This chapter tries to answer the real-life questions of users and presents the most common scenario use-cases I could come up with. The use-cases are presented from most likely to less likely. Clearly, this is one of the most common question people ask: they just want to crank code and implement a subclass of a GObject. Sometimes because they want to create their own class hierarchy, sometimes because they want to subclass one of GTK+'s widget. This chapter will focus on the implementation of a subtype of GObject. The sample source code associated to this section can be found in the documentation's source tarball, in the sample/gobject directory: maman-bar.{h|c}: this is the source for a object which derives from GObject and which shows how to declare different types of methods on the object. maman-subbar.{h|c}: this is the source for a object which derives from MamanBar and which shows how to override some of its parent's methods. maman-foo.{h|c}: this is the source for an object which derives from GObject and which declares a signal. test.c: this is the main source which instantiates an instance of type and exercises their API. The first step before writing the code for your GObject is to write the type's header which contains the needed type, function and macro definitions. Each of these elements is nothing but a convention which is followed not only by GTK+'s code but also by most users of GObject. If you feel the need not to obey the rules stated below, think about it twice: If your users are a bit accustomed to GTK+ code or any Glib code, they will be a bit surprised and getting used to the conventions you decided upon will take time (money) and will make them grumpy (not a good thing) You must assess the fact that these conventions might have been designed by both smart and experienced people: maybe they were at least partly right. Try to put your ego aside. Pick a name convention for your headers and source code and stick to it: use a dash to separate the prefix from the typename: maman-bar.h and maman-bar.c (this is the convention used by Nautilus and most GNOME libraries). use an underscore to separate the prefix from the typename: maman_bar.h and maman_bar.c. Do not separate the prefix from the typename: mamanbar.h and mamanbar.c. (this is the convention used by GTK+) I personally like the first solution better: it makes reading file names easier for those with poor eyesight like me. When you need some private (internal) declarations in several (sub)classes, you can define them in a private header file which is often named by appending the private keyword to the public header name. For example, one could use maman-bar-private.h, maman_bar_private.h or mamanbarprivate.h. Typically, such private header files are not installed. The basic conventions for any header which exposes a GType are described in . Most GObject-based code also obeys onf of the following conventions: pick one and stick to it. If you want to declare a type named bar with prefix maman, name the type instance MamanBar and its class MamanBarClass (name is case-sensitive). It is customary to declare them with code similar to the following: /* * Copyright/Licensing information. */ #ifndef MAMAN_BAR_H #define MAMAN_BAR_H /* * Potentially, include other headers on which this header depends. */ /* * Type macros. */ typedef struct _MamanBar MamanBar; typedef struct _MamanBarClass MamanBarClass; struct _MamanBar { GObject parent; /* instance members */ }; struct _MamanBarClass { GObjectClass parent; /* class members */ }; /* used by MAMAN_BAR_TYPE */ GType maman_bar_get_type (void); /* * Method definitions. */ #endif Most GTK+ types declare their private fields in the public header with a /* private */ comment, relying on their user's intelligence not to try to play with these fields. Fields not marked private are considered public by default. The /* protected */ comment (same semantics as those of C++) is also used, mainly in the GType library, in code written by Tim Janik. struct _MamanBar { GObject parent; /* private */ int hsize; }; All of Nautilus code and a lot of GNOME libraries use private indirection members, as described by Herb Sutter in his Pimpl articles (see Compilation Firewalls and The Fast Pimpl Idiom : he summarizes the different issues better than I will). typedef struct _MamanBarPrivate MamanBarPrivate; struct _MamanBar { GObject parent; /* private */ MamanBarPrivate *priv; }; Do not call this private, as that is a registered c++ keyword. The private structure is then defined in the .c file, instantiated in the object's init function and destroyed in the object's finalize function. static void maman_bar_finalize (GObject *object) { MamanBar *self = MAMAN_BAR (object); /* do stuff */ g_free (self->priv); } static void maman_bar_init (GTypeInstance *instance, gpointer g_class) { MamanBar *self = MAMAN_BAR (instance); self->priv = g_new0 (MamanBarPrivate,1); /* do stuff */ } A similar alternative, available since Glib version 2.4, is to define a private structure in the .c file, declare it as a private structure in maman_bar_class_init using g_type_class_add_private. Instead of allocating memory in maman_bar_init a pointer to the private memory area is stored in the instance to allow convenient access to this structure. A private structure will then be attached to each newly created object by the GObject system. You dont need to free or allocate the private structure, only the objects or pointers that it may contain. Another advantage of this to the previous version is that is lessens memory fragmentation, as the public and private parts of the instance memory are allocated at once. typedef struct _MamanBarPrivate MamanBarPrivate; struct _MamanBarPrivate { int private_field; }; static void maman_bar_class_init (MamanBarClass *klass) { ... g_type_class_add_private (klass, sizeof (MamanBarPrivate)); ... } static void maman_bar_init (GTypeInstance *instance, gpointer g_class) { MamanBar *self = MAMAN_BAR (instance); self->priv = G_TYPE_INSTANCE_GET_PRIVATE (self, MAMAN_BAR_TYPE, MamanBarPrivate); /* do stuff */ } Finally, there are different header include conventions. Again, pick one and stick to it. I personally use indifferently any of the two, depending on the codebase I work on: the rule is consistency. Some people add at the top of their headers a number of #include directives to pull in all the headers needed to compile client code. This allows client code to simply #include "maman-bar.h". Other do not #include anything and expect the client to #include themselves the headers they need before including your header. This speeds up compilation because it minimizes the amount of pre-processor work. This can be used in conjunction with the re-declaration of certain unused types in the client code to minimize compile-time dependencies and thus speed up compilation. In your code, the first step is to #include the needed headers: depending on your header include strategy, this can be as simple as #include "maman-bar.h" or as complicated as tens of #include lines ending with #include "maman-bar.h": /* * Copyright information */ #include "maman-bar.h" /* If you use Pimpls, include the private structure * definition here. Some people create a maman-bar-private.h header * which is included by the maman-bar.c file and which contains the * definition for this private structure. */ struct _MamanBarPrivate { int member_1; /* stuff */ }; /* * forward definitions */ Implement maman_bar_get_type and make sure the code compiles: GType maman_bar_get_type (void) { static GType type = 0; if (type == 0) { static const GTypeInfo info = { sizeof (MamanBarClass), NULL, /* base_init */ NULL, /* base_finalize */ NULL, /* class_init */ NULL, /* class_finalize */ NULL, /* class_data */ sizeof (MamanBar), 0, /* n_preallocs */ NULL /* instance_init */ }; type = g_type_register_static (G_TYPE_OBJECT, "MamanBarType", &info, 0); } return type; } People often get confused when trying to construct their GObjects because of the sheer number of different ways to hook into the objects's construction process: it is difficult to figure which is the correct, recommended way. shows what user-provided functions are invoked during object instanciation and in which order they are invoked. A user looking for the equivalent of the simple C++ constructor function should use the instance_init method. It will be invoked after all the parent's instance_init functions have been invoked. It cannot take arbitrary construction parameters (as in C++) but if your object needs arbitrary parameters to complete initialization, you can use construction properties. Construction properties will be set only after all instance_init functions have run. No object reference will be returned to the client of g_object_new> until all the construction properties have been set. As such, I would recommend writing the following code first: static void maman_bar_init (GTypeInstance *instance, gpointer g_class) { MamanBar *self = (MamanBar *)instance; self->private = g_new0 (MamanBarPrivate, 1); /* initialize all public and private members to reasonable default values. */ /* If you need specific consruction properties to complete initialization, * delay initialization completion until the property is set. */ } And make sure that you set maman_bar_init as the type's instance_init function in maman_bar_get_type. Make sure the code builds and runs: create an instance of the object and make sure maman_bar_init is called (add a g_print call in it). Now, if you need special construction properties, install the properties in the class_init function, override the set and get methods and implement the get and set methods as described in . Make sure that these properties use a construct only GParamSpec by setting the param spec's flag field to G_PARAM_CONSTRUCT_ONLY: this helps GType ensure that these properties are not set again later by malicious user code. static void bar_class_init (MamanBarClass *klass) { GObjectClass *gobject_class = G_OBJECT_CLASS (klass); GParamSpec *maman_param_spec; gobject_class->set_property = bar_set_property; gobject_class->get_property = bar_get_property; maman_param_spec = g_param_spec_string ("maman", "Maman construct prop", "Set maman's name", "no-name-set" /* default value */, G_PARAM_CONSTRUCT_ONLY |G_PARAM_READWRITE); g_object_class_install_property (gobject_class, PROP_MAMAN, maman_param_spec); } If you need this, make sure you can build and run code similar to the code shown above. Make sure your construct properties can set correctly during construction, make sure you cannot set them afterwards and make sure that if your users do not call g_object_new with the required construction properties, these will be initialized with the default values. I consider good taste to halt program execution if a construction property is set its default value. This allows you to catch client code which does not give a reasonable value to the construction properties. Of course, you are free to disagree but you should have a good reason to do so. Some people sometimes need to construct their object but only after the construction properties have been set. This is possible through the use of the constructor class method as described in . However, I have yet to see any reasonable use of this feature. As such, to initialize your object instances, use by default the base_init function and construction properties. Again, it is often difficult to figure out which mechanism to use to hook into the object's destruction process: when the last g_object_unref function call is made, a lot of things happen as described in . The destruction process of your object must be split is two different phases: you must override both the dispose and the finalize class methods. struct _MamanBarPrivate { gboolean dispose_has_run; }; static GObjectClass parent_class = NULL; static void bar_dispose (GObject *obj) { MamanBar *self = (MamanBar *)obj; if (self->private->dispose_has_run) { /* If dispose did already run, return. */ return; } /* Make sure dispose does not run twice. */ object->private->dispose_has_run = TRUE; /* * In dispose, you are supposed to free all types referenced from this * object which might themselves hold a reference to self. Generally, * the most simple solution is to unref all members on which you own a * reference. */ /* Chain up to the parent class */ G_OBJECT_CLASS (parent_class)->dispose (obj); } static void bar_finalize (GObject *obj) { MamanBar *self = (MamanBar *)obj; /* * Here, complete object destruction. * You might not need to do much... */ g_free (self->private); /* Chain up to the parent class */ G_OBJECT_CLASS (parent_class)->finalize (obj); } static void bar_class_init (BarClass *klass) { GObjectClass *gobject_class = G_OBJECT_CLASS (klass); gobject_class->dispose = bar_dispose; gobject_class->finalize = bar_finalize; } static void maman_bar_init (GTypeInstance *instance, gpointer g_class) { MamanBar *self = (MamanBar *)instance; self->private = g_new0 (MamanBarPrivate, 1); self->private->dispose_has_run = FALSE; parent_class = g_type_class_peek_parent (klass); } Add similar code to your GObject, make sure the code still builds and runs: dispose and finalize must be called during the last unref. It is possible that object methods might be invoked after dispose is run and before finalize runs. GObject does not consider this to be a program error: you must gracefully detect this and neither crash nor warn the user. To do this, you need something like the following code at the start of each object method, to make sure the object's data is still valid before manipulating it: if (self->private->dispose_has_run) { /* Dispose has run. Data is not valid anymore. */ return; } Just as with C++, there are many different ways to define object methods and extend them: the following list and sections draw on C++ vocabulary. (Readers are expected to know basic C++ buzzwords. Those who have not had to write C++ code recently can refer to e.g. to refresh their memories.) non-virtual public methods, virtual public methods and virtual private methods These are the simplest: you want to provide a simple method which can act on your object. All you need to do is to provide a function prototype in the header and an implementation of that prototype in the source file. /* declaration in the header. */ void maman_bar_do_action (MamanBar *self, /* parameters */); /* implementation in the source file */ void maman_bar_do_action (MamanBar *self, /* parameters */) { /* do stuff here. */ } There is really nothing scary about this. This is the preferred way to create polymorphic GObjects. All you need to do is to define the common method and its class function in the public header, implement the common method in the source file and re-implement the class function in each object which inherits from you. /* declaration in maman-bar.h. */ struct _MamanBarClass { GObjectClass parent; /* stuff */ void (*do_action) (MamanBar *self, /* parameters */); }; void maman_bar_do_action (MamanBar *self, /* parameters */); /* implementation in maman-bar.c */ void maman_bar_do_action (MamanBar *self, /* parameters */) { MAMAN_BAR_GET_CLASS (self)->do_action (self, /* parameters */); } The code above simply redirects the do_action call to the relevant class function. Some users, concerned about performance, do not provide the maman_bar_do_action wrapper function and require users to dereference the class pointer themselves. This is not such a great idea in terms of encapsulation and makes it difficult to change the object's implementation afterwards, should this be needed. Other users, also concerned by performance issues, declare the maman_bar_do_action function inline in the header file. This, however, makes it difficult to change the object's implementation later (although easier than requiring users to directly dereference the class function) and is often difficult to write in a portable way (the inline keyword is not part of the C standard). In doubt, unless a user shows you hard numbers about the performance cost of the function call, just maman_bar_do_action in the source file. Please, note that it is possible for you to provide a default implementation for this class method in the object's class_init function: initialize the klass->do_action field to a pointer to the actual implementation. You can also make this class method pure virtual by initializing the klass->do_action field to NULL: static void maman_bar_real_do_action_two (MamanBar *self, /* parameters */) { /* Default implementation for the virtual method. */ } static void maman_bar_class_init (BarClass *klass) { /* pure virtual method: mandates implementation in children. */ klass->do_action_one = NULL; /* merely virtual method. */ klass->do_action_two = maman_bar_real_do_action_two; } void maman_bar_do_action_one (MamanBar *self, /* parameters */) { MAMAN_BAR_GET_CLASS (self)->do_action_one (self, /* parameters */); } void maman_bar_do_action_two (MamanBar *self, /* parameters */) { MAMAN_BAR_GET_CLASS (self)->do_action_two (self, /* parameters */); } These are very similar to Virtual Public methods. They just don't have a public function to call the function directly. The header file contains only a declaration of the class function: /* declaration in maman-bar.h. */ struct _MamanBarClass { GObjectClass parent; /* stuff */ void (*helper_do_specific_action) (MamanBar *self, /* parameters */); }; void maman_bar_do_any_action (MamanBar *self, /* parameters */); These class functions are often used to delegate part of the job to child classes: /* this accessor function is static: it is not exported outside of this file. */ static void maman_bar_do_specific_action (MamanBar *self, /* parameters */) { MAMAN_BAR_GET_CLASS (self)->do_specific_action (self, /* parameters */); } void maman_bar_do_any_action (MamanBar *self, /* parameters */) { /* random code here */ /* * Try to execute the requested action. Maybe the requested action cannot be implemented * here. So, we delegate its implementation to the child class: */ maman_bar_do_specific_action (self, /* parameters */); /* other random code here */ } Again, it is possible to provide a default implementation for this private virtual class function: static void maman_bar_class_init (MamanBarClass *klass) { /* pure virtual method: mandates implementation in children. */ klass->do_specific_action_one = NULL; /* merely virtual method. */ klass->do_specific_action_two = maman_bar_real_do_specific_action_two; } Children can then implement the subclass with code such as: static void maman_bar_subtype_class_init (MamanBarSubTypeClass *klass) { MamanBarClass *bar_class = MAMAN_BAR_CLASS (klass); /* implement pure virtual class function. */ bar_class->do_specific_action_one = maman_bar_subtype_do_specific_action_one; } Chaining up is often loosely defined by the following set of conditions: Parent class A defines a public virtual method named foo and provides a default implementation. Child class B re-implements method foo. In the method B::foo, the child class B calls its parent class method A::foo. There are many uses to this idiom: You need to change the behaviour of a class without modifying its code. You create a subclass to inherit its implementation, re-implement a public virtual method to modify the behaviour slightly and chain up to ensure that the previous behaviour is not really modifed, just extended. You are lazy, you have access to the source code of the parent class but you don't want to modify it to add method calls to new specialized method calls: it is faster to hack the child class to chain up than to modify the parent to call down. You need to implement the Chain Of Responsability pattern: each object of the inheritance tree chains up to its parent (typically, at the begining or the end of the method) to ensure that they each handler is run in turn. I am personally not really convinced any of the last two uses are really a good idea but since this programming idiom is often used, this section attemps to explain how to implement it. To explicitely chain up to the implementation of the virtual method in the parent class, you first need a handle to the original parent class structure. This pointer can then be used to access the original class function pointer and invoke it directly. The original adjective used in this sentence is not innocuous. To fully understand its meaning, you need to recall how class structures are initialized: for each object type, the class structure associated to this object is created by first copying the class structure of its parent type (a simple memcpy) and then by invoking the class_init callback on the resulting class structure. Since the class_init callback is responsible for overwriting the class structure with the user re-implementations of the class methods, we cannot merely use the modified copy of the parent class structure stored in our derived instance. We want to get a copy of the class structure of an instance of the parent class. The function g_type_class_peek_parent is used to access the original parent class structure. Its input is a pointer to the class of the derived object and it returns a pointer to the original parent class structure. The code below shows how you could use it: static void b_method_to_call (B *obj, int a) { BClass *klass; AClass *parent_class; klass = B_GET_CLASS (obj); parent_class = g_type_class_peek_parent (klass); /* do stuff before chain up */ parent_class->method_to_call (obj, a); /* do stuff after chain up */ } A lot of people who use this idiom in GTK+ store the parent class structure pointer in a global static variable to avoid the costly call to g_type_class_peek_parent for each function call. Typically, the class_init callback initializes the global static variable. gtk/gtkhscale.c does this. The bulk of interface definition has already been shown in but I feel it is needed to show exactly how to create an interface. The sample source code associated to this section can be found in the documentation's source tarball, in the sample/interface/maman-ibaz.{h|c} file. As above, the first step is to get the header right: #ifndef MAMAN_IBAZ_H #define MAMAN_IBAZ_H #include <glib-object.h> #define MAMAN_TYPE_IBAZ (maman_ibaz_get_type ()) #define MAMAN_IBAZ(obj) (G_TYPE_CHECK_INSTANCE_CAST ((obj), MAMAN_TYPE_IBAZ, MamanIbaz)) #define MAMAN_IS_IBAZ(obj) (G_TYPE_CHECK_INSTANCE_TYPE ((obj), MAMAN_TYPE_IBAZ)) #define MAMAN_IBAZ_GET_INTERFACE(inst) (G_TYPE_INSTANCE_GET_INTERFACE ((inst), MAMAN_TYPE_IBAZ, MamanIbazInterface)) typedef struct _MamanIbaz MamanIbaz; /* dummy object */ typedef struct _MamanIbazInterface MamanIbazInterface; struct _MamanIbazInterface { GTypeInterface parent; void (*do_action) (MamanIbaz *self); }; GType maman_ibaz_get_type (void); void maman_ibaz_do_action (MamanIbaz *self); #endif /*MAMAN_IBAZ_H*/ This code is the same as the code for a normal GType which derives from a GObject except for a few details: The _GET_CLASS macro is called _GET_INTERFACE and not implemented with G_TYPE_INSTANCE_GET_CLASS but with G_TYPE_INSTANCE_GET_INTERFACE. The instance type, MamanIbaz is not fully defined: it is used merely as an abstract type which represents an instance of whatever object which implements the interface. The implementation of the MamanIbaz type itself is trivial: maman_ibaz_get_type registers the type in the type system. maman_ibaz_base_init is expected to register the interface's signals if there are any (we will see a bit (later how to use them). Make sure to use a static local boolean variable to make sure not to run the initialization code twice (as described in , base_init is run once for each interface implementation instanciation) maman_ibaz_do_action dereferences the class structure to access its associated class function and calls it. static void maman_ibaz_base_init (gpointer g_class) { static gboolean initialized = FALSE; if (!initialized) { /* create interface signals here. */ initialized = TRUE; } } GType maman_ibaz_get_type (void) { static GType type = 0; if (type == 0) { static const GTypeInfo info = { sizeof (MamanIbazInterface), maman_ibaz_base_init, /* base_init */ NULL, /* base_finalize */ NULL, /* class_init */ NULL, /* class_finalize */ NULL, /* class_data */ 0, 0, /* n_preallocs */ NULL /* instance_init */ }; type = g_type_register_static (G_TYPE_INTERFACE, "MamanIbaz", &info, 0); } return type; } void maman_ibaz_do_action (MamanIbaz *self) { MAMAN_IBAZ_GET_INTERFACE (self)->do_action (self); } Once the interface is defined, implementing it is rather trivial. Source code showing how to do this for the IBaz interface defined in the previous section is located in sample/interface/maman-baz.{h|c}. The first step is to define a normal GType. Here, we have decided to use a GType which derives from GObject. Its name is MamanBaz: #ifndef MAMAN_BAZ_H #define MAMAN_BAZ_H #include <glib-object.h> #define MAMAN_TYPE_BAZ (maman_baz_get_type ()) #define MAMAN_BAZ(obj) (G_TYPE_CHECK_INSTANCE_CAST ((obj), MAMAN_TYPE_BAZ, Mamanbaz)) #define MAMAN_BAZ_CLASS(vtable) (G_TYPE_CHECK_CLASS_CAST ((vtable), MAMAN_TYPE_BAZ, MamanbazClass)) #define MAMAN_IS_BAZ(obj) (G_TYPE_CHECK_INSTANCE_TYPE ((obj), MAMAN_TYPE_BAZ)) #define MAMAN_IS_BAZ_CLASS(vtable) (G_TYPE_CHECK_CLASS_TYPE ((vtable), MAMAN_TYPE_BAZ)) #define MAMAN_BAZ_GET_CLASS(inst) (G_TYPE_INSTANCE_GET_CLASS ((inst), MAMAN_TYPE_BAZ, MamanbazClass)) typedef struct _MamanBaz MamanBaz; typedef struct _MamanBazClass MamanBazClass; struct _MamanBaz { GObject parent; int instance_member; }; struct _MamanBazClass { GObjectClass parent; }; GType maman_baz_get_type (void); #endif //MAMAN_BAZ_H There is clearly nothing specifically weird or scary about this header: it does not define any weird API or derives from a weird type. The second step is to implement maman_baz_get_type: GType maman_baz_get_type (void) { static GType type = 0; if (type == 0) { static const GTypeInfo info = { sizeof (MamanBazClass), NULL, /* base_init */ NULL, /* base_finalize */ NULL, /* class_init */ NULL, /* class_finalize */ NULL, /* class_data */ sizeof (MamanBaz), 0, /* n_preallocs */ baz_instance_init /* instance_init */ }; static const GInterfaceInfo ibaz_info = { (GInterfaceInitFunc) baz_interface_init, /* interface_init */ NULL, /* interface_finalize */ NULL /* interface_data */ }; type = g_type_register_static (G_TYPE_OBJECT, "MamanBazType", &info, 0); g_type_add_interface_static (type, MAMAN_TYPE_IBAZ, &ibaz_info); } return type; } This function is very much like all the similar functions we looked at previously. The only interface-specific code present here is the call to g_type_add_interface_static which is used to inform the type system that this just-registered GType also implements the interface MAMAN_TYPE_IBAZ. baz_interface_init, the interface initialization function, is also pretty simple: static void baz_do_action (MamanBaz *self) { g_print ("Baz implementation of IBaz interface Action: 0x%x.\n", self->instance_member); } static void baz_interface_init (gpointer g_iface, gpointer iface_data) { MamanIbazInteface *iface = (MamanIbazInteface *)g_iface; iface->do_action = (void (*) (MamanIbaz *self))baz_do_action; } static void baz_instance_init (GTypeInstance *instance, gpointer g_class) { MamanBaz *self = MAMAN_BAZ(instance); self->instance_member = 0xdeadbeaf; } baz_interface_init merely initializes the interface methods to the implementations defined by MamanBaz: maman_baz_do_action does nothing very useful but it could :) To specify that an interface requires the presence of other interfaces when implemented, GObject introduces the concept of prerequisites: it is possible to associate a list of prerequisite interfaces to an interface. For example, if object A wishes to implement interface I1, and if interface I1 has a prerequisite on interface I2, A has to implement both I1 and I2. The mechanism described above is, in practice, very similar to Java's interface I1 extends interface I2. The example below shows the GObject equivalent: type = g_type_register_static (G_TYPE_INTERFACE, "MamanIbar", &info, 0); /* Make the MamanIbar interface require MamanIbaz interface. */ g_type_interface_add_prerequisite (type, MAMAN_TYPE_IBAZ); The code shown above adds the MamanIbaz interface to the list of prerequisites of MamanIbar while the code below shows how an implementation can implement both interfaces and register their implementations: static void ibar_do_another_action (MamanBar *self) { g_print ("Bar implementation of IBar interface Another Action: 0x%x.\n", self->instance_member); } static void ibar_interface_init (gpointer g_iface, gpointer iface_data) { MamanIbarInterface *iface = (MamanIbarInterface *)g_iface; iface->do_another_action = (void (*) (MamanIbar *self))ibar_do_another_action; } static void ibaz_do_action (MamanBar *self) { g_print ("Bar implementation of IBaz interface Action: 0x%x.\n", self->instance_member); } static void ibaz_interface_init (gpointer g_iface, gpointer iface_data) { MamanIbazInterface *iface = (MamanIbazInterface *)g_iface; iface->do_action = (void (*) (MamanIbaz *self))ibaz_do_action; } static void bar_instance_init (GTypeInstance *instance, gpointer g_class) { MamanBar *self = (MamanBar *)instance; self->instance_member = 0x666; } GType maman_bar_get_type (void) { static GType type = 0; if (type == 0) { static const GTypeInfo info = { sizeof (MamanBarClass), NULL, /* base_init */ NULL, /* base_finalize */ NULL, /* class_init */ NULL, /* class_finalize */ NULL, /* class_data */ sizeof (MamanBar), 0, /* n_preallocs */ bar_instance_init /* instance_init */ }; static const GInterfaceInfo ibar_info = { (GInterfaceInitFunc) ibar_interface_init, /* interface_init */ NULL, /* interface_finalize */ NULL /* interface_data */ }; static const GInterfaceInfo ibaz_info = { (GInterfaceInitFunc) ibaz_interface_init, /* interface_init */ NULL, /* interface_finalize */ NULL /* interface_data */ }; type = g_type_register_static (G_TYPE_OBJECT, "MamanBarType", &info, 0); g_type_add_interface_static (type, MAMAN_TYPE_IBAZ, &ibaz_info); g_type_add_interface_static (type, MAMAN_TYPE_IBAR, &ibar_info); } return type; } It is very important to notice that the order in which interface implementations are added to the main object is not random: g_type_add_interface_static must be invoked first on the interfaces which have no prerequisites and then on the others. Complete source code showing how to define the MamanIbar interface which requires MamanIbaz and how to implement the MamanIbar interface is located in sample/interface/maman-ibar.{h|c} and sample/interface/maman-bar.{h|c}. Starting from version 2.4 of glib, gobject interfaces can also have properties. Declaration of the interface properties is similar to declaring the properties of ordinary gobject types as explained in , except that g_object_interface_install_property is used to declare the properties instead of g_object_class_install_property. To include a property named 'name' of type string in the maman_ibaz interface example code above, we only need to add one That really is one line extended to six for the sake of clarity line in the maman_ibaz_base_init The g_object_interface_install_property can also be called from class_init but it must not be called after that point. as shown below: static void maman_ibaz_base_init (gpointer g_iface) { static gboolean initialized = FALSE; if (!initialized) { /* create interface signals here. */ g_object_interface_install_property (g_iface, g_param_spec_string ("name", "maman_ibaz_name", "Name of the MamanIbaz", "maman", G_PARAM_READWRITE)); initialized = TRUE; } } One point worth noting is that the declared property wasn't assigned an integer ID. The reason being that integer IDs of properities are utilized only inside the get and set methods and since interfaces do not implement properties, there is no need to assign integer IDs to interface properties. The story for the implementers of the interface is also quite trivial. An implementer shall declare and define it's properties in the usual way as explained in , except for one small change: it shall declare the properties of the interface it implements using g_object_class_override_property instead of g_object_class_install_property. The following code snipet shows the modifications needed in the MamanBaz declaration and implementation above: struct _MamanBaz { GObject parent; gint instance_member; gchar *name; /* placeholder for property */ }; enum { ARG_0, ARG_NAME }; GType maman_baz_get_type (void) { static GType type = 0; if (type == 0) { static const GTypeInfo info = { sizeof (MamanBazClass), NULL, /* base_init */ NULL, /* base_finalize */ baz_class_init, /* class_init */ NULL, /* class_finalize */ NULL, /* class_data */ sizeof (MamanBaz), 0, /* n_preallocs */ baz_instance_init /* instance_init */ }; static const GInterfaceInfo ibaz_info = { (GInterfaceInitFunc) baz_interface_init, /* interface_init */ NULL, /* interface_finalize */ NULL /* interface_data */ }; type = g_type_register_static (G_TYPE_OBJECT, "MamanBazType", &info, 0); g_type_add_interface_static (type, MAMAN_TYPE_IBAZ, &ibaz_info); } return type; } static void maman_baz_class_init (MamanBazClass * klass) { GObjectClass *gobject_class; gobject_class = (GObjectClass *) klass; parent_class = g_type_class_ref (G_TYPE_OBJECT); gobject_class->set_property = maman_baz_set_property; gobject_class->get_property = maman_baz_get_property; g_object_class_override_property (gobject_class, ARG_NAME, "name"); } static void maman_baz_set_property (GObject * object, guint prop_id, const GValue * value, GParamSpec * pspec) { MamanBaz *baz; GObject *obj; /* it's not null if we got it, but it might not be ours */ g_return_if_fail (G_IS_MAMAN_BAZ (object)); baz = MAMAN_BAZ (object); switch (prop_id) { case ARG_NAME: baz->name = g_value_get_string (value); break; default: G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec); break; } } static void maman_baz_get_property (GObject * object, guint prop_id, GValue * value, GParamSpec * pspec) { MamanBaz *baz; /* it's not null if we got it, but it might not be ours */ g_return_if_fail (G_IS_TEXT_PLUGIN (object)); baz = MAMAN_BAZ (object); switch (prop_id) { case ARG_NAME: g_value_set_string (value, baz->name); break; default: G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec); break; } } The signal system which was built in GType is pretty complex and flexible: it is possible for its users to connect at runtime any number of callbacks (implemented in any language for which a binding exists) A python callback can be connected to any signal on any C-based GObject. to any signal and to stop the emission of any signal at any state of the signal emission process. This flexibility makes it possible to use GSignal for much more than just emit events which can be received by numerous clients. The most basic use of signals is to implement simple event notification: for example, if we have a MamanFile object, and if this object has a write method, we might wish to be notified whenever someone uses this method. The code below shows how the user can connect a callback to the write signal. Full code for this simple example is located in sample/signal/maman-file.{h|c} and in sample/signal/test.c file = g_object_new (MAMAN_FILE_TYPE, NULL); g_signal_connect (G_OBJECT (file), "write", (GCallback)write_event, NULL); maman_file_write (file, buffer, 50); The MamanFile signal is registered in the class_init function: klass->write_signal_id = g_signal_newv ("write", G_TYPE_FROM_CLASS (g_class), G_SIGNAL_RUN_LAST | G_SIGNAL_NO_RECURSE | G_SIGNAL_NO_HOOKS, NULL /* class closure */, NULL /* accumulator */, NULL /* accu_data */, g_cclosure_marshal_VOID__VOID, G_TYPE_NONE /* return_type */, 0 /* n_params */, NULL /* param_types */); and the signal is emited in maman_file_write: void maman_file_write (MamanFile *self, guint8 *buffer, guint32 size) { /* First write data. */ /* Then, notify user of data written. */ g_signal_emit (self, MAMAN_FILE_GET_CLASS (self)->write_signal_id, 0 /* details */, NULL); } As shown above, you can safely set the details parameter to zero if you do not know what it can be used for. For a discussion of what you could used it for, see The signature of the signal handler in the above example is defined as g_cclosure_marshal_VOID__VOID. Its name follows a simple convention which encodes the function parameter and return value types in the function name. Specifically, the value infront of the double underscore is the type of the return value, while the value(s) after the double underscore denote the parameter types. The header gobject/gmarshal.h defines a set of commonly needed closures that one can use. The previous implementation does the job but the signal facility of GObject can be used to provide even more flexibility to this file change notification mechanism. One of the key ideas is to make the process of writing data to the file part of the signal emission process to allow users to be notified either before or after the data is written to the file. To integrate the process of writing the data to the file into the signal emission mechanism, we can register a default class closure for this signal which will be invoked during the signal emission, just like any other user-connected signal handler. The first step to implement this idea is to change the signature of the signal: we need to pass around the buffer to write and its size. To do this, we use our own marshaller which will be generated through glib's genmarshall tool. We thus create a file named marshall.list which contains the following single line: VOID:POINTER,UINT and use the Makefile provided in sample/signal/Makefile to generate the file named maman-file-complex-marshall.c. This C file is finally included in maman-file-complex.c. Once the marshaller is present, we register the signal and its marshaller in the class_init function of the object MamanFileComplex (full source for this object is included in sample/signal/maman-file-complex.{h|c}): GClosure *default_closure; GType param_types[2]; default_closure = g_cclosure_new (G_CALLBACK (default_write_signal_handler), (gpointer)0xdeadbeaf /* user_data */, NULL /* destroy_data */); param_types[0] = G_TYPE_POINTER; param_types[1] = G_TYPE_UINT; klass->write_signal_id = g_signal_newv ("write", G_TYPE_FROM_CLASS (g_class), G_SIGNAL_RUN_LAST | G_SIGNAL_NO_RECURSE | G_SIGNAL_NO_HOOKS, default_closure /* class closure */, NULL /* accumulator */, NULL /* accu_data */, maman_file_complex_VOID__POINTER_UINT, G_TYPE_NONE /* return_type */, 2 /* n_params */, param_types /* param_types */); The code shown above first creates the closure which contains the code to complete the file write. This closure is registered as the default class_closure of the newly created signal. Of course, you need to implement completely the code for the default closure since I just provided a skeleton: static void default_write_signal_handler (GObject *obj, guint8 *buffer, guint size, gpointer user_data) { g_assert (user_data == (gpointer)0xdeadbeaf); /* Here, we trigger the real file write. */ g_print ("default signal handler: 0x%x %u\n", buffer, size); } Finally, the client code must invoke the maman_file_complex_write function which triggers the signal emission: void maman_file_complex_write (MamanFileComplex *self, guint8 *buffer, guint size) { /* trigger event */ g_signal_emit (self, MAMAN_FILE_COMPLEX_GET_CLASS (self)->write_signal_id, 0, /* details */ buffer, size); } The client code (as shown in sample/signal/test.c and below) can now connect signal handlers before and after the file write is completed: since the default signal handler which does the write itself runs during the RUN_LAST phase of the signal emission, it will run after all handlers connected with g_signal_connect and before all handlers connected with g_signal_connect_after. If you intent to write a GObject which emits signals, I would thus urge you to create all your signals with the G_SIGNAL_RUN_LAST such that your users have a maximum of flexibility as to when to get the event. Here, we combined it with G_SIGNAL_NO_RECURSE and G_SIGNAL_NO_HOOKS to ensure our users will not try to do really weird things with our GObject. I strongly advise you to do the same unless you really know why (in which case you really know the inner workings of GSignal by heart and you are not reading this). static void complex_write_event_before (GObject *file, guint8 *buffer, guint size, gpointer user_data) { g_assert (user_data == NULL); g_print ("Complex Write event before: 0x%x, %u\n", buffer, size); } static void complex_write_event_after (GObject *file, guint8 *buffer, guint size, gpointer user_data) { g_assert (user_data == NULL); g_print ("Complex Write event after: 0x%x, %u\n", buffer, size); } static void test_file_complex (void) { guint8 buffer[100]; GObject *file; file = g_object_new (MAMAN_FILE_COMPLEX_TYPE, NULL); g_signal_connect (G_OBJECT (file), "write", (GCallback)complex_write_event_before, NULL); g_signal_connect_after (G_OBJECT (file), "write", (GCallback)complex_write_event_after, NULL); maman_file_complex_write (MAMAN_FILE_COMPLEX (file), buffer, 50); g_object_unref (G_OBJECT (file)); } The code above generates the following output on my machine: Complex Write event before: 0xbfffe280, 50 default signal handler: 0xbfffe280 50 Complex Write event after: 0xbfffe280, 50 For many historic reasons related to how the ancestor of GObject used to work in GTK+ 1.x versions, there is a much simpler I personally think that this method is horribly mind-twisting: it adds a new indirection which unecessarily complicates the overall code path. However, because this method is widely used by all of GTK+ and GObject code, readers need to understand it. The reason why this is done that way in most of GTK+ is related to the fact that the ancestor of GObject did not provide any other way to create a signal with a default handler than this one. Some people have tried to justify that it is done that way because it is better, faster (I am extremly doubtfull about the faster bit. As a matter of fact, the better bit also mystifies me ;-). I have the feeling no one really knows and everyone does it because they copy/pasted code from code which did the same. It is probably better to leave this specific trivia to hacker legends domain... way to create a signal with a default handler than to create a closure by hand and to use the g_signal_newv. For example, g_signal_new can be used to create a signal which uses a default handler which is stored in the class structure of the object. More specifically, the class structure contains a function pointer which is accessed during signal emission to invoke the default handler and the user is expected to provide to g_signal_new the offset from the start of the class structure to the function pointer. I would like to point out here that the reason why the default handler of a signal is named everywhere a class_closure is probably related to the fact that it used to be really a function pointer stored in the class structure. The following code shows the declaration of the MamanFileSimple class structure which contains the write function pointer. struct _MamanFileSimpleClass { GObjectClass parent; guint write_signal_id; /* signal default handlers */ void (*write) (MamanFileSimple *self, guint8 *buffer, guint size); }; The write function pointer is initialied in the class_init function of the object to default_write_signal_handler: static void maman_file_simple_class_init (gpointer g_class, gpointer g_class_data) { GObjectClass *gobject_class = G_OBJECT_CLASS (g_class); MamanFileSimpleClass *klass = MAMAN_FILE_SIMPLE_CLASS (g_class); klass->write = default_write_signal_handler; Finally, the signal is created with g_signal_new in the same class_init function: klass->write_signal_id = g_signal_new ("write", G_TYPE_FROM_CLASS (g_class), G_SIGNAL_RUN_LAST | G_SIGNAL_NO_RECURSE | G_SIGNAL_NO_HOOKS, G_STRUCT_OFFSET (MamanFileSimpleClass, write), NULL /* accumulator */, NULL /* accu_data */, maman_file_complex_VOID__POINTER_UINT, G_TYPE_NONE /* return_type */, 2 /* n_params */, G_TYPE_POINTER, G_TYPE_UINT); Of note, here, is the 4th argument to the function: it is an integer calculated by the G_STRUCT_OFFSET macro which indicates the offset of the member write from the start of the MamanFileSimpleClass class structure. GSignal uses this offset to create a special wrapper closure which first retrieves the target function pointer before calling it. While the complete code for this type of default handler looks less clutered as shown in sample/signal/maman-file-simple.{h|c}, it contains numerous subtleties. The main subtle point which everyone must be aware of is that the signature of the default handler created that way does not have a user_data argument: default_write_signal_handler is different in sample/signal/maman-file-complex.c and in sample/signal/maman-file-simple.c. If you have doubts about which method to use, I would advise you to use the second one which involves g_signal_new rather than g_signal_newv: it is better to write code which looks like the vast majority of other GTK+/Gobject code than to do it your own way. However, now, you know why. Now that you know how to create signals to which the users can connect easily and at any point in the signal emission process thanks to g_signal_connect, g_signal_connect_after and G_SIGNAL_RUN_LAST, it is time to look into how your users can and will screw you. This is also interesting to know how you too, can screw other people. This will make you feel good and eleet. The users can: stop the emission of the signal at anytime override the default handler of the signal if it is stored as a function pointer in the class structure (which is the prefered way to create a default signal handler, as discussed in the previous section). In both cases, the original programmer should be as careful as possible to write code which is resistant to the fact that the default handler of the signal might not able to run. This is obviously not the case in the example used in the previous sections since the write to the file depends on whether or not the default handler runs (however, this might be your goal: to allow the user to prevent the file write if he wishes to). If all you want to do is to stop the signal emission from one of the callbacks you connected yourself, you can call g_signal_stop_by_name. Its use is very simple which is why I won't detail it further. If the signal's default handler is just a class function pointer, it is also possible to override it yourself from the class_init function of a type which derives from the parent. That way, when the signal is emitted, the parent class will use the function provided by the child as a signal default handler. Of course, it is also possible (and recommended) to chain up from the child to the parent's default signal handler to ensure the integrity of the parent object. Overriding a class method and chaining up was demonstrated in which is why I won't bother to show exactly how to do it here again.