Line Breaking Algorithm: handle the Object Replacement Character

See http://www.unicode.org/reports/tr14/#CB
and http://www.unicode.org/reports/tr14/#LB20 for details

Change-Id: Ice0aa2b2ce81f6e39839a353240420436eddd754
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>

diff --git a/src/corelib/tools/qunicodetools.cpp b/src/corelib/tools/qunicodetools.cpp
index c038690..76d66bf 100644 (file)
--- a/src/corelib/tools/qunicodetools.cpp
+++ b/src/corelib/tools/qunicodetools.cpp
@@ -369,36 +369,37 @@ enum Action {
     CombiningProhibitedBreak, CP = CombiningProhibitedBreak
 };
 
-static const uchar breakTable[QUnicodeTables::LineBreak_JT + 1][QUnicodeTables::LineBreak_JT + 1] = {
-/*         OP  CL  CP  QU  GL  NS  EX  SY  IS  PR  PO  NU  AL  HL  ID  IN  HY  BA  BB  B2  ZW  CM  WJ  H2  H3  JL  JV  JT */
-/* OP */ { PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, CP, PB, PB, PB, PB, PB, PB },
-/* CL */ { DB, PB, PB, IB, IB, PB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* CP */ { DB, PB, PB, IB, IB, PB, PB, PB, PB, DB, DB, IB, IB, IB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* QU */ { PB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB },
-/* GL */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB },
-/* NS */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* EX */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* SY */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* IS */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, IB, IB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* PR */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, IB, DB, IB, IB, DB, DB, PB, CI, PB, IB, IB, IB, IB, IB },
-/* PO */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* NU */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* AL */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* HL */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* ID */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* IN */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* HY */ { DB, PB, PB, IB, DB, IB, PB, PB, PB, DB, DB, IB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* BA */ { DB, PB, PB, IB, DB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* BB */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB },
-/* B2 */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, PB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* ZW */ { DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, PB, DB, DB, DB, DB, DB, DB, DB },
-/* CM */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB },
-/* WJ */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB },
-/* H2 */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, IB, IB },
-/* H3 */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, IB },
-/* JL */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, IB, IB, IB, IB, DB },
-/* JV */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, IB, IB },
-/* JT */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, IB }
+static const uchar breakTable[QUnicodeTables::LineBreak_CB + 1][QUnicodeTables::LineBreak_CB + 1] = {
+/*         OP  CL  CP  QU  GL  NS  EX  SY  IS  PR  PO  NU  AL  HL  ID  IN  HY  BA  BB  B2  ZW  CM  WJ  H2  H3  JL  JV  JT  CB */
+/* OP */ { PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, PB, CP, PB, PB, PB, PB, PB, PB, PB },
+/* CL */ { DB, PB, PB, IB, IB, PB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* CP */ { DB, PB, PB, IB, IB, PB, PB, PB, PB, DB, DB, IB, IB, IB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* QU */ { PB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB, IB },
+/* GL */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB, IB },
+/* NS */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* EX */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* SY */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* IS */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, IB, IB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* PR */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, IB, DB, IB, IB, DB, DB, PB, CI, PB, IB, IB, IB, IB, IB, DB },
+/* PO */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* NU */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* AL */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* HL */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* ID */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* IN */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* HY */ { DB, PB, PB, IB, DB, IB, PB, PB, PB, DB, DB, IB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* BA */ { DB, PB, PB, IB, DB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* BB */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB, DB },
+/* B2 */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, IB, IB, DB, PB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* ZW */ { DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, PB, DB, DB, DB, DB, DB, DB, DB, DB },
+/* CM */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, DB, DB, IB, IB, IB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB },
+/* WJ */ { IB, PB, PB, IB, IB, IB, PB, PB, PB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, IB, PB, CI, PB, IB, IB, IB, IB, IB, IB },
+/* H2 */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, IB, IB, DB },
+/* H3 */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, IB, DB },
+/* JL */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, IB, IB, IB, IB, DB, DB },
+/* JV */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, IB, IB, DB },
+/* JT */ { DB, PB, PB, IB, IB, IB, PB, PB, PB, DB, IB, DB, DB, DB, DB, IB, IB, IB, DB, DB, PB, CI, PB, DB, DB, DB, DB, IB, DB },
+/* CB */ { DB, PB, PB, IB, IB, DB, PB, PB, PB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, DB, PB, CI, PB, DB, DB, DB, DB, DB, DB }
 };
 
 } // namespace LB

diff --git a/tests/auto/corelib/tools/qchar/tst_qchar.cpp b/tests/auto/corelib/tools/qchar/tst_qchar.cpp
index 1a9d10a..bf7fc98 100644 (file)
--- a/tests/auto/corelib/tools/qchar/tst_qchar.cpp
+++ b/tests/auto/corelib/tools/qchar/tst_qchar.cpp
@@ -698,6 +698,7 @@ void tst_QChar::lineBreakClass()
     QVERIFY(QUnicodeTables::lineBreakClass(0x0033u) == QUnicodeTables::LineBreak_NU);
     QVERIFY(QUnicodeTables::lineBreakClass(0x00adu) == QUnicodeTables::LineBreak_BA);
     QVERIFY(QUnicodeTables::lineBreakClass(0x05d0u) == QUnicodeTables::LineBreak_HL);
+    QVERIFY(QUnicodeTables::lineBreakClass(0xfffcu) == QUnicodeTables::LineBreak_CB);
     QVERIFY(QUnicodeTables::lineBreakClass(0xe0164u) == QUnicodeTables::LineBreak_CM);
     QVERIFY(QUnicodeTables::lineBreakClass(0x2f9a4u) == QUnicodeTables::LineBreak_ID);
     QVERIFY(QUnicodeTables::lineBreakClass(0x10000u) == QUnicodeTables::LineBreak_AL);

diff --git a/tests/auto/corelib/tools/qtextboundaryfinder/tst_qtextboundaryfinder.cpp b/tests/auto/corelib/tools/qtextboundaryfinder/tst_qtextboundaryfinder.cpp
index d445f47..eef2931 100644 (file)
--- a/tests/auto/corelib/tools/qtextboundaryfinder/tst_qtextboundaryfinder.cpp
+++ b/tests/auto/corelib/tools/qtextboundaryfinder/tst_qtextboundaryfinder.cpp
@@ -285,9 +285,6 @@ void tst_QTextBoundaryFinder::lineBoundariesDefault()
     QFETCH(QString, testString);
     QFETCH(QList<int>, expectedBreakPositions);
 
-    if (testString.contains(QChar::ObjectReplacementCharacter))
-        QSKIP("QTextBoundaryFinder doesn't handle Object Replacement character");
-
     expectedBreakPositions.prepend(0); // ### QTBF generates a boundary at start of text
     doTestData(testString, expectedBreakPositions, QTextBoundaryFinder::Line, true);
 }

diff --git a/util/unicode/main.cpp b/util/unicode/main.cpp
index ec4f4e0..aefe690 100644 (file)
--- a/util/unicode/main.cpp
+++ b/util/unicode/main.cpp
@@ -320,15 +320,15 @@ static void initSentenceBreak()
 
 static const char *line_break_class_string =
     "    // see http://www.unicode.org/reports/tr14/tr14-28.html\n"
-    "    // we don't use the XX, AI, and CB classes and map them to AL instead.\n"
+    "    // we don't use the XX and AI classes and map them to AL instead.\n"
     "    enum LineBreakClass {\n"
     "        LineBreak_OP, LineBreak_CL, LineBreak_CP, LineBreak_QU, LineBreak_GL,\n"
     "        LineBreak_NS, LineBreak_EX, LineBreak_SY, LineBreak_IS, LineBreak_PR,\n"
     "        LineBreak_PO, LineBreak_NU, LineBreak_AL, LineBreak_HL, LineBreak_ID,\n"
     "        LineBreak_IN, LineBreak_HY, LineBreak_BA, LineBreak_BB, LineBreak_B2,\n"
     "        LineBreak_ZW, LineBreak_CM, LineBreak_WJ, LineBreak_H2, LineBreak_H3,\n"
-    "        LineBreak_JL, LineBreak_JV, LineBreak_JT, LineBreak_SA, LineBreak_SG,\n"
-    "        LineBreak_SP, LineBreak_CR, LineBreak_LF, LineBreak_BK\n"
+    "        LineBreak_JL, LineBreak_JV, LineBreak_JT, LineBreak_CB, LineBreak_SA,\n"
+    "        LineBreak_SG, LineBreak_SP, LineBreak_CR, LineBreak_LF, LineBreak_BK\n"
     "    };\n\n";
 
 enum LineBreakClass {
@@ -337,8 +337,8 @@ enum LineBreakClass {
     LineBreak_PO, LineBreak_NU, LineBreak_AL, LineBreak_HL, LineBreak_ID,
     LineBreak_IN, LineBreak_HY, LineBreak_BA, LineBreak_BB, LineBreak_B2,
     LineBreak_ZW, LineBreak_CM, LineBreak_WJ, LineBreak_H2, LineBreak_H3,
-    LineBreak_JL, LineBreak_JV, LineBreak_JT, LineBreak_SA, LineBreak_SG,
-    LineBreak_SP, LineBreak_CR, LineBreak_LF, LineBreak_BK
+    LineBreak_JL, LineBreak_JV, LineBreak_JT, LineBreak_CB, LineBreak_SA,
+    LineBreak_SG, LineBreak_SP, LineBreak_CR, LineBreak_LF, LineBreak_BK
 
     , LineBreak_Unassigned
 };
@@ -348,7 +348,6 @@ static QHash<QByteArray, LineBreakClass> line_break_map;
 static void initLineBreak()
 {
     // ### Classes XX and AI are left out and mapped to AL for now.
-    // ### Class CB is unsupported for now and mapped to AL as well.
     // ### Class NL is mapped to BK.
     // ### Treating characters of class CJ as class NS will give CSS strict line breaking;
     //     treating them as class ID will give CSS normal breaking.
@@ -370,7 +369,7 @@ static void initLineBreak()
         { LineBreak_BA, "BA" },
         { LineBreak_BB, "BB" },
         { LineBreak_HY, "HY" },
-        { LineBreak_AL, "CB" }, // ###
+        { LineBreak_CB, "CB" },
         { LineBreak_NS, "CJ" },
         { LineBreak_CL, "CL" },
         { LineBreak_CP, "CP" },