diff options
Diffstat (limited to 'examples/corelib/serialization/convert/doc/src/convert.qdoc')
| -rw-r--r-- | examples/corelib/serialization/convert/doc/src/convert.qdoc | 156 |
1 files changed, 156 insertions, 0 deletions
diff --git a/examples/corelib/serialization/convert/doc/src/convert.qdoc b/examples/corelib/serialization/convert/doc/src/convert.qdoc new file mode 100644 index 0000000000..187e81a85e --- /dev/null +++ b/examples/corelib/serialization/convert/doc/src/convert.qdoc @@ -0,0 +1,156 @@ +// Copyright (C) 2022 The Qt Company Ltd. +// SPDX-License-Identifier: LicenseRef-Qt-Commercial OR GFDL-1.3-no-invariants-only + +/*! + \example serialization/convert + \examplecategory {Data Processing & I/O} + \meta tag {network} + \title Serialization Converter + + \brief How to convert between different serialization formats. + + This example converts between JSON, CBOR, XML, QDataStream and some simple + text formats. It can auto-detect the format being used, or be told which + format to use. Not all formats support both input and output, and they have + different sets of which content datatypes they support. QDataStream and XML + are the richest, followed by CBOR, then JSON, and then the plain text + formats. Conversion via the less capable formats is apt to lose structure + from the data. + + \image convert.png + + \sa {Parsing and displaying CBOR data}, {Saving and Loading a Game} + + \section1 The Converter Class + + The Converter class is the abstract superclass for all the converters to and + from all the formats. They all convert from or to the QVariant class, which + is used to represent all the datastructures internally. + + \snippet serialization/convert/converter.h 0 + + The Converter constructor and destructor manage a list of available + converters used by the main program so that it knows what converters are + available. Each converter type defines a static instance that ensures it is + constructed and thus available to the main program via this list. The \c + allConverters() method provides \c main()'s code with access to the list. + + \snippet serialization/convert/converter.cpp 0 + + The name() function returns the name of the converter. The directions() + function is used to determine if a converter can be used for input, output, + or both. These enable the main program to report what converters are + available in its help text for the command-line options to select input and + output formats. + + \snippet serialization/convert/main.cpp 0 + + The optionsHelp() function is used to report the various command-line + options supported by the available formats, when queried using its \c + {--format-options <format>} command-line option. + + \snippet serialization/convert/main.cpp 1 + + The outputOptions() function reports the output capabilities of a converter. + At present the only optional feature is support for arbitrary keys in + mappings from keys to values. An input converter's loadFile() can use this + information to tailor the form in which it presents the data it has read, to + be as faithfully represented by the output converter as its capabilities + permit. + + The probeFile() function is used to determine if a file matches the format + of the converter. The main program uses this to determine what format to use + when reading or writing a file, based on its name and potentially content, + when the user has not specified the format to use on the command-line. + + The loadFile() function deserializes data. The caller tells loadFile() which + serializer it intends to use, so that loadFile() can query its + outputOptions() to determine the form in which to represent the loaded data. + If the caller hasn't settled on a choice of output converter, loadFile() + supplies it with a default output converter suitable to the data it is + returning. + + The saveFile() function serializes data. It is passed options from the + command-line, as described by loadHelp(), that can tune the details of how + it represents the data when saving to file. + + Both loadFile() and saveFile() can be used with an arbitrary \l QIODevice. + This means that a Converter could also be used with a network socket or + other source of data, to read from or write to. In the present program, the + main program always passes a \l QFile, accessing either a file on disk or + one of the standard streams of the process. + + \section2 The Available Converters + + Several converters are supported, illustrating how the converter program + could be adapted to other formats, should the need arise. See the source + code for each for its details. The CBOR converters serve as a relatively + full-featured illustration of the ways converters can work, that we'll look + into in more detail below. This table summarizes the available converters: + + \table + \header \li Class \li mode \li format + \row \li CborConverter \li In/Out \li CBOR + \row \li CborDiagnosticDumper \li Out \li CBOR diagnostic + \row \li DataStreamConverter \li In/Out \li QDataStream + \row \li DebugTextDumper \li Out \li Lossless, non-standard, human-readable + \row \li JsonConverter \li In/Out \li JSON + \row \li NullConverter \li Out \li No output + \row \li TextConverter \li In/Out \li Structured plain text + \row \li XmlConverter \li In/Out \li XML + \endtable + + Those that support input use themselves as loadFile()'s fallback converter, + except for the CBOR and QDataStream converters, which use their respective + output-only dumper companion classes. The null converter can be used as + output converter when running the program for the sake of any validation or + verification that an input converter may perform. + + \section2 The CborConverter and CborDiagnosticDumper Classes + + The CborConverter class supports serializing to and from the CBOR format. + It supports various options to configure the output of floating point values + and a \c{signature} option to determine whether to start its output with a + CBOR tag that serves as a file header, identifying the file as containing + CBOR data. + + There is also a CborDiagnosticDumper class to output in CBOR diagnostic + notation. It does not support loading data. The form of its output can be + configured using two options. One selects whether to use the (more verbose) + extended CBOR diagnostic format. The other control whether each CBOR value + appears on a separate line. + + The plain diagnostic notation is similar to JSON, but not exactly, because + it supports displaying the contents of a CBOR stream losslessly, while a + conversion to JSON can be lossy. CborConverter's loadFile() uses + CborDiagnosticDumper for the fallback output converter, if its caller hasn't + determined the output format for itself. + + The convertCborValue(), convertCborMap() and convertCborArray() helper + functions are used to convert a QCborValue to a QVariant, for the benefit of + CborConverter::loadFile(). + + \snippet serialization/convert/cborconverter.cpp 0 + + The convertFromVariant() function is used to convert a QVariant to a + QCborValue for output by the \c saveFile() of either class. + + \snippet serialization/convert/cborconverter.cpp 1 + + \sa {CBOR Support in Qt} + + \section1 The convert program + + The \c main() function sets up a \l QApplication and a \l QCommandLineParser + to make sense of the options the user has specified and provide help if the + user asks for it. It uses the values obtained for the various \l + QCommandLineOption instances describing the user's choices, plus the + positional arguments for file names, to prepare the converters it will use. + + It then uses its input converter to load data (and possibly resolve its + choice of output converter, if it hasn't selected one yet) and its output + converter to serialize that data, taking account of any output options the + user has supplied on the command-line. + + \snippet serialization/convert/main.cpp 2 +*/ |