lupdate: Escape special characters when processing QML

The transcode function is now in a parser tool class.
It is used by the cpp and qml parser.

Fixes: QTBUG-40639
Change-Id: I3c9c9d137669900642a077cf81a739bfc31be800
Reviewed-by: Kai Koehne <kai.koehne@qt.io>

4 files changed, 69 insertions, 65 deletions

diff --git a/src/linguist/lupdate/cpp.cpp b/src/linguist/lupdate/cpp.cpp
index d8a4f9a37..e96e102af 100644
--- a/src/linguist/lupdate/cpp.cpp
+++ b/src/linguist/lupdate/cpp.cpp
@@ -250,7 +250,6 @@ private:
     bool matchStringOrNull(QString *s);
     bool matchExpression();
 
-    QString transcode(const QString &str);
     void recordMessage(
         int line, const QString &context, const QString &text, const QString &comment,
         const QString &extracomment, const QString &msgid, const TranslatorMessage::ExtraData &extra,
@@ -1469,7 +1468,7 @@ bool CppParser::matchString(QString *s)
             return matches;
         matches = true;
         if (yyTok == Tok_String)
-            *s += transcode(yyWord);
+            *s += ParserTool::transcode(yyWord);
         else
             *s += yyWord;
         s->detach();
@@ -1550,65 +1549,14 @@ bool CppParser::matchExpression()
     return true;
 }
 
-QString CppParser::transcode(const QString &str)
-{
-    static const char tab[] = "abfnrtv";
-    static const char backTab[] = "\a\b\f\n\r\t\v";
-    // This function has to convert back to bytes, as C's \0* sequences work at that level.
-    const QByteArray in = str.toUtf8();
-    QByteArray out;
-
-    out.reserve(in.length());
-    for (int i = 0; i < in.length();) {
-        uchar c = in[i++];
-        if (c == '\\') {
-            if (i >= in.length())
-                break;
-            c = in[i++];
-
-            if (c == '\n')
-                continue;
-
-            if (c == 'x' || c == 'u' || c == 'U') {
-                const bool unicode = (c != 'x');
-                QByteArray hex;
-                while (i < in.length() && isxdigit((c = in[i]))) {
-                    hex += c;
-                    i++;
-                }
-                if (unicode)
-                    out += QString(QChar(hex.toUInt(nullptr, 16))).toUtf8();
-                else
-                    out += hex.toUInt(nullptr, 16);
-            } else if (c >= '0' && c < '8') {
-                QByteArray oct;
-                int n = 0;
-                oct += c;
-                while (n < 2 && i < in.length() && (c = in[i]) >= '0' && c < '8') {
-                    i++;
-                    n++;
-                    oct += c;
-                }
-                out += oct.toUInt(0, 8);
-            } else {
-                const char *p = strchr(tab, c);
-                out += !p ? c : backTab[p - tab];
-            }
-        } else {
-            out += c;
-        }
-    }
-    return QString::fromUtf8(out.constData(), out.length());
-}
-
 void CppParser::recordMessage(int line, const QString &context, const QString &text, const QString &comment,
     const QString &extracomment, const QString &msgid, const TranslatorMessage::ExtraData &extra, bool plural)
 {
     TranslatorMessage msg(
-        transcode(context), text, transcode(comment), QString(),
+        ParserTool::transcode(context), text, ParserTool::transcode(comment), QString(),
         yyFileName, line, QStringList(),
         TranslatorMessage::Unfinished, plural);
-    msg.setExtraComment(transcode(extracomment.simplified()));
+    msg.setExtraComment(ParserTool::transcode(extracomment.simplified()));
     msg.setId(msgid);
     msg.setExtras(extra);
     tor->append(msg);
@@ -1778,7 +1726,7 @@ void CppParser::handleTrId(bool plural)
     yyTok = getToken();
     if (matchString(&msgid) && !msgid.isEmpty()) {
         plural |= match(Tok_Comma);
-        recordMessage(line, QString(), transcode(sourcetext), QString(), extracomment,
+        recordMessage(line, QString(), ParserTool::transcode(sourcetext), QString(), extracomment,
                       msgid, extra, plural);
     }
     sourcetext.clear();
@@ -2335,11 +2283,11 @@ void CppParser::processComment()
                 context = comment.left(k);
                 comment.remove(0, k + 1);
                 TranslatorMessage msg(
-                        transcode(context), QString(),
-                        transcode(comment), QString(),
+                        ParserTool::transcode(context), QString(),
+                        ParserTool::transcode(comment), QString(),
                         yyFileName, yyLineNo, QStringList(),
                         TranslatorMessage::Finished, false);
-                msg.setExtraComment(transcode(extracomment.simplified()));
+                msg.setExtraComment(ParserTool::transcode(extracomment.simplified()));
                 extracomment.clear();
                 tor->append(msg);
                 tor->setExtras(extra);
diff --git a/src/linguist/lupdate/lupdate.h b/src/linguist/lupdate/lupdate.h
index 0d98379d0..bb58cbd1e 100644
--- a/src/linguist/lupdate/lupdate.h
+++ b/src/linguist/lupdate/lupdate.h
@@ -99,6 +99,11 @@ bool loadQml(Translator &translator, const QString &filename, ConversionData &cd
     UNARY_MACRO(qsTranslate) \
     /*end*/
 
+class ParserTool
+{
+public:
+    static QString transcode(const QString &str);
+};
 
 class TrFunctionAliasManager {
 public:
diff --git a/src/linguist/lupdate/main.cpp b/src/linguist/lupdate/main.cpp
index ba1126aa7..c6b026dc7 100644
--- a/src/linguist/lupdate/main.cpp
+++ b/src/linguist/lupdate/main.cpp
@@ -136,6 +136,57 @@ QStringList TrFunctionAliasManager::availableFunctionsWithAliases() const
 
 TrFunctionAliasManager trFunctionAliasManager;
 
+QString ParserTool::transcode(const QString &str)
+{
+    static const char tab[] = "abfnrtv";
+    static const char backTab[] = "\a\b\f\n\r\t\v";
+    // This function has to convert back to bytes, as C's \0* sequences work at that level.
+    const QByteArray in = str.toUtf8();
+    QByteArray out;
+
+    out.reserve(in.length());
+    for (int i = 0; i < in.length();) {
+        uchar c = in[i++];
+        if (c == '\\') {
+            if (i >= in.length())
+                break;
+            c = in[i++];
+
+            if (c == '\n')
+                continue;
+
+            if (c == 'x' || c == 'u' || c == 'U') {
+                const bool unicode = (c != 'x');
+                QByteArray hex;
+                while (i < in.length() && isxdigit((c = in[i]))) {
+                    hex += c;
+                    i++;
+                }
+                if (unicode)
+                    out += QString(QChar(hex.toUInt(nullptr, 16))).toUtf8();
+                else
+                    out += hex.toUInt(nullptr, 16);
+            } else if (c >= '0' && c < '8') {
+                QByteArray oct;
+                int n = 0;
+                oct += c;
+                while (n < 2 && i < in.length() && (c = in[i]) >= '0' && c < '8') {
+                    i++;
+                    n++;
+                    oct += c;
+                }
+                out += oct.toUInt(0, 8);
+            } else {
+                const char *p = strchr(tab, c);
+                out += !p ? c : backTab[p - tab];
+            }
+        } else {
+            out += c;
+        }
+    }
+    return QString::fromUtf8(out.constData(), out.length());
+}
+
 static QString m_defaultExtensions;
 
 static void printOut(const QString & out)
diff --git a/src/linguist/lupdate/qdeclarative.cpp b/src/linguist/lupdate/qdeclarative.cpp
index 69164add2..fa8ad260e 100644
--- a/src/linguist/lupdate/qdeclarative.cpp
+++ b/src/linguist/lupdate/qdeclarative.cpp
@@ -130,11 +130,11 @@ protected:
                 if (!sourcetext.isEmpty())
                     yyMsg(identLineNo) << qPrintable(LU::tr("//% cannot be used with %1(). Ignoring\n").arg(name));
 
-                TranslatorMessage msg(m_component, source,
+                TranslatorMessage msg(m_component, ParserTool::transcode(source),
                     comment, QString(), m_fileName,
                     node->firstSourceLocation().startLine, QStringList(),
                     TranslatorMessage::Unfinished, plural);
-                msg.setExtraComment(extracomment.simplified());
+                msg.setExtraComment(ParserTool::transcode(extracomment.simplified()));
                 msg.setId(msgid);
                 msg.setExtras(extra);
                 m_translator->extend(msg, m_cd);
@@ -171,11 +171,11 @@ protected:
                         plural = true;
                 }
 
-                TranslatorMessage msg(context, source,
+                TranslatorMessage msg(context, ParserTool::transcode(source),
                     comment, QString(), m_fileName,
                     node->firstSourceLocation().startLine, QStringList(),
                     TranslatorMessage::Unfinished, plural);
-                msg.setExtraComment(extracomment.simplified());
+                msg.setExtraComment(ParserTool::transcode(extracomment.simplified()));
                 msg.setId(msgid);
                 msg.setExtras(extra);
                 m_translator->extend(msg, m_cd);
@@ -199,11 +199,11 @@ protected:
 
                 bool plural = node->arguments->next;
 
-                TranslatorMessage msg(QString(), sourcetext,
+                TranslatorMessage msg(QString(), ParserTool::transcode(sourcetext),
                     QString(), QString(), m_fileName,
                     node->firstSourceLocation().startLine, QStringList(),
                     TranslatorMessage::Unfinished, plural);
-                msg.setExtraComment(extracomment.simplified());
+                msg.setExtraComment(ParserTool::transcode(extracomment.simplified()));
                 msg.setId(id);
                 msg.setExtras(extra);
                 m_translator->extend(msg, m_cd);