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.

g_variant_unref ()

void
g_variant_unref (GVariant *value);

Decreases the reference count of value . When its reference count drops to 0, the memory used by the variant is freed.

Since: 2.24

g_variant_ref ()

GVariant *
g_variant_ref (GVariant *value);

Increases the reference count of value .

Since: 2.24

g_variant_ref_sink ()

GVariant *
g_variant_ref_sink (GVariant *value);

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.

Since: 2.24

g_variant_is_floating ()

gboolean
g_variant_is_floating (GVariant *value);

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.

Since: 2.26

g_variant_take_ref ()

GVariant *
g_variant_take_ref (GVariant *value);

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.

g_variant_get_type ()

const GVariantType *
g_variant_get_type (GVariant *value);

Determines the type of value .

The return value is valid for the lifetime of value and must not be freed.

Since: 2.24

g_variant_get_type_string ()

const gchar *
g_variant_get_type_string (GVariant *value);

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.

Since: 2.24

g_variant_is_of_type ()

gboolean
g_variant_is_of_type (GVariant *value,
                      const GVariantType *type);

Checks if a value has a type matching the provided type.

Since: 2.24

g_variant_is_container ()

gboolean
g_variant_is_container (GVariant *value);

Checks if value is a container.

Since: 2.24

g_variant_compare ()

gint
g_variant_compare (gconstpointer one,
                   gconstpointer two);

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.

Since: 2.26

g_variant_classify ()

GVariantClass
g_variant_classify (GVariant *value);

Classifies value according to its top-level type.

Since: 2.24

g_variant_check_format_string ()

gboolean
g_variant_check_format_string (GVariant *value,
                               const gchar *format_string,
                               gboolean copy_only);

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()).

Since: 2.34

g_variant_get ()

void
g_variant_get (GVariant *value,
               const gchar *format_string,
               ...);

Deconstructs a GVariant instance.

Think of this function as an analogue to scanf().

The arguments that are expected by this function are entirely determined by format_string . format_string also restricts the permissible types of value . It is an error to give a value with an incompatible type. See the section on GVariant format strings. Please note that the syntax of the format string is very likely to be extended in the future.

format_string determines the C types that are used for unpacking the values and also determines if the values are copied or borrowed, see the section on GVariant format strings.

[skip]

Since: 2.24

g_variant_get_va ()

void
g_variant_get_va (GVariant *value,
                  const gchar *format_string,
                  const gchar **endptr,
                  va_list *app);

This function is intended to be used by libraries based on GVariant that want to provide g_variant_get()-like functionality to their users.

The API is more general than g_variant_get() to allow a wider range of possible uses.

format_string must still point to a valid format string, but it only need to be nul-terminated if endptr is NULL. If endptr is non-NULL then it is updated to point to the first character past the end of the format string.

app is a pointer to a va_list. The arguments, according to format_string , are collected from this va_list and the list is left pointing to the argument following the last.

These two generalisations allow mixing of multiple calls to g_variant_new_va() and g_variant_get_va() within a single actual varargs call by the user.

format_string determines the C types that are used for unpacking the values and also determines if the values are copied or borrowed, see the section on GVariant format strings.

[skip]

Since: 2.24

g_variant_new ()

GVariant *
g_variant_new (const gchar *format_string,
               ...);

Creates a new GVariant instance.

Think of this function as an analogue to g_strdup_printf().

The type of the created instance and the arguments that are expected by this function are determined by format_string . See the section on GVariant format strings. Please note that the syntax of the format string is very likely to be extended in the future.

The first character of the format string must not be '*' '?' '@' or 'r'; in essence, a new GVariant must always be constructed by this function (and not merely passed through it unmodified).

Note that the arguments must be of the correct width for their types specified in format_string . This can be achieved by casting them. See the GVariant varargs documentation.

1
2
3
4
5
6
7
8

MyFlags some_flags = FLAG_ONE | FLAG_TWO;
const gchar *some_strings[] = { "a", "b", "c", NULL };
GVariant *new_variant;

new_variant = g_variant_new ("(t^as)",
                             // This cast is required.
                             (guint64) some_flags,
                             some_strings);

[skip]

Since: 2.24

g_variant_new_va ()

GVariant *
g_variant_new_va (const gchar *format_string,
                  const gchar **endptr,
                  va_list *app);

This function is intended to be used by libraries based on GVariant that want to provide g_variant_new()-like functionality to their users.

The API is more general than g_variant_new() to allow a wider range of possible uses.

format_string must still point to a valid format string, but it only needs to be nul-terminated if endptr is NULL. If endptr is non-NULL then it is updated to point to the first character past the end of the format string.

app is a pointer to a va_list. The arguments, according to format_string , are collected from this va_list and the list is left pointing to the argument following the last.

Note that the arguments in app must be of the correct width for their types specified in format_string when collected into the va_list. See the GVariant varargs documentation.

These two generalisations allow mixing of multiple calls to g_variant_new_va() and g_variant_get_va() within a single actual varargs call by the user.

The return value will be floating if it was a newly created GVariant instance (for example, if the format string was "(ii)"). In the case that the format_string was '*', '?', 'r', or a format starting with '@' then the collected GVariant pointer will be returned unmodified, without adding any additional references.

In order to behave correctly in all cases it is necessary for the calling function to g_variant_ref_sink() the return result before returning control to the user that originally provided the pointer. At this point, the caller will have their own full reference to the result. This can also be done by adding the result to a container, or by passing it to another g_variant_new() call.

[skip]

Since: 2.24

g_variant_new_boolean ()

GVariant *
g_variant_new_boolean (gboolean value);

Creates a new boolean GVariant instance -- either TRUE or FALSE.

Since: 2.24

g_variant_new_byte ()

GVariant *
g_variant_new_byte (guint8 value);

Creates a new byte GVariant instance.

Since: 2.24

g_variant_new_int16 ()

GVariant *
g_variant_new_int16 (gint16 value);

Creates a new int16 GVariant instance.

Since: 2.24

g_variant_new_uint16 ()

GVariant *
g_variant_new_uint16 (guint16 value);

Creates a new uint16 GVariant instance.

Since: 2.24

g_variant_new_int32 ()

GVariant *
g_variant_new_int32 (gint32 value);

Creates a new int32 GVariant instance.

Since: 2.24

g_variant_new_uint32 ()

GVariant *
g_variant_new_uint32 (guint32 value);

Creates a new uint32 GVariant instance.

Since: 2.24

g_variant_new_int64 ()

GVariant *
g_variant_new_int64 (gint64 value);

Creates a new int64 GVariant instance.

Since: 2.24

g_variant_new_uint64 ()

GVariant *
g_variant_new_uint64 (guint64 value);

Creates a new uint64 GVariant instance.

Since: 2.24

g_variant_new_handle ()

GVariant *
g_variant_new_handle (gint32 value);

Creates a new handle GVariant instance.

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.

Since: 2.24

g_variant_new_double ()

GVariant *
g_variant_new_double (gdouble value);

Creates a new double GVariant instance.

Since: 2.24

g_variant_new_string ()

GVariant *
g_variant_new_string (const gchar *string);

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.

Since: 2.24

g_variant_new_take_string ()

GVariant *
g_variant_new_take_string (gchar *string);

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 this with g_variant_new_maybe().

This function consumes string . g_free() will be called on string when it is no longer required.

You must not modify or access string in any other way after passing it to this function. It is even possible that string is immediately freed.

[skip]

Since: 2.38

g_variant_new_printf ()

GVariant *
g_variant_new_printf (const gchar *format_string,
                      ...);

Creates a string-type GVariant using printf formatting.

This is similar to calling g_strdup_printf() and then g_variant_new_string() but it saves a temporary variable and an unnecessary copy.

[skip]

Since: 2.38

g_variant_new_object_path ()

GVariant *
g_variant_new_object_path (const gchar *object_path);

Creates a D-Bus object path GVariant with the contents of string . string must be a valid D-Bus object path. Use g_variant_is_object_path() if you're not sure.

Since: 2.24

g_variant_is_object_path ()

gboolean
g_variant_is_object_path (const gchar *string);

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.

Since: 2.24

g_variant_new_signature ()

GVariant *
g_variant_new_signature (const gchar *signature);

Creates a D-Bus type signature GVariant with the contents of string . string must be a valid D-Bus type signature. Use g_variant_is_signature() if you're not sure.

Since: 2.24

g_variant_is_signature ()

gboolean
g_variant_is_signature (const gchar *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.

Since: 2.24

g_variant_new_variant ()

GVariant *
g_variant_new_variant (GVariant *value);

Boxes value . The result is a GVariant instance representing a variant containing the original value.

If child is a floating reference (see g_variant_ref_sink()), the new instance takes ownership of child .

[constructor]

Since: 2.24

g_variant_new_strv ()

GVariant *
g_variant_new_strv (const gchar * const *strv,
                    gssize length);

Constructs an array of strings GVariant from the given array of strings.

If length is -1 then strv is NULL-terminated.

Since: 2.24

g_variant_new_objv ()

GVariant *
g_variant_new_objv (const gchar * const *strv,
                    gssize length);

Constructs an array of object paths GVariant from the given array of strings.

Each string must be a valid GVariant object path; see g_variant_is_object_path().

If length is -1 then strv is NULL-terminated.

Since: 2.30

g_variant_new_bytestring ()

GVariant *
g_variant_new_bytestring (const gchar *string);

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.

Since: 2.26

g_variant_new_bytestring_array ()

GVariant *
g_variant_new_bytestring_array (const gchar * const *strv,
                                gssize length);

Constructs an array of bytestring GVariant from the given array of strings.

If length is -1 then strv is NULL-terminated.

Since: 2.26

g_variant_get_boolean ()

gboolean
g_variant_get_boolean (GVariant *value);

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.

Since: 2.24

g_variant_get_byte ()

guint8
g_variant_get_byte (GVariant *value);

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.

Since: 2.24

g_variant_get_int16 ()

gint16
g_variant_get_int16 (GVariant *value);

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.

Since: 2.24

g_variant_get_uint16 ()

guint16
g_variant_get_uint16 (GVariant *value);

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.

Since: 2.24

g_variant_get_int32 ()

gint32
g_variant_get_int32 (GVariant *value);

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.

Since: 2.24

g_variant_get_uint32 ()

guint32
g_variant_get_uint32 (GVariant *value);

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.

Since: 2.24

g_variant_get_int64 ()

gint64
g_variant_get_int64 (GVariant *value);

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.

Since: 2.24

g_variant_get_uint64 ()

guint64
g_variant_get_uint64 (GVariant *value);

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.

Since: 2.24

g_variant_get_handle ()

gint32
g_variant_get_handle (GVariant *value);

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.

Since: 2.24

g_variant_get_double ()

gdouble
g_variant_get_double (GVariant *value);

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.

Since: 2.24

g_variant_get_string ()

const gchar *
g_variant_get_string (GVariant *value,
                      gsize *length);

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.

Since: 2.24

g_variant_dup_string ()

gchar *
g_variant_dup_string (GVariant *value,
                      gsize *length);

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().

Since: 2.24

g_variant_get_variant ()

GVariant *
g_variant_get_variant (GVariant *value);

Unboxes value . The result is the GVariant instance that was contained in value .

Since: 2.24

g_variant_get_strv ()

const gchar **
g_variant_get_strv (GVariant *value,
                    gsize *length);

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.

Since: 2.24

g_variant_dup_strv ()

gchar **
g_variant_dup_strv (GVariant *value,
                    gsize *length);

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.

Since: 2.24

g_variant_get_objv ()

const gchar **
g_variant_get_objv (GVariant *value,
                    gsize *length);

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.

Since: 2.30

g_variant_dup_objv ()

gchar **
g_variant_dup_objv (GVariant *value,
                    gsize *length);

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.

Since: 2.30

g_variant_get_bytestring ()

const gchar *
g_variant_get_bytestring (GVariant *value);

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.

Since: 2.26

g_variant_dup_bytestring ()

gchar *
g_variant_dup_bytestring (GVariant *value,
                          gsize *length);

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().

Since: 2.26

g_variant_get_bytestring_array ()

const gchar **
g_variant_get_bytestring_array (GVariant *value,
                                gsize *length);

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.

Since: 2.26

g_variant_dup_bytestring_array ()

gchar **
g_variant_dup_bytestring_array (GVariant *value,
                                gsize *length);

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.

Since: 2.26

g_variant_new_maybe ()

GVariant *
g_variant_new_maybe (const GVariantType *child_type,
                     GVariant *child);

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 .

Since: 2.24

g_variant_new_array ()

GVariant *
g_variant_new_array (const GVariantType *child_type,
                     GVariant * const *children,
                     gsize n_children);

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().

Since: 2.24

g_variant_new_tuple ()

GVariant *
g_variant_new_tuple (GVariant * const *children,
                     gsize n_children);

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().

Since: 2.24

g_variant_new_dict_entry ()

GVariant *
g_variant_new_dict_entry (GVariant *key,
                          GVariant *value);

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().

[constructor]

Since: 2.24

g_variant_new_fixed_array ()

GVariant *
g_variant_new_fixed_array (const GVariantType *element_type,
                           gconstpointer elements,
                           gsize n_elements,
                           gsize element_size);

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.

Since: 2.32

g_variant_get_maybe ()

GVariant *
g_variant_get_maybe (GVariant *value);

Given a maybe-typed GVariant instance, extract its value. If the value is Nothing, then this function returns NULL.

Since: 2.24

g_variant_n_children ()

gsize
g_variant_n_children (GVariant *value);

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).

Since: 2.24

g_variant_get_child_value ()

GVariant *
g_variant_get_child_value (GVariant *value,
                           gsize index_);

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).

Since: 2.24

g_variant_get_child ()

void
g_variant_get_child (GVariant *value,
                     gsize index_,
                     const gchar *format_string,
                     ...);

Reads a child item out of a container GVariant instance and deconstructs it according to format_string . This call is essentially a combination of g_variant_get_child_value() and g_variant_get().

format_string determines the C types that are used for unpacking the values and also determines if the values are copied or borrowed, see the section on GVariant format strings.

[skip]

Since: 2.24

g_variant_lookup_value ()

GVariant *
g_variant_lookup_value (GVariant *dictionary,
                        const gchar *key,
                        const GVariantType *expected_type);

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.

Since: 2.28

g_variant_lookup ()

gboolean
g_variant_lookup (GVariant *dictionary,
                  const gchar *key,
                  const gchar *format_string,
                  ...);

Looks up a value in a dictionary GVariant.

This function is a wrapper around g_variant_lookup_value() and g_variant_get(). In the case that NULL would have been returned, this function returns FALSE. Otherwise, it unpacks the returned value and returns TRUE.

format_string determines the C types that are used for unpacking the values and also determines if the values are copied or borrowed, see the section on GVariant format strings.

This function is currently implemented with a linear scan. If you plan to do many lookups then GVariantDict may be more efficient.

[skip]

Since: 2.28

g_variant_get_fixed_array ()

gconstpointer
g_variant_get_fixed_array (GVariant *value,
                           gsize *n_elements,
                           gsize element_size);

Provides access to the serialized data for an array of fixed-sized items.

value 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, as given by the section on serialized data memory.

In particular, arrays of these fixed-sized types can be interpreted as an array of the given C type, with element_size set to the size the appropriate type:

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 , which must be non-NULL, is set equal to the number of items in the array.

Since: 2.24

g_variant_get_size ()

gsize
g_variant_get_size (GVariant *value);

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.

Since: 2.24

g_variant_get_data ()

gconstpointer
g_variant_get_data (GVariant *value);

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).

Since: 2.24

g_variant_get_data_as_bytes ()

GBytes *
g_variant_get_data_as_bytes (GVariant *value);

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.

Since: 2.36

g_variant_store ()

void
g_variant_store (GVariant *value,
                 gpointer data);

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 .

Since: 2.24

g_variant_new_from_data ()

GVariant *
g_variant_new_from_data (const GVariantType *type,
                         gconstpointer data,
                         gsize size,
                         gboolean trusted,
                         GDestroyNotify notify,
                         gpointer user_data);

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.

Since: 2.24

g_variant_new_from_bytes ()

GVariant *
g_variant_new_from_bytes (const GVariantType *type,
                          GBytes *bytes,
                          gboolean trusted);

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.

Since: 2.36

g_variant_byteswap ()

GVariant *
g_variant_byteswap (GVariant *value);

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).

While this function can safely handle untrusted, non-normal data, it is recommended to check whether the input is in normal form beforehand, using g_variant_is_normal_form(), and to reject non-normal inputs if your application can be strict about what inputs it rejects.

The returned value is always in normal form and is marked as trusted.

Since: 2.24

g_variant_get_normal_form ()

GVariant *
g_variant_get_normal_form (GVariant *value);

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 . The non-normal parts of value will be replaced with default values which are guaranteed to be in normal form.

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.

Since: 2.24

g_variant_is_normal_form ()

gboolean
g_variant_is_normal_form (GVariant *value);

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.

Since: 2.24

g_variant_hash ()

guint
g_variant_hash (gconstpointer value);

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.

Since: 2.24

g_variant_equal ()

gboolean
g_variant_equal (gconstpointer one,
                 gconstpointer two);

Checks if one and two have the same type and value.

The types of one and two are gconstpointer only to allow use of this function with GHashTable. They must each be a GVariant.

Since: 2.24

g_variant_print ()

gchar *
g_variant_print (GVariant *value,
                 gboolean type_annotate);

Pretty-prints value in the format understood by g_variant_parse().

The format is described here.

If type_annotate is TRUE, then type information is included in the output.

Since: 2.24

g_variant_print_string ()

GString *
g_variant_print_string (GVariant *value,
                        GString *string,
                        gboolean type_annotate);

Behaves as g_variant_print(), but operates on a GString.

If string is non-NULL then it is appended to and returned. Else, a new empty GString is allocated and it is returned.

[skip]

Since: 2.24

g_variant_iter_copy ()

GVariantIter *
g_variant_iter_copy (GVariantIter *iter);

Creates a new heap-allocated GVariantIter to iterate over the container that was being iterated over by iter . Iteration begins on the new iterator from the current position of the old iterator but the two copies are independent past that point.

Use g_variant_iter_free() to free the return value when you no longer need it.

A reference is taken to the container that iter is iterating over and will be related only when g_variant_iter_free() is called.

Since: 2.24

g_variant_iter_free ()

void
g_variant_iter_free (GVariantIter *iter);

Frees a heap-allocated GVariantIter. Only call this function on iterators that were returned by g_variant_iter_new() or g_variant_iter_copy().

Since: 2.24

g_variant_iter_init ()

gsize
g_variant_iter_init (GVariantIter *iter,
                     GVariant *value);

Initialises (without allocating) a GVariantIter. iter may be completely uninitialised prior to this call; its old value is ignored.

The iterator remains valid for as long as value exists, and need not be freed in any way.

[skip]

Since: 2.24

g_variant_iter_n_children ()

gsize
g_variant_iter_n_children (GVariantIter *iter);

Queries the number of child items in the container that we are iterating over. This is the total number of items -- not the number of items remaining.

This function might be useful for preallocation of arrays.

Since: 2.24

g_variant_iter_new ()

GVariantIter *
g_variant_iter_new (GVariant *value);

Creates a heap-allocated GVariantIter for iterating over the items in value .

Use g_variant_iter_free() to free the return value when you no longer need it.

A reference is taken to value and will be released only when g_variant_iter_free() is called.

Since: 2.24

g_variant_iter_next_value ()

GVariant *
g_variant_iter_next_value (GVariantIter *iter);

Gets the next item in the container. If no more items remain then NULL is returned.

Use g_variant_unref() to drop your reference on the return value when you no longer need it.

Here is an example for iterating with g_variant_iter_next_value():

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

// recursively iterate a container
void
iterate_container_recursive (GVariant *container)
{
  GVariantIter iter;
  GVariant *child;

  g_variant_iter_init (&iter, container);
  while ((child = g_variant_iter_next_value (&iter)))
    {
      g_print ("type '%s'\n", g_variant_get_type_string (child));

      if (g_variant_is_container (child))
        iterate_container_recursive (child);

      g_variant_unref (child);
    }
}

Since: 2.24

g_variant_iter_next ()

gboolean
g_variant_iter_next (GVariantIter *iter,
                     const gchar *format_string,
                     ...);

Gets the next item in the container and unpacks it into the variable argument list according to format_string , returning TRUE.

If no more items remain then FALSE is returned.

All of the pointers given on the variable arguments list of this function are assumed to point at uninitialised memory. It is the responsibility of the caller to free all of the values returned by the unpacking process.

Here is an example for memory management with g_variant_iter_next():

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

// Iterates a dictionary of type 'a{sv}'
void
iterate_dictionary (GVariant *dictionary)
{
  GVariantIter iter;
  GVariant *value;
  gchar *key;

  g_variant_iter_init (&iter, dictionary);
  while (g_variant_iter_next (&iter, "{sv}", &key, &value))
    {
      g_print ("Item '%s' has type '%s'\n", key,
               g_variant_get_type_string (value));

      // must free data for ourselves
      g_variant_unref (value);
      g_free (key);
    }
}

For a solution that is likely to be more convenient to C programmers when dealing with loops, see g_variant_iter_loop().

format_string determines the C types that are used for unpacking the values and also determines if the values are copied or borrowed.

See the section on GVariant format strings.

[skip]

Since: 2.24

g_variant_iter_loop ()

gboolean
g_variant_iter_loop (GVariantIter *iter,
                     const gchar *format_string,
                     ...);

Gets the next item in the container and unpacks it into the variable argument list according to format_string , returning TRUE.

If no more items remain then FALSE is returned.

On the first call to this function, the pointers appearing on the variable argument list are assumed to point at uninitialised memory. On the second and later calls, it is assumed that the same pointers will be given and that they will point to the memory as set by the previous call to this function. This allows the previous values to be freed, as appropriate.

This function is intended to be used with a while loop as demonstrated in the following example. This function can only be used when iterating over an array. It is only valid to call this function with a string constant for the format string and the same string constant must be used each time. Mixing calls to this function and g_variant_iter_next() or g_variant_iter_next_value() on the same iterator causes undefined behavior.

If you break out of a such a while loop using g_variant_iter_loop() then you must free or unreference all the unpacked values as you would with g_variant_get(). Failure to do so will cause a memory leak.

Here is an example for memory management with g_variant_iter_loop():

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

// Iterates a dictionary of type 'a{sv}'
void
iterate_dictionary (GVariant *dictionary)
{
  GVariantIter iter;
  GVariant *value;
  gchar *key;

  g_variant_iter_init (&iter, dictionary);
  while (g_variant_iter_loop (&iter, "{sv}", &key, &value))
    {
      g_print ("Item '%s' has type '%s'\n", key,
               g_variant_get_type_string (value));

      // no need to free 'key' and 'value' here
      // unless breaking out of this loop
    }
}

For most cases you should use g_variant_iter_next().

This function is really only useful when unpacking into GVariant or GVariantIter in order to allow you to skip the call to g_variant_unref() or g_variant_iter_free().

For example, if you are only looping over simple integer and string types, g_variant_iter_next() is definitely preferred. For string types, use the '&' prefix to avoid allocating any memory at all (and thereby avoiding the need to free anything as well).

format_string determines the C types that are used for unpacking the values and also determines if the values are copied or borrowed.

See the section on GVariant format strings.

[skip]

Since: 2.24

G_VARIANT_BUILDER_INIT()

#define             G_VARIANT_BUILDER_INIT(variant_type)

A stack-allocated GVariantBuilder must be initialized if it is used together with g_auto() to avoid warnings or crashes if function returns before g_variant_builder_init() is called on the builder.

This macro can be used as initializer instead of an explicit zeroing a variable when declaring it and a following g_variant_builder_init(), but it cannot be assigned to a variable.

The passed variant_type should be a static GVariantType to avoid lifetime issues, as copying the variant_type does not happen in the G_VARIANT_BUILDER_INIT() call, but rather in functions that make sure that GVariantBuilder is valid.

1

g_auto(GVariantBuilder) builder = G_VARIANT_BUILDER_INIT (G_VARIANT_TYPE_BYTESTRING);

Since: 2.50

g_variant_builder_unref ()

void
g_variant_builder_unref (GVariantBuilder *builder);

Decreases the reference count on builder .

In the event that there are no more references, releases all memory associated with the GVariantBuilder.

Don't call this on stack-allocated GVariantBuilder instances or bad things will happen.

Since: 2.24

g_variant_builder_ref ()

GVariantBuilder *
g_variant_builder_ref (GVariantBuilder *builder);

Increases the reference count on builder .

Don't call this on stack-allocated GVariantBuilder instances or bad things will happen.

Since: 2.24

g_variant_builder_new ()

GVariantBuilder *
g_variant_builder_new (const GVariantType *type);

Allocates and initialises a new GVariantBuilder.

You should call g_variant_builder_unref() on the return value when it is no longer needed. The memory will not be automatically freed by any other call.

In most cases it is easier to place a GVariantBuilder directly on the stack of the calling function and initialise it with g_variant_builder_init().

Since: 2.24

g_variant_builder_init ()

void
g_variant_builder_init (GVariantBuilder *builder,
                        const GVariantType *type);

Initialises a GVariantBuilder structure.

type must be non-NULL. It specifies the type of container to construct. It can be an indefinite type such as G_VARIANT_TYPE_ARRAY or a definite type such as "as" or "(ii)". Maybe, array, tuple, dictionary entry and variant-typed values may be constructed.

After the builder is initialised, values are added using g_variant_builder_add_value() or g_variant_builder_add().

After all the child values are added, g_variant_builder_end() frees the memory associated with the builder and returns the GVariant that was created.

This function completely ignores the previous contents of builder . On one hand this means that it is valid to pass in completely uninitialised memory. On the other hand, this means that if you are initialising over top of an existing GVariantBuilder you need to first call g_variant_builder_clear() in order to avoid leaking memory.

You must not call g_variant_builder_ref() or g_variant_builder_unref() on a GVariantBuilder that was initialised with this function. If you ever pass a reference to a GVariantBuilder outside of the control of your own code then you should assume that the person receiving that reference may try to use reference counting; you should use g_variant_builder_new() instead of this function.

[skip]

Since: 2.24

g_variant_builder_clear ()

void
g_variant_builder_clear (GVariantBuilder *builder);

Releases all memory associated with a GVariantBuilder without freeing the GVariantBuilder structure itself.

It typically only makes sense to do this on a stack-allocated GVariantBuilder if you want to abort building the value part-way through. This function need not be called if you call g_variant_builder_end() and it also doesn't need to be called on builders allocated with g_variant_builder_new() (see g_variant_builder_unref() for that).

This function leaves the GVariantBuilder structure set to all-zeros. It is valid to call this function on either an initialised GVariantBuilder or one that is set to all-zeros but it is not valid to call this function on uninitialised memory.

[skip]

Since: 2.24

g_variant_builder_add_value ()

void
g_variant_builder_add_value (GVariantBuilder *builder,
                             GVariant *value);

Adds value to builder .

It is an error to call this function in any way that would create an inconsistent value to be constructed. Some examples of this are putting different types of items into an array, putting the wrong types or number of items in a tuple, putting more than one value into a variant, etc.

If value is a floating reference (see g_variant_ref_sink()), the builder instance takes ownership of value .

Since: 2.24

g_variant_builder_add ()

void
g_variant_builder_add (GVariantBuilder *builder,
                       const gchar *format_string,
                       ...);

Adds to a GVariantBuilder.

This call is a convenience wrapper that is exactly equivalent to calling g_variant_new() followed by g_variant_builder_add_value().

Note that the arguments must be of the correct width for their types specified in format_string . This can be achieved by casting them. See the GVariant varargs documentation.

This function might be used as follows:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

GVariant *
make_pointless_dictionary (void)
{
  GVariantBuilder builder;
  int i;

  g_variant_builder_init (&builder, G_VARIANT_TYPE_ARRAY);
  for (i = 0; i < 16; i++)
    {
      gchar buf[3];

      sprintf (buf, "%d", i);
      g_variant_builder_add (&builder, "{is}", i, buf);
    }

  return g_variant_builder_end (&builder);
}

[skip]

Since: 2.24

g_variant_builder_add_parsed ()

void
g_variant_builder_add_parsed (GVariantBuilder *builder,
                              const gchar *format,
                              ...);

Adds to a GVariantBuilder.

This call is a convenience wrapper that is exactly equivalent to calling g_variant_new_parsed() followed by g_variant_builder_add_value().

Note that the arguments must be of the correct width for their types specified in format_string . This can be achieved by casting them. See the GVariant varargs documentation.

This function might be used as follows:

1
2
3
4
5
6
7
8
9
10
11
12

GVariant *
make_pointless_dictionary (void)
{
  GVariantBuilder builder;
  int i;

  g_variant_builder_init (&builder, G_VARIANT_TYPE_ARRAY);
  g_variant_builder_add_parsed (&builder, "{'width', <%i>}", 600);
  g_variant_builder_add_parsed (&builder, "{'title', <%s>}", "foo");
  g_variant_builder_add_parsed (&builder, "{'transparency', <0.5>}");
  return g_variant_builder_end (&builder);
}

Since: 2.26

g_variant_builder_end ()

GVariant *
g_variant_builder_end (GVariantBuilder *builder);

Ends the builder process and returns the constructed value.

It is not permissible to use builder in any way after this call except for reference counting operations (in the case of a heap-allocated GVariantBuilder) or by reinitialising it with g_variant_builder_init() (in the case of stack-allocated). This means that for the stack-allocated builders there is no need to call g_variant_builder_clear() after the call to g_variant_builder_end().

It is an error to call this function in any way that would create an inconsistent value to be constructed (ie: insufficient number of items added to a container with a specific number of children required). It is also an error to call this function if the builder was created with an indefinite array or maybe type and no children have been added; in this case it is impossible to infer the type of the empty array.

Since: 2.24

g_variant_builder_open ()

void
g_variant_builder_open (GVariantBuilder *builder,
                        const GVariantType *type);

Opens a subcontainer inside the given builder . When done adding items to the subcontainer, g_variant_builder_close() must be called. type is the type of the container: so to build a tuple of several values, type must include the tuple itself.

It is an error to call this function in any way that would cause an inconsistent value to be constructed (ie: adding too many values or a value of an incorrect type).

Example of building a nested variant:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24