qpdf.cpp does not compile with -qreal float on ARM linux with gcc 4.9

Build on ARM linux with gcc 4.9 fails while using -qreal float:

painting/qpdf.cpp: In member function 'int QPdfEnginePrivate::generateRadialGradientShader(const QRadialGradient*, const QTransform&, bool)':
painting/qpdf.cpp:2157:41: error: ambiguous overload for 'operator<<' (operand types are 'QPdf::ByteStream' and 'double')
"/Coords [" << p0.x() << p0.y() << r0 << p1.x() << p1.y() << r1 << "]\n"
^
painting/qpdf.cpp:2157:41: note: candidates are:
painting/qpdf.cpp:180:17: note: QPdf::ByteStream& QPdf::ByteStream::operator<<(char)
ByteStream &ByteStream::operator <<(char chr)
^
painting/qpdf.cpp:218:17: note: QPdf::ByteStream& QPdf::ByteStream::operator<<(qreal)
ByteStream &ByteStream::operator <<(qreal val) {
^
painting/qpdf.cpp:225:17: note: QPdf::ByteStream& QPdf::ByteStream::operator<<(int)
ByteStream &ByteStream::operator <<(int val) {
^