|
Serialization
Archive Class Usage
|
- Archive Classes
- Library Interface
- Saving
- Loading
- Details by Type
- Primitive Types
- Class Types
- Base Classes
- Pointers
- References
- Character Sets
An archive is defined by two complementary classes. One is for saving data while
the other is for loading it.
The system includes a number of archive implementations "ready to go" for the
most common requirements. These can be used "as is" or as a basis for developing
one's own particular type of archive. To invoke serialization using one of
these archives, one or more of the following header files must be
included in the code module containing the serialization code.
// a portable text archive
boost::archive::text_oarchive(ostream &s) // saving
boost::archive::text_iarchive(istream &s) // loading
// a portable text archive using a wide character stream
boost::archive::text_woarchive(wostream &s) // saving
boost::archive::text_wiarchive(wistream &s) // loading
// a non-portable native binary archive
boost::archive::binary_oarchive(ostream &s) // saving
boost::archive::binary_iarchive(istream &s) // loading
// a portable XML archive
boost::archive::xml_oarchive(ostream &s) // saving
boost::archive::xml_iarchive(istream &s) // loading
// a portable XML archive which uses wide characters - use for utf-8 output
boost::archive::xml_woarchive(wostream &s) // saving
boost::archive::xml_wiarchive(wistream &s) // loading
An archive contains a sequence of bytes created from
an arbitrary nested set of C++ data structures. Archives are implemented as a
hierarchy of classes. However, the interface to all archives included with
the library can best be represented by the following public interface.
namespace boost {
namespace archive {
enum archive_flags {
no_header = 1, // suppress archive header info
no_codecvt = 2, // suppress alteration of codecvt facet
no_xml_tag_checking = 4 // suppress checking of xml tags - igored on saving
};
template<class OStream>
class oarchive : ...
{
...
public:
// called to save objects
template<class T>
oarchive & operator<<(const T & t);
template<class T>
oarchive & operator&(T & t)
{
return *this << t;
}
void save_binary(const void *address, std::size_t count);
template<class T>
register_type(T * t = NULL);
unsigned int library_version() const;
struct is_saving {
typedef mpl::bool_<true> type;
BOOST_STATIC_CONSTANT(bool, value=true);
};
struct is_loading {
typedef mpl::bool_<false> type;
BOOST_STATIC_CONSTANT(bool, value=false);
};
oarchive(OStream & os, unsigned int flags = 0);
~oarchive();
};
template<class T>
oarchive & operator<<(const T & t);
template<class T>
oarchive & operator&(T & t);
-
Appends an object of type T to the archive. The object may be
- A primitive data type such as int, char, float, etc.
- A class or struct for which a
serialize
function has been defined.
- A pointer to a serializable object.
void save_binary(const void *address, std::size_t count);
-
Appends to the archive
count
bytes found at
address
.
template<class T>
register_type(T * t = NULL);
-
Appends a sequential integer to the archive. This integer becomes the "key" used
to look up the class type when the archive is later loaded. This process is
referred to as "class registration". It is only necessary to invoke this function for
objects of derived classes which are serialized through a base class pointer. This
is explained in detail in
Special Considerations - Derived Pointers
unsigned int library_version() const;
-
Returns the version number of the serialization library that created the archive.
This number will be incremented each time the library is altered in such a way
that serialization could be altered for some type. For example, suppose the type
used for a count of collection members is changed. The code that loads collections
might be conditioned on the library version to make sure that libraries created by
previous versions of the library can still be read.
is_saving::type = mpl::bool<true>;
is_saving::value= true;
is_loading::type = mpl::bool<false>;
is_loading::value= false;
-
These integral constants permit archive attributes to be queried at compiler
or execution time. They can used to generate code with boost
mpl
. For and example
showing how these can beused, see the implementation of
split_free.hpp.
oarchive(OStream & os, unsigned int flags = 0);
-
Contructs and archive given an open
stream
as
an argument and optional flags. For most applications there will be no need to use
flags. Flags are taken from the following table and can be combined with the
|
operator.By default, archives prepend
output with initial data which helps identify them as archives produced by this system.
This permits a more graceful in the case where attempt is made to load an archive
from an invalid file format. In addition to this, each type of archive might have
its own information. For example, native binary archives include information about
sizes of native types and endianess to gracefully handle the case where it has been
erroneously assumed that such an archive is portable across platforms. In some cases
where this extra overhead might be considered objectionable, it can be suppressed with the
no_header
flag.
In some cases, an archive may alter (and later restore)
the codecvt facet of the stream locale. To suppress this action,
include the no_codecvt
flag.
XML archives contain nested tags signifying the start and end of data fields.
These tags are normally checked for aggreement with the object name when
data is loaded. If a mismatch occurs an exception is thrown. Its possible
that this may not be desired behavior. To suppress this checking of XML
tags, use no_xml_tag_checking
flag.
~oarchive();
-
Destructor for an archive. This should be called before the stream is
closed. It restores any altered stream facets to thier state before the
the archive was opened.
template<class IStream>
class iarchive : ...
{
...
public:
// called to load objects
template<class T>
iarchive & operator>>(T & t);
template<class T>
iarchive & operator&(T & t)
{
return *this >> t;
}
void delete_created_pointers();
void load_binary(void *address, std::size_t count);
template<class T>
register_type(T * t = NULL);
unsigned int library_version() const;
struct is_saving {
typedef mpl::bool_<false> type;
BOOST_STATIC_CONSTANT(bool, value=false);
};
struct is_loading {
typedef mpl::bool_<true> type;
BOOST_STATIC_CONSTANT(bool, value=true);
};
iarchive(IStream & is, unsigned int flags = 0);
~iarchive();
};
} //namespace archive
) //namespace boost
template<class T>
iarchive & operator>>(T & t);
template<class T>
iarchive & operator&(T & t);
-
Retrieves an object of type T from the archive. The object may be
- A primitive data type such as int, char, float, etc.
- A class or struct for which a
serialize
function has been defined.
- A pointer to a serializable object.
void load_binary(void *address, std::size_t count);
-
Retrieves from the archive
count
bytes and stores
them in memory starting at address
.
void delete_created_pointers();
-
Deletes all objects created by the loading of pointers. This can be used to
avoid memory leaks that might otherwise occur if pointers are being loaded
and the archive load encounters an exception.
template<class T>
register_type(T * t = NULL);
-
Retrieves the next integer from the archive and adds an entry to a table which
relates the integer to the type T. When pointers are loaded, this integer is
used to indicate which object type should be created. This process is
referred to as "class registration". It is only necessary to invoke this function for
objects of derived classes which are serialized through a base class pointer. If this
function is called during the saving of data to the archive, it should be called during the
loading of the data from the archive at the same point in the serialization process.
This is explained in detail in
Special Considerations - Derived Pointers
unsigned int library_version() const;
-
Returns the version number of the serialization library that created the archive.
This number will be incremented each time the library is altered in such a way
that serialization could be altered for some type. For example, suppose the type
used for a count of collection members is changed. The code that loads collections
might be conditioned on the library version to make sure that libraries created by
previous versions of the library can still be read.
is_saving::type = mpl::bool<false>;
is_saving::value= false;
is_loading::type = mpl::bool<true>;
is_loading::value= true;
-
These integral constants permit archive attributes to be queried at compiler
or execution time. They can used to generate code with boost
mpl
. For and example
showing how these can beused, see the implementation of
split_free.hpp.
iarchive(IStream & is, unsigned int flags = 0);
-
Contructs an archive given an open
stream
as
an argument and optional flags. If flags are used, they should be the same
as those used when the archive was created. Function and usage of flags is described
above.
~iarchive();
-
Destructor for an archive. This should be called before the stream is
closed. It restores any altered stream facets to thier state before the
the archive was opened.
There are archives based on text, binary and XML file
formats but all have the above interface. Given that all archives present
the same public interface, specifcation of serialization is exactly the same
for all archives. Archive classes have other members not mentioned here.
However they are related to the internal functioning of the library and
are not meant to be called by users of an archive. Implementation of new
archives is discussed in
New Archives - Implementation.
The existence of the <<
and >>
suggest
a relationship between archives and C++ i/o streams. Archives are not
C++ i/o streams. All the archives included with this system take a stream
as an argument in the constructor and that stream is used for output or input.
However, this is not a requirement of the serialization functions or the
archive interface. It just turns out that the archives written so far have
found it useful to base their implementation on streams.
In this document, we use the term primitive type to mean
types whose data is simply saved/loaded to/from an archive
with no further processing. By default, arithmetic (including characters),
bool, and enum types are primitive types. Using
serialization traits,
any user type can also be designated as "primitive"
so that it is handled in this way.
The template operators &, <<, and >> of the archive classes
described above will generate code to save/load all primitive types
to/from an archive. This code will usually just add the
data to the archive according to the archive format.
For example, a four byte integer is appended to a binary archive
as 4 binary bytes while a to a text archive it would be
rendered as a space followed by a string representation.
For class/struct types, the template operators &, <<, and >>
will generate code that invokes the programmer's serialization code for the
particular data type. There is no default. An attempt to serialize a
class/struct for which no serialization has been explicitly specified
will result in a compile time error. Specification of serialization
for user defined types is explained in detail in the next section
of this manual.
The header file
base_object.hpp
includes the template:
template<class Base, class Derived>
Base & base_object(Derived &d);
which should be used to create a reference to an object of the base
which can be used as an argument to the archive save/load operators:
ar & boost::serialization::base_object<Base>(*this);
Resist the temptation to just cast *this
to the base class.
This might seem to work but may fail to invoke code necessary for
proper serialization.
A pointer to any class instance can be serialized with any of the archive
save/load operators.
To properly save and restore an object through a pointer the
following situations must be addressed:
- If the same object is saved multiple times through different
pointers, only one copy of the object need be saved.
- If an object is loaded multiple times through different pointers,
only one new object should be created and all returned pointers
should point to it.
- The system must detect the case where an object is first
saved through a pointer then the object itself is saved.
Without taking extra precautions, loading would result in the
creation of multiple copies of the original object. This system detects
this case when saving and throws an exception - see below.
- An object of a derived class may be stored through a
pointer to the base class. The true type of the object must
be determined and saved. Upon restoration the correct type
must be created and its address correctly cast to the base
class. That is, polymorphic pointers have to be considered.
- NULL pointers must be dectected when saved and restored
to NULL when deserialized.
This serialization library addresses all of the above
considerations. The process of saving and loading an object
through a pointer is non-trivial. It can be summarized as
follows:
Saving a pointer:
- determine the true type of the object being pointed to.
- write a special tag to the archive
- if the object pointed to has not already been written
to the archive, do so now
Loading a pointer:
- read a tag from the archive.
- determine the type of object to be created
- if the object has already been loaded, return it's address.
- otherwise, create a new instance of the object
- read the data back in using the operators described above
- return the address of the newly created object.
Given that class instances are saved/loaded to/from the archive
only once, regardless of how many times they are serialized with
the <<
and >>
operators
- Loading the same pointer object multiple times
results in only one object being created, thereby replicating
the original pointer configuration.
- Structures such as collections of polymorphic pointers,
are handled with no special effort on the part of users of this library.
Serialization of pointers of derived types through a pointer to the
base class may require a little extra "help". Also, the programmer
may desire to modify the process described above for his own reasons.
For example, it might be desired to suppress the tracking of objects
as it is known a priori that the application in question can never
create duplicate objects. Serialization of pointers can be "fine tuned"
via the specification of Class Serialization Traits
as described in
another section of this manual
In general, references are serialized just as any other objects are.
However, references have the property that several references may
refer to the same object - much like pointers. So, there exists
the opportunity to gain some storage efficiency by storing only
one copy. This subject is addressed the chapter Class Serialization Traits -
Object Tracking.
This library includes two archive classes for XML. The wide character
version (xml_w?archive
) renders it output as UTF-8 which can
handle any wide character without loss of information.
std::string
data is converted from multi-byte format to wide
character format using the current
locale
. Hence this version should give a fair rendering of all
C++ data for all cases. This could result in some unexpected behavior.
Suppose an std::string
is created with the locale
character
set to hebrew characters. On output this is converted to wide characters.
On input however, there could be a problem if the locale
is
not set the same as when the archive is created.
The normal character version (xml_?archive
) renders
std::string
output without any conversion. Though this may work
fine for serialization, it may create difficulties if the XML archive is used
for some other purpose.
© Copyright Robert Ramey 2002-2004.
Distributed under the Boost Software License, Version 1.0. (See
accompanying file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)