Use UTF-8 when converting 8 bit data in QTextStream

This was overlooked when doing the conversion to use UTF-8 as the
standard 8 bit encoding for text.

Fixes: QTBUG-54942
Change-Id: Ib7b1b75b4d694648ab7143f6930b6bb1dcad19c9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

4 files changed, 29 insertions, 16 deletions

diff --git a/src/corelib/serialization/qtextstream.cpp b/src/corelib/serialization/qtextstream.cpp
index 68ead1068b..3c69f3398f 100644
--- a/src/corelib/serialization/qtextstream.cpp
+++ b/src/corelib/serialization/qtextstream.cpp
@@ -945,6 +945,11 @@ void QTextStreamPrivate::putString(QLatin1String data, bool number)
     }
 }
 
+void QTextStreamPrivate::putString(QUtf8StringView data, bool number)
+{
+    putString(data.toString(), number);
+}
+
 /*!
     Constructs a QTextStream. Before you can use it for reading or
     writing, you must assign a device or a string.
@@ -2186,7 +2191,7 @@ QTextStream &QTextStream::operator>>(QString &str)
 /*!
     \overload
 
-    Converts the word to ISO-8859-1, then stores it in \a array.
+    Converts the word to UTF-8, then stores it in \a array.
 
     \sa QString::toLatin1()
 */
@@ -2195,7 +2200,6 @@ QTextStream &QTextStream::operator>>(QByteArray &array)
     Q_D(QTextStream);
     CHECK_VALID_STREAM(*this);
 
-    array.clear();
     d->scan(nullptr, nullptr, 0, QTextStreamPrivate::NotSpace);
     d->consumeLastToken();
 
@@ -2203,11 +2207,11 @@ QTextStream &QTextStream::operator>>(QByteArray &array)
     int length;
     if (!d->scan(&ptr, &length, 0, QTextStreamPrivate::Space)) {
         setStatus(ReadPastEnd);
+        array.clear();
         return *this;
     }
 
-    for (int i = 0; i < length; ++i)
-        array += ptr[i].toLatin1();
+    array = QStringView(ptr, length).toUtf8();
 
     d->consumeLastToken();
     return *this;
@@ -2216,13 +2220,14 @@ QTextStream &QTextStream::operator>>(QByteArray &array)
 /*!
     \overload
 
-    Stores the word in \a c, terminated by a '\\0' character. If no word is
-    available, only the '\\0' character is stored.
+    Converts the word to UTF-8 and stores it in \a c, terminated by a '\\0'
+    character. If no word is available, only the '\\0' character is stored.
 
     Warning: Although convenient, this operator is dangerous and must
     be used with care. QTextStream assumes that \a c points to a
     buffer with enough space to hold the word. If the buffer is too
-    small, your application may crash.
+    small, your application may crash. For a word consisting of \c{n} QChars,
+    the buffer needs to be at least \c{3*n+1} characters long.
 
     If possible, use the QByteArray operator instead.
 */
@@ -2241,9 +2246,9 @@ QTextStream &QTextStream::operator>>(char *c)
         return *this;
     }
 
-    for (int i = 0; i < length; ++i)
-        *c++ = ptr[i].toLatin1();
-    *c = '\0';
+    QStringEncoder encoder(QStringConverter::Utf8);
+    char *e = encoder.appendToBuffer(c, ptr, length);
+    *e = '\0';
     d->consumeLastToken();
     return *this;
 }
@@ -2559,7 +2564,7 @@ QTextStream &QTextStream::operator<<(const QByteArray &array)
     \overload
 
     Writes the constant string pointed to by \a string to the stream. \a
-    string is assumed to be in ISO-8859-1 encoding. This operator
+    string is assumed to be in UTF-8 encoding. This operator
     is convenient when working with constant string data. Example:
 
     \snippet code/src_corelib_io_qtextstream.cpp 8
@@ -2572,8 +2577,7 @@ QTextStream &QTextStream::operator<<(const char *string)
 {
     Q_D(QTextStream);
     CHECK_VALID_STREAM(*this);
-    // ### Qt6: consider changing to UTF-8
-    d->putString(QLatin1String(string));
+    d->putString(QUtf8StringView(string));
     return *this;
 }
 
diff --git a/src/corelib/serialization/qtextstream_p.h b/src/corelib/serialization/qtextstream_p.h
index 067a5ec94a..7d6889aa70 100644
--- a/src/corelib/serialization/qtextstream_p.h
+++ b/src/corelib/serialization/qtextstream_p.h
@@ -177,6 +177,7 @@ public:
     inline void putString(const QString &ch, bool number = false) { putString(ch.constData(), ch.length(), number); }
     void putString(const QChar *data, int len, bool number = false);
     void putString(QLatin1String data, bool number = false);
+    void putString(QUtf8StringView data, bool number = false);
     inline void putChar(QChar ch);
     void putNumber(qulonglong number, bool negative);
 
diff --git a/tests/auto/corelib/serialization/qtextstream/tst_qtextstream.cpp b/tests/auto/corelib/serialization/qtextstream/tst_qtextstream.cpp
index e1fd7a7147..f9de9eeebf 100644
--- a/tests/auto/corelib/serialization/qtextstream/tst_qtextstream.cpp
+++ b/tests/auto/corelib/serialization/qtextstream/tst_qtextstream.cpp
@@ -2125,9 +2125,9 @@ void tst_QTextStream::generateStringData(bool for_QString)
 
     if (!for_QString) {
         QTest::newRow("utf16-BE (empty)") << QByteArray("\xff\xfe", 2) << QByteArray() << QString();
-        QTest::newRow("utf16-BE (corrupt)") << QByteArray("\xff", 1) << QByteArray("\xff") << QString::fromLatin1("\xff");
+        QTest::newRow("utf16-BE (corrupt)") << QByteArray("\xff", 1) << QByteArray("\xc3\xbf") << QString::fromUtf8("\xc3\xbf");
         QTest::newRow("utf16-LE (empty)") << QByteArray("\xfe\xff", 2) << QByteArray() << QString();
-        QTest::newRow("utf16-LE (corrupt)") << QByteArray("\xfe", 1) << QByteArray("\xfe") << QString::fromLatin1("\xfe");
+        QTest::newRow("utf16-LE (corrupt)") << QByteArray("\xfe", 1) << QByteArray("\xc3\xbe") << QString::fromUtf8("\xc3\xbe");
     }
 }
 
diff --git a/tests/auto/corelib/text/qstring/tst_qstring.cpp b/tests/auto/corelib/text/qstring/tst_qstring.cpp
index 8f77d03226..6e1b47a839 100644
--- a/tests/auto/corelib/text/qstring/tst_qstring.cpp
+++ b/tests/auto/corelib/text/qstring/tst_qstring.cpp
@@ -3963,8 +3963,16 @@ void tst_QString::check_QTextIOStream()
     {
         a="";
         QTextStream ts(&a);
+        // invalid Utf8
         ts << "pi \261= " << 3.125;
-        QCOMPARE(a, QString::fromLatin1("pi \261= 3.125"));
+        QCOMPARE(a, QString::fromUtf16(u"pi \xfffd= 3.125"));
+    }
+    {
+        a="";
+        QTextStream ts(&a);
+        // valid Utf8
+        ts << "pi ø= " << 3.125;
+        QCOMPARE(a, QString::fromUtf16(u"pi ø= 3.125"));
     }
     {
         a="123 456";