// Copyright (C) 2016 The Qt Company Ltd. // SPDX-License-Identifier: LicenseRef-Qt-Commercial OR LGPL-3.0-only OR GPL-2.0-only OR GPL-3.0-only #include "qdatastream.h" #include "qdatastream_p.h" #if !defined(QT_NO_DATASTREAM) || defined(QT_BOOTSTRAPPED) #include "qbuffer.h" #include "qfloat16.h" #include "qstring.h" #include #include #include #include "qendian.h" QT_BEGIN_NAMESPACE constexpr quint32 QDataStream::NullCode; constexpr quint32 QDataStream::ExtendedSize; /*! \class QDataStream \inmodule QtCore \ingroup qtserialization \reentrant \brief The QDataStream class provides serialization of binary data to a QIODevice. \ingroup io A data stream is a binary stream of encoded information which is 100% independent of the host computer's operating system, CPU or byte order. For example, a data stream that is written by a PC under Windows can be read by a Sun SPARC running Solaris. You can also use a data stream to read/write \l{raw}{raw unencoded binary data}. If you want a "parsing" input stream, see QTextStream. The QDataStream class implements the serialization of C++'s basic data types, like \c char, \c short, \c int, \c{char *}, etc. Serialization of more complex data is accomplished by breaking up the data into primitive units. A data stream cooperates closely with a QIODevice. A QIODevice represents an input/output medium one can read data from and write data to. The QFile class is an example of an I/O device. Example (write binary data to a stream): \snippet code/src_corelib_io_qdatastream.cpp 0 Example (read binary data from a stream): \snippet code/src_corelib_io_qdatastream.cpp 1 Each item written to the stream is written in a predefined binary format that varies depending on the item's type. Supported Qt types include QBrush, QColor, QDateTime, QFont, QPixmap, QString, QVariant and many others. For the complete list of all Qt types supporting data streaming see \l{Serializing Qt Data Types}. For integers it is best to always cast to a Qt integer type for writing, and to read back into the same Qt integer type. This ensures that you get integers of the size you want and insulates you from compiler and platform differences. Enumerations can be serialized through QDataStream without the need of manually defining streaming operators. Enum classes are serialized using the declared size. The initial I/O device is usually set in the constructor, but can be changed with setDevice(). If you've reached the end of the data (or if there is no I/O device set) atEnd() will return true. \section1 Serializing containers and strings The serialization format is a length specifier first, then \a l bytes of data. The length specifier is one quint32 if the version is less than 6.7 or if the number of elements is less than 0xfffffffe (2^32 -2). Otherwise there is an extend value 0xfffffffe followed by one quint64 with the actual value. In addition for containers that support isNull(), it is encoded as a single quint32 with all bits set and no data. To take one example, if the string size fits into 32 bits, a \c{char *} string is written as a 32-bit integer equal to the length of the string, including the '\\0' byte, followed by all the characters of the string, including the '\\0' byte. If the string size is greater, the value 0xffffffffe is written as a marker of an extended size, followed by 64 bits of the actual size. When reading a \c {char *} string, 4 bytes are read first. If the value is not equal to 0xffffffffe (the marker of extended size), then these 4 bytes are treated as the 32 bit size of the string. Otherwise, the next 8 bytes are read and treated as a 64 bit size of the string. Then, all the characters for the \c {char *} string, including the '\\0' terminator, are read. \section1 Versioning QDataStream's binary format has evolved since Qt 1.0, and is likely to continue evolving to reflect changes done in Qt. When inputting or outputting complex types, it's very important to make sure that the same version of the stream (version()) is used for reading and writing. If you need both forward and backward compatibility, you can hardcode the version number in the application: \snippet code/src_corelib_io_qdatastream.cpp 2 If you are producing a new binary data format, such as a file format for documents created by your application, you could use a QDataStream to write the data in a portable format. Typically, you would write a brief header containing a magic string and a version number to give yourself room for future expansion. For example: \snippet code/src_corelib_io_qdatastream.cpp 3 Then read it in with: \snippet code/src_corelib_io_qdatastream.cpp 4 You can select which byte order to use when serializing data. The default setting is big-endian (MSB first). Changing it to little-endian breaks the portability (unless the reader also changes to little-endian). We recommend keeping this setting unless you have special requirements. \target raw \section1 Reading and Writing Raw Binary Data You may wish to read/write your own raw binary data to/from the data stream directly. Data may be read from the stream into a preallocated \c{char *} using readRawData(). Similarly data can be written to the stream using writeRawData(). Note that any encoding/decoding of the data must be done by you. A similar pair of functions is readBytes() and writeBytes(). These differ from their \e raw counterparts as follows: readBytes() reads a quint32 which is taken to be the length of the data to be read, then that number of bytes is read into the preallocated \c{char *}; writeBytes() writes a quint32 containing the length of the data, followed by the data. Note that any encoding/decoding of the data (apart from the length quint32) must be done by you. \section1 Reading and Writing Qt Collection Classes The Qt container classes can also be serialized to a QDataStream. These include QList, QSet, QHash, and QMap. The stream operators are declared as non-members of the classes. \target Serializing Qt Classes \section1 Reading and Writing Other Qt Classes In addition to the overloaded stream operators documented here, any Qt classes that you might want to serialize to a QDataStream will have appropriate stream operators declared as non-member of the class: \snippet code/src_corelib_serialization_qdatastream.cpp 0 For example, here are the stream operators declared as non-members of the QImage class: \snippet code/src_corelib_serialization_qdatastream.cpp 1 To see if your favorite Qt class has similar stream operators defined, check the \b {Related Non-Members} section of the class's documentation page. \section1 Using Read Transactions When a data stream operates on an asynchronous device, the chunks of data can arrive at arbitrary points in time. The QDataStream class implements a transaction mechanism that provides the ability to read the data atomically with a series of stream operators. As an example, you can handle incomplete reads from a socket by using a transaction in a slot connected to the readyRead() signal: \snippet code/src_corelib_io_qdatastream.cpp 6 If no full packet is received, this code restores the stream to the initial position, after which you need to wait for more data to arrive. \section1 Corruption and Security QDataStream is not resilient against corrupted data inputs and should therefore not be used for security-sensitive situations, even when using transactions. Transactions will help determine if a valid input can currently be decoded with the data currently available on an asynchronous device, but will assume that the data that is available is correctly formed. Additionally, many QDataStream demarshalling operators will allocate memory based on information found in the stream. Those operators perform no verification on whether the requested amount of memory is reasonable or if it is compatible with the amount of data available in the stream (example: demarshalling a QByteArray or QString may see the request for allocation of several gigabytes of data). QDataStream should not be used on content whose provenance cannot be trusted. Applications should be designed to attempt to decode only streams whose provenance is at least as trustworthy as that of the application itself or its plugins. \sa QTextStream, QVariant */ /*! \enum QDataStream::ByteOrder The byte order used for reading/writing the data. \value BigEndian Most significant byte first (the default) \value LittleEndian Least significant byte first */ /*! \enum QDataStream::FloatingPointPrecision The precision of floating point numbers used for reading/writing the data. This will only have an effect if the version of the data stream is Qt_4_6 or higher. \warning The floating point precision must be set to the same value on the object that writes and the object that reads the data stream. \value SinglePrecision All floating point numbers in the data stream have 32-bit precision. \value DoublePrecision All floating point numbers in the data stream have 64-bit precision. \sa setFloatingPointPrecision(), floatingPointPrecision() */ /*! \enum QDataStream::Status This enum describes the current status of the data stream. \value Ok The data stream is operating normally. \value ReadPastEnd The data stream has read past the end of the data in the underlying device. \value ReadCorruptData The data stream has read corrupt data. \value WriteFailed The data stream cannot write to the underlying device. \value [since 6.7] SizeLimitExceeded The data stream cannot read or write the data because its size is larger than supported by the current platform. This can happen, for example, when trying to read more that 2 GiB of data on a 32-bit platform. */ /***************************************************************************** QDataStream member functions *****************************************************************************/ #define Q_VOID #undef CHECK_STREAM_PRECOND #ifndef QT_NO_DEBUG #define CHECK_STREAM_PRECOND(retVal) \ if (!dev) { \ qWarning("QDataStream: No device"); \ return retVal; \ } #else #define CHECK_STREAM_PRECOND(retVal) \ if (!dev) { \ return retVal; \ } #endif #define CHECK_STREAM_WRITE_PRECOND(retVal) \ CHECK_STREAM_PRECOND(retVal) \ if (q_status != Ok) \ return retVal; #define CHECK_STREAM_TRANSACTION_PRECOND(retVal) \ if (transactionDepth == 0) { \ qWarning("QDataStream: No transaction in progress"); \ return retVal; \ } /*! Constructs a data stream that has no I/O device. \sa setDevice() */ QDataStream::QDataStream() { } /*! Constructs a data stream that uses the I/O device \a d. \sa setDevice(), device() */ QDataStream::QDataStream(QIODevice *d) { dev = d; // set device } /*! \fn QDataStream::QDataStream(QByteArray *a, OpenMode mode) Constructs a data stream that operates on a byte array, \a a. The \a mode describes how the device is to be used. Alternatively, you can use QDataStream(const QByteArray &) if you just want to read from a byte array. Since QByteArray is not a QIODevice subclass, internally a QBuffer is created to wrap the byte array. */ QDataStream::QDataStream(QByteArray *a, OpenMode flags) { QBuffer *buf = new QBuffer(a); #ifndef QT_NO_QOBJECT buf->blockSignals(true); #endif buf->open(flags); dev = buf; owndev = true; } /*! Constructs a read-only data stream that operates on byte array \a a. Use QDataStream(QByteArray*, int) if you want to write to a byte array. Since QByteArray is not a QIODevice subclass, internally a QBuffer is created to wrap the byte array. */ QDataStream::QDataStream(const QByteArray &a) { QBuffer *buf = new QBuffer; #ifndef QT_NO_QOBJECT buf->blockSignals(true); #endif buf->setData(a); buf->open(QIODevice::ReadOnly); dev = buf; owndev = true; } /*! Destroys the data stream. The destructor will not affect the current I/O device, unless it is an internal I/O device (e.g. a QBuffer) processing a QByteArray passed in the \e constructor, in which case the internal I/O device is destroyed. */ QDataStream::~QDataStream() { if (owndev) delete dev; } /*! \fn QIODevice *QDataStream::device() const Returns the I/O device currently set, or \nullptr if no device is currently set. \sa setDevice() */ /*! void QDataStream::setDevice(QIODevice *d) Sets the I/O device to \a d, which can be \nullptr to unset to current I/O device. \sa device() */ void QDataStream::setDevice(QIODevice *d) { if (owndev) { delete dev; owndev = false; } dev = d; } /*! \fn bool QDataStream::atEnd() const Returns \c true if the I/O device has reached the end position (end of the stream or file) or if there is no I/O device set; otherwise returns \c false. \sa QIODevice::atEnd() */ bool QDataStream::atEnd() const { return dev ? dev->atEnd() : true; } /*! \fn QDataStream::FloatingPointPrecision QDataStream::floatingPointPrecision() const Returns the floating point precision of the data stream. \since 4.6 \sa FloatingPointPrecision, setFloatingPointPrecision() */ /*! Sets the floating point precision of the data stream to \a precision. If the floating point precision is DoublePrecision and the version of the data stream is Qt_4_6 or higher, all floating point numbers will be written and read with 64-bit precision. If the floating point precision is SinglePrecision and the version is Qt_4_6 or higher, all floating point numbers will be written and read with 32-bit precision. For versions prior to Qt_4_6, the precision of floating point numbers in the data stream depends on the stream operator called. The default is DoublePrecision. Note that this property does not affect the serialization or deserialization of \c qfloat16 instances. \warning This property must be set to the same value on the object that writes and the object that reads the data stream. \since 4.6 */ void QDataStream::setFloatingPointPrecision(QDataStream::FloatingPointPrecision precision) { fpPrecision = precision; } /*! \fn QDataStream::status() const Returns the status of the data stream. \sa Status, setStatus(), resetStatus() */ /*! Resets the status of the data stream. \sa Status, status(), setStatus() */ void QDataStream::resetStatus() { q_status = Ok; } /*! Sets the status of the data stream to the \a status given. Subsequent calls to setStatus() are ignored until resetStatus() is called. \sa Status, status(), resetStatus() */ void QDataStream::setStatus(Status status) { if (q_status == Ok) q_status = status; } /*! \fn int QDataStream::byteOrder() const Returns the current byte order setting -- either BigEndian or LittleEndian. \sa setByteOrder() */ /*! Sets the serialization byte order to \a bo. The \a bo parameter can be QDataStream::BigEndian or QDataStream::LittleEndian. The default setting is big-endian. We recommend leaving this setting unless you have special requirements. \sa byteOrder() */ void QDataStream::setByteOrder(ByteOrder bo) { #if QT_VERSION < QT_VERSION_CHECK(7, 0, 0) && !defined(QT_BOOTSTRAPPED) // accessed by inline byteOrder() prior to Qt 6.8 byteorder = bo; #endif if (QSysInfo::ByteOrder == QSysInfo::BigEndian) noswap = (bo == BigEndian); else noswap = (bo == LittleEndian); } /*! \enum QDataStream::Version This enum provides symbolic synonyms for the data serialization format version numbers. \value Qt_1_0 Version 1 (Qt 1.x) \value Qt_2_0 Version 2 (Qt 2.0) \value Qt_2_1 Version 3 (Qt 2.1, 2.2, 2.3) \value Qt_3_0 Version 4 (Qt 3.0) \value Qt_3_1 Version 5 (Qt 3.1, 3.2) \value Qt_3_3 Version 6 (Qt 3.3) \value Qt_4_0 Version 7 (Qt 4.0, Qt 4.1) \value Qt_4_1 Version 7 (Qt 4.0, Qt 4.1) \value Qt_4_2 Version 8 (Qt 4.2) \value Qt_4_3 Version 9 (Qt 4.3) \value Qt_4_4 Version 10 (Qt 4.4) \value Qt_4_5 Version 11 (Qt 4.5) \value Qt_4_6 Version 12 (Qt 4.6, Qt 4.7, Qt 4.8) \value Qt_4_7 Same as Qt_4_6. \value Qt_4_8 Same as Qt_4_6. \value Qt_4_9 Same as Qt_4_6. \value Qt_5_0 Version 13 (Qt 5.0) \value Qt_5_1 Version 14 (Qt 5.1) \value Qt_5_2 Version 15 (Qt 5.2) \value Qt_5_3 Same as Qt_5_2 \value Qt_5_4 Version 16 (Qt 5.4) \value Qt_5_5 Same as Qt_5_4 \value Qt_5_6 Version 17 (Qt 5.6) \value Qt_5_7 Same as Qt_5_6 \value Qt_5_8 Same as Qt_5_6 \value Qt_5_9 Same as Qt_5_6 \value Qt_5_10 Same as Qt_5_6 \value Qt_5_11 Same as Qt_5_6 \value Qt_5_12 Version 18 (Qt 5.12) \value Qt_5_13 Version 19 (Qt 5.13) \value Qt_5_14 Same as Qt_5_13 \value Qt_5_15 Same as Qt_5_13 \value Qt_6_0 Version 20 (Qt 6.0) \value Qt_6_1 Same as Qt_6_0 \value Qt_6_2 Same as Qt_6_0 \value Qt_6_3 Same as Qt_6_0 \value Qt_6_4 Same as Qt_6_0 \value Qt_6_5 Same as Qt_6_0 \value Qt_6_6 Version 21 (Qt 6.6) \value Qt_6_7 Version 22 (Qt 6.7) \value Qt_6_8 Same as Qt_6_7 \omitvalue Qt_DefaultCompiledVersion \sa setVersion(), version() */ /*! \fn int QDataStream::version() const Returns the version number of the data serialization format. \sa setVersion(), Version */ /*! \fn void QDataStream::setVersion(int v) Sets the version number of the data serialization format to \a v, a value of the \l Version enum. You don't \e have to set a version if you are using the current version of Qt, but for your own custom binary formats we recommend that you do; see \l{Versioning} in the Detailed Description. To accommodate new functionality, the datastream serialization format of some Qt classes has changed in some versions of Qt. If you want to read data that was created by an earlier version of Qt, or write data that can be read by a program that was compiled with an earlier version of Qt, use this function to modify the serialization format used by QDataStream. The \l Version enum provides symbolic constants for the different versions of Qt. For example: \snippet code/src_corelib_io_qdatastream.cpp 5 \sa version(), Version */ /*! \since 5.7 Starts a new read transaction on the stream. Defines a restorable point within the sequence of read operations. For sequential devices, read data will be duplicated internally to allow recovery in case of incomplete reads. For random-access devices, this function saves the current position of the stream. Call commitTransaction(), rollbackTransaction(), or abortTransaction() to finish the current transaction. Once a transaction is started, subsequent calls to this function will make the transaction recursive. Inner transactions act as agents of the outermost transaction (i.e., report the status of read operations to the outermost transaction, which can restore the position of the stream). \note Restoring to the point of the nested startTransaction() call is not supported. When an error occurs during a transaction (including an inner transaction failing), reading from the data stream is suspended (all subsequent read operations return empty/zero values) and subsequent inner transactions are forced to fail. Starting a new outermost transaction recovers from this state. This behavior makes it unnecessary to error-check every read operation separately. \sa commitTransaction(), rollbackTransaction(), abortTransaction() */ void QDataStream::startTransaction() { CHECK_STREAM_PRECOND(Q_VOID) if (++transactionDepth == 1) { dev->startTransaction(); resetStatus(); } } /*! \since 5.7 Completes a read transaction. Returns \c true if no read errors have occurred during the transaction; otherwise returns \c false. If called on an inner transaction, committing will be postponed until the outermost commitTransaction(), rollbackTransaction(), or abortTransaction() call occurs. Otherwise, if the stream status indicates reading past the end of the data, this function restores the stream data to the point of the startTransaction() call. When this situation occurs, you need to wait for more data to arrive, after which you start a new transaction. If the data stream has read corrupt data or any of the inner transactions was aborted, this function aborts the transaction. \sa startTransaction(), rollbackTransaction(), abortTransaction() */ bool QDataStream::commitTransaction() { CHECK_STREAM_TRANSACTION_PRECOND(false) if (--transactionDepth == 0) { CHECK_STREAM_PRECOND(false) if (q_status == ReadPastEnd) { dev->rollbackTransaction(); return false; } dev->commitTransaction(); } return q_status == Ok; } /*! \since 5.7 Reverts a read transaction. This function is commonly used to rollback the transaction when an incomplete read was detected prior to committing the transaction. If called on an inner transaction, reverting is delegated to the outermost transaction, and subsequently started inner transactions are forced to fail. For the outermost transaction, restores the stream data to the point of the startTransaction() call. If the data stream has read corrupt data or any of the inner transactions was aborted, this function aborts the transaction. If the preceding stream operations were successful, sets the status of the data stream to \value ReadPastEnd. \sa startTransaction(), commitTransaction(), abortTransaction() */ void QDataStream::rollbackTransaction() { setStatus(ReadPastEnd); CHECK_STREAM_TRANSACTION_PRECOND(Q_VOID) if (--transactionDepth != 0) return; CHECK_STREAM_PRECOND(Q_VOID) if (q_status == ReadPastEnd) dev->rollbackTransaction(); else dev->commitTransaction(); } /*! \since 5.7 Aborts a read transaction. This function is commonly used to discard the transaction after higher-level protocol errors or loss of stream synchronization. If called on an inner transaction, aborting is delegated to the outermost transaction, and subsequently started inner transactions are forced to fail. For the outermost transaction, discards the restoration point and any internally duplicated data of the stream. Will not affect the current read position of the stream. Sets the status of the data stream to \value ReadCorruptData. \sa startTransaction(), commitTransaction(), rollbackTransaction() */ void QDataStream::abortTransaction() { q_status = ReadCorruptData; CHECK_STREAM_TRANSACTION_PRECOND(Q_VOID) if (--transactionDepth != 0) return; CHECK_STREAM_PRECOND(Q_VOID) dev->commitTransaction(); } /*! \internal */ bool QDataStream::isDeviceTransactionStarted() const { return dev && dev->isTransactionStarted(); } /***************************************************************************** QDataStream read functions *****************************************************************************/ /*! \internal */ qint64 QDataStream::readBlock(char *data, qint64 len) { // Disable reads on failure in transacted stream if (q_status != Ok && dev->isTransactionStarted()) return -1; const qint64 readResult = dev->read(data, len); if (readResult != len) setStatus(ReadPastEnd); return readResult; } /*! \fn QDataStream &QDataStream::operator>>(std::nullptr_t &ptr) \since 5.9 \overload Simulates reading a \c{std::nullptr_t} from the stream into \a ptr and returns a reference to the stream. This function does not actually read anything from the stream, as \c{std::nullptr_t} values are stored as 0 bytes. */ /*! \fn QDataStream &QDataStream::operator>>(quint8 &i) \overload Reads an unsigned byte from the stream into \a i, and returns a reference to the stream. */ /*! Reads a signed byte from the stream into \a i, and returns a reference to the stream. */ QDataStream &QDataStream::operator>>(qint8 &i) { i = 0; CHECK_STREAM_PRECOND(*this) char c; if (readBlock(&c, 1) == 1) i = qint8(c); return *this; } /*! \fn QDataStream &QDataStream::operator>>(quint16 &i) \overload Reads an unsigned 16-bit integer from the stream into \a i, and returns a reference to the stream. */ /*! \overload Reads a signed 16-bit integer from the stream into \a i, and returns a reference to the stream. */ QDataStream &QDataStream::operator>>(qint16 &i) { i = 0; CHECK_STREAM_PRECOND(*this) if (readBlock(reinterpret_cast(&i), 2) != 2) { i = 0; } else { if (!noswap) { i = qbswap(i); } } return *this; } /*! \fn QDataStream &QDataStream::operator>>(quint32 &i) \overload Reads an unsigned 32-bit integer from the stream into \a i, and returns a reference to the stream. */ /*! \overload Reads a signed 32-bit integer from the stream into \a i, and returns a reference to the stream. */ QDataStream &QDataStream::operator>>(qint32 &i) { i = 0; CHECK_STREAM_PRECOND(*this) if (readBlock(reinterpret_cast(&i), 4) != 4) { i = 0; } else { if (!noswap) { i = qbswap(i); } } return *this; } /*! \fn QDataStream &QDataStream::operator>>(quint64 &i) \overload Reads an unsigned 64-bit integer from the stream, into \a i, and returns a reference to the stream. */ /*! \overload Reads a signed 64-bit integer from the stream into \a i, and returns a reference to the stream. */ QDataStream &QDataStream::operator>>(qint64 &i) { i = qint64(0); CHECK_STREAM_PRECOND(*this) if (version() < 6) { quint32 i1, i2; *this >> i2 >> i1; i = ((quint64)i1 << 32) + i2; } else { if (readBlock(reinterpret_cast(&i), 8) != 8) { i = qint64(0); } else { if (!noswap) { i = qbswap(i); } } } return *this; } /*! Reads a boolean value from the stream into \a i. Returns a reference to the stream. */ QDataStream &QDataStream::operator>>(bool &i) { qint8 v; *this >> v; i = !!v; return *this; } /*! \overload Reads a floating point number from the stream into \a f, using the standard IEEE 754 format. Returns a reference to the stream. \sa setFloatingPointPrecision() */ QDataStream &QDataStream::operator>>(float &f) { if (version() >= QDataStream::Qt_4_6 && floatingPointPrecision() == QDataStream::DoublePrecision) { double d; *this >> d; f = d; return *this; } f = 0.0f; CHECK_STREAM_PRECOND(*this) if (readBlock(reinterpret_cast(&f), 4) != 4) { f = 0.0f; } else { if (!noswap) { union { float val1; quint32 val2; } x; x.val2 = qbswap(*reinterpret_cast(&f)); f = x.val1; } } return *this; } /*! \overload Reads a floating point number from the stream into \a f, using the standard IEEE 754 format. Returns a reference to the stream. \sa setFloatingPointPrecision() */ QDataStream &QDataStream::operator>>(double &f) { if (version() >= QDataStream::Qt_4_6 && floatingPointPrecision() == QDataStream::SinglePrecision) { float d; *this >> d; f = d; return *this; } f = 0.0; CHECK_STREAM_PRECOND(*this) if (readBlock(reinterpret_cast(&f), 8) != 8) { f = 0.0; } else { if (!noswap) { union { double val1; quint64 val2; } x; x.val2 = qbswap(*reinterpret_cast(&f)); f = x.val1; } } return *this; } /*! \overload Reads string \a s from the stream and returns a reference to the stream. The string is deserialized using \c{readBytes()} where the serialization format is a \c quint32 length specifier first, followed by that many bytes of data. The resulting string is always '\\0'-terminated. Space for the string is allocated using \c{new []} -- the caller must destroy it with \c{delete []}. \sa readBytes(), readRawData() */ QDataStream &QDataStream::operator>>(char *&s) { qint64 len = 0; return readBytes(s, len); } /*! \overload \since 6.0 Reads a 16bit wide char from the stream into \a c and returns a reference to the stream. */ QDataStream &QDataStream::operator>>(char16_t &c) { quint16 u; *this >> u; c = char16_t(u); return *this; } /*! \overload \since 6.0 Reads a 32bit wide character from the stream into \a c and returns a reference to the stream. */ QDataStream &QDataStream::operator>>(char32_t &c) { quint32 u; *this >> u; c = char32_t(u); return *this; } #if QT_DEPRECATED_SINCE(6, 11) /* \deprecated [6.11] Use an overload that takes qint64 length instead. */ QDataStream &QDataStream::readBytes(char *&s, uint &l) { qint64 length = 0; (void)readBytes(s, length); if (length != qint64(uint(length))) { setStatus(SizeLimitExceeded); // Cannot store length in l delete[] s; l = 0; return *this; } l = uint(length); return *this; } #endif // QT_DEPRECATED_SINCE(6, 11) /*! \since 6.7 Reads the buffer \a s from the stream and returns a reference to the stream. The buffer \a s is allocated using \c{new []}. Destroy it with the \c{delete []} operator. The \a l parameter is set to the length of the buffer. If the string read is empty, \a l is set to 0 and \a s is set to \nullptr. The serialization format is a length specifier first, then \a l bytes of data. The length specifier is one quint32 if the version is less than 6.7 or if the number of elements is less than 0xfffffffe (2^32 -2), otherwise there is an extend value 0xfffffffe followed by one quint64 with the actual value. In addition for containers that support isNull(), it is encoded as a single quint32 with all bits set and no data. \sa readRawData(), writeBytes() */ QDataStream &QDataStream::readBytes(char *&s, qint64 &l) { s = nullptr; l = 0; CHECK_STREAM_PRECOND(*this) qint64 length = readQSizeType(*this); if (length == 0) return *this; qsizetype len = qsizetype(length); if (length != len || length < 0) { setStatus(SizeLimitExceeded); // Cannot store len return *this; } qsizetype step = (dev->bytesAvailable() >= len) ? len : 1024 * 1024; qsizetype allocated = 0; std::unique_ptr curBuf = nullptr; constexpr qsizetype StepIncreaseThreshold = std::numeric_limits::max() / 2; do { qsizetype blockSize = qMin(step, len - allocated); const qsizetype n = allocated + blockSize + 1; if (const auto prevBuf = std::exchange(curBuf, std::make_unique(n))) memcpy(curBuf.get(), prevBuf.get(), allocated); if (readBlock(curBuf.get() + allocated, blockSize) != blockSize) return *this; allocated += blockSize; if (step <= StepIncreaseThreshold) step *= 2; } while (allocated < len); s = curBuf.release(); s[len] = '\0'; l = len; return *this; } /*! Reads at most \a len bytes from the stream into \a s and returns the number of bytes read. If an error occurs, this function returns -1. The buffer \a s must be preallocated. The data is \e not decoded. \sa readBytes(), QIODevice::read(), writeRawData() */ qint64 QDataStream::readRawData(char *s, qint64 len) { CHECK_STREAM_PRECOND(-1) return readBlock(s, len); } /*! \fn template QDataStream &operator>>(QDataStream &in, std::pair &pair) \since 6.0 \relates QDataStream Reads a pair from stream \a in into \a pair. This function requires the T1 and T2 types to implement \c operator>>(). \sa {Serializing Qt Data Types} */ /***************************************************************************** QDataStream write functions *****************************************************************************/ /*! \fn QDataStream &QDataStream::operator<<(std::nullptr_t ptr) \since 5.9 \overload Simulates writing a \c{std::nullptr_t}, \a ptr, to the stream and returns a reference to the stream. This function does not actually write anything to the stream, as \c{std::nullptr_t} values are stored as 0 bytes. */ /*! \fn QDataStream &QDataStream::operator<<(quint8 i) \overload Writes an unsigned byte, \a i, to the stream and returns a reference to the stream. */ /*! Writes a signed byte, \a i, to the stream and returns a reference to the stream. */ QDataStream &QDataStream::operator<<(qint8 i) { CHECK_STREAM_WRITE_PRECOND(*this) if (!dev->putChar(i)) q_status = WriteFailed; return *this; } /*! \fn QDataStream &QDataStream::operator<<(quint16 i) \overload Writes an unsigned 16-bit integer, \a i, to the stream and returns a reference to the stream. */ /*! \overload Writes a signed 16-bit integer, \a i, to the stream and returns a reference to the stream. */ QDataStream &QDataStream::operator<<(qint16 i) { CHECK_STREAM_WRITE_PRECOND(*this) if (!noswap) { i = qbswap(i); } if (dev->write((char *)&i, sizeof(qint16)) != sizeof(qint16)) q_status = WriteFailed; return *this; } /*! \overload Writes a signed 32-bit integer, \a i, to the stream and returns a reference to the stream. */ QDataStream &QDataStream::operator<<(qint32 i) { CHECK_STREAM_WRITE_PRECOND(*this) if (!noswap) { i = qbswap(i); } if (dev->write((char *)&i, sizeof(qint32)) != sizeof(qint32)) q_status = WriteFailed; return *this; } /*! \fn QDataStream &QDataStream::operator<<(quint64 i) \overload Writes an unsigned 64-bit integer, \a i, to the stream and returns a reference to the stream. */ /*! \overload Writes a signed 64-bit integer, \a i, to the stream and returns a reference to the stream. */ QDataStream &QDataStream::operator<<(qint64 i) { CHECK_STREAM_WRITE_PRECOND(*this) if (version() < 6) { quint32 i1 = i & 0xffffffff; quint32 i2 = i >> 32; *this << i2 << i1; } else { if (!noswap) { i = qbswap(i); } if (dev->write((char *)&i, sizeof(qint64)) != sizeof(qint64)) q_status = WriteFailed; } return *this; } /*! \fn QDataStream &QDataStream::operator<<(quint32 i) \overload Writes an unsigned integer, \a i, to the stream as a 32-bit unsigned integer (quint32). Returns a reference to the stream. */ /*! \fn QDataStream &QDataStream::operator<<(bool i) \overload Writes a boolean value, \a i, to the stream. Returns a reference to the stream. */ /*! \overload Writes a floating point number, \a f, to the stream using the standard IEEE 754 format. Returns a reference to the stream. \sa setFloatingPointPrecision() */ QDataStream &QDataStream::operator<<(float f) { if (version() >= QDataStream::Qt_4_6 && floatingPointPrecision() == QDataStream::DoublePrecision) { *this << double(f); return *this; } CHECK_STREAM_WRITE_PRECOND(*this) float g = f; // fixes float-on-stack problem if (!noswap) { union { float val1; quint32 val2; } x; x.val1 = g; x.val2 = qbswap(x.val2); if (dev->write((char *)&x.val2, sizeof(float)) != sizeof(float)) q_status = WriteFailed; return *this; } if (dev->write((char *)&g, sizeof(float)) != sizeof(float)) q_status = WriteFailed; return *this; } /*! \overload Writes a floating point number, \a f, to the stream using the standard IEEE 754 format. Returns a reference to the stream. \sa setFloatingPointPrecision() */ QDataStream &QDataStream::operator<<(double f) { if (version() >= QDataStream::Qt_4_6 && floatingPointPrecision() == QDataStream::SinglePrecision) { *this << float(f); return *this; } CHECK_STREAM_WRITE_PRECOND(*this) if (noswap) { if (dev->write((char *)&f, sizeof(double)) != sizeof(double)) q_status = WriteFailed; } else { union { double val1; quint64 val2; } x; x.val1 = f; x.val2 = qbswap(x.val2); if (dev->write((char *)&x.val2, sizeof(double)) != sizeof(double)) q_status = WriteFailed; } return *this; } /*! \overload Writes the '\\0'-terminated string \a s to the stream and returns a reference to the stream. The string is serialized using \c{writeBytes()}. \sa writeBytes(), writeRawData() */ QDataStream &QDataStream::operator<<(const char *s) { // Include null terminator, unless s itself is null const qint64 len = s ? qint64(qstrlen(s)) + 1 : 0; writeBytes(s, len); return *this; } /*! \overload \since 6.0 Writes a character, \a c, to the stream. Returns a reference to the stream */ QDataStream &QDataStream::operator<<(char16_t c) { return *this << qint16(c); } /*! \overload \since 6.0 Writes a character, \a c, to the stream. Returns a reference to the stream */ QDataStream &QDataStream::operator<<(char32_t c) { return *this << qint32(c); } /*! Writes the length specifier \a len and the buffer \a s to the stream and returns a reference to the stream. The \a len is serialized as a quint32 and an optional quint64, followed by \a len bytes from \a s. Note that the data is \e not encoded. \sa writeRawData(), readBytes() */ QDataStream &QDataStream::writeBytes(const char *s, qint64 len) { if (len < 0) { q_status = WriteFailed; return *this; } CHECK_STREAM_WRITE_PRECOND(*this) // Write length then, if any, content if (writeQSizeType(*this, len) && len > 0) writeRawData(s, len); return *this; } /*! Writes \a len bytes from \a s to the stream. Returns the number of bytes actually written, or -1 on error. The data is \e not encoded. \sa writeBytes(), QIODevice::write(), readRawData() */ qint64 QDataStream::writeRawData(const char *s, qint64 len) { CHECK_STREAM_WRITE_PRECOND(-1) qint64 ret = dev->write(s, len); if (ret != len) q_status = WriteFailed; return ret; } /*! \since 4.1 Skips \a len bytes from the device. Returns the number of bytes actually skipped, or -1 on error. This is equivalent to calling readRawData() on a buffer of length \a len and ignoring the buffer. \sa QIODevice::seek() */ qint64 QDataStream::skipRawData(qint64 len) { CHECK_STREAM_PRECOND(-1) if (q_status != Ok && dev->isTransactionStarted()) return -1; const qint64 skipResult = dev->skip(len); if (skipResult != len) setStatus(ReadPastEnd); return skipResult; } /*! \fn template QDataStream &operator<<(QDataStream &out, const std::pair &pair) \since 6.0 \relates QDataStream Writes the pair \a pair to stream \a out. This function requires the T1 and T2 types to implement \c operator<<(). \sa {Serializing Qt Data Types} */ QT_END_NAMESPACE #endif // QT_NO_DATASTREAM