Widget

class Widget(**properties: Any)

Superclasses: InitiallyUnowned, Object

Subclasses: ActionBar, AppChooserButton, AppChooserWidget, AspectFrame, Box, Button, Calendar, CellView, CenterBox, CheckButton, ColorButton, ColorChooserWidget, ColorDialogButton, ColumnView, ComboBox, DragIcon, DrawingArea, DropDown, EditableLabel, Entry, Expander, FileChooserWidget, Fixed, FlowBox, FlowBoxChild, FontButton, FontChooserWidget, FontDialogButton, Frame, GLArea, GraphicsOffload, Grid, HeaderBar, IconView, Image, InfoBar, Inscription, Label, LevelBar, ListBase, ListBox, ListBoxRow, MediaControls, MenuButton, Notebook, Overlay, Paned, PasswordEntry, Picture, Popover, PopoverBin, PopoverMenuBar, ProgressBar, Range, Revealer, ScaleButton, Scrollbar, ScrolledWindow, SearchBar, SearchEntry, Separator, ShortcutLabel, ShortcutsShortcut, SpinButton, Spinner, Stack, StackSidebar, StackSwitcher, Statusbar, Switch, Text, TextView, TreeExpander, TreeView, Video, Viewport, Window, WindowControls, WindowHandle

Implemented Interfaces: Accessible, Buildable, ConstraintTarget

The base class for all widgets.

It manages the widget lifecycle, layout, states and style.

Height-for-width Geometry Management

GTK uses a height-for-width (and width-for-height) geometry management system. Height-for-width means that a widget can change how much vertical space it needs, depending on the amount of horizontal space that it is given (and similar for width-for-height). The most common example is a label that reflows to fill up the available width, wraps to fewer lines, and therefore needs less height.

Height-for-width geometry management is implemented in GTK by way of two virtual methods:

• get_request_mode

• measure

There are some important things to keep in mind when implementing height-for-width and when using it in widget implementations.

If you implement a direct GtkWidget subclass that supports height-for-width or width-for-height geometry management for itself or its child widgets, the get_request_mode virtual function must be implemented as well and return the widget’s preferred request mode. The default implementation of this virtual function returns CONSTANT_SIZE, which means that the widget will only ever get -1 passed as the for_size value to its measure implementation.

The geometry management system will query a widget hierarchy in only one orientation at a time. When widgets are initially queried for their minimum sizes it is generally done in two initial passes in the SizeRequestMode chosen by the toplevel.

For example, when queried in the normal HEIGHT_FOR_WIDTH mode:

First, the default minimum and natural width for each widget in the interface will be computed using measure with an orientation of HORIZONTAL and a for_size of -1. Because the preferred widths for each widget depend on the preferred widths of their children, this information propagates up the hierarchy, and finally a minimum and natural width is determined for the entire toplevel. Next, the toplevel will use the minimum width to query for the minimum height contextual to that width using measure with an orientation of VERTICAL and a for_size of the just computed width. This will also be a highly recursive operation. The minimum height for the minimum width is normally used to set the minimum size constraint on the toplevel.

After the toplevel window has initially requested its size in both dimensions it can go on to allocate itself a reasonable size (or a size previously specified with set_default_size). During the recursive allocation process it’s important to note that request cycles will be recursively executed while widgets allocate their children. Each widget, once allocated a size, will go on to first share the space in one orientation among its children and then request each child’s height for its target allocated width or its width for allocated height, depending. In this way a widget will typically be requested its size a number of times before actually being allocated a size. The size a widget is finally allocated can of course differ from the size it has requested. For this reason, GtkWidget caches a small number of results to avoid re-querying for the same sizes in one allocation cycle.

If a widget does move content around to intelligently use up the allocated size then it must support the request in both GtkSizeRequestModes even if the widget in question only trades sizes in a single orientation.

For instance, a Label that does height-for-width word wrapping will not expect to have measure with an orientation of VERTICAL called because that call is specific to a width-for-height request. In this case the label must return the height required for its own minimum possible width. By following this rule any widget that handles height-for-width or width-for-height requests will always be allocated at least enough space to fit its own content.

Here are some examples of how a HEIGHT_FOR_WIDTH widget generally deals with width-for-height requests:

static void
foo_widget_measure (GtkWidget      *widget,
                    GtkOrientation  orientation,
                    int             for_size,
                    int            *minimum_size,
                    int            *natural_size,
                    int            *minimum_baseline,
                    int            *natural_baseline)
{
  if (orientation == GTK_ORIENTATION_HORIZONTAL)
    {
      // Calculate minimum and natural width
    }
  else // VERTICAL
    {
      if (i_am_in_height_for_width_mode)
        {
          int min_width, dummy;

          // First, get the minimum width of our widget
          GTK_WIDGET_GET_CLASS (widget)->measure (widget, GTK_ORIENTATION_HORIZONTAL, -1,
                                                  &min_width, &dummy, &dummy, &dummy);

          // Now use the minimum width to retrieve the minimum and natural height to display
          // that width.
          GTK_WIDGET_GET_CLASS (widget)->measure (widget, GTK_ORIENTATION_VERTICAL, min_width,
                                                  minimum_size, natural_size, &dummy, &dummy);
        }
      else
        {
          // ... some widgets do both.
        }
    }
}

Often a widget needs to get its own request during size request or allocation. For example, when computing height it may need to also compute width. Or when deciding how to use an allocation, the widget may need to know its natural size. In these cases, the widget should be careful to call its virtual methods directly, like in the code example above.

It will not work to use the wrapper function measure inside your own size_allocate implementation. These return a request adjusted by SizeGroup, the widget’s align and expand flags, as well as its CSS style.

If a widget used the wrappers inside its virtual method implementations, then the adjustments (such as widget margins) would be applied twice. GTK therefore does not allow this and will warn if you try to do it.

Of course if you are getting the size request for another widget, such as a child widget, you must use measure; otherwise, you would not properly consider widget margins, SizeGroup, and so forth.

GTK also supports baseline vertical alignment of widgets. This means that widgets are positioned such that the typographical baseline of widgets in the same row are aligned. This happens if a widget supports baselines, has a vertical alignment using baselines, and is inside a widget that supports baselines and has a natural “row” that it aligns to the baseline, or a baseline assigned to it by the grandparent.

Baseline alignment support for a widget is also done by the measure virtual function. It allows you to report both a minimum and natural size.

If a widget ends up baseline aligned it will be allocated all the space in the parent as if it was FILL, but the selected baseline can be found via get_baseline. If the baseline has a value other than -1 you need to align the widget such that the baseline appears at the position.

GtkWidget as GtkBuildable

The GtkWidget implementation of the GtkBuildable interface supports various custom elements to specify additional aspects of widgets that are not directly expressed as properties.

If the widget uses a LayoutManager, GtkWidget supports a custom <layout> element, used to define layout properties:

<object class="GtkGrid" id="my_grid">
  <child>
    <object class="GtkLabel" id="label1">
      <property name="label">Description</property>
      <layout>
        <property name="column">0</property>
        <property name="row">0</property>
        <property name="row-span">1</property>
        <property name="column-span">1</property>
      </layout>
    </object>
  </child>
  <child>
    <object class="GtkEntry" id="description_entry">
      <layout>
        <property name="column">1</property>
        <property name="row">0</property>
        <property name="row-span">1</property>
        <property name="column-span">1</property>
      </layout>
    </object>
  </child>
</object>

GtkWidget allows style information such as style classes to be associated with widgets, using the custom <style> element:

<object class="GtkButton" id="button1">
  <style>
    <class name="my-special-button-class"/>
    <class name="dark-button"/>
  </style>
</object>

GtkWidget allows defining accessibility information, such as properties, relations, and states, using the custom <accessibility> element:

<object class="GtkButton" id="button1">
  <accessibility>
    <property name="label">Download</property>
    <relation name="labelled-by">label1</relation>
  </accessibility>
</object>

Building composite widgets from template XML

GtkWidgetexposes some facilities to automate the procedure of creating composite widgets using “templates”.

To create composite widgets with GtkBuilder XML, one must associate the interface description with the widget class at class initialization time using set_template.

The interface description semantics expected in composite template descriptions is slightly different from regular Builder XML.

Unlike regular interface descriptions, set_template will expect a <template> tag as a direct child of the toplevel <interface> tag. The <template> tag must specify the “class” attribute which must be the type name of the widget. Optionally, the “parent” attribute may be specified to specify the direct parent type of the widget type; this is ignored by GtkBuilder but can be used by UI design tools to introspect what kind of properties and internal children exist for a given type when the actual type does not exist.

The XML which is contained inside the <template> tag behaves as if it were added to the <object> tag defining the widget itself. You may set properties on a widget by inserting <property> tags into the <template> tag, and also add <child> tags to add children and extend a widget in the normal way you would with <object> tags.

Additionally, <object> tags can also be added before and after the initial <template> tag in the normal way, allowing one to define auxiliary objects which might be referenced by other widgets declared as children of the <template> tag.

Since, unlike the <object> tag, the <template> tag does not contain an “id” attribute, if you need to refer to the instance of the object itself that the template will create, simply refer to the template class name in an applicable element content.

Here is an example of a template definition, which includes an example of this in the <signal> tag:

<interface>
  <template class="FooWidget" parent="GtkBox">
    <property name="orientation">horizontal</property>
    <property name="spacing">4</property>
    <child>
      <object class="GtkButton" id="hello_button">
        <property name="label">Hello World</property>
        <signal name="clicked" handler="hello_button_clicked" object="FooWidget" swapped="yes"/>
      </object>
    </child>
    <child>
      <object class="GtkButton" id="goodbye_button">
        <property name="label">Goodbye World</property>
      </object>
    </child>
  </template>
</interface>

Typically, you’ll place the template fragment into a file that is bundled with your project, using GResource. In order to load the template, you need to call set_template_from_resource from the class initialization of your GtkWidget type:

static void
foo_widget_class_init (FooWidgetClass *klass)
{
  // ...

  gtk_widget_class_set_template_from_resource (GTK_WIDGET_CLASS (klass),
                                               "/com/example/ui/foowidget.ui");
}

You will also need to call init_template from the instance initialization function:

static void
foo_widget_init (FooWidget *self)
{
  gtk_widget_init_template (GTK_WIDGET (self));

  // Initialize the rest of the widget...
}

as well as calling dispose_template from the dispose function:

static void
foo_widget_dispose (GObject *gobject)
{
  FooWidget *self = FOO_WIDGET (gobject);

  // Dispose objects for which you have a reference...

  // Clear the template children for this widget type
  gtk_widget_dispose_template (GTK_WIDGET (self), FOO_TYPE_WIDGET);

  G_OBJECT_CLASS (foo_widget_parent_class)->dispose (gobject);
}

You can access widgets defined in the template using the get_template_child function, but you will typically declare a pointer in the instance private data structure of your type using the same name as the widget in the template definition, and call bind_template_child_full (or one of its wrapper macros widget_class_bind_template_child and widget_class_bind_template_child_private) with that name, e.g.

typedef struct {
  GtkWidget *hello_button;
  GtkWidget *goodbye_button;
} FooWidgetPrivate;

G_DEFINE_TYPE_WITH_PRIVATE (FooWidget, foo_widget, GTK_TYPE_BOX)

static void
foo_widget_dispose (GObject *gobject)
{
  gtk_widget_dispose_template (GTK_WIDGET (gobject), FOO_TYPE_WIDGET);

  G_OBJECT_CLASS (foo_widget_parent_class)->dispose (gobject);
}

static void
foo_widget_class_init (FooWidgetClass *klass)
{
  // ...
  G_OBJECT_CLASS (klass)->dispose = foo_widget_dispose;

  gtk_widget_class_set_template_from_resource (GTK_WIDGET_CLASS (klass),
                                               "/com/example/ui/foowidget.ui");
  gtk_widget_class_bind_template_child_private (GTK_WIDGET_CLASS (klass),
                                                FooWidget, hello_button);
  gtk_widget_class_bind_template_child_private (GTK_WIDGET_CLASS (klass),
                                                FooWidget, goodbye_button);
}

static void
foo_widget_init (FooWidget *widget)
{
  gtk_widget_init_template (GTK_WIDGET (widget));
}

You can also use bind_template_callback_full (or is wrapper macro widget_class_bind_template_callback) to connect a signal callback defined in the template with a function visible in the scope of the class, e.g.

// the signal handler has the instance and user data swapped
// because of the swapped="yes" attribute in the template XML
static void
hello_button_clicked (FooWidget *self,
                      GtkButton *button)
{
  g_print ("Hello, world!\n");
}

static void
foo_widget_class_init (FooWidgetClass *klass)
{
  // ...
  gtk_widget_class_set_template_from_resource (GTK_WIDGET_CLASS (klass),
                                               "/com/example/ui/foowidget.ui");
  gtk_widget_class_bind_template_callback (GTK_WIDGET_CLASS (klass), hello_button_clicked);
}

Methods

class Widget

action_set_enabled(action_name: str, enabled: bool) → None

Enables or disables an action installed with install_action.

Parameters:

• action_name – action name, such as “clipboard.paste”

• enabled – whether the action is now enabled

activate() → bool

Activates the widget.

The activation will emit the signal set using set_activate_signal during class initialization.

Activation is what happens when you press Enter on a widget.

If you wish to handle the activation keybinding yourself, it is recommended to use add_shortcut with an action created with new.

If widget is not activatable, the function returns false.

activate_action(name: str, args: Variant | None = None) → bool

Activates an action for the widget.

The action is looked up in the action groups associated with widget and its ancestors.

This is a wrapper around activate_action_variant that constructs the args variant according to format_string.

Parameters:

• name – the name of the action to activate

• args

activate_default() → None

Activates the default.activate action for the widget.

The action is looked up in the same was as for activate_action.

add_controller(controller: EventController) → None

Adds an event controller to the widget.

The event controllers of a widget handle the events that are propagated to the widget.

You will usually want to call this function right after creating any kind of EventController.

Parameters:

controller – an event controller that hasn’t been added to a widget yet

add_css_class(css_class: str) → None

Adds a style class to the widget.

After calling this function, the widget’s style will match for css_class, according to CSS matching rules.

Use remove_css_class to remove the style again.

Parameters:

css_class – style class to add to widget, without the leading period

add_mnemonic_label(label: Widget) → None

Adds a widget to the list of mnemonic labels for this widget.

See list_mnemonic_labels.

Note that the list of mnemonic labels for the widget is cleared when the widget is destroyed, so the caller must make sure to update its internal state at this point as well.

Parameters:

label – a widget that acts as a mnemonic label for widget

classmethod add_shortcut(shortcut: Shortcut) → None

Parameters:

shortcut

add_tick_callback(callback: Callable[[Widget, FrameClock, Any], bool], user_data: Any = None) → int

Queues an animation frame update and adds a callback to be called before each frame.

Until the tick callback is removed, it will be called frequently (usually at the frame rate of the output device or as quickly as the application can be repainted, whichever is slower). For this reason, is most suitable for handling graphics that change every frame or every few frames.

The tick callback does not automatically imply a relayout or repaint. If you want a repaint or relayout, and aren’t changing widget properties that would trigger that (for example, changing the text of a label), then you will have to call queue_resize or queue_draw yourself.

get_frame_time should generally be used for timing continuous animations and get_predicted_presentation_time should be used if you are trying to display isolated frames at particular times.

This is a more convenient alternative to connecting directly to the update signal of the frame clock, since you don’t have to worry about when a frame clock is assigned to a widget.

To remove a tick callback, pass the ID that is returned by this function to remove_tick_callback.

Parameters:

• callback – function to call for updating animations

• user_data – data to pass to callback

allocate(width: int, height: int, baseline: int, transform: Transform | None = None) → None

Assigns size, position, (optionally) a baseline and transform to a child widget.

In this function, the allocation and baseline may be adjusted. The given allocation will be forced to be bigger than the widget’s minimum size, as well as at least 0×0 in size.

This function is only used by widget implementations.

For a version that does not take a transform, see size_allocate.

Parameters:

• width – new width

• height – new height

• baseline – new baseline, or -1

• transform – transformation to be applied

classmethod bind_template_callback_full(callback_name: str, callback_symbol: Callable[[], None]) → None

Parameters:

• callback_name

• callback_symbol

classmethod bind_template_child_full(name: str, internal_child: bool, struct_offset: int) → None

Parameters:

• name

• internal_child

• struct_offset

child_focus(direction: DirectionType) → bool

Called by widgets as the user moves around the window using keyboard shortcuts.

The direction argument indicates what kind of motion is taking place (up, down, left, right, tab forward, tab backward).

This function calls the focus virtual function; widgets can override the virtual function in order to implement appropriate focus behavior.

The default focus() virtual function for a widget should return true if moving in direction left the focus on a focusable location inside that widget, and false if moving in direction moved the focus outside the widget. When returning true, widgets normally call grab_focus to place the focus accordingly; when returning false, they don’t modify the current focus location.

This function is used by custom widget implementations; if you’re writing an app, you’d use grab_focus to move the focus to a particular widget.

Parameters:

direction – direction of focus movement

compute_bounds(target: Widget) → tuple[bool, Rect]

Computes the bounds for widget in the coordinate space of target.

The bounds of widget are (the bounding box of) the region that it is expected to draw in. See the coordinate system overview to learn more.

If the operation is successful, true is returned. If widget has no bounds or the bounds cannot be expressed in target's coordinate space (for example if both widgets are in different windows), false is returned and bounds is set to the zero rectangle.

It is valid for widget and target to be the same widget.

Parameters:

target – the target widget

compute_expand(orientation: Orientation) → bool

Computes whether a parent widget should give this widget extra space when possible.

Widgets with children should check this, rather than looking at get_hexpand or get_vexpand.

This function already checks whether the widget is visible, so visibility does not need to be checked separately. Non-visible widgets are not expanded.

The computed expand value uses either the expand setting explicitly set on the widget itself, or, if none has been explicitly set, the widget may expand if some of its children do.

Parameters:

orientation – expand direction

compute_point(target: Widget, point: Point) → tuple[bool, Point]

Translates the given point in widget's coordinates to coordinates in target’s coordinate system.

In order to perform this operation, both widgets must share a a common ancestor. If that is not the case, out_point is set to (0, 0) and false is returned.

Parameters:

• target – the widget to transform into

• point – a point in widget's coordinate system

compute_transform(target: Widget) → tuple[bool, Matrix]

Computes a matrix suitable to describe a transformation from widget's coordinate system into target's coordinate system.

The transform can not be computed in certain cases, for example when widget and target do not share a common ancestor. In that case out_transform gets set to the identity matrix.

To learn more about widget coordinate systems, see the coordinate system overview.

Parameters:

target – the target widget that the matrix will transform to

contains(x: float, y: float) → bool

Tests if a given point is contained in the widget.

The coordinates for (x, y) must be in widget coordinates, so (0, 0) is assumed to be the top left of widget's content area.

Parameters:

• x – X coordinate to test, relative to widget's origin

• y – Y coordinate to test, relative to widget's origin

create_pango_context() → Context

Creates a new PangoContext that is configured for the widget.

The PangoContext will have the appropriate font map, font options, font description, and base direction set.

See also get_pango_context.

create_pango_layout(text: str | None = None) → Layout

Creates a new PangoLayout that is configured for the widget.

The PangoLayout will have the appropriate font map, font description, and base direction set.

If you keep a PangoLayout created in this way around, you need to re-create it when the widgets PangoContext is replaced. This can be tracked by listening to changes of the root property on the widget.

Parameters:

text – text to set on the layout

dispose_template(widget_type: GType) → None

Clears the template children for the widget.

This function is the opposite of init_template, and it is used to clear all the template children from a widget instance. If you bound a template child to a field in the instance structure, or in the instance private data structure, the field will be set to NULL after this function returns.

You should call this function inside the GObjectClass.dispose() implementation of any widget that called init_template. Typically, you will want to call this function last, right before chaining up to the parent type’s dispose implementation, e.g.

static void
some_widget_dispose (GObject *gobject)
{
  SomeWidget *self = SOME_WIDGET (gobject);

  // Clear the template data for SomeWidget
  gtk_widget_dispose_template (GTK_WIDGET (self), SOME_TYPE_WIDGET);

  G_OBJECT_CLASS (some_widget_parent_class)->dispose (gobject);
}

Added in version 4.8.

Parameters:

widget_type – the type of the widget to finalize the template for

do_compute_expand(self, hexpand_p: bool, vexpand_p: bool) → None

Parameters:

• hexpand_p

• vexpand_p

do_contains(self, x: float, y: float) → bool

Parameters:

• x

• y

do_css_changed(self, change: CssStyleChange) → None

Parameters:

change

do_direction_changed(self, previous_direction: TextDirection) → None

Parameters:

previous_direction

do_focus(self, direction: DirectionType) → bool

Parameters:

direction

do_get_request_mode(self) → SizeRequestMode

do_grab_focus(self) → bool

do_hide(self) → None

do_keynav_failed(self, direction: DirectionType) → bool

Parameters:

direction

do_map(self) → None

do_measure(self, orientation: Orientation, for_size: int) → tuple[int, int, int, int]

Parameters:

• orientation

• for_size

do_mnemonic_activate(self, group_cycling: bool) → bool

Parameters:

group_cycling

do_move_focus(self, direction: DirectionType) → None

Parameters:

direction

do_query_tooltip(self, x: int, y: int, keyboard_tooltip: bool, tooltip: Tooltip) → bool

Parameters:

• x

• y

• keyboard_tooltip

• tooltip

do_realize(self) → None

do_root(self) → None

do_set_focus_child(self, child: Widget | None = None) → None

Parameters:

child

do_show(self) → None

do_size_allocate(self, width: int, height: int, baseline: int) → None

Parameters:

• width

• height

• baseline

do_snapshot(self, snapshot: Snapshot) → None

Parameters:

snapshot

do_state_flags_changed(self, previous_state_flags: StateFlags) → None

Parameters:

previous_state_flags

do_system_setting_changed(self, settings: SystemSetting) → None

Parameters:

settings

do_unmap(self) → None

do_unrealize(self) → None

do_unroot(self) → None

drag_check_threshold(start_x: int, start_y: int, current_x: int, current_y: int) → bool

Checks to see if a drag movement has passed the GTK drag threshold.

Parameters:

• start_x – X coordinate of start of drag

• start_y – Y coordinate of start of drag

• current_x – current X coordinate

• current_y – current Y coordinate

error_bell() → None

Notifies the user about an input-related error on the widget.

If the gtk_error_bell setting is true, it calls beep, otherwise it does nothing.

Note that the effect of beep can be configured in many ways, depending on the windowing backend and the desktop environment or window manager that is used.

classmethod get_activate_signal() → int

get_allocated_baseline() → int

Returns the baseline that has currently been allocated to the widget.

This function is intended to be used when implementing handlers for the GtkWidgetClass.snapshot() function, and when allocating child widgets in GtkWidgetClass.size_allocate().

Deprecated since version 4.12: Use get_baseline instead

get_allocated_height() → int

Returns the height that has currently been allocated to the widget.

To learn more about widget sizes, see the coordinate system overview.

Deprecated since version 4.12: Use get_height instead

get_allocated_width() → int

Returns the width that has currently been allocated to the widget.

To learn more about widget sizes, see the coordinate system overview.

Deprecated since version 4.12: Use get_width instead

get_allocation() → Rectangle

Retrieves the widget’s allocation.

Note, when implementing a layout widget: a widget’s allocation will be its “adjusted” allocation, that is, the widget’s parent typically calls size_allocate with an allocation, and that allocation is then adjusted (to handle margin and alignment for example) before assignment to the widget. get_allocation returns the adjusted allocation that was actually assigned to the widget. The adjusted allocation is guaranteed to be completely contained within the size_allocate allocation, however.

So a layout widget is guaranteed that its children stay inside the assigned bounds, but not that they have exactly the bounds the widget assigned.

Deprecated since version 4.12: Use compute_bounds, get_width or get_height instead.

get_ancestor(widget_type: GType) → Widget | None

Gets the first ancestor of the widget with type widget_type.

For example, gtk_widget_get_ancestor (widget, GTK_TYPE_BOX) gets the first GtkBox that’s an ancestor of widget. No reference will be added to the returned widget; it should not be unreferenced.

Note that unlike is_ancestor, this function considers widget to be an ancestor of itself.

Parameters:

widget_type – ancestor type

get_baseline() → int

Returns the baseline that has currently been allocated to the widget.

This function is intended to be used when implementing handlers for the GtkWidgetClass.snapshot() function, and when allocating child widgets in GtkWidgetClass.size_allocate().

Added in version 4.12.

get_can_focus() → bool

Determines whether the input focus can enter the widget or any of its children.

See set_can_focus.

get_can_target() → bool

Queries whether the widget can be the target of pointer events.

get_child_visible() → bool

Gets the value set with set_child_visible.

If you feel a need to use this function, your code probably needs reorganization.

This function is only useful for widget implementations and should never be called by an application.

get_clipboard() → Clipboard

Gets the clipboard object for the widget.

This is a utility function to get the clipboard object for the display that widget is using.

Note that this function always works, even when widget is not realized yet.

get_color() → RGBA

Gets the current foreground color for the widget’s style.

This function should only be used in snapshot implementations that need to do custom drawing with the foreground color.

Added in version 4.10.

get_css_classes() → list[str]

Returns the list of style classes applied to the widget.

get_css_name() → str

Returns the CSS name of the widget.

get_cursor() → Cursor | None

Gets the cursor set on the widget.

See set_cursor for details.

classmethod get_default_direction() → TextDirection

Obtains the default reading direction.

See set_default_direction.

get_direction() → TextDirection

Gets the reading direction for the widget.

See set_direction.

get_display() → Display

Get the display for the window that the widget belongs to.

This function can only be called after the widget has been added to a widget hierarchy with a GtkRoot at the top.

In general, you should only create display-specific resources when a widget has been realized, and you should free those resources when the widget is unrealized.

get_first_child() → Widget | None

Returns the widget’s first child.

This function is primarily meant for widget implementations.

get_focus_child() → Widget | None

Returns the focus child of the widget.

get_focus_on_click() → bool

Returns whether the widget should grab focus when it is clicked with the mouse.

See set_focus_on_click.

get_focusable() → bool

Determines whether the widget can own the input focus.

See set_focusable.

get_font_map() → FontMap | None

Gets the font map of the widget.

See set_font_map.

get_font_options() → FontOptions | None

Returns the cairo_font_options_t of the widget.

Seee set_font_options.

Deprecated since version 4.16: Please do not use it in newly written code

get_frame_clock() → FrameClock | None

Obtains the frame clock for a widget.

The frame clock is a global “ticker” that can be used to drive animations and repaints. The most common reason to get the frame clock is to call get_frame_time, in order to get a time to use for animating. For example you might record the start of the animation with an initial value from get_frame_time, and then update the animation by calling get_frame_time again during each repaint.

request_phase will result in a new frame on the clock, but won’t necessarily repaint any widgets. To repaint a widget, you have to use queue_draw which invalidates the widget (thus scheduling it to receive a draw on the next frame). queue_draw will also end up requesting a frame on the appropriate frame clock.

A widget’s frame clock will not change while the widget is mapped. Reparenting a widget (which implies a temporary unmap) can change the widget’s frame clock.

Unrealized widgets do not have a frame clock.

get_halign() → Align

Gets the horizontal alignment of the widget.

For backwards compatibility reasons this method will never return one of the baseline alignments, but instead it will convert it to fill or center.

Baselines are not supported for horizontal alignment.

get_has_tooltip() → bool

Returns the current value of the has-tooltip property.

get_height() → int

Returns the content height of the widget.

This function returns the height passed to its size-allocate implementation, which is the height you should be using in snapshot.

For pointer events, see contains.

To learn more about widget sizes, see the coordinate system overview.

get_hexpand() → bool

Gets whether the widget would like any available extra horizontal space.

When a user resizes a window, widgets with expand set to true generally receive the extra space. For example, a list or scrollable area or document in your window would often be set to expand.

Widgets with children should use compute_expand rather than this function, to see whether any of its children, has the expand flag set. If any child of a widget wants to expand, the parent may ask to expand also.

This function only looks at the widget’s own hexpand flag, rather than computing whether the entire widget tree rooted at this widget wants to expand.

get_hexpand_set() → bool

Gets whether the hexpand flag has been explicitly set.

If hexpand property is set, then it overrides any computed expand value based on child widgets. If hexpand is not set, then the expand value depends on whether any children of the widget would like to expand.

There are few reasons to use this function, but it’s here for completeness and consistency.

get_last_child() → Widget | None

Returns the widget’s last child.

This function is primarily meant for widget implementations.

get_layout_manager() → LayoutManager | None

Retrieves the layout manager of the widget.

See set_layout_manager.

classmethod get_layout_manager_type() → GType

get_limit_events() → bool

Gets the value of the limit_events property.

Added in version 4.18.

get_mapped() → bool

Returns whether the widget is mapped.

get_margin_bottom() → int

Gets the bottom margin of the widget.

get_margin_end() → int

Gets the end margin of the widget.

get_margin_start() → int

Gets the start margin of the widget.

get_margin_top() → int

Gets the top margin of the widget.

get_name() → str

Retrieves the name of a widget.

See set_name for the significance of widget names.

get_native() → Native | None

Returns the nearest GtkNative ancestor of the widget.

This function will return NULL if the widget is not contained inside a widget tree with a native ancestor.

GtkNative widgets will return themselves here.

get_next_sibling() → Widget | None

Returns the widget’s next sibling.

This function is primarily meant for widget implementations.

get_opacity() → float

Fetches the requested opacity for the widget.

See set_opacity.

get_overflow() → Overflow

Returns the widget’s overflow value.

get_pango_context() → Context

Gets a PangoContext that is configured for the widget.

The PangoContext will have the appropriate font map, font description, and base direction set.

Unlike the context returned by create_pango_context, this context is owned by the widget (it can be used until the screen for the widget changes or the widget is removed from its toplevel), and will be updated to match any changes to the widget’s attributes. This can be tracked by listening to changes of the root property on the widget.

get_parent() → Widget | None

Returns the parent widget of the widget.

get_preferred_size() → tuple[Requisition, Requisition]

Retrieves the minimum and natural size of a widget, taking into account the widget’s preference for height-for-width management.

This is used to retrieve a suitable size by container widgets which do not impose any restrictions on the child placement. It can be used to deduce toplevel window and menu sizes as well as child widgets in free-form containers such as GtkFixed.

Handle with care. Note that the natural height of a height-for-width widget will generally be a smaller size than the minimum height, since the required height for the natural width is generally smaller than the required height for the minimum width.

Use measure if you want to support baseline alignment.

get_prev_sibling() → Widget | None

Returns the widget’s previous sibling.

This function is primarily meant for widget implementations.

get_primary_clipboard() → Clipboard

Gets the primary clipboard of the widget.

This is a utility function to get the primary clipboard object for the display that widget is using.

Note that this function always works, even when widget is not realized yet.

get_realized() → bool

Determines whether the widget is realized.

get_receives_default() → bool

Determines whether the widget is always treated as the default widget within its toplevel when it has the focus, even if another widget is the default.

See set_receives_default.

get_request_mode() → SizeRequestMode

Gets whether the widget prefers a height-for-width layout or a width-for-height layout.

Single-child widgets generally propagate the preference of their child, more complex widgets need to request something either in context of their children or in context of their allocation capabilities.

get_root() → Root | None

Returns the GtkRoot widget of the widget.

This function will return NULL if the widget is not contained inside a widget tree with a root widget.

GtkRoot widgets will return themselves here.

get_scale_factor() → int

Retrieves the internal scale factor that maps from window coordinates to the actual device pixels.

On traditional systems this is 1, on high density outputs, it can be a higher value (typically 2).

See get_scale_factor.

Note that modern systems may support fractional scaling, where the scale factor is not an integer. On such systems, this function will return the next higher integer value, but you probably want to use get_scale to get the fractional scale value.

get_sensitive() → bool

Returns the widget’s sensitivity.

This function returns the value that has been set using set_sensitive).

The effective sensitivity of a widget is however determined by both its own and its parent widget’s sensitivity. See is_sensitive.

get_settings() → Settings

Gets the settings object holding the settings used for the widget.

Note that this function can only be called when the GtkWidget is attached to a toplevel, since the settings object is specific to a particular display. If you want to monitor the widget for changes in its settings, connect to the notify::display signal.

get_size(orientation: Orientation) → int

Returns the content width or height of the widget.

Which dimension is returned depends on orientation.

This is equivalent to calling get_width for horizontal or get_height for vertical, but can be used when writing orientation-independent code, such as when implementing Orientable widgets.

To learn more about widget sizes, see the coordinate system overview.

Parameters:

orientation – the orientation to query

get_size_request() → tuple[int, int]

Gets the size request that was explicitly set for the widget.

A value of -1 stored in width or height indicates that that dimension has not been set explicitly and the natural requisition of the widget will be used instead.

See set_size_request.

To get the size a widget will actually request, call measure instead of this function.

get_state_flags() → StateFlags

Returns the widget state as a flag set.

It is worth mentioning that the effective insensitive state will be returned, that is, also based on parent insensitivity, even if widget itself is sensitive.

Also note that if you are looking for a way to obtain the StateFlags to pass to a StyleContext method, you should look at get_state.

get_style_context() → StyleContext

Returns the style context associated to the widget.

The returned object is guaranteed to be the same for the lifetime of widget.

Deprecated since version 4.10: Style contexts will be removed in GTK 5

get_template_child(widget_type: GType, name: str) → Object

Fetches an object build from the template XML for widget_type in the widget.

This will only report children which were previously declared with bind_template_child_full or one of its variants.

This function is only meant to be called for code which is private to the widget_type which declared the child and is meant for language bindings which cannot easily make use of the GObject structure offsets.

Parameters:

• widget_type – The type of the widget class that defines the child in the template

• name – ID of the child defined in the template XML

get_tooltip_markup() → str | None

Gets the contents of the tooltip for the widget.

If the tooltip has not been set using set_tooltip_markup, this function returns NULL.

get_tooltip_text() → str | None

Gets the contents of the tooltip for the widget.

If the widget's tooltip was set using set_tooltip_markup, this function will return the escaped text.

get_valign() → Align

Gets the vertical alignment of the widget.

get_vexpand() → bool

Gets whether the widget would like any available extra vertical space.

See get_hexpand for more detail.

get_vexpand_set() → bool

Gets whether the vexpand flag has been explicitly set.

See get_hexpand_set for more detail.

get_visible() → bool

Determines whether the widget is visible.

If you want to take into account whether the widget’s parent is also marked as visible, use is_visible instead.

This function does not check if the widget is obscured in any way.

See set_visible.

get_width() → int

Returns the content width of the widget.

This function returns the width passed to its size-allocate implementation, which is the width you should be using in snapshot.

For pointer events, see contains.

To learn more about widget sizes, see the coordinate system overview.

grab_focus() → bool

Causes widget to have the keyboard focus for the window that it belongs to.

If widget is not focusable, or its grab_focus implementation cannot transfer the focus to a descendant of widget that is focusable, it will not take focus and false will be returned.

Calling grab_focus on an already focused widget is allowed, should not have an effect, and return true.

has_css_class(css_class: str) → bool

Returns whether a style class is currently applied to the widget.

Parameters:

css_class – style class, without the leading period

has_default() → bool

Determines whether the widget is the current default widget within its toplevel.

has_focus() → bool

Determines if the widget has the global input focus.

See is_focus for the difference between having the global input focus, and only having the focus within a toplevel.

has_visible_focus() → bool

Determines if the widget should show a visible indication that it has the global input focus.

This is a convenience function that takes into account whether focus indication should currently be shown in the toplevel window of widget. See get_focus_visible for more information about focus indication.

To find out if the widget has the global input focus, use has_focus.

hide() → None

Reverses the effects of [method.Gtk.Widget.show].

This is causing the widget to be hidden (invisible to the user).

Deprecated since version 4.10: Use set_visible instead

in_destruction() → bool

Returns whether the widget is currently being destroyed.

This information can sometimes be used to avoid doing unnecessary work.

init_template() → None

Creates and initializes child widgets defined in templates.

This function must be called in the instance initializer for any class which assigned itself a template using set_template.

It is important to call this function in the instance initializer of a widget subclass and not in GObject.constructed() or GObject.constructor() for two reasons:

• derived widgets will assume that the composite widgets

  defined by its parent classes have been created in their relative instance initializers

• when calling g_object_new() on a widget with composite templates,

  it’s important to build the composite widgets before the construct properties are set. Properties passed to g_object_new() should take precedence over properties set in the private template XML

A good rule of thumb is to call this function as the first thing in an instance initialization function.

insert_action_group(name: str, group: ActionGroup | None = None) → None

Inserts an action group into the widget’s actions.

Children of widget that implement Actionable can then be associated with actions in group by setting their “action-name” to prefix.``action-name``.

Note that inheritance is defined for individual actions. I.e. even if you insert a group with prefix prefix, actions with the same prefix will still be inherited from the parent, unless the group contains an action with the same name.

If group is NULL, a previously inserted group for name is removed from widget.

Parameters:

• name – the prefix for actions in group

• group – an action group

insert_after(parent: Widget, previous_sibling: Widget | None = None) → None

Sets the parent widget of the widget.

In contrast to set_parent, this function inserts widget at a specific position into the list of children of the parent widget.

It will be placed after previous_sibling, or at the beginning if previous_sibling is NULL.

After calling this function, gtk_widget_get_prev_sibling (widget) will return previous_sibling.

If parent is already set as the parent widget of widget, this function can also be used to reorder widget in the child widget list of parent.

This function is primarily meant for widget implementations; if you are just using a widget, you must use its own API for adding children.

Parameters:

• parent – the parent widget to insert widget into

• previous_sibling – the new previous sibling of widget

insert_before(parent: Widget, next_sibling: Widget | None = None) → None

Sets the parent widget of the widget.

In contrast to set_parent, this function inserts widget at a specific position into the list of children of the parent widget.

It will be placed before next_sibling, or at the end if next_sibling is NULL.

After calling this function, gtk_widget_get_next_sibling (widget) will return next_sibling.

If parent is already set as the parent widget of widget, this function can also be used to reorder widget in the child widget list of parent.

This function is primarily meant for widget implementations; if you are just using a widget, you must use its own API for adding children.

Parameters:

• parent – the parent widget to insert widget into

• next_sibling – the new next sibling of widget

classmethod install_action(action_name: str, parameter_type: str | None, activate: Callable[[Widget, str, Variant | None], None]) → None

Parameters:

• action_name

• parameter_type

• activate

classmethod install_property_action(action_name: str, property_name: str) → None

Parameters:

• action_name

• property_name

is_ancestor(ancestor: Widget) → bool

Determines whether the widget is a descendent of ancestor.

Parameters:

ancestor – another GtkWidget

is_drawable() → bool

Determines whether the widget can be drawn to.

A widget can be drawn if it is mapped and visible.

is_focus() → bool

Determines if the widget is the focus widget within its toplevel.

This does not mean that the has_focus property is necessarily set; has_focus will only be set if the toplevel widget additionally has the global input focus.

is_sensitive() → bool

Returns the widget’s effective sensitivity.

This means it is sensitive itself and also its parent widget is sensitive.

is_visible() → bool

Determines whether the widget and all its parents are marked as visible.

This function does not check if the widget is obscured in any way.

See also get_visible and set_visible.

keynav_failed(direction: DirectionType) → bool

Emits the keynav_failed signal on the widget.

This function should be called whenever keyboard navigation within a single widget hits a boundary.

The return value of this function should be interpreted in a way similar to the return value of child_focus. When true is returned, stay in the widget, the failed keyboard navigation is ok and/or there is nowhere we can/should move the focus to. When false is returned, the caller should continue with keyboard navigation outside the widget, e.g. by calling child_focus on the widget’s toplevel.

The default keynav_failed handler returns false for tab-forward and tab-backward. For the other values of DirectionType it returns true.

Whenever the default handler returns true, it also calls error_bell to notify the user of the failed keyboard navigation.

A use case for providing an own implementation of ::keynav-failed (either by connecting to it or by overriding it) would be a row of Entry widgets where the user should be able to navigate the entire row with the cursor keys, as e.g. known from user interfaces that require entering license keys.

Parameters:

direction – direction of focus movement

list_mnemonic_labels() → list[Widget]

Returns the widgets for which this widget is the target of a mnemonic.

Typically, these widgets will be labels. See, for example, set_mnemonic_widget.

The widgets in the list are not individually referenced. If you want to iterate through the list and perform actions involving callbacks that might destroy the widgets, you must call g_list_foreach (result, (GFunc)g_object_ref, NULL) first, and then unref all the widgets afterwards.

map() → None

Causes a widget to be mapped if it isn’t already.

This function is only for use in widget implementations.

measure(orientation: Orientation, for_size: int) → tuple[int, int, int, int]

Measures widget in the orientation orientation and for the given for_size.

As an example, if orientation is HORIZONTAL and for_size is 300, this functions will compute the minimum and natural width of widget if it is allocated at a height of 300 pixels.

See GtkWidget’s geometry management section for a more details on implementing GtkWidgetClass.measure().

Parameters:

• orientation – the orientation to measure

• for_size – Size for the opposite of orientation, i.e. if orientation is HORIZONTAL, this is the height the widget should be measured with. The VERTICAL case is analogous. This way, both height-for-width and width-for-height requests can be implemented. If no size is known, -1 can be passed.

mnemonic_activate(group_cycling: bool) → bool

Emits the mnemonic_activate signal.

Parameters:

group_cycling – true if there are other widgets with the same mnemonic

observe_children() → ListModel

Returns a list model to track the children of the widget.

Calling this function will enable extra internal bookkeeping to track children and emit signals on the returned listmodel. It may slow down operations a lot.

Applications should try hard to avoid calling this function because of the slowdowns.

observe_controllers() → ListModel

Returns a list model to track the event controllers of the widget.

Calling this function will enable extra internal bookkeeping to track controllers and emit signals on the returned listmodel. It may slow down operations a lot.

Applications should try hard to avoid calling this function because of the slowdowns.

pick(x: float, y: float, flags: PickFlags) → Widget | None

Finds the descendant of the widget closest to a point.

The point (x, y) must be given in widget coordinates, so (0, 0) is assumed to be the top left of widget's content area.

Usually widgets will return NULL if the given coordinate is not contained in widget checked via contains. Otherwise they will recursively try to find a child that does not return NULL. Widgets are however free to customize their picking algorithm.

This function is used on the toplevel to determine the widget below the mouse cursor for purposes of hover highlighting and delivering events.

Parameters:

• x – x coordinate to test, relative to widget's origin

• y – y coordinate to test, relative to widget's origin

• flags – flags to influence what is picked

classmethod query_action(index_: int) → tuple[bool, GType, str, VariantType | None, str | None]

Parameters:

index

queue_allocate() → None

Flags the widget for a rerun of the size_allocate function.

Use this function instead of queue_resize when the widget's size request didn’t change but it wants to reposition its contents.

An example user of this function is set_halign.

This function is only for use in widget implementations.

queue_draw() → None

Schedules this widget to be redrawn.

The redraw will happen in the paint phase of the current or the next frame.

This means widget's snapshot implementation will be called.

queue_resize() → None

Flags a widget to have its size renegotiated.

This should be called when a widget for some reason has a new size request. For example, when you change the text in a Label, the label queues a resize to ensure there’s enough space for the new text.

Note that you cannot call queue_resize() on a widget from inside its implementation of the size_allocate virtual method. Calls to queue_resize() from inside size_allocate will be silently ignored.

This function is only for use in widget implementations.

realize() → None

Creates the GDK resources associated with a widget.

Normally realization happens implicitly; if you show a widget and all its parent containers, then the widget will be realized and mapped automatically.

Realizing a widget requires all the widget’s parent widgets to be realized; calling this function realizes the widget’s parents in addition to widget itself. If a widget is not yet inside a toplevel window when you realize it, bad things will happen.

This function is primarily used in widget implementations, and isn’t very useful otherwise. Many times when you think you might need it, a better approach is to connect to a signal that will be called after the widget is realized automatically, such as realize.

remove_controller(controller: EventController) → None

Removes an event controller from the widget.

The removed event controller will not receive any more events, and should not be used again.

Widgets will remove all event controllers automatically when they are destroyed, there is normally no need to call this function.

Parameters:

controller – an event controller

remove_css_class(css_class: str) → None

Removes a style from the widget.

After this, the style of widget will stop matching for css_class.

Parameters:

css_class – style class to remove from widget, without the leading period

remove_mnemonic_label(label: Widget) → None

Removes a widget from the list of mnemonic labels for this widget.

See list_mnemonic_labels.

The widget must have previously been added to the list with add_mnemonic_label.

Parameters:

label – a widget that is a mnemonic label for widget

remove_tick_callback(id: int) → None

Removes a tick callback previously registered with add_tick_callback.

Parameters:

id – an ID returned by add_tick_callback

classmethod set_accessible_role(accessible_role: AccessibleRole) → None

Parameters:

accessible_role

classmethod set_activate_signal(signal_id: int) → None

Parameters:

signal_id

classmethod set_activate_signal_from_name(signal_name: str) → None

Parameters:

signal_name

set_can_focus(can_focus: bool) → None

Sets whether the input focus can enter the widget or any of its children.

Applications should set can_focus to false to mark a widget as for pointer/touch use only.

Note that having can_focus be true is only one of the necessary conditions for being focusable. A widget must also be sensitive and focusable and not have an ancestor that is marked as not can-focus in order to receive input focus.

See grab_focus for actually setting the input focus on a widget.

Parameters:

can_focus – whether the input focus can enter the widget or any of its children

set_can_target(can_target: bool) → None

Sets whether the widget can be the target of pointer events.

Parameters:

can_target – whether this widget should be able to receive pointer events

set_child_visible(child_visible: bool) → None

Sets whether the widget should be mapped along with its parent.

The child visibility can be set for widget before it is added to a container with set_parent, to avoid mapping children unnecessary before immediately unmapping them. However it will be reset to its default state of true when the widget is removed from a container.

Note that changing the child visibility of a widget does not queue a resize on the widget. Most of the time, the size of a widget is computed from all visible children, whether or not they are mapped. If this is not the case, the container can queue a resize itself.

This function is only useful for widget implementations and should never be called by an application.

Parameters:

child_visible – whether widget should be mapped along with its parent

set_css_classes(classes: list[str]) → None

Replaces the current style classes of the widget with classes.

Parameters:

classes – NULL-terminated list of style classes

classmethod set_css_name(name: str) → None

Parameters:

name

set_cursor(cursor: Cursor | None = None) → None

Sets the cursor to be shown when the pointer hovers over the widget.

If the cursor is NULL, widget will use the cursor inherited from its parent.

Parameters:

cursor – the new cursor

set_cursor_from_name(name: str | None = None) → None

Sets the cursor to be shown when the pointer hovers over the widget.

This is a utility function that creates a cursor via new_from_name and then sets it on widget with set_cursor. See those functions for details.

On top of that, this function allows name to be NULL, which will do the same as calling set_cursor with a NULL cursor.

Parameters:

name – the name of the cursor

classmethod set_default_direction(dir: TextDirection) → None

Sets the default reading direction for widgets.

See set_direction.

Parameters:

dir – the new default direction, either ltr or rtl

set_direction(dir: TextDirection) → None

Sets the reading direction on the widget.

This direction controls the primary direction for widgets containing text, and also the direction in which the children of a container are packed. The ability to set the direction is present in order so that correct localization into languages with right-to-left reading directions can be done.

Generally, applications will let the default reading direction prevail, except for widgets where the children are arranged in an order that is explicitly visual rather than logical (such as buttons for text justification).

If the direction is set to none, then the value set by set_default_direction will be used.

Parameters:

dir – the new direction

set_focus_child(child: Widget | None = None) → None

Set the focus child of the widget.

This function is only suitable for widget implementations. If you want a certain widget to get the input focus, call grab_focus on it.

Parameters:

child – a direct child widget of widget or NULL to unset the focus child

set_focus_on_click(focus_on_click: bool) → None

Sets whether the widget should grab focus when it is clicked with the mouse.

Making mouse clicks not grab focus is useful in places like toolbars where you don’t want the keyboard focus removed from the main area of the application.

Parameters:

focus_on_click – whether the widget should grab focus when clicked with the mouse

set_focusable(focusable: bool) → None

Sets whether the widget can own the input focus.

Widget implementations should set focusable to true in their init() function if they want to receive keyboard input.

Note that having focusable be true is only one of the necessary conditions for being focusable. A widget must also be sensitive and can-focus and not have an ancestor that is marked as not can-focus in order to receive input focus.

See grab_focus for actually setting the input focus on a widget.

Parameters:

focusable – whether or not widget can own the input focus

set_font_map(font_map: FontMap | None = None) → None

Sets the font map to use for text rendering in the widget.

The font map is the object that is used to look up fonts. Setting a custom font map can be useful in special situations, e.g. when you need to add application-specific fonts to the set of available fonts.

When not set, the widget will inherit the font map from its parent.

Parameters:

font_map – a PangoFontMap

set_font_options(options: FontOptions | None = None) → None

Sets the cairo_font_options_t used for text rendering in the widget.

When not set, the default font options for the GdkDisplay will be used.

Deprecated since version 4.16: Please do not use it in newly written code

Parameters:

options – a cairo_font_options_t struct to unset any previously set default font options

set_halign(align: Align) → None

Sets the horizontal alignment of the widget.

Parameters:

align – the horizontal alignment

set_has_tooltip(has_tooltip: bool) → None

Sets the has-tooltip property on the widget.

Parameters:

has_tooltip – whether or not widget has a tooltip

set_hexpand(expand: bool) → None

Sets whether the widget would like any available extra horizontal space.

When a user resizes a window, widgets with expand set to true generally receive the extra space. For example, a list or scrollable area or document in your window would often be set to expand.

Call this function to set the expand flag if you would like your widget to become larger horizontally when the window has extra room.

By default, widgets automatically expand if any of their children want to expand. (To see if a widget will automatically expand given its current children and state, call compute_expand. A widget can decide how the expandability of children affects its own expansion by overriding the compute_expand virtual method on GtkWidget.).

Setting hexpand explicitly with this function will override the automatic expand behavior.

This function forces the widget to expand or not to expand, regardless of children. The override occurs because set_hexpand sets the hexpand-set property (see set_hexpand_set) which causes the widget’s hexpand value to be used, rather than looking at children and widget state.

Parameters:

expand – whether to expand

set_hexpand_set(set: bool) → None

Sets whether the hexpand flag will be used.

The hexpand_set property will be set automatically when you call set_hexpand to set hexpand, so the most likely reason to use this function would be to unset an explicit expand flag.

If hexpand is set, then it overrides any computed expand value based on child widgets. If hexpand is not set, then the expand value depends on whether any children of the widget would like to expand.

There are few reasons to use this function, but it’s here for completeness and consistency.

Parameters:

set – value for hexpand-set property

set_layout_manager(layout_manager: LayoutManager | None = None) → None

Sets the layout manager to use for measuring and allocating children of the widget.

Parameters:

layout_manager – a layout manager

classmethod set_layout_manager_type(type: GType) → None

Parameters:

type

set_limit_events(limit_events: bool) → None

Sets whether the widget acts like a modal dialog, with respect to event delivery.

Added in version 4.18.

Parameters:

limit_events – whether to limit events

set_margin_bottom(margin: int) → None

Sets the bottom margin of the widget.

Parameters:

margin – the bottom margin

set_margin_end(margin: int) → None

Sets the end margin of the widget.

Parameters:

margin – the end margin

set_margin_start(margin: int) → None

Sets the start margin of the widget.

Parameters:

margin – the start margin

set_margin_top(margin: int) → None

Sets the top margin of the widget.

Parameters:

margin – the top margin

set_name(name: str) → None

Sets a widgets name.

Setting a name allows you to refer to the widget from a CSS file. You can apply a style to widgets with a particular name in the CSS file. See the documentation for the CSS syntax (on the same page as the docs for StyleContext.

Note that the CSS syntax has certain special characters to delimit and represent elements in a selector (period, #, >, *…), so using these will make your widget impossible to match by name. Any combination of alphanumeric symbols, dashes and underscores will suffice.

Parameters:

name – name for the widget

set_opacity(opacity: float) → None

Requests the widget to be rendered partially transparent.

An opacity of 0 is fully transparent and an opacity of 1 is fully opaque.

Opacity works on both toplevel widgets and child widgets, although there are some limitations: For toplevel widgets, applying opacity depends on the capabilities of the windowing system. On X11, this has any effect only on X displays with a compositing manager, see is_composited. On Windows and Wayland it will always work, although setting a window’s opacity after the window has been shown may cause some flicker.

Note that the opacity is inherited through inclusion — if you set a toplevel to be partially translucent, all of its content will appear translucent, since it is ultimatively rendered on that toplevel. The opacity value itself is not inherited by child widgets (since that would make widgets deeper in the hierarchy progressively more translucent). As a consequence, Popover instances and other Native widgets with their own surface will use their own opacity value, and thus by default appear non-translucent, even if they are attached to a toplevel that is translucent.

Parameters:

opacity – desired opacity, between 0 and 1

set_overflow(overflow: Overflow) → None

Sets how the widget treats content that is drawn outside the it’s content area.

See the definition of Overflow for details.

This setting is provided for widget implementations and should not be used by application code.

The default value is visible.

Parameters:

overflow – desired overflow value

set_parent(parent: Widget) → None

Sets the parent widget of the widget.

This takes care of details such as updating the state and style of the child to reflect its new location and resizing the parent. The opposite function is unparent.

This function is useful only when implementing subclasses of GtkWidget.

Parameters:

parent – parent widget

set_receives_default(receives_default: bool) → None

Sets whether the widget will be treated as the default widget within its toplevel when it has the focus, even if another widget is the default.

Parameters:

receives_default – whether or not widget can be a default widget

set_sensitive(sensitive: bool) → None

Sets the sensitivity of the widget.

A widget is sensitive if the user can interact with it. Insensitive widgets are “grayed out” and the user can’t interact with them. Insensitive widgets are known as “inactive”, “disabled”, or “ghosted” in some other toolkits.

Parameters:

sensitive – true to make the widget sensitive

set_size_request(width: int, height: int) → None

Sets the minimum size of the widget.

That is, the widget’s size request will be at least width by height. You can use this function to force a widget to be larger than it normally would be.

In most cases, set_default_size is a better choice for toplevel windows than this function; setting the default size will still allow users to shrink the window. Setting the size request will force them to leave the window at least as large as the size request.

Note the inherent danger of setting any fixed size - themes, translations into other languages, different fonts, and user action can all change the appropriate size for a given widget. So, it is basically impossible to hardcode a size that will always work.

The size request of a widget is the smallest size a widget can accept while still functioning well and drawing itself correctly. However in some strange cases a widget may be allocated less than its requested size, and in many cases a widget may be allocated more space than it requested.

If the size request in a given direction is -1 (unset), then the “natural” size request of the widget will be used instead.

The size request set here does not include any margin from the properties margin_start, margin_end, margin_top, and margin_bottom, but it does include pretty much all other padding or border properties set by any subclass of GtkWidget.

Parameters:

• width – width widget should request, or -1 to unset

• height – height widget should request, or -1 to unset

set_state_flags(flags: StateFlags, clear: bool) → None

Turns on flag values in the current widget state.

Typical widget states are insensitive, prelighted, etc.

This function accepts the values dir-ltr and dir-rtl but ignores them. If you want to set the widget’s direction, use set_direction.

This function is for use in widget implementations.

Parameters:

• flags – state flags to turn on

• clear – whether to clear state before turning on flags

classmethod set_template(template_bytes: Bytes) → None

Parameters:

template_bytes

classmethod set_template_from_resource(resource_name: str) → None

Parameters:

resource_name

classmethod set_template_scope(scope: BuilderScope) → None

Parameters:

scope

set_tooltip_markup(markup: str | None = None) → None

Sets the contents of the tooltip for widget.

markup must contain Pango markup.

This function will take care of setting the has_tooltip as a side effect, and of the default handler for the query_tooltip signal.

See also set_markup.

Parameters:

markup – the contents of the tooltip for widget

set_tooltip_text(text: str | None = None) → None

Sets the contents of the tooltip for the widget.

If text contains any markup, it will be escaped.

This function will take care of setting has_tooltip as a side effect, and of the default handler for the query_tooltip signal.

See also set_text.

Parameters:

text – the contents of the tooltip for widget

set_valign(align: Align) → None

Sets the vertical alignment of the widget.

Parameters:

align – the vertical alignment

set_vexpand(expand: bool) → None

Sets whether the widget would like any available extra vertical space.

See set_hexpand for more detail.

Parameters:

expand – whether to expand

set_vexpand_set(set: bool) → None

Sets whether the vexpand flag will be used.

See set_hexpand_set for more detail.

Parameters:

set – value for vexpand-set property

set_visible(visible: bool) → None

Sets the visibility state of widget.

Note that setting this to true doesn’t mean the widget is actually viewable, see get_visible.

Parameters:

visible – whether the widget should be shown or not

should_layout() → bool

Returns whether the widget should contribute to the measuring and allocation of its parent.

This is false for invisible children, but also for children that have their own surface, such as Popover instances.

show() → None

Flags a widget to be displayed.

Any widget that isn’t shown will not appear on the screen.

Remember that you have to show the containers containing a widget, in addition to the widget itself, before it will appear onscreen.

When a toplevel widget is shown, it is immediately realized and mapped; other shown widgets are realized and mapped when their toplevel widget is realized and mapped.

Deprecated since version 4.10: Use set_visible instead

size_allocate(allocation: Rectangle, baseline: int) → None

Allocates widget with a transformation that translates the origin to the position in allocation.

This is a simple form of allocate.

Parameters:

• allocation – position and size to be allocated to widget

• baseline – the baseline of the child, or -1

snapshot_child(child: Widget, snapshot: Snapshot) → None

Snapshots a child of the widget.

When a widget receives a call to the snapshot function, it must send synthetic snapshot calls to all children. This function provides a convenient way of doing this. A widget, when it receives a call to its snapshot function, calls snapshot_child() once for each child, passing in the snapshot the widget received.

This function takes care of translating the origin of snapshot, and deciding whether the child needs to be snapshot.

It does nothing for children that implement GtkNative.

Parameters:

• child – a child of widget

• snapshot – snapshot as passed to the widget. In particular, no calls to translate or other transform calls should have been made

translate_coordinates(dest_widget: Widget, src_x: float, src_y: float) → tuple[bool, float, float]

Translates coordinates relative to src_widget’s allocation to coordinates relative to dest_widget’s allocations.

In order to perform this operation, both widget must share a common ancestor. If that is not the case, dest_x and dest_y are set to 0 and false is returned.

Deprecated since version 4.12: Use compute_point instead

Parameters:

• dest_widget – another widget

• src_x – X position in widget coordinates of src_widget

• src_y – Y position in widget coordinates of src_widget

trigger_tooltip_query() → None

Triggers a tooltip query on the display of the widget.

unmap() → None

Causes a widget to be unmapped if it’s currently mapped.

This function is only for use in widget implementations.

unparent() → None

Removes widget from its parent.

This function is only for use in widget implementations, typically in dispose.

unrealize() → None

Causes a widget to be unrealized.

This frees all GDK resources associated with the widget.

This function is only useful in widget implementations.

unset_state_flags(flags: StateFlags) → None

Turns off flag values for the current widget state.

See set_state_flags.

This function is for use in widget implementations.

Parameters:

flags – state flags to turn off

Properties

class Widget

props.can_focus: bool

Whether the widget or any of its descendents can accept the input focus.

This property is meant to be set by widget implementations, typically in their instance init function.

props.can_target: bool

Whether the widget can receive pointer events.

props.css_classes: list[str]

A list of css classes applied to this widget.

props.css_name: str

The name of this widget in the CSS tree.

This property is meant to be set by widget implementations, typically in their instance init function.

props.cursor: Cursor

The cursor used by widget.

props.focus_on_click: bool

Whether the widget should grab focus when it is clicked with the mouse.

This property is only relevant for widgets that can take focus.

props.focusable: bool

Whether this widget itself will accept the input focus.

props.halign: Align

How to distribute horizontal space if widget gets extra space.

props.has_default: bool

Whether the widget is the default widget.

props.has_focus: bool

Whether the widget has the input focus.

props.has_tooltip: bool

Enables or disables the emission of the query_tooltip signal on widget.

A true value indicates that widget can have a tooltip, in this case the widget will be queried using query_tooltip to determine whether it will provide a tooltip or not.

props.height_request: int

Overrides for height request of the widget.

If this is -1, the natural request will be used.

props.hexpand: bool

Whether to expand horizontally.

props.hexpand_set: bool

Whether to use the hexpand property.

props.layout_manager: LayoutManager

The LayoutManager instance to use to compute the preferred size of the widget, and allocate its children.

This property is meant to be set by widget implementations, typically in their instance init function.

props.limit_events: bool

Makes this widget act like a modal dialog, with respect to event delivery.

Global event controllers will not handle events with targets inside the widget, unless they are set up to ignore propagation limits. See set_propagation_limit.

Added in version 4.18.

props.margin_bottom: int

Margin on bottom side of widget.

This property adds margin outside of the widget’s normal size request, the margin will be added in addition to the size from set_size_request for example.

props.margin_end: int

Margin on end of widget, horizontally.

This property supports left-to-right and right-to-left text directions.

This property adds margin outside of the widget’s normal size request, the margin will be added in addition to the size from set_size_request for example.

props.margin_start: int

Margin on start of widget, horizontally.

This property supports left-to-right and right-to-left text directions.

This property adds margin outside of the widget’s normal size request, the margin will be added in addition to the size from set_size_request for example.

props.margin_top: int

Margin on top side of widget.

This property adds margin outside of the widget’s normal size request, the margin will be added in addition to the size from set_size_request for example.

props.name: str

The name of the widget.

props.opacity: float

The requested opacity of the widget.

props.overflow: Overflow

How content outside the widget’s content area is treated.

This property is meant to be set by widget implementations, typically in their instance init function.

props.parent: Widget

The parent widget of this widget.

props.receives_default: bool

Whether the widget will receive the default action when it is focused.

props.root: Root

The GtkRoot widget of the widget tree containing this widget.

This will be NULL if the widget is not contained in a root widget.

props.scale_factor: int

The scale factor of the widget.

props.sensitive: bool

Whether the widget responds to input.

props.tooltip_markup: str

Sets the text of tooltip to be the given string, which is marked up with Pango markup.

Also see set_markup.

This is a convenience property which will take care of getting the tooltip shown if the given string is not NULL: has_tooltip will automatically be set to true and there will be taken care of query_tooltip in the default signal handler.

Note that if both tooltip_text and tooltip_markup are set, the last one wins.

props.tooltip_text: str

Sets the text of tooltip to be the given string.

Also see set_text.

This is a convenience property which will take care of getting the tooltip shown if the given string is not NULL: has_tooltip will automatically be set to true and there will be taken care of query_tooltip in the default signal handler.

Note that if both tooltip_text and tooltip_markup are set, the last one wins.

props.valign: Align

How to distribute vertical space if widget gets extra space.

props.vexpand: bool

Whether to expand vertically.

props.vexpand_set: bool

Whether to use the vexpand property.

props.visible: bool

Whether the widget is visible.

props.width_request: int

Overrides for width request of the widget.

If this is -1, the natural request will be used.

Signals

class Widget.signals

destroy() → None

Signals that all holders of a reference to the widget should release the reference that they hold.

May result in finalization of the widget if all references are released.

This signal is not suitable for saving widget state.

direction_changed(previous_direction: TextDirection) → None

Emitted when the text direction of a widget changes.

Parameters:

previous_direction – the previous text direction

hide() → None

Emitted when widget is hidden.

keynav_failed(direction: DirectionType) → bool

Emitted if keyboard navigation fails.

See keynav_failed for details.

Parameters:

direction – the direction of movement

map() → None

Emitted when widget is going to be mapped.

A widget is mapped when the widget is visible (which is controlled with visible) and all its parents up to the toplevel widget are also visible.

The ::map signal can be used to determine whether a widget will be drawn, for instance it can resume an animation that was stopped during the emission of unmap.

mnemonic_activate(group_cycling: bool) → bool

Emitted when a widget is activated via a mnemonic.

The default handler for this signal activates widget if group_cycling is false, or just makes widget grab focus if group_cycling is true.

Parameters:

group_cycling – true if there are other widgets with the same mnemonic

move_focus(direction: DirectionType) → None

Emitted when the focus is moved.

The ::move-focus signal is a keybinding signal.

The default bindings for this signal are Tab to move forward, and Shift+:kbd:Tab to move backward.

Parameters:

direction – the direction of the focus move

query_tooltip(x: int, y: int, keyboard_mode: bool, tooltip: Tooltip) → bool

Emitted when the widget’s tooltip is about to be shown.

This happens when the has_tooltip property is true and the hover timeout has expired with the cursor hovering above widget; or emitted when widget got focus in keyboard mode.

Using the given coordinates, the signal handler should determine whether a tooltip should be shown for widget. If this is the case true should be returned, false otherwise. Note that if keyboard_mode is true, the values of x and y are undefined and should not be used.

The signal handler is free to manipulate tooltip with the therefore destined function calls.

Parameters:

• x – the x coordinate of the cursor position in widget coordinates

• y – the y coordinate of the cursor position in widget coordinates

• keyboard_mode – true if the tooltip was triggered using the keyboard

• tooltip – a GtkTooltip

realize() → None

Emitted when widget is associated with a GdkSurface.

This means that realize has been called or the widget has been mapped (that is, it is going to be drawn).

show() → None

Emitted when widget is shown.

state_flags_changed(flags: StateFlags) → None

Emitted when the widget state changes.

See get_state_flags.

Parameters:

flags – the previous state flags

unmap() → None

Emitted when widget is going to be unmapped.

A widget is unmapped when either it or any of its parents up to the toplevel widget have been set as hidden.

As ::unmap indicates that a widget will not be shown any longer, it can be used to, for example, stop an animation on the widget.

unrealize() → None

Emitted when the GdkSurface associated with widget is destroyed.

This means that unrealize has been called or the widget has been unmapped (that is, it is going to be hidden).

Virtual Methods

class Widget

do_compute_expand(hexpand_p: bool, vexpand_p: bool) → None

Computes whether a container should give this widget extra space when possible.

Parameters:

• hexpand_p

• vexpand_p

do_contains(x: float, y: float) → bool

Tests if a given point is contained in the widget.

The coordinates for (x, y) must be in widget coordinates, so (0, 0) is assumed to be the top left of widget's content area.

Parameters:

• x – X coordinate to test, relative to widget's origin

• y – Y coordinate to test, relative to widget's origin

do_css_changed(change: CssStyleChange) → None

Vfunc called when the CSS used by widget was changed. Widgets should then discard their caches that depend on CSS and queue resizes or redraws accordingly. The default implementation will take care of this for all the default CSS properties, so implementations must chain up.

Parameters:

change

do_direction_changed(previous_direction: TextDirection) → None

Signal emitted when the text direction of a widget changes.

Parameters:

previous_direction

do_focus(direction: DirectionType) → bool

Vfunc for child_focus()

Parameters:

direction

do_get_request_mode() → SizeRequestMode

Gets whether the widget prefers a height-for-width layout or a width-for-height layout.

Single-child widgets generally propagate the preference of their child, more complex widgets need to request something either in context of their children or in context of their allocation capabilities.

do_grab_focus() → bool

Causes widget to have the keyboard focus for the window that it belongs to.

If widget is not focusable, or its grab_focus implementation cannot transfer the focus to a descendant of widget that is focusable, it will not take focus and false will be returned.

Calling grab_focus on an already focused widget is allowed, should not have an effect, and return true.

do_hide() → None

Reverses the effects of [method.Gtk.Widget.show].

This is causing the widget to be hidden (invisible to the user).

Deprecated since version 4.10: Use set_visible instead

do_keynav_failed(direction: DirectionType) → bool

Emits the keynav_failed signal on the widget.

This function should be called whenever keyboard navigation within a single widget hits a boundary.

The return value of this function should be interpreted in a way similar to the return value of child_focus. When true is returned, stay in the widget, the failed keyboard navigation is ok and/or there is nowhere we can/should move the focus to. When false is returned, the caller should continue with keyboard navigation outside the widget, e.g. by calling child_focus on the widget’s toplevel.

The default keynav_failed handler returns false for tab-forward and tab-backward. For the other values of DirectionType it returns true.

Whenever the default handler returns true, it also calls error_bell to notify the user of the failed keyboard navigation.

A use case for providing an own implementation of ::keynav-failed (either by connecting to it or by overriding it) would be a row of Entry widgets where the user should be able to navigate the entire row with the cursor keys, as e.g. known from user interfaces that require entering license keys.

Parameters:

direction – direction of focus movement

do_map() → None

Causes a widget to be mapped if it isn’t already.

This function is only for use in widget implementations.

do_measure(orientation: Orientation, for_size: int) → tuple[int, int, int, int]

Measures widget in the orientation orientation and for the given for_size.

As an example, if orientation is HORIZONTAL and for_size is 300, this functions will compute the minimum and natural width of widget if it is allocated at a height of 300 pixels.

See GtkWidget’s geometry management section for a more details on implementing GtkWidgetClass.measure().

Parameters:

• orientation – the orientation to measure

• for_size – Size for the opposite of orientation, i.e. if orientation is HORIZONTAL, this is the height the widget should be measured with. The VERTICAL case is analogous. This way, both height-for-width and width-for-height requests can be implemented. If no size is known, -1 can be passed.

do_mnemonic_activate(group_cycling: bool) → bool

Emits the mnemonic_activate signal.

Parameters:

group_cycling – true if there are other widgets with the same mnemonic

do_move_focus(direction: DirectionType) → None

Signal emitted when a change of focus is requested

Parameters:

direction

do_query_tooltip(x: int, y: int, keyboard_tooltip: bool, tooltip: Tooltip) → bool

Signal emitted when “has-tooltip” is True and the hover timeout has expired with the cursor hovering “above” widget; or emitted when widget got focus in keyboard mode.

Parameters:

• x

• y

• keyboard_tooltip

• tooltip

do_realize() → None

Creates the GDK resources associated with a widget.

Normally realization happens implicitly; if you show a widget and all its parent containers, then the widget will be realized and mapped automatically.

Realizing a widget requires all the widget’s parent widgets to be realized; calling this function realizes the widget’s parents in addition to widget itself. If a widget is not yet inside a toplevel window when you realize it, bad things will happen.

This function is primarily used in widget implementations, and isn’t very useful otherwise. Many times when you think you might need it, a better approach is to connect to a signal that will be called after the widget is realized automatically, such as realize.

do_root() → None

Called when the widget gets added to a GtkRoot widget. Must chain up

do_set_focus_child(child: Widget | None = None) → None

Set the focus child of the widget.

This function is only suitable for widget implementations. If you want a certain widget to get the input focus, call grab_focus on it.

Parameters:

child – a direct child widget of widget or NULL to unset the focus child

do_show() → None

Flags a widget to be displayed.

Any widget that isn’t shown will not appear on the screen.

Remember that you have to show the containers containing a widget, in addition to the widget itself, before it will appear onscreen.

When a toplevel widget is shown, it is immediately realized and mapped; other shown widgets are realized and mapped when their toplevel widget is realized and mapped.

Deprecated since version 4.10: Use set_visible instead

do_size_allocate(width: int, height: int, baseline: int) → None

Called to set the allocation, if the widget does not have a layout manager.

Parameters:

• width

• height

• baseline

do_snapshot(snapshot: Snapshot) → None

Vfunc called when a new snapshot of the widget has to be taken.

Parameters:

snapshot

do_state_flags_changed(previous_state_flags: StateFlags) → None

Signal emitted when the widget state changes, see get_state_flags().

Parameters:

previous_state_flags

do_system_setting_changed(settings: SystemSetting) → None

Emitted when a system setting was changed. Must chain up.

Parameters:

settings

do_unmap() → None

Causes a widget to be unmapped if it’s currently mapped.

This function is only for use in widget implementations.

do_unrealize() → None

Causes a widget to be unrealized.

This frees all GDK resources associated with the widget.

This function is only useful in widget implementations.

do_unroot() → None

Called when the widget is about to be removed from its GtkRoot widget. Must chain up

Fields

class Widget

parent_instance

priv