[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-commits
Subject:    KDE/kdelibs/khtml
From:       Maks Orlovich <maksim () kde ! org>
Date:       2010-10-30 20:26:09
Message-ID: 20101030202609.0F439AC89B () svn ! kde ! org
[Download RAW message or body]

SVN commit 1191378 by orlovich:

- Better escaping in wildcard mode
- Be consistent in excluding stuff we don't handle.


 M  +34 -19    khtml_filter.cpp  
 M  +1 -4      khtml_settings.cpp  


--- trunk/KDE/kdelibs/khtml/khtml_filter.cpp #1191377:1191378
@@ -31,11 +31,39 @@
 
 namespace khtml {
 
+// We only want a subset of features of wildcards -- just the 
+// star, so we escape the rest before passing to QRegExp. 
+// The \ is escaped due to a QRegExp bug.
+// ### we really should rather parse it ourselves, in order to 
+// handle adblock-special things like | and ^ properly.
+static QRegExp fromAdBlockWildcard(const QString& wcStr) {
+    QRegExp rx;
+    rx.setPatternSyntax(QRegExp::Wildcard);
+
+    QString out;
+    for (int p = 0; p < wcStr.length(); ++p) {
+        QChar c = wcStr[p];
+        if (c == QLatin1Char('?'))
+            out += QLatin1String("[?]");
+        else if (c == QLatin1Char('['))
+            out += QLatin1String("[[]");
+        else if (c == QLatin1Char('\\'))
+            out += QLatin1String("[\\]");
+        else
+            out += c;
+    }
+    
+    rx.setPattern(out);
+    return rx;
+}
+
 void FilterSet::addFilter(const QString& filterStr)
 {
     QString filter = filterStr;
 
-    if (filter.startsWith(QLatin1Char('!')))
+    /** ignore special lines starting with "[", "!", "&", or "#" or contain "#" \
(comments or features are not supported by KHTML's AdBlock */ +    QChar firstChar = \
filter.at(0); +    if (firstChar == QLatin1Char('[') || firstChar == QLatin1Char('!') \
|| firstChar == QLatin1Char('&') || firstChar == QLatin1Char('#') || \
filter.contains(QLatin1Char('#')))  return;
 
     // Strip leading @@
@@ -86,29 +114,16 @@
         // Now, do we still have any wildcard stuff left?
         if (filter.contains("*"))
         {
-//             qDebug() << "W:" << filter;
             // check if we can use RK first (and then check full RE for the rest) \
for better performance  int aPos = filter.indexOf('*');
             if (aPos < 0)
                 aPos = filter.length();
-            int qPos = filter.indexOf('?');
-            if (qPos < 0)
-                qPos = filter.length();
-            int pos = qMin(aPos, qPos);
-            if (pos > 7) {
-                QRegExp rx;
-
-                rx.setPatternSyntax(QRegExp::Wildcard);
-                // Pad the final r.e. with * so we can check for an exact match
-                rx.setPattern(filter.mid(pos) + QLatin1Char('*'));
-
-                stringFiltersMatcher.addWildedString(filter.mid(0, pos), rx);
-
+            if (aPos > 7) {
+                QRegExp rx = fromAdBlockWildcard(filter.mid(aPos) + \
QLatin1Char('*')); +                // We pad the final r.e. with * so we can check \
for an exact match +                \
stringFiltersMatcher.addWildedString(filter.mid(0, aPos), rx);  } else {
-                QRegExp rx;
-
-                rx.setPatternSyntax(QRegExp::Wildcard);
-                rx.setPattern(filter);
+                QRegExp rx = fromAdBlockWildcard(filter);
                 reFilters.append(rx);
             }
         }
--- trunk/KDE/kdelibs/khtml/khtml_settings.cpp #1191377:1191378
@@ -138,7 +138,6 @@
             int whiteCounter = 0, blackCounter = 0;
 #endif // NDEBUG
             while (!line.isEmpty()) {
-                QChar firstChar = line.at(0);
                 /** white list lines start with "@@" */
                 if (line.startsWith(QLatin1String("@@")))
                 {
@@ -147,9 +146,7 @@
 #endif // NDEBUG
                     adWhiteList.addFilter(line);
                 }
-                /** ignore special lines starting with "[", "!", "&", or "#" or \
                contain "#"
-                    (those features are not supported by KHTML's AdBlock */
-                else if (firstChar != QLatin1Char('[') && firstChar != \
QLatin1Char('!') && firstChar != QLatin1Char('&') && firstChar != QLatin1Char('#') && \
!line.contains(QLatin1Char('#'))) +                else
                 {
 #ifndef NDEBUG
                     ++blackCounter;


[prev in list] [next in list] [prev in thread] [next in thread]