QDebug: pretty-print QStrings and QStringRefs

[ChangeLog][QtCore][QDebug] Printing of QStrings and QStringRefs
whenever "noquote" is not active now prints the strings in a format that
can be copied back to C++ code. All characters that aren't printable in
US-ASCII are escaped (this includes printable Unicode characters outside
of US-ASCII). Pretty-printing will not respect QTextFormat padding or
field widths.

Change-Id: I169a8a0508e24693f5652f0129defe7f709e5d08
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

src/corelib/io/qdebug.cpp

@@ -1,6 +1,7 @@
 /****************************************************************************
 **
 ** Copyright (C) 2014 Digia Plc and/or its subsidiary(-ies).
+** Copyright (C) 2014 Intel Corporation.
 ** Contact: http://www.qt-project.org/legal
 **
 ** This file is part of the QtCore module of the Qt Toolkit.
@@ -40,9 +41,12 @@
 
 #include "qdebug.h"
 #include <private/qtextstream_p.h>
+#include <private/qtools_p.h>
 
 QT_BEGIN_NAMESPACE
 
+using QtMiscUtils::toHexUpper;
+
 // This file is needed to force compilation of QDebug into the kernel library.
 
 /*!
@@ -168,6 +172,105 @@ void QDebug::putUcs4(uint ucs4)
     maybeQuote('\'');
 }
 
+template <typename Char>
+static inline void putEscapedString(QTextStreamPrivate *d, const Char *begin, int length)
+{
+    QChar quote(QLatin1Char('"'));
+    d->write(&quote, 1);
+
+    const Char *end = begin + length;
+    for (const Char *p = begin; p != end; ++p) {
+        if (sizeof(Char) == sizeof(QChar)) {
+            int runLength = 0;
+            while (p + runLength != end &&
+                   p[runLength] < 0x7f && p[runLength] >= 0x20 && p[runLength] != '\\' && p[runLength] != '"')
+                ++runLength;
+            if (runLength) {
+                d->write(reinterpret_cast<const QChar *>(p), runLength);
+                p += runLength - 1;
+                continue;
+            }
+        }
+
+        // print as an escape sequence
+        int buflen = 2;
+        ushort buf[sizeof "\\U12345678" - 1];
+        buf[0] = '\\';
+
+        switch (*p) {
+        case '"':
+        case '\\':
+            buf[1] = *p;
+            break;
+        case '\b':
+            buf[1] = 'b';
+            break;
+        case '\f':
+            buf[1] = 'f';
+            break;
+        case '\n':
+            buf[1] = 'n';
+            break;
+        case '\r':
+            buf[1] = 'r';
+            break;
+        case '\t':
+            buf[1] = 't';
+            break;
+        default:
+            if (QChar::isHighSurrogate(*p)) {
+                if ((p + 1) != end && QChar::isLowSurrogate(p[1])) {
+                    // properly-paired surrogates
+                    uint ucs4 = QChar::surrogateToUcs4(*p, p[1]);
+                    ++p;
+                    buf[1] = 'U';
+                    buf[2] = '0'; // toHexUpper(ucs4 >> 32);
+                    buf[3] = '0'; // toHexUpper(ucs4 >> 28);
+                    buf[4] = toHexUpper(ucs4 >> 20);
+                    buf[5] = toHexUpper(ucs4 >> 16);
+                    buf[6] = toHexUpper(ucs4 >> 12);
+                    buf[7] = toHexUpper(ucs4 >> 8);
+                    buf[8] = toHexUpper(ucs4 >> 4);
+                    buf[9] = toHexUpper(ucs4);
+                    buflen = 10;
+                    break;
+                }
+                // improperly-paired surrogates, fall through
+            }
+            buf[1] = 'u';
+            if (sizeof(Char) == 1) {
+                buf[2] = buf[3] = '0';
+            } else {
+                buf[2] = toHexUpper(*p >> 12);
+                buf[3] = toHexUpper(*p >> 8);
+            }
+            buf[4] = toHexUpper(*p >> 4);
+            buf[5] = toHexUpper(*p);
+            buflen = 6;
+        }
+        d->write(reinterpret_cast<QChar *>(buf), buflen);
+    }
+
+    d->write(&quote, 1);
+}
+
+/*!
+    \internal
+    Duplicated from QtTest::toPrettyUnicode().
+*/
+void QDebug::putString(const QChar *begin, size_t length)
+{
+    if (stream->testFlag(Stream::NoQuotes)) {
+        // no quotes, write the string directly too (no pretty-printing)
+        // this respects the QTextStream state, though
+        stream->ts.d_ptr->putString(begin, length);
+    } else {
+        // we'll reset the QTextStream formatting mechanisms, so save the state
+        QDebugStateSaver saver(*this);
+        stream->ts.d_ptr->params.reset();
+        putEscapedString(stream->ts.d_ptr.data(), reinterpret_cast<const ushort *>(begin), length);
+    }
+}
 
 /*!
     \fn QDebug::swap(QDebug &other)

src/corelib/io/qdebug.h

@@ -78,6 +78,7 @@ class Q_CORE_EXPORT QDebug
     } *stream;
 
     void putUcs4(uint ucs4);
+    void putString(const QChar *begin, size_t length);
 public:
     inline QDebug(QIODevice *device) : stream(new Stream(device)) {}
     inline QDebug(QString *string) : stream(new Stream(string)) {}
@@ -118,8 +119,8 @@ public:
     inline QDebug &operator<<(float t) { stream->ts << t; return maybeSpace(); }
     inline QDebug &operator<<(double t) { stream->ts << t; return maybeSpace(); }
     inline QDebug &operator<<(const char* t) { stream->ts << QString::fromUtf8(t); return maybeSpace(); }
-    inline QDebug &operator<<(const QString & t) { maybeQuote(); stream->ts << t; maybeQuote(); return maybeSpace(); }
+    inline QDebug &operator<<(const QString & t) { putString(t.constData(), uint(t.length())); return maybeSpace(); }
-    inline QDebug &operator<<(const QStringRef & t) { return operator<<(t.toString()); }
+    inline QDebug &operator<<(const QStringRef & t) { putString(t.constData(), uint(t.length())); return maybeSpace(); }
     inline QDebug &operator<<(QLatin1String t) { maybeQuote(); stream->ts << t; maybeQuote(); return maybeSpace(); }
     inline QDebug &operator<<(const QByteArray & t) { maybeQuote(); stream->ts << t; maybeQuote(); return maybeSpace(); }
     inline QDebug &operator<<(const void * t) { stream->ts << t; return maybeSpace(); }

src/corelib/io/qtextstream.cpp

@@ -806,7 +806,7 @@ inline void QTextStreamPrivate::restoreToSavedConverterState()
 /*!
     \internal
 */
-inline void QTextStreamPrivate::write(const QChar *data, int len)
+void QTextStreamPrivate::write(const QChar *data, int len)
 {
     if (string) {
         // ### What about seek()??
@@ -885,7 +885,7 @@ inline void QTextStreamPrivate::putChar(QChar ch)
 /*!
     \internal
 */
-inline void QTextStreamPrivate::putString(const QChar *data, int len, bool number)
+void QTextStreamPrivate::putString(const QChar *data, int len, bool number)
 {
     QString pad;
     int padLeft = 0, padRight = 0;

src/corelib/io/qtextstream.h

@@ -185,6 +185,7 @@ public:
 private:
     Q_DISABLE_COPY(QTextStream)
     friend class QDebugStateSaverPrivate;
+    friend class QDebug;
 
     QScopedPointer<QTextStreamPrivate> d_ptr;
 };

src/testlib/qtestcase.cpp

@@ -2146,6 +2146,8 @@ char *toHexRepresentation(const char *ba, int length)
     \internal
     Returns the same QString but with only the ASCII characters still shown;
     everything else is replaced with \c {\uXXXX}.
+
+    Similar to QDebug::putString().
 */
 char *toPrettyUnicode(const ushort *p, int length)
 {

tests/auto/corelib/io/qdebug/tst_qdebug.cpp

@@ -52,6 +52,7 @@ private slots:
     void stateSaver() const;
     void veryLongWarningMessage() const;
     void qDebugQChar() const;
+    void qDebugQString() const;
     void qDebugQStringRef() const;
     void qDebugQLatin1String() const;
     void qDebugQByteArray() const;
@@ -344,6 +345,54 @@ void tst_QDebug::qDebugQChar() const
 
 }
 
+void tst_QDebug::qDebugQString() const
+{
+    /* Use a basic string. */
+    {
+        QString file, function;
+        int line = 0;
+        const QString in(QLatin1String("input"));
+        const QStringRef inRef(&in);
+
+        MessageHandlerSetter mhs(myMessageHandler);
+        { qDebug() << inRef; }
+#ifndef QT_NO_MESSAGELOGCONTEXT
+        file = __FILE__; line = __LINE__ - 2; function = Q_FUNC_INFO;
+#endif
+        QCOMPARE(s_msgType, QtDebugMsg);
+        QCOMPARE(s_msg, QString::fromLatin1("\"input\""));
+        QCOMPARE(QString::fromLatin1(s_file), file);
+        QCOMPARE(s_line, line);
+        QCOMPARE(QString::fromLatin1(s_function), function);
+    }
+
+    /* simpler tests from now on */
+    MessageHandlerSetter mhs(myMessageHandler);
+
+    QString string = "Hello";
+    qDebug() << string;
+    QCOMPARE(s_msg, QString("\"Hello\""));
+
+    qDebug().noquote().nospace() << string;
+    QCOMPARE(s_msg, string);
+
+    qDebug().noquote().nospace() << qSetFieldWidth(8) << string;
+    QCOMPARE(s_msg, "   " + string);
+
+    string = QLatin1String("\nSm\xF8rg\xE5sbord\\");
+    qDebug().noquote().nospace() << string;
+    QCOMPARE(s_msg, string);
+
+    qDebug() << string;
+    QCOMPARE(s_msg, QString("\"\\nSm\\u00F8rg\\u00E5sbord\\\\\""));
+
+    // surrogate pairs (including broken pairings)
+    ushort utf16[] = { 0xDC00, 0xD800, 0xDC00, 'x', 0xD800, 0xDC00, 0xD800, 0 };
+    string = QString::fromUtf16(utf16);
+    qDebug() << string;
+    QCOMPARE(s_msg, QString("\"\\uDC00\\U00010000x\\U00010000\\uD800\""));
+}
+
 void tst_QDebug::qDebugQStringRef() const
 {
     /* Use a basic string. */
@@ -403,6 +452,24 @@ void tst_QDebug::qDebugQLatin1String() const
     QCOMPARE(QString::fromLatin1(s_file), file);
     QCOMPARE(s_line, line);
     QCOMPARE(QString::fromLatin1(s_function), function);
+
+    /* simpler tests from now on */
+    QLatin1String string("\"Hello\"");
+    qDebug() << string;
+    QCOMPARE(s_msg, QString("\"\\\"Hello\\\"\""));
+
+    qDebug().noquote().nospace() << string;
+    QCOMPARE(s_msg, QString(string));
+
+    qDebug().noquote().nospace() << qSetFieldWidth(8) << string;
+    QCOMPARE(s_msg, " " + QString(string));
+
+    string = QLatin1String("\nSm\xF8rg\xE5sbord\\");
+    qDebug().noquote().nospace() << string;
+    QCOMPARE(s_msg, QString(string));
+
+    qDebug() << string;
+    QCOMPARE(s_msg, QString("\"\\nSm\\u00F8rg\\u00E5sbord\\\\\""));
 }
 
 void tst_QDebug::qDebugQByteArray() const

tests/auto/testlib/selftests/expected_badxml.lightxml

@@ -113,7 +113,7 @@
 </TestFunction>
 <TestFunction name="encoding">
 <Message type="qdebug" file="" line="0">
-    <Description><![CDATA["Ülrich Ümläut"]]></Description>
+    <Description><![CDATA["\u00DClrich \u00DCml\u00E4ut"]]></Description>
 </Message>
 <Incident type="pass" file="" line="0" />
     <Duration msecs="0"/>

tests/auto/testlib/selftests/expected_badxml.xml

@@ -115,7 +115,7 @@
 </TestFunction>
 <TestFunction name="encoding">
 <Message type="qdebug" file="" line="0">
-    <Description><![CDATA["Ülrich Ümläut"]]></Description>
+    <Description><![CDATA["\u00DClrich \u00DCml\u00E4ut"]]></Description>
 </Message>
 <Incident type="pass" file="" line="0" />
     <Duration msecs="0"/>

tests/auto/testlib/selftests/expected_badxml.xunitxml

@@ -30,7 +30,7 @@
     <failure message="failure message" result="fail"/>
   </testcase>
   <testcase result="pass" name="encoding">
-    <!-- message="&quot;Ülrich Ümläut&quot;" type="qdebug" -->
+    <!-- message="&quot;\u00DClrich \u00DCml\u00E4ut&quot;" type="qdebug" -->
   </testcase>
   <testcase result="pass" name="cleanupTestCase"/>
   <system-err>
@@ -46,6 +46,6 @@
 <![CDATA[quotes " text" more text]]>
 <![CDATA[xml close > open < tags < text]]>
 <![CDATA[all > " mixed ]]]><![CDATA[]> up > " in < the ]]]><![CDATA[]> hopes < of triggering "< ]]]><![CDATA[]> bugs]]>
-<![CDATA["Ülrich Ümläut"]]>
+<![CDATA["\u00DClrich \u00DCml\u00E4ut"]]>
   </system-err>
 </testsuite>