AuroraMiddleware/gtk2
1
0
Fork 0
forked from AuroraMiddleware/gtk
Code Issues Pull Requests Projects Releases Wiki Activity

Update drawing model docs

Browse Source 
Include material from Alex' blog post about the modern GTK+
rendering model.
This commit is contained in:
Matthias Clasen 2013-11-10 03:25:48 -05:00
parent eae9513cbd
commit 7806684592
1 changed files with 211 additions and 310 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

521
docs/reference/gtk/drawing-model.xml
View File

@ -30,77 +30,172 @@
background color of all widgets with the same method.
</para>
<para>
Programs that run in a windowing system generally create
rectangular regions in the screen called
<firstterm>windows</firstterm>. 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 <firstterm>exposure event</firstterm> to the
program for that window.
</para>
<para>
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.
</para>
<para>
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 <link
linkend="GtkEntry"><classname>GtkEntry</classname></link>
widget, the entry asks GTK+ to queue a redraw operation for
itself.
</para>
<para>
The following sections describe how GTK+ decides which widgets
need to be repainted, and how widgets work internally in terms
of the resources they use from the windowing system.
</para>
<para>
A <link linkend="GdkWindow"><classname>GdkWindow</classname></link>
represents a window from the underlying windowing system on which GTK+
is running. For example, on X11 it corresponds to a
<type>Window</type>; on Win32, it corresponds to a <type>HANDLE</type>.
The windowing system generates events for these windows. The GDK
interface to the windowing system translates such native events into
<link linkend="GdkEvent"><structname>GdkEvent</structname></link>
structures and sends them on to the GTK layer. In turn, the GTK layer
finds the widget that corresponds to a particular
<classname>GdkWindow</classname> and emits the corresponding event
signals on that widget.
</para>
<refsect2 id="emission of the draw event">
<title>Emission of the draw event</title>
<refsect2 id="drawing model windows">
<title>Windows and events</title>
<para>
When the program needs to redraw a region of a
<classname>GdkWindow</classname>, generates an event of
type <link
linkend="GDK_EVENT_EXPOSE"><constant>GDK_EVENT_EXPOSE</constant></link>
for that window, specifying the region to redraw in the process.
Programs that run in a windowing system generally create
rectangular regions in the screen called
<firstterm>windows</firstterm>. 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 <firstterm>exposure event</firstterm> to the
program for that window.
</para>
<para>
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.
</para>
<para>
In practice, most windows in modern GTK+ application are client-side
constructs. Only few windows (in particular toplevel windows) are
<emphasis>native</emphasis>, 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 <type>Window</type>; on Win32,
it corresponds to a <type>HANDLE</type>.
</para>
<para>
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 <link
linkend="GtkEntry"><classname>GtkEntry</classname></link>
widget, the entry asks GTK+ to queue a redraw operation for
itself.
</para>
<para>
The windowing system generates events for native windows. The GDK
interface to the windowing system translates such native events into
<link linkend="GdkEvent"><structname>GdkEvent</structname></link>
structures and sends them on to the GTK layer. In turn, the GTK layer
finds the widget that corresponds to a particular
<classname>GdkWindow</classname> and emits the corresponding event
signals on that widget.
</para>
<para>
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.
</para>
</refsect2>
<refsect2 id="frameclock">
<title>The frame clock</title>
<para>
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:
<itemizedlist>
<listitem><para>Events</para></listitem>
<listitem><para>Update</para></listitem>
<listitem><para>Layout</para></listitem>
<listitem><para>Paint</para></listitem>
</itemizedlist>
The phases happens in this order and we will always run each
phase through before going back to the start.
</para>
<para>
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.
</para>
<para>
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.
</para>
<para>
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).
</para>
<para>
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:
<itemizedlist>
<listitem><para>
If you are doing an animation, you can use
gtk_widget_add_tick_callback() which will cause a regular
beating of the clock with a callback in the Update phase
until you stop the tick.
</para></listitem>
<listitem><para>
If some state changes that causes the size of your widget
to change you call gtk_widget_queue_resize() which will
request a Layout phase and mark your widget as needing
relayout.
</para></listitem>
<listitem><para>
If some state changes so you need to redraw some area of
your widget you use the normal gtk_widget_queue_draw()
set of functions. These will request a Paint phase and
mark the region as needing redraw.
</para></listitem>
</itemizedlist>
There are also a lot of implicit triggers of these from the
CSS layer (which does animations, resizes and repaints as needed).
</para>
</refsect2>
<refsect2 id="hierarchical-drawing">
<title>Hierarchical drawing</title>
<para>
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.
</para>
<para>
@ -111,218 +206,55 @@
</para>
<para>
When the GTK+ widget layer receives the event, it finds the widget that
corresponds to the window, and causes it to render itself using the
widget's #GtkWidget::draw signal. For this purpose it creates a
<link linkend="#cairo_t">cairo context</link>. It then clips the context
to the area that needs to be drawn. This makes sure that the minimal
amount of work is done if only a small part of the widget needs to be
repainted. After translating the context so that its (0, 0) coordinate
corresponds to the top left corner of the widget, it effectively calls
the widget's <function>gtk_widget_draw</function> function.
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.
</para>
<para>
<function>gtk_widget_draw</function> takes care of drawing the widget
to the cairo context. It first checks that the widget actually needs to
be drawn. Widgets might for example be empty or outside of the cairo
context's clipped area, which would make drawing them not do anything.
Usually they will need to be drawn. In this case, the context will be
clipped to the widget's allocated size and the
<link linkend="GtkWidget::draw">draw signal</link> will be emitted on
the widget which will finally draw the widget.
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.)
</para>
<para>
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.
</para>
</refsect2>
<refsect2 id="window-no-window-widgets">
<title>Window and no-window widgets</title>
<refsect2 id="scrolling drawing model">
<title>Scrolling</title>
<para>
In principle, each widget could have a
<classname>GdkWindow</classname> of its own. With such a
scheme, the drawing cycle would be trivial: when GDK notifies
the GTK layer about an exposure event for a
<classname>GdkWindow</classname>, the GTK layer would simply
emit the #GtkWidget::draw signal for that widget. The signal
handler would subsequently repaint the widget. No further
work would be necessary; the windowing system would generate
exposure events for each window that needs it, and then each
corresponding widget would draw itself in turn.
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).
</para>
<para>
However, in practice it is convenient to have widgets which do
not have a <classname>GdkWindow</classname> of their own, but
rather share the one from their parent widget. Such widgets
have called <function>gtk_widget_set_has_window</function> to
disable it; this can be tested easily with the <link
linkend="gtk-widget-get-has-window"><function>gtk_widget_get_has_window()</function></link>
function. As such, these are called <firstterm>no-window
widgets</firstterm>.
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.
</para>
<para>
No-window widgets are useful for various reasons:
</para>
<itemizedlist>
<listitem>
<para>
Some widgets may want the parent's background to show through, even
when they draw on parts of it. For example, consider a theme that
uses textured backgrounds, such as gradients or repeating
patterns. If each widget had its own window, and in turn its own
gradient background, labels would look bad because there would be a
visible break with respect to their surroundings. <xref
linkend="figure-windowed-label"/> shows this undesirable effect.
</para>
<figure id="figure-windowed-label">
<title>Windowed label vs. no-window label</title>
<graphic fileref="figure-windowed-label.png" format="png"/>
</figure>
</listitem>
<listitem>
<para>
Reducing the number of windows creates less traffic between GTK+ and
the underlying windowing system, especially when getting events.
</para>
</listitem>
</itemizedlist>
<para>
On the other hand, widgets that would benefit from having a "hard"
clipping region may find it more convenient to create their own
windows. Also, widgets which want to receive events resulting from
user interaction may find it convenient to use windows of their own as
well. Widgets may have more than one window if they want to
define different regions for capturing events.
</para>
</refsect2>
<refsect2 id="hierarchical-drawing">
<title>Hierarchical drawing</title>
<para>
When the GTK layer receives an exposure event from GDK, it
finds the widget that corresponds to the window which received
the event. By definition, this corresponds to a widget that
has the <constant>GTK_NO_WINDOW</constant> flag turned
<emphasis>off</emphasis> (otherwise, the widget wouldn't own
the window!). First this widget paints its background, and
then, if it is a container widget, it tells each of its
<constant>GTK_NO_WINDOW</constant> children to paint
themselves. This process is applied recursively for all the
<constant>GTK_NO_WINDOW</constant> descendants of the original
widget.
</para>
<para>
Note that this process does not get propagated to widgets
which have windows of their own, that is, to widgets which
have the <constant>GTK_NO_WINDOW</constant> flag turned off.
If such widgets require redrawing, then the windowing system
will already have sent exposure events to their corresponding
windows. As such, there is no need to
<firstterm>propagate</firstterm> the exposure to them on the
GTK+ side.
</para>
<para>
<xref
linkend="figure-hierarchical-drawing"/> shows how a simple toplevel window would
paint itself when it contains only <constant>GTK_NO_WINDOW</constant> descendants:
<orderedlist>
<listitem>
<para>
The outermost, thick rectangle is a toplevel <link
linkend="GtkWindow"><classname>GtkWindow</classname></link>,
which is not a <constant>GTK_NO_WINDOW</constant> widget &mdash;
as such, it does receive its exposure event as it comes from GDK.
First the <classname>GtkWindow</classname> would paint its own
background. Then, it would ask its only child to paint itself,
numbered 2.
</para>
</listitem>
<listitem>
<para>
The dotted rectangle represents a <link
linkend="GtkVBox"><classname>GtkVBox</classname></link>, which
has been made the sole child of the
<classname>GtkWindow</classname>. Boxes are just layout
containers that do not paint anything by themselves, so this
<classname>GtkVBox</classname> would draw nothing, but rather ask
its children to draw themselves. The children are numbered 3 and
6.
</para>
</listitem>
<listitem>
<para>
The thin rectangle is a <link
linkend="GtkFrame"><classname>GtkFrame</classname></link>,
which has two children: a label for the frame, numbered 4, and
another label inside, numbered 5. First the frame would draw its
own beveled box, then ask the frame label and its internal child to
draw themselves.
</para>
</listitem>
<listitem>
<para>
The frame label has no children, so it just draws its text: "Frame&nbsp;Label".
</para>
</listitem>
<listitem>
<para>
The internal label has no children, so it just draws its text: "This
is some text inside the frame!".
</para>
</listitem>
<listitem>
<para>
The dotted rectangle represents a <link
linkend="GtkHBox"><classname>GtkHBox</classname></link>. Again,
this does not draw anything by itself, but rather asks its children
to draw themselves. The children are numbered 7 and 9.
</para>
</listitem>
<listitem>
<para>
The thin rectangle is a <link
linkend="GtkButton"><classname>GtkButton</classname></link> with
a single child, numbered 8. First the button would draw its
beveled box, and then it would ask its child to draw itself.
</para>
</listitem>
<listitem>
<para>
This is a text label which has no children, so it just draws its
own text: "Cancel".
</para>
</listitem>
<listitem>
<para>
Similar to number 7, this is a button with a single child, numbered
10. First the button would draw its beveled box, and then it would
ask its child to draw itself.
</para>
</listitem>
<listitem>
<para>
Similar to number 8, this is a text label which has no children,
so it just draws its own text: "OK".
</para>
</listitem>
</orderedlist>
</para>
<figure id="figure-hierarchical-drawing">
<title>Hierarchical drawing order</title>
<graphic fileref="figure-hierarchical-drawing.png" format="png"/>
</figure>
</refsect2>
</refsect1>
@ -330,16 +262,6 @@
<refsect1 id="double-buffering">
<title>Double buffering</title>
<para>
When the GTK layer receives an exposure event from GDK, it first finds
the <literal>!<constant>GTK_NO_WINDOW</constant></literal> widget that
corresponds to the event's window. Then, it emits the
#GtkWidget::draw signal for that
widget. As described above, that widget will first draw its background,
and then ask each of its <constant>GTK_NO_WINDOW</constant> children to
draw themselves.
</para>
<para>
If each of the drawing calls made by each subwidget's
<literal>draw</literal> handler were sent directly to the
@ -352,26 +274,6 @@
when all drawing operations are done.
</para>
<!-- FIXME: figure with a timeline of non-double-buffered and
double-buffered paints:
onscreen:
[garbage]
[background]
[button-frame]
[icon]
[label]
onscreen: offscreen:
[garbage]
[background]
[button-frame]
[icon]
[label]
[final result]
-->
<para>
Two basic functions in GDK form the core of the double-buffering
mechanism: <link
@ -396,14 +298,11 @@
</para>
<para>
To make this easier, most GTK+ widgets have the
<constant>GTK_DOUBLE_BUFFERED</constant> <link
linkend="GtkWidgetFlags">widget flag</link> turned on by
default. When GTK+ encounters such a widget, it automatically
calls <function>gdk_window_begin_paint_region()</function>
before emitting the #GtkWidget::draw signal for the widget, and
To make this easier, GTK+ normally calls
<function>gdk_window_begin_paint_region()</function>
before emitting the #GtkWidget::draw signal, and
then it calls <function>gdk_window_end_paint()</function>
after the signal has been emitted. This is convenient for
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.
@ -411,10 +310,12 @@
<para>
However, some widgets may prefer to disable this kind of
automatic double buffering and do things on their own. To do
this, call the
<function>gtk_widget_set_double_buffered()</function> function
in your widget's constructor.
automatic double buffering and do things on their own.
To do this, call the
<function>gtk_widget_set_double_buffered()</function>
function in your widget's constructor. Double buffering
can only be turned off for widgets that have a native
window.
</para>
<example id="disabling-double-buffering">