Let the user type the hyphen if he wants when searching

It happens that sometimes the hypen is actually "part of the word" like
in one-third, so if there's one- at the end of a line
and third at the beginning of the next, we should still match and not
force the user to type onethird, even we will also match onethird since
there's no way to know if "hyphen at end of line" is supposed to be part
of the word or not

BUGS: 418520

autotests/searchtest.cpp

@@ -288,6 +288,22 @@ void SearchTest::testHyphenAtEndOfLineWithoutYOverlap()
     QCOMPARE(*result, expected);
     delete result;
 
+    result = tp->findText(0, QStringLiteral("supercalifragilisticexpialidocious"),
+                                                 Okular::FromBottom, Qt::CaseSensitive, nullptr);
+    QVERIFY(result);
+    QCOMPARE(*result, expected);
+    delete result;
+
+    // If the user is looking for the text explicitely with the hyphen also find it
+    result = tp->findText(0, QStringLiteral("super-cali-fragilistic"), Okular::FromTop, Qt::CaseSensitive, nullptr);
+    QVERIFY(result);
+    delete result;
+
+    // If the user is looking for the text explicitely with the hyphen also find it
+    result = tp->findText(0, QStringLiteral("super-cali-fragilistic"), Okular::FromBottom, Qt::CaseSensitive, nullptr);
+    QVERIFY(result);
+    delete result;
+
     delete page;
 }
 

core/textpage.cpp

@@ -54,11 +54,17 @@ class SearchPoint
 
 static bool CaseInsensitiveCmpFn( const QStringRef & from, const QStringRef & to )
 {
+#ifdef DEBUG_TEXTPAGE
+    qDebug(OkularCoreDebug) << from << ":" << to << "(case insensitive)";
+#endif
     return from.compare( to, Qt::CaseInsensitive ) == 0;
 }
 
 static bool CaseSensitiveCmpFn( const QStringRef & from, const QStringRef & to )
 {
+#ifdef DEBUG_TEXTPAGE
+    qDebug(OkularCoreDebug) << from << ":" << to << "(case sensitive)";
+#endif
     return from.compare( to, Qt::CaseSensitive ) == 0;
 }
 
@@ -780,7 +786,7 @@ RegularAreaRect* TextPage::findText( int searchID, const QString &query, SearchD
 // if the '-' is the last entry
 static int stringLengthAdaptedWithHyphen(const QString &str, const TextList::ConstIterator &it, const TextList::ConstIterator &textListEnd)
 {
-    int len = str.length();
+    const int len = str.length();
     
     // hyphenated '-' must be at the end of a word, so hyphenation means
     // we have a '-' just followed by a '\n' character
@@ -795,10 +801,9 @@ static int stringLengthAdaptedWithHyphen(const QString &str, const TextList::Con
             const QString &lookahedStr = (*(it+1))->text();
             if (lookahedStr.startsWith(QLatin1Char('\n')))
             {
-                len -= 1;
+                return len - 1;
             }
-            else
+
-            {
             // 2. if the next word is in a different line or not
             const NormalizedRect& hyphenArea = (*it)->area;
             const NormalizedRect& lookaheadArea = (*(it + 1))->area;
@@ -806,15 +811,14 @@ static int stringLengthAdaptedWithHyphen(const QString &str, const TextList::Con
             // lookahead to check whether both the '-' rect and next character rect overlap
             if( !doesConsumeY( hyphenArea, lookaheadArea, 70 ) )
             {
-                    len -= 1;
+                return len - 1;
-                }
             }
         }
     }
     // else if it is the second last entry - for example in pdf format
     else if (str.endsWith(QLatin1String("-\n")))
     {
-        len -= 2;
+        return len - 2;
     }
     
     return len;
@@ -850,8 +854,7 @@ RegularAreaRect* TextPagePrivate::findTextInternalForward( int searchID, const Q
     const QString query = _query.normalized(QString::NormalizationForm_KC);
 
     // j is the current position in our query
-    // len is the length of the string in TextEntity
+    // queryLeft is the length of the query we have left to match
-    // queryLeft is the length of the query we have left
     int j=0, queryLeft=query.length();
 
     TextList::ConstIterator it = start;
@@ -864,9 +867,11 @@ RegularAreaRect* TextPagePrivate::findTextInternalForward( int searchID, const Q
     {
         const TinyTextEntity* curEntity = *it;
         const QString& str = curEntity->text();
-        int len = stringLengthAdaptedWithHyphen(str, it, m_words.constEnd());
+        const int strLen = str.length();
+        const int adjustedLen = stringLengthAdaptedWithHyphen(str, it, m_words.constEnd());
+        // adjustedLen <= strLen
 
-        if (offset >= len)
+        if (offset >= strLen)
         {
             it++;
             offset = 0;
@@ -879,15 +884,21 @@ RegularAreaRect* TextPagePrivate::findTextInternalForward( int searchID, const Q
             offset_begin = offset;
         }
 
-        int min=qMin(queryLeft,len-offset);
+        // Let the user write the hyphen or not when searching for text
+        int matchedLen = -1;
+        for (int matchingLen = strLen; matchingLen >= adjustedLen; matchingLen--)
         {
-#ifdef DEBUG_TEXTPAGE
-            qCDebug(OkularCoreDebug) << str.midRef(offset, min) << ":" << _query.midRef(j, min);
-#endif
             // we have equal (or less than) area of the query left as the length of the current
             // entity
+            const int min = qMin(queryLeft, matchingLen - offset);
+            if ( comparer( str.midRef( offset, min ), query.midRef( j, min ) ) )
+            {
+                matchedLen = min;
+                break;
+            }
+        }
 
-            if ( !comparer( str.midRef( offset, min ), query.midRef( j, min ) ) )
+        if ( matchedLen == -1 )
         {
             // we have not matched
             // this means we do not have a complete match
@@ -911,10 +922,10 @@ RegularAreaRect* TextPagePrivate::findTextInternalForward( int searchID, const Q
             // subtract the length of the current entity from
             // the left length of the query
 #ifdef DEBUG_TEXTPAGE
-            qCDebug(OkularCoreDebug) << "\tmatched";
+            qCDebug(OkularCoreDebug) << "\tmatched" << matchedLen;
 #endif
-                    j += min;
+            j += matchedLen;
-                    queryLeft -= min;
+            queryLeft -= matchedLen;
 
             if (queryLeft==0)
             {
@@ -928,7 +939,7 @@ RegularAreaRect* TextPagePrivate::findTextInternalForward( int searchID, const Q
                 sp->it_begin = it_begin;
                 sp->it_end = it;
                 sp->offset_begin = offset_begin;
-                        sp->offset_end = offset + min;
+                sp->offset_end = offset + matchedLen;
                 return searchPointToArea(sp);
             }
 
@@ -936,7 +947,6 @@ RegularAreaRect* TextPagePrivate::findTextInternalForward( int searchID, const Q
             offset = 0;
         }
     }
-    }
     // end of loop - it means that we've ended the textentities
 
     const QMap< int, SearchPoint* >::iterator sIt = m_searchPoints.find( searchID );
@@ -982,11 +992,13 @@ RegularAreaRect* TextPagePrivate::findTextInternalBackward( int searchID, const
 
         const TinyTextEntity* curEntity = *it;
         const QString& str = curEntity->text();
-        int len = stringLengthAdaptedWithHyphen(str, it, m_words.constEnd());
+        const int strLen = str.length();
+        const int adjustedLen = stringLengthAdaptedWithHyphen(str, it, m_words.constEnd());
+        // adjustedLen <= strLen
 
         if (offset <= 0)
         {
-            offset = len;
+            offset = strLen;
         }
 
         if ( it_begin == TextList::ConstIterator() )
@@ -995,16 +1007,22 @@ RegularAreaRect* TextPagePrivate::findTextInternalBackward( int searchID, const
             offset_begin = offset;
         }
 
-        int min=qMin(queryLeft,offset);
+        // Let the user write the hyphen or not when searching for text
-        {
+        int matchedLen = -1;
-#ifdef DEBUG_TEXTPAGE
-            qCDebug(OkularCoreDebug) << str.midRef(offset-min, min) << " : " << _query.midRef(j-min, min);
-#endif
         // we have equal (or less than) area of the query left as the length of the current
         // entity
+        for (int matchingLen = strLen; matchingLen >= adjustedLen; matchingLen--)
+        {
+            const int hyphenOffset = (strLen - matchingLen);
+            const int min = qMin(queryLeft + hyphenOffset, offset);
+            if ( comparer( str.midRef( offset - min, min - hyphenOffset ), query.midRef( j - min + hyphenOffset, min - hyphenOffset ) ) )
+            {
+                matchedLen = min - hyphenOffset;
+                break;
+            }
+        }
 
-            // Note len is not str.length() so we can't use rightRef here
+        if ( matchedLen == -1 )
-            if ( !comparer( str.midRef(offset-min, min ), query.midRef( j - min, min ) ) )
         {
             // we have not matched
             // this means we do not have a complete match
@@ -1031,8 +1049,8 @@ RegularAreaRect* TextPagePrivate::findTextInternalBackward( int searchID, const
 #ifdef DEBUG_TEXTPAGE
             qCDebug(OkularCoreDebug) << "\tmatched";
 #endif
-                    j -= min;
+            j -= matchedLen;
-                    queryLeft -= min;
+            queryLeft -= matchedLen;
 
             if ( queryLeft == 0 )
             {
@@ -1045,16 +1063,13 @@ RegularAreaRect* TextPagePrivate::findTextInternalBackward( int searchID, const
                 SearchPoint* sp = *sIt;
                 sp->it_begin = it;
                 sp->it_end = it_begin;
-                        sp->offset_begin = offset - min;
+                sp->offset_begin = offset - matchedLen;
                 sp->offset_end = offset_begin;
                 return searchPointToArea(sp);
             }
 
             offset = 0;
         }
-
-        }
-
     }
     // end of loop - it means that we've ended the textentities