Shift positions for lineBreakType

to keep them consistent with positions for all other flags.
This changes the internal behavior so that attributes[0].lineBreakType now means
"break opportunity at start of the text (before the first character in the string)"
and is always assigned with HB_NoBreak to conform rule LB2
(see http://www.unicode.org/reports/tr14/#LB2).

The current implementation is based on the sample implementation from tr14
that aimed to be as simple as possible rather than to be optimal.
From now, we can use pieces of the attributes array "as is"
without having to adjust some positions. Or we can analize some long text
by chunks (e.g. paragraph by paragraph) and consume less memory.

This introduces a minor overhead that will be eliminated shortly.

Change-Id: Ic873a05a9d5203b1c3d5aff2e4445a3f034c4bd2
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>

6 files changed, 17 insertions, 16 deletions

diff --git a/src/3rdparty/harfbuzz/src/harfbuzz-myanmar.c b/src/3rdparty/harfbuzz/src/harfbuzz-myanmar.c
index 4b68e64d85..1327b185b0 100644
--- a/src/3rdparty/harfbuzz/src/harfbuzz-myanmar.c
+++ b/src/3rdparty/harfbuzz/src/harfbuzz-myanmar.c
@@ -521,8 +521,7 @@ void HB_MyanmarAttributes(HB_Script script, const HB_UChar16 *text, hb_uint32 fr
 	hb_uint32 boundary = myanmar_nextSyllableBoundary(text, from+i, end, &invalid) - from;
 
 	attributes[i].charStop = TRUE;
-        if (i)
-            attributes[i-1].lineBreakType = HB_Break;
+    attributes[i].lineBreakType = HB_Break;
 
 	if (boundary > len-1)
             boundary = len;
diff --git a/src/3rdparty/harfbuzz/src/harfbuzz-thai.c b/src/3rdparty/harfbuzz/src/harfbuzz-thai.c
index 039394a5a2..54fb9537a4 100644
--- a/src/3rdparty/harfbuzz/src/harfbuzz-thai.c
+++ b/src/3rdparty/harfbuzz/src/harfbuzz-thai.c
@@ -414,8 +414,7 @@ static void HB_ThaiAssignAttributes(const HB_UChar16 *string, hb_uint32 len, HB_
         numbreaks = th_brk((const unsigned char *)cstr, break_positions, brp_size);
         for (i = 0; i < numbreaks; ++i) {
             attributes[break_positions[i]].wordBoundary = TRUE;
-            if (break_positions[i] > 0)
-                attributes[break_positions[i]-1].lineBreakType = HB_Break;
+            attributes[break_positions[i]].lineBreakType = HB_Break;
         }
 
         if (break_positions != brp)
diff --git a/src/corelib/tools/qtextboundaryfinder.cpp b/src/corelib/tools/qtextboundaryfinder.cpp
index 1e12d6e4a3..042f92f1e7 100644
--- a/src/corelib/tools/qtextboundaryfinder.cpp
+++ b/src/corelib/tools/qtextboundaryfinder.cpp
@@ -374,8 +374,7 @@ int QTextBoundaryFinder::toNextBoundary()
             ++pos;
         break;
     case Line:
-        Q_ASSERT(pos);
-        while (pos < length && d->attributes[pos-1].lineBreakType == HB_NoBreak)
+        while (pos < length && d->attributes[pos].lineBreakType == HB_NoBreak)
             ++pos;
         break;
     }
@@ -417,7 +416,7 @@ int QTextBoundaryFinder::toPreviousBoundary()
             --pos;
         break;
     case Line:
-        while (pos > 0 && d->attributes[pos-1].lineBreakType == HB_NoBreak)
+        while (pos > 0 && d->attributes[pos].lineBreakType == HB_NoBreak)
             --pos;
         break;
     }
@@ -442,7 +441,7 @@ bool QTextBoundaryFinder::isAtBoundary() const
     case Word:
         return d->attributes[pos].wordBoundary;
     case Line:
-        return (pos > 0) ? d->attributes[pos-1].lineBreakType != HB_NoBreak : true;
+        return pos == 0 || d->attributes[pos].lineBreakType != HB_NoBreak;
     case Sentence:
         return d->attributes[pos].sentenceBoundary;
     }
@@ -458,7 +457,7 @@ QTextBoundaryFinder::BoundaryReasons QTextBoundaryFinder::boundaryReasons() cons
         return NotAtBoundary;
     if (! isAtBoundary())
         return NotAtBoundary;
-    if (t == Line && pos < length && d->attributes[pos-1].lineBreakType == HB_SoftHyphen)
+    if (t == Line && pos < length && d->attributes[pos].lineBreakType == HB_SoftHyphen)
         return SoftHyphen;
     if (pos == 0) {
         if (d->attributes[pos].whiteSpace)
diff --git a/src/corelib/tools/qunicodetools.cpp b/src/corelib/tools/qunicodetools.cpp
index a311213ede..5845765e46 100644
--- a/src/corelib/tools/qunicodetools.cpp
+++ b/src/corelib/tools/qunicodetools.cpp
@@ -239,7 +239,10 @@ static void calcGraphemeAndLineBreaks(const ushort *string, quint32 len, HB_Char
         grapheme = ngrapheme;
         attributes[i-1].lineBreakType = lineBreakType;
     }
-    attributes[len-1].lineBreakType = HB_ForcedBreak;
+
+    for (quint32 i = len - 1; i > 0; --i)
+        attributes[i].lineBreakType = attributes[i - 1].lineBreakType;
+    attributes[0].lineBreakType = HB_NoBreak; // LB2
 }
 
 
diff --git a/src/gui/text/qtextlayout.cpp b/src/gui/text/qtextlayout.cpp
index 3cdd54627e..3250e2831c 100644
--- a/src/gui/text/qtextlayout.cpp
+++ b/src/gui/text/qtextlayout.cpp
@@ -1867,7 +1867,7 @@ void QTextLine::layout_helper(int maxGlyphs)
 
                 if (lbh.currentPosition >= eng->layoutData->string.length()
                     || attributes[lbh.currentPosition].whiteSpace
-                    || attributes[lbh.currentPosition-1].lineBreakType != HB_NoBreak) {
+                    || attributes[lbh.currentPosition].lineBreakType != HB_NoBreak) {
                     sb_or_ws = true;
                     break;
                 } else if (breakany && attributes[lbh.currentPosition].charStop) {
@@ -1876,7 +1876,8 @@ void QTextLine::layout_helper(int maxGlyphs)
             } while (lbh.currentPosition < end);
             lbh.minw = qMax(lbh.tmpData.textWidth, lbh.minw);
 
-            if (lbh.currentPosition && attributes[lbh.currentPosition - 1].lineBreakType == HB_SoftHyphen) {
+            if (lbh.currentPosition > 0 && lbh.currentPosition < end
+                && attributes[lbh.currentPosition].lineBreakType == HB_SoftHyphen) {
                 // if we are splitting up a word because of
                 // a soft hyphen then we ...
                 //
diff --git a/tests/auto/gui/text/qtextlayout/tst_qtextlayout.cpp b/tests/auto/gui/text/qtextlayout/tst_qtextlayout.cpp
index d20fe83f2f..62cbc1fa5f 100644
--- a/tests/auto/gui/text/qtextlayout/tst_qtextlayout.cpp
+++ b/tests/auto/gui/text/qtextlayout/tst_qtextlayout.cpp
@@ -274,15 +274,15 @@ void tst_QTextLayout::lineBreaking()
         QString str = QString::fromUtf8(b->utf8);
         QTextEngine engine(str, QFont());
         const HB_CharAttributes *attrs = engine.attributes();
+        QVERIFY(attrs[0].lineBreakType == HB_NoBreak);
         int i;
         for (i = 0; i < (int)str.length() - 1; ++i) {
             QVERIFY(b->breaks[i] != 0xff);
-            if ( (attrs[i].lineBreakType != HB_NoBreak) != (bool)b->breaks[i] ) {
-                qDebug("test case \"%s\" failed at char %d; break type: %d", b->utf8, i, attrs[i].lineBreakType);
-                QCOMPARE( (attrs[i].lineBreakType != HB_NoBreak), (bool)b->breaks[i] );
+            if ( (attrs[i + 1].lineBreakType != HB_NoBreak) != (bool)b->breaks[i] ) {
+                qDebug("test case \"%s\" failed at char %d; break type: %d", b->utf8, i, attrs[i + 1].lineBreakType);
+                QCOMPARE( (attrs[i + 1].lineBreakType != HB_NoBreak), (bool)b->breaks[i] );
             }
         }
-        QVERIFY(attrs[i].lineBreakType == HB_ForcedBreak);
         QCOMPARE(b->breaks[i], (uchar)0xff);
         ++b;
     }