Performs a byteswapping operation on the contents of value
. The
result is that all multi-byte numeric data contained in value
is
byteswapped. That includes 16, 32, and 64bit signed and unsigned
integers as well as file handles and double precision floating point
values.
This function is an identity mapping on any value that does not contain multi-byte numeric data. That include strings, booleans, bytes and containers containing only these things (recursively).
The returned value is always in normal form and is marked as trusted.
Checks if calling g_variant_get() with format_string
on value
would
be valid from a type-compatibility standpoint. format_string
is
assumed to be a valid format string (from a syntactic standpoint).
If copy_only
is %TRUE then this function additionally checks that it
would be safe to call g_variant_unref() on value
immediately after
the call to g_variant_get() without invalidating the result. This is
only possible if deep copies are made (ie: there are no pointers to
the data inside of the soon-to-be-freed #GVariant instance). If this
check fails then a g_critical() is printed and %FALSE is returned.
This function is meant to be used by functions that wish to provide varargs accessors to #GVariant values of uncertain values (eg: g_variant_lookup() or g_menu_model_get_item_attribute()).
a valid #GVariant format string
%TRUE to ensure the format string makes deep copies
Classifies value
according to its top-level type.
Compares one
and two
.
The types of one
and two
are #gconstpointer only to allow use of
this function with #GTree, #GPtrArray, etc. They must each be a
#GVariant.
Comparison is only defined for basic types (ie: booleans, numbers, strings). For booleans, %FALSE is less than %TRUE. Numbers are ordered in the usual way. Strings are in ASCII lexographical order.
It is a programmer error to attempt to compare container values or two values that have types that are not exactly equal. For example, you cannot compare a 32-bit signed integer with a 32-bit unsigned integer. Also note that this function is not particularly well-behaved when it comes to comparison of doubles; in particular, the handling of incomparable values (ie: NaN) is undefined.
If you only require an equality comparison, g_variant_equal() is more general.
Similar to g_variant_get_bytestring() except that instead of returning a constant string, the string is duplicated.
The return value must be freed using g_free().
Gets the contents of an array of array of bytes #GVariant. This call makes a deep copy; the return result should be released with g_strfreev().
If length
is non-%NULL then the number of elements in the result is
stored there. In any case, the resulting array will be
%NULL-terminated.
For an empty array, length
will be set to 0 and a pointer to a
%NULL pointer will be returned.
Gets the contents of an array of object paths #GVariant. This call makes a deep copy; the return result should be released with g_strfreev().
If length
is non-%NULL then the number of elements in the result
is stored there. In any case, the resulting array will be
%NULL-terminated.
For an empty array, length
will be set to 0 and a pointer to a
%NULL pointer will be returned.
Similar to g_variant_get_string() except that instead of returning a constant string, the string is duplicated.
The string will always be UTF-8 encoded.
The return value must be freed using g_free().
Gets the contents of an array of strings #GVariant. This call makes a deep copy; the return result should be released with g_strfreev().
If length
is non-%NULL then the number of elements in the result
is stored there. In any case, the resulting array will be
%NULL-terminated.
For an empty array, length
will be set to 0 and a pointer to a
%NULL pointer will be returned.
Returns the boolean value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_BOOLEAN.
Returns the byte value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_BYTE.
Returns the string value of a #GVariant instance with an array-of-bytes type. The string has no particular encoding.
If the array does not end with a nul terminator character, the empty string is returned. For this reason, you can always trust that a non-%NULL nul-terminated string will be returned by this function.
If the array contains a nul terminator character somewhere other than the last byte then the returned string is the string, up to the first such nul character.
g_variant_get_fixed_array() should be used instead if the array contains arbitrary data that could not be nul-terminated or could contain nul bytes.
It is an error to call this function with a value
that is not an
array of bytes.
The return value remains valid as long as value
exists.
Gets the contents of an array of array of bytes #GVariant. This call makes a shallow copy; the return result should be released with g_free(), but the individual strings must not be modified.
If length
is non-%NULL then the number of elements in the result is
stored there. In any case, the resulting array will be
%NULL-terminated.
For an empty array, length
will be set to 0 and a pointer to a
%NULL pointer will be returned.
Reads a child item out of a container #GVariant instance. This includes variants, maybes, arrays, tuples and dictionary entries. It is an error to call this function on any other type of #GVariant.
It is an error if index_
is greater than the number of child items
in the container. See g_variant_n_children().
The returned value is never floating. You should free it with g_variant_unref() when you're done with it.
Note that values borrowed from the returned child are not guaranteed to
still be valid after the child is freed even if you still hold a reference
to value,
if value
has not been serialized at the time this function is
called. To avoid this, you can serialize value
by calling
g_variant_get_data() and optionally ignoring the return value.
There may be implementation specific restrictions on deeply nested values, which would result in the unit tuple being returned as the child value, instead of further nested children. #GVariant is guaranteed to handle nesting up to at least 64 levels.
This function is O(1).
the index of the child to fetch
Returns a pointer to the serialized form of a #GVariant instance.
The returned data may not be in fully-normalised form if read from an
untrusted source. The returned data must not be freed; it remains
valid for as long as value
exists.
If value
is a fixed-sized value that was deserialized from a
corrupted serialized container then %NULL may be returned. In this
case, the proper thing to do is typically to use the appropriate
number of nul bytes in place of value
. If value
is not fixed-sized
then %NULL is never returned.
In the case that value
is already in serialized form, this function
is O(1). If the value is not already in serialized form,
serialization occurs implicitly and is approximately O(n) in the size
of the result.
To deserialize the data returned by this function, in addition to the serialized data, you must know the type of the #GVariant, and (if the machine might be different) the endianness of the machine that stored it. As a result, file formats or network messages that incorporate serialized #GVariants must include this information either implicitly (for instance "the file always contains a %G_VARIANT_TYPE_VARIANT and it is always in little-endian order") or explicitly (by storing the type and/or endianness in addition to the serialized data).
Returns a pointer to the serialized form of a #GVariant instance. The semantics of this function are exactly the same as g_variant_get_data(), except that the returned #GBytes holds a reference to the variant data.
Returns the double precision floating point value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_DOUBLE.
Returns the 32-bit signed integer value of value
.
It is an error to call this function with a value
of any type other
than %G_VARIANT_TYPE_HANDLE.
By convention, handles are indexes into an array of file descriptors that are sent alongside a D-Bus message. If you're not interacting with D-Bus, you probably don't need them.
Returns the 16-bit signed integer value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_INT16.
Returns the 32-bit signed integer value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_INT32.
Returns the 64-bit signed integer value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_INT64.
Gets a #GVariant instance that has the same value as value
and is
trusted to be in normal form.
If value
is already trusted to be in normal form then a new
reference to value
is returned.
If value
is not already trusted, then it is scanned to check if it
is in normal form. If it is found to be in normal form then it is
marked as trusted and a new reference to it is returned.
If value
is found not to be in normal form then a new trusted
#GVariant is created with the same value as value
.
It makes sense to call this function if you've received #GVariant data from untrusted sources and you want to ensure your serialized output is definitely in normal form.
If value
is already in normal form, a new reference will be returned
(which will be floating if value
is floating). If it is not in normal form,
the newly created #GVariant will be returned with a single non-floating
reference. Typically, g_variant_take_ref() should be called on the return
value from this function to guarantee ownership of a single non-floating
reference to it.
Gets the contents of an array of object paths #GVariant. This call makes a shallow copy; the return result should be released with g_free(), but the individual strings must not be modified.
If length
is non-%NULL then the number of elements in the result
is stored there. In any case, the resulting array will be
%NULL-terminated.
For an empty array, length
will be set to 0 and a pointer to a
%NULL pointer will be returned.
Determines the number of bytes that would be required to store value
with g_variant_store().
If value
has a fixed-sized type then this function always returned
that fixed size.
In the case that value
is already in serialized form or the size has
already been calculated (ie: this function has been called before)
then this function is O(1). Otherwise, the size is calculated, an
operation which is approximately O(n) in the number of values
involved.
Returns the string value of a #GVariant instance with a string type. This includes the types %G_VARIANT_TYPE_STRING, %G_VARIANT_TYPE_OBJECT_PATH and %G_VARIANT_TYPE_SIGNATURE.
The string will always be UTF-8 encoded, will never be %NULL, and will never contain nul bytes.
If length
is non-%NULL then the length of the string (in bytes) is
returned there. For trusted values, this information is already
known. Untrusted values will be validated and, if valid, a strlen() will be
performed. If invalid, a default value will be returned — for
%G_VARIANT_TYPE_OBJECT_PATH, this is "/"
, and for other types it is the
empty string.
It is an error to call this function with a value
of any type
other than those three.
The return value remains valid as long as value
exists.
Gets the contents of an array of strings #GVariant. This call makes a shallow copy; the return result should be released with g_free(), but the individual strings must not be modified.
If length
is non-%NULL then the number of elements in the result
is stored there. In any case, the resulting array will be
%NULL-terminated.
For an empty array, length
will be set to 0 and a pointer to a
%NULL pointer will be returned.
Determines the type of value
.
The return value is valid for the lifetime of value
and must not
be freed.
Returns the type string of value
. Unlike the result of calling
g_variant_type_peek_string(), this string is nul-terminated. This
string belongs to #GVariant and must not be freed.
Returns the 16-bit unsigned integer value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_UINT16.
Returns the 32-bit unsigned integer value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_UINT32.
Returns the 64-bit unsigned integer value of value
.
It is an error to call this function with a value
of any type
other than %G_VARIANT_TYPE_UINT64.
Generates a hash value for a #GVariant instance.
The output of this function is guaranteed to be the same for a given value only per-process. It may change between different processor architectures or even different versions of GLib. Do not use this function as a basis for building protocols or file formats.
The type of value
is #gconstpointer only to allow use of this
function with #GHashTable. value
must be a #GVariant.
Checks if value
is a container.
Checks whether value
has a floating reference count.
This function should only ever be used to assert that a given variant is or is not floating, or for debug purposes. To acquire a reference to a variant that might be floating, always use g_variant_ref_sink() or g_variant_take_ref().
See g_variant_ref_sink() for more information about floating reference counts.
Checks if value
is in normal form.
The main reason to do this is to detect if a given chunk of serialized data is in normal form: load the data into a #GVariant using g_variant_new_from_data() and then use this function to check.
If value
is found to be in normal form then it will be marked as
being trusted. If the value was already marked as being trusted then
this function will immediately return %TRUE.
There may be implementation specific restrictions on deeply nested values. GVariant is guaranteed to handle nesting up to at least 64 levels.
Checks if a value has a type matching the provided type.
a #GVariantType
Looks up a value in a dictionary #GVariant.
This function works with dictionaries of the type a{s*} (and equally well with type a{o*}, but we only further discuss the string case for sake of clarity).
In the event that dictionary
has the type a{sv}, the expected_type
string specifies what type of value is expected to be inside of the
variant. If the value inside the variant has a different type then
%NULL is returned. In the event that dictionary
has a value type other
than v then expected_type
must directly match the value type and it is
used to unpack the value directly or an error occurs.
In either case, if key
is not found in dictionary,
%NULL is returned.
If the key is found and the value has the correct type, it is
returned. If expected_type
was specified then any non-%NULL return
value will have this type.
This function is currently implemented with a linear scan. If you plan to do many lookups then #GVariantDict may be more efficient.
the key to look up in the dictionary
a #GVariantType, or %NULL
Determines the number of children in a container #GVariant instance. This includes variants, maybes, arrays, tuples and dictionary entries. It is an error to call this function on any other type of #GVariant.
For variants, the return value is always 1. For values with maybe types, it is always zero or one. For arrays, it is the length of the array. For tuples it is the number of tuple items (which depends only on the type). For dictionary entries, it is always 2
This function is O(1).
Pretty-prints value
in the format understood by g_variant_parse().
The format is described [here][gvariant-text].
If type_annotate
is %TRUE, then type information is included in
the output.
%TRUE if type information should be included in the output
#GVariant uses a floating reference count system. All functions with
names starting with g_variant_new_
return floating
references.
Calling g_variant_ref_sink() on a #GVariant with a floating reference will convert the floating reference into a full reference. Calling g_variant_ref_sink() on a non-floating #GVariant results in an additional normal reference being added.
In other words, if the value
is floating, then this call "assumes
ownership" of the floating reference, converting it to a normal
reference. If the value
is not floating, then this call adds a
new normal reference increasing the reference count by one.
All calls that result in a #GVariant instance being inserted into a container will call g_variant_ref_sink() on the instance. This means that if the value was just created (and has only its floating reference) then the container will assume sole ownership of the value at that point and the caller will not need to unreference it. This makes certain common styles of programming much easier while still maintaining normal refcounting semantics in situations where values are not floating.
Stores the serialized form of value
at data
. data
should be
large enough. See g_variant_get_size().
The stored data is in machine native byte order but may not be in fully-normalised form if read from an untrusted source. See g_variant_get_normal_form() for a solution.
As with g_variant_get_data(), to be able to deserialize the serialized variant successfully, its type and (if the destination machine might be different) its endianness must also be available.
This function is approximately O(n) in the size of data
.
the location to store the serialized data at
If value
is floating, sink it. Otherwise, do nothing.
Typically you want to use g_variant_ref_sink() in order to automatically do the correct thing with respect to floating or non-floating references, but there is one specific scenario where this function is helpful.
The situation where this function is helpful is when creating an API that allows the user to provide a callback function that returns a #GVariant. We certainly want to allow the user the flexibility to return a non-floating reference from this callback (for the case where the value that is being returned already exists).
At the same time, the style of the #GVariant API makes it likely that for newly-created #GVariant instances, the user can be saved some typing if they are allowed to return a #GVariant with a floating reference.
Using this function on the return value of the user's callback allows the user to do whichever is more convenient for them. The caller will always receives exactly one full reference to the value: either the one that was returned in the first place, or a floating reference that has been converted to a full reference.
This function has an odd interaction when combined with g_variant_ref_sink() running at the same time in another thread on the same #GVariant instance. If g_variant_ref_sink() runs first then the result will be that the floating reference is converted to a hard reference. If g_variant_take_ref() runs first then the result will be that the floating reference is converted to a hard reference and an additional reference on top of that one is added. It is best to avoid this situation.
Decreases the reference count of value
. When its reference count
drops to 0, the memory used by the variant is freed.
Determines if a given string is a valid D-Bus object path. You should ensure that a string is a valid D-Bus object path before passing it to g_variant_new_object_path().
A valid object path starts with /
followed by zero or more
sequences of characters separated by /
characters. Each sequence
must contain only the characters [A-Z][a-z][0-9]_
. No sequence
(including the one following the final /
character) may be empty.
a normal C nul-terminated string
Determines if a given string is a valid D-Bus type signature. You should ensure that a string is a valid D-Bus type signature before passing it to g_variant_new_signature().
D-Bus type signatures consist of zero or more definite #GVariantType strings in sequence.
a normal C nul-terminated string
Creates a new #GVariant array from children
.
child_type
must be non-%NULL if n_children
is zero. Otherwise, the
child type is determined by inspecting the first element of the
children
array. If child_type
is non-%NULL then it must be a
definite type.
The items of the array are taken from the children
array. No entry
in the children
array may be %NULL.
All items in the array must have the same type, which must be the
same as child_type,
if given.
If the children
are floating references (see g_variant_ref_sink()), the
new instance takes ownership of them as if via g_variant_ref_sink().
the element type of the new array
an array of #GVariant pointers, the children
Creates an array-of-bytes #GVariant with the contents of string
.
This function is just like g_variant_new_string() except that the
string need not be valid UTF-8.
The nul terminator character at the end of the string is stored in the array.
a normal nul-terminated string in no particular encoding
Creates a new dictionary entry #GVariant. key
and value
must be
non-%NULL. key
must be a value of a basic type (ie: not a container).
If the key
or value
are floating references (see g_variant_ref_sink()),
the new instance takes ownership of them as if via g_variant_ref_sink().
Constructs a new array #GVariant instance, where the elements are
of element_type
type.
elements
must be an array with fixed-sized elements. Numeric types are
fixed-size as are tuples containing only other fixed-sized types.
element_size
must be the size of a single element in the array.
For example, if calling this function for an array of 32-bit integers,
you might say sizeof(gint32). This value isn't used except for the purpose
of a double-check that the form of the serialized data matches the caller's
expectation.
n_elements
must be the length of the elements
array.
the #GVariantType of each element
a pointer to the fixed array of contiguous elements
the number of elements
the size of each element
Constructs a new serialized-mode #GVariant instance. This is the inner interface for creation of new serialized values that gets called from various functions in gvariant.c.
A reference is taken on bytes
.
The data in bytes
must be aligned appropriately for the type
being loaded.
Otherwise this function will internally create a copy of the memory (since
GLib 2.60) or (in older versions) fail and exit the process.
a #GVariantType
a #GBytes
if the contents of bytes
are trusted
Creates a new #GVariant instance from serialized data.
type
is the type of #GVariant instance that will be constructed.
The interpretation of data
depends on knowing the type.
data
is not modified by this function and must remain valid with an
unchanging value until such a time as notify
is called with
user_data
. If the contents of data
change before that time then
the result is undefined.
If data
is trusted to be serialized data in normal form then
trusted
should be %TRUE. This applies to serialized data created
within this process or read from a trusted location on the disk (such
as a file installed in /usr/lib alongside your application). You
should set trusted to %FALSE if data
is read from the network, a
file in the user's home directory, etc.
If data
was not stored in this machine's native endianness, any multi-byte
numeric values in the returned variant will also be in non-native
endianness. g_variant_byteswap() can be used to recover the original values.
notify
will be called with user_data
when data
is no longer
needed. The exact time of this call is unspecified and might even be
before this function returns.
Note: data
must be backed by memory that is aligned appropriately for the
type
being loaded. Otherwise this function will internally create a copy of
the memory (since GLib 2.60) or (in older versions) fail and exit the
process.
a definite #GVariantType
the serialized data
%TRUE if data
is definitely in normal form
function to call when data
is no longer needed
data for notify
Depending on if child
is %NULL, either wraps child
inside of a
maybe container or creates a Nothing instance for the given type
.
At least one of child_type
and child
must be non-%NULL.
If child_type
is non-%NULL then it must be a definite type.
If they are both non-%NULL then child_type
must be the type
of child
.
If child
is a floating reference (see g_variant_ref_sink()), the new
instance takes ownership of child
.
the #GVariantType of the child, or %NULL
the child value, or %NULL
Creates a string #GVariant with the contents of string
.
string
must be valid UTF-8, and must not be %NULL. To encode
potentially-%NULL strings, use g_variant_new() with ms
as the
[format string][gvariant-format-strings-maybe-types].
a normal UTF-8 nul-terminated string
Creates a new tuple #GVariant out of the items in children
. The
type is determined from the types of children
. No entry in the
children
array may be %NULL.
If n_children
is 0 then the unit tuple is constructed.
If the children
are floating references (see g_variant_ref_sink()), the
new instance takes ownership of them as if via g_variant_ref_sink().
Parses a #GVariant from a text representation.
A single #GVariant is parsed from the content of text
.
The format is described [here][gvariant-text].
The memory at limit
will never be accessed and the parser behaves as
if the character at limit
is the nul terminator. This has the
effect of bounding text
.
If endptr
is non-%NULL then text
is permitted to contain data
following the value that this function parses and endptr
will be
updated to point to the first character past the end of the text
parsed by this function. If endptr
is %NULL and there is extra data
then an error is returned.
If type
is non-%NULL then the value will be parsed to have that
type. This may result in additional parse errors (in the case that
the parsed value doesn't fit the type) but may also result in fewer
errors (in the case that the type would have been ambiguous, such as
with empty arrays).
In the event that the parsing is successful, the resulting #GVariant is returned. It is never floating, and must be freed with g_variant_unref().
In case of any error, %NULL will be returned. If error
is non-%NULL
then it will be set to reflect the error that occurred.
Officially, the language understood by the parser is "any string produced by g_variant_print()".
There may be implementation specific restrictions on deeply nested values, which would result in a %G_VARIANT_PARSE_ERROR_RECURSION error. #GVariant is guaranteed to handle nesting up to at least 64 levels.
a #GVariantType, or %NULL
a string containing a GVariant in text form
a pointer to the end of text,
or %NULL
a location to store the end pointer, or %NULL
Pretty-prints a message showing the context of a #GVariant parse error within the string for which parsing was attempted.
The resulting string is suitable for output to the console or other monospace media where newlines are treated in the usual way.
The message will typically look something like one of the following:
|[ unterminated string constant: (1, 2, 3, 'abc ^^^^
or
|[
unable to find a common type:
[1, 2, 3, 'str']
^ ^^^^^
The format of the message may change in a future version.
error
must have come from a failed attempt to g_variant_parse() and
source_str
must be exactly the same string that caused the error.
If source_str
was not nul-terminated when you passed it to
g_variant_parse() then you must add nul termination before using this
function.
a #GError from the #GVariantParseError domain
the string that was given to the parser
Same as g_variant_error_quark().
#GVariant is a variant datatype; it can contain one or more values along with information about the type of the values.
A #GVariant may contain simple types, like an integer, or a boolean value; or complex types, like an array of two strings, or a dictionary of key value pairs. A #GVariant is also immutable: once it's been created neither its type nor its content can be modified further.
GVariant is useful whenever data needs to be serialized, for example when sending method parameters in D-Bus, or when saving settings using GSettings.
When creating a new #GVariant, you pass the data you want to store in it along with a string representing the type of data you wish to pass to it.
For instance, if you want to create a #GVariant holding an integer value you can use:
The string "u" in the first argument tells #GVariant that the data passed to the constructor (40) is going to be an unsigned integer.
More advanced examples of #GVariant in use can be found in documentation for [GVariant format strings][gvariant-format-strings-pointers].
The range of possible values is determined by the type.
The type system used by #GVariant is #GVariantType.
#GVariant instances always have a type and a value (which are given at construction time). The type and value of a #GVariant instance can never change other than by the #GVariant itself being destroyed. A #GVariant cannot contain a pointer.
#GVariant is reference counted using g_variant_ref() and g_variant_unref(). #GVariant also has floating reference counts -- see g_variant_ref_sink().
#GVariant is completely threadsafe. A #GVariant instance can be concurrently accessed in any way from any number of threads without problems.
#GVariant is heavily optimised for dealing with data in serialized form. It works particularly well with data located in memory-mapped files. It can perform nearly all deserialization operations in a small constant time, usually touching only a single memory page. Serialized #GVariant data can also be sent over the network.
#GVariant is largely compatible with D-Bus. Almost all types of #GVariant instances can be sent over D-Bus. See #GVariantType for exceptions. (However, #GVariant's serialization format is not the same as the serialization format of a D-Bus message body: use #GDBusMessage, in the gio library, for those.)
For space-efficiency, the #GVariant serialization format does not automatically include the variant's length, type or endianness, which must either be implied from context (such as knowledge that a particular file format always contains a little-endian %G_VARIANT_TYPE_VARIANT which occupies the whole length of the file) or supplied out-of-band (for instance, a length, type and/or endianness indicator could be placed at the beginning of a file, network message or network stream).
A #GVariant's size is limited mainly by any lower level operating system constraints, such as the number of bits in #gsize. For example, it is reasonable to have a 2GB file mapped into memory with #GMappedFile, and call g_variant_new_from_data() on it.
For convenience to C programmers, #GVariant features powerful varargs-based value construction and destruction. This feature is designed to be embedded in other libraries.
There is a Python-inspired text language for describing #GVariant values. #GVariant includes a printer for this language and a parser with type inferencing.
Memory Use
#GVariant tries to be quite efficient with respect to memory use. This section gives a rough idea of how much memory is used by the current implementation. The information here is subject to change in the future.
The memory allocated by #GVariant can be grouped into 4 broad purposes: memory for serialized data, memory for the type information cache, buffer management memory and memory for the #GVariant structure itself.
Serialized Data Memory
This is the memory that is used for storing GVariant data in serialized form. This is what would be sent over the network or what would end up on disk, not counting any indicator of the endianness, or of the length or type of the top-level variant.
The amount of memory required to store a boolean is 1 byte. 16, 32 and 64 bit integers and double precision floating point numbers use their "natural" size. Strings (including object path and signature strings) are stored with a nul terminator, and as such use the length of the string plus 1 byte.
Maybe types use no space at all to represent the null value and use the same amount of space (sometimes plus one byte) as the equivalent non-maybe-typed value to represent the non-null case.
Arrays use the amount of space required to store each of their members, concatenated. Additionally, if the items stored in an array are not of a fixed-size (ie: strings, other arrays, etc) then an additional framing offset is stored for each item. The size of this offset is either 1, 2 or 4 bytes depending on the overall size of the container. Additionally, extra padding bytes are added as required for alignment of child values.
Tuples (including dictionary entries) use the amount of space required to store each of their members, concatenated, plus one framing offset (as per arrays) for each non-fixed-sized item in the tuple, except for the last one. Additionally, extra padding bytes are added as required for alignment of child values.
Variants use the same amount of space as the item inside of the variant, plus 1 byte, plus the length of the type string for the item inside the variant.
As an example, consider a dictionary mapping strings to variants. In the case that the dictionary is empty, 0 bytes are required for the serialization.
If we add an item "width" that maps to the int32 value of 500 then we will use 4 byte to store the int32 (so 6 for the variant containing it) and 6 bytes for the string. The variant must be aligned to 8 after the 6 bytes of the string, so that's 2 extra bytes. 6 (string) + 2 (padding) + 6 (variant) is 14 bytes used for the dictionary entry. An additional 1 byte is added to the array as a framing offset making a total of 15 bytes.
If we add another entry, "title" that maps to a nullable string that happens to have a value of null, then we use 0 bytes for the null value (and 3 bytes for the variant to contain it along with its type string) plus 6 bytes for the string. Again, we need 2 padding bytes. That makes a total of 6 + 2 + 3 = 11 bytes.
We now require extra padding between the two items in the array. After the 14 bytes of the first item, that's 2 bytes required. We now require 2 framing offsets for an extra two bytes. 14 + 2 + 11 + 2 = 29 bytes to encode the entire two-item dictionary.
Type Information Cache
For each GVariant type that currently exists in the program a type information structure is kept in the type information cache. The type information structure is required for rapid deserialization.
Continuing with the above example, if a #GVariant exists with the type "a{sv}" then a type information struct will exist for "a{sv}", "{sv}", "s", and "v". Multiple uses of the same type will share the same type information. Additionally, all single-digit types are stored in read-only static memory and do not contribute to the writable memory footprint of a program using #GVariant.
Aside from the type information structures stored in read-only memory, there are two forms of type information. One is used for container types where there is a single element type: arrays and maybe types. The other is used for container types where there are multiple element types: tuples and dictionary entries.
Array type info structures are 6 * sizeof (void *), plus the memory required to store the type string itself. This means that on 32-bit systems, the cache entry for "a{sv}" would require 30 bytes of memory (plus malloc overhead).
Tuple type info structures are 6 * sizeof (void *), plus 4 * sizeof (void *) for each item in the tuple, plus the memory required to store the type string itself. A 2-item tuple, for example, would have a type information structure that consumed writable memory in the size of 14 * sizeof (void *) (plus type string) This means that on 32-bit systems, the cache entry for "{sv}" would require 61 bytes of memory (plus malloc overhead).
This means that in total, for our "a{sv}" example, 91 bytes of type information would be allocated.
The type information cache, additionally, uses a #GHashTable to store and look up the cached items and stores a pointer to this hash table in static storage. The hash table is freed when there are zero items in the type cache.
Although these sizes may seem large it is important to remember that a program will probably only have a very small number of different types of values in it and that only one type information structure is required for many different values of the same type.
Buffer Management Memory
#GVariant uses an internal buffer management structure to deal with the various different possible sources of serialized data that it uses. The buffer is responsible for ensuring that the correct call is made when the data is no longer in use by #GVariant. This may involve a g_free() or a g_slice_free() or even g_mapped_file_unref().
One buffer management structure is used for each chunk of serialized data. The size of the buffer management structure is 4 * (void *). On 32-bit systems, that's 16 bytes.
GVariant structure
The size of a #GVariant structure is 6 * (void *). On 32-bit systems, that's 24 bytes.
#GVariant structures only exist if they are explicitly created with API calls. For example, if a #GVariant is constructed out of serialized data for the example given above (with the dictionary) then although there are 9 individual values that comprise the entire dictionary (two keys, two values, two variants containing the values, two dictionary entries, plus the dictionary itself), only 1 #GVariant instance exists -- the one referring to the dictionary.
If calls are made to start accessing the other values then #GVariant instances will exist for those values only for as long as they are in use (ie: until you call g_variant_unref()). The type information is shared. The serialized data and the buffer management structure for that serialized data is shared by the child.
Summary
To put the entire example together, for our dictionary mapping strings to variants (with two entries, as given above), we are using 91 bytes of memory for type information, 29 bytes of memory for the serialized data, 16 bytes for buffer management and 24 bytes for the #GVariant instance, or a total of 160 bytes, plus malloc overhead. If we were to use g_variant_get_child_value() to access the two dictionary entries, we would use an additional 48 bytes. If we were to have other dictionaries of the same type, we would use more memory for the serialized data and buffer management for those dictionaries, but the type information would be shared.