QString("a") + QChar(0x2069) + QString::fromUtf16(unicode1, 2)) << true;
+    QTest::newRow("rli") << (QString("12345") + QChar(0x2067) + QString::fromUtf16(unicode1, 2) + QChar(0x2069) + QString("a")) << false;
+    QTest::newRow("fsi1") << (QString("12345") + QChar(0x2068) + QString("a") + QChar(0x2069) + QString::fromUtf16(unicode1, 2)) << true;
+    QTest::newRow("fsi2") << (QString("12345") + QChar(0x2068) + QString::fromUtf16(unicode1, 2) + QChar(0x2069) + QString("a")) << false;
 }
 
 void tst_QString::isRightToLeft()