fakevim: Emulate Vim regexp more precisely

Change-Id: Ia951f6e10d0e8d19e8d7d4b2a696eb9277ce6201
Reviewed-by: hjk <qthjk@ovi.com>

src/plugins/fakevim/fakevimhandler.cpp

@@ -3916,22 +3916,121 @@ bool FakeVimHandler::Private::handleExPluginCommand(const ExCommand &cmd)
     return handled;
 }
 
-static QRegExp vimPatternToQtPattern(QString needle, QTextDocument::FindFlags *flags)
+static QRegExp vimPatternToQtPattern(QString needle)
 {
-    // FIXME: Rough mapping of a common case.
+    /* Trasformations (Vim regexp -> QRegExp):
-    if (needle.startsWith(_("\\<")) && needle.endsWith(_("\\>")))
+     *   \a -> [A-Za-z]
-        (*flags) |= QTextDocument::FindWholeWords;
+     *   \A -> [^A-Za-z]
-    needle.remove(_("\\<")); // start of word
+     *   \h -> [A-Za-z_]
-    needle.remove(_("\\>")); // end of word
+     *   \H -> [^A-Za-z_]
+     *   \l -> [a-z]
+     *   \L -> [^a-z]
+     *   \o -> [0-7]
+     *   \O -> [^0-7]
+     *   \u -> [A-Z]
+     *   \U -> [^A-Z]
+     *   \x -> [0-9A-Fa-f]
+     *   \X -> [^0-9A-Fa-f]
+     *
+     *   \< -> \b
+     *   \> -> \b
+     *   [] -> \[\]
+     *   \= -> ?
+     *
+     *   (...)  <-> \(...\)
+     *   {...}  <-> \{...\}
+     *   |      <-> \|
+     *   ?      <-> \?
+     *   +      <-> \+
+     *   \{...} -> {...}
+     *
+     *   \c - set ignorecase for rest
+     *   \C - set noignorecase for rest
+     */
+    // TODO: Set initial value and handle smartcase option.
+    bool ignorecase = false;
+    QString pattern;
+    pattern.reserve(2 * needle.size());
 
-    // QRegExp's | and \| have the opposite meaning of vim's.
+    bool escape = false;
-    QString dummy(QLatin1Char(1));
+    bool brace = false;
-    needle.replace(_("\\|"), dummy);
+    bool curly = false;
-    needle.replace(_("|"), _("\\|"));
+    foreach (const QChar &c, needle) {
-    needle.replace(dummy, _("|"));
+        if (brace) {
+            brace = false;
+            if (c == ']') {
+                pattern.append(_("\\[\\]"));
+                continue;
+            } else {
+                pattern.append('[');
+            }
+        }
+        if (QString("(){}+|?").indexOf(c) != -1) {
+            if (c == '{') {
+                curly = escape;
+            } else if (c == '}' && curly) {
+                curly = false;
+                escape = true;
+            }
 
-    //qDebug() << "NEEDLE " << needle << needle;
+            if (escape)
-    return QRegExp(needle);
+                escape = false;
+            else
+                pattern.append('\\');
+            pattern.append(c);
+        } else if (escape) {
+            // escape expression
+            escape = false;
+            if (c == '<' || c == '>')
+                pattern.append(_("\\b"));
+            else if (c == 'a')
+                pattern.append(_("[a-zA-Z]"));
+            else if (c == 'A')
+                pattern.append(_("[^a-zA-Z]"));
+            else if (c == 'h')
+                pattern.append(_("[A-Za-z_]"));
+            else if (c == 'H')
+                pattern.append(_("[^A-Za-z_]"));
+            else if (c == 'c' || c == 'C')
+                ignorecase = (c == 'c');
+            else if (c == 'l')
+                pattern.append(_("[a-z]"));
+            else if (c == 'L')
+                pattern.append(_("[^a-z]"));
+            else if (c == 'o')
+                pattern.append(_("[0-7]"));
+            else if (c == 'O')
+                pattern.append(_("[^0-7]"));
+            else if (c == 'u')
+                pattern.append(_("[A-Z]"));
+            else if (c == 'U')
+                pattern.append(_("[^A-Z]"));
+            else if (c == 'x')
+                pattern.append(_("[0-9A-Fa-f]"));
+            else if (c == 'X')
+                pattern.append(_("[^0-9A-Fa-f]"));
+            else if (c == '=')
+                pattern.append(_("?"));
+            else
+                pattern.append('\\' + c);
+        } else {
+            // unescaped expression
+            if (c == '\\')
+                escape = true;
+            else if (c == '[')
+                brace = true;
+            else if (c.isLetter() && ignorecase)
+                pattern.append('[' + c.toLower() + c.toUpper() + ']');
+            else
+                pattern.append(c);
+        }
+    }
+    if (escape)
+        pattern.append('\\');
+    else if (brace)
+        pattern.append('[');
+
+    return QRegExp(pattern);
 }
 
 void FakeVimHandler::Private::searchBalanced(bool forward, QChar needle, QChar other)
@@ -3975,7 +4074,7 @@ void FakeVimHandler::Private::search(const SearchData &sd)
     if (!sd.forward)
         flags |= QTextDocument::FindBackward;
 
-    QRegExp needleExp = vimPatternToQtPattern(sd.needle, &flags);
+    QRegExp needleExp = vimPatternToQtPattern(sd.needle);
 
     const int oldLine = cursorLine() - cursorLineOnScreen();
 
@@ -4036,10 +4135,16 @@ void FakeVimHandler::Private::highlightMatches(const QString &needle)
         QTextCursor tc = cursor();
         tc.movePosition(StartOfDocument, MoveAnchor);
 
-        QTextDocument::FindFlags flags = QTextDocument::FindCaseSensitively;
+        QRegExp needleExp = vimPatternToQtPattern(needle);
-        QRegExp needleExp = vimPatternToQtPattern(needle, &flags);
+        if (!needleExp.isValid()) {
+            QString error = needleExp.errorString();
+            showRedMessage(
+                FakeVimHandler::tr("Invalid regular expression: %1").arg(error));
+            return;
+        }
+
         while (!tc.atEnd()) {
-            tc = tc.document()->find(needleExp, tc.position(), flags);
+            tc = tc.document()->find(needleExp, tc.position());
             if (tc.isNull())
                 break;
             if (!tc.hasSelection())