The GTK+ Drawing Model

Windows and events

Programs that run in a windowing system generally create rectangular regions in the screen called windows. Traditional windowing systems do not automatically save the graphical content of windows, and instead ask client programs to repaint those windows whenever it is needed. For example, if a window that is stacked below other windows gets raised to the top, then a client program has to repaint the area that was previously obscured. When the windowing system asks a client program to redraw part of a window, it sends an exposure event to the program for that window.

Here, "windows" means "rectangular regions with automatic clipping", instead of "toplevel application windows". Most windowing systems support nested windows, where the contents of child windows get clipped by the boundaries of their parents. Although GTK+ and GDK in particular may run on a windowing system with no such notion of nested windows, GDK presents the illusion of being under such a system. A toplevel window may contain many subwindows and sub-subwindows, for example, one for the menu bar, one for the document area, one for each scrollbar, and one for the status bar. In addition, controls that receive user input, such as clickable buttons, are likely to have their own subwindows as well.

In practice, most windows in modern GTK+ application are client-side constructs. Only few windows (in particular toplevel windows) are native, which means that they represent a window from the underlying windowing system on which GTK+ is running. For example, on X11 it corresponds to a Window; on Win32, it corresponds to a HANDLE.

Generally, the drawing cycle begins when GTK+ receives an exposure event from the underlying windowing system: if the user drags a window over another one, the windowing system will tell the underlying window that it needs to repaint itself. The drawing cycle can also be initiated when a widget itself decides that it needs to update its display. For example, when the user types a character in a GtkEntry widget, the entry asks GTK+ to queue a redraw operation for itself.

The windowing system generates events for native windows. The GDK interface to the windowing system translates such native events into GdkEvent structures and sends them on to the GTK layer. In turn, the GTK layer finds the widget that corresponds to a particular GdkWindow and emits the corresponding event signals on that widget.

The following sections describe how GTK+ decides which widgets need to be repainted in response to such events, and how widgets work internally in terms of the resources they use from the windowing system.

The frame clock

All GTK+ applications are mainloop-driven, which means that most of the time the app is idle inside a loop that just waits for something to happen and then calls out to the right place when it does. On top of this GTK+ has a frame clock that gives a “pulse” to the application. This clock beats at a steady rate, which is tied to the framerate of the output (this is synced to the monitor via the window manager/compositor). The clock has several phases:

The phases happens in this order and we will always run each phase through before going back to the start.

The Events phase is a long stretch of time between each redraw where we get input events from the user and other events (like e.g. network I/O). Some events, like mouse motion are compressed so that we only get a single mouse motion event per clock cycle.

Once the Events phase is over we pause all external events and run the redraw loop. First is the Update phase, where all animations are run to calculate the new state based on the estimated time the next frame will be visible (available via the frame clock). This often involves geometry changes which drives the next phase, Layout. If there are any changes in widget size requirements we calculate a new layout for the widget hierarchy (i.e. we assign sizes and positions). Then we go to the Paint phase where we redraw the regions of the window that need redrawing.

If nothing requires the Update/Layout/Paint phases we will stay in the Events phase forever, as we don’t want to redraw if nothing changes. Each phase can request further processing in the following phases (e.g. the Update phase will cause there to be layout work, and layout changes cause repaints).

There are multiple ways to drive the clock, at the lowest level you can request a particular phase with gdk_frame_clock_request_phase() which will schedule a clock beat as needed so that it eventually reaches the requested phase. However, in practice most things happen at higher levels:

There are also a lot of implicit triggers of these from the CSS layer (which does animations, resizes and repaints as needed).

Hierarchical drawing

During the Paint phase we will send a single expose event to the toplevel window. The event handler will create a cairo context for the window and emit a GtkWidget::draw() signal on it, which will propagate down the entire widget hierarchy in back-to-front order, using the clipping and transform of the cairo context. This lets each widget draw its content at the right place and time, correctly handling things like partial transparencies and overlapping widgets.

When generating the event, GDK also sets up double buffering to avoid the flickering that would result from each widget drawing itself in turn. the section called “Double buffering” describes the double buffering mechanism in detail.

Normally, there is only a single cairo context which is used in the entire repaint, rather than one per GdkWindow. This means you have to respect (and not reset) existing clip and transformations set on it.

Most widgets, including those that create their own GdkWindows have a transparent background, so they draw on top of whatever widgets are below them. This was not the case in GTK+ 2 where the theme set the background of most widgets to the default background color. (In fact, transparent GdkWindows used to be impossible.)

The whole rendering hierarchy is captured in the call stack, rather than having multiple separate draw emissions, so you can use effects like e.g. cairo_push/pop_group() which will affect all the widgets below you in the hierarchy. This makes it possible to have e.g. partially transparent containers.

Scrolling

Traditionally, GTK+ has used self-copy operations to implement scrolling with native windows. With transparent backgrounds, this no longer works. Instead, we just mark the entire affected area for repainting when these operations are used. This allows (partially) transparent backgrounds, and it also more closely models modern hardware where self-copy operations are problematic (they break the rendering pipeline).

Since the above causes some overhead, we introduce a caching mechanism. Containers that scroll a lot (GtkViewport, GtkTextView, GtkTreeView, etc) allocate an offscreen image during scrolling and render their children to it (which is possible since drawing is fully hierarchical). The offscreen image is a bit larger than the visible area, so most of the time when scrolling it just needs to draw the offscreen in a different position. This matches contemporary graphics hardware much better, as well as allowing efficient transparent backgrounds. In order for this to work such containers need to detect when child widgets are redrawn so that it can update the offscreen. This can be done with the new gdk_window_set_invalidate_handler() function.

Automatic double buffering

It would be inconvenient for all widgets to call gdk_window_begin_paint_region() and gdk_window_end_paint() at the beginning and end of their draw handlers.

To make this easier, GTK+ normally calls gdk_window_begin_paint_region() before emitting the #GtkWidget::draw signal, and then it calls gdk_window_end_paint() after the signal has been emitted. This is convenient for most widgets, as they do not need to worry about creating their own temporary drawing buffers or about calling those functions.

However, some widgets may prefer to disable this kind of automatic double buffering and do things on their own. To do this, call the gtk_widget_set_double_buffered() function in your widget's constructor. Double buffering can only be turned off for widgets that have a native window.

static void
my_widget_init (MyWidget *widget)
{
  ...

  gtk_widget_set_double_buffered (widget, FALSE);

  ...
}

When is it convenient to disable double buffering? Generally, this is the case only if your widget gets drawn in such a way that the different drawing operations do not overlap each other. For example, this may be the case for a simple image viewer: it can just draw the image in a single operation. This would not be the case with a word processor, since it will need to draw and over-draw the page's background, then the background for highlighted text, and then the text itself.

Even if you turn off double buffering on a widget, you can still call gdk_window_begin_paint_region() and gdk_window_end_paint() by hand to use temporary drawing buffers.