In case a toytoolkit application manages to schedule resizes constantly,
throttle them to the main surface display.
When resizing, all surfaces are updated synchronously, so it also makes
sense to synchronize on the main surface's frame callback particularly.
Rendering any faster will not make sense.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Mesa's eglSwapBuffers() waits for the frame event from the previous
swapBuffers, before it returns. Apparently eglSwapInterval(), which
should be able to disable the wait, is unimplemented for now.
When a sub-surface contains an EGL widget, and the commit mode is
synchronized, the frame events will not be delivered to EGL until the
parent surface gets committed. Therefore rendering the EGL widget twice
would lead to a deadlock.
When the window is being resized, we need to force a repaint of the EGL
widget, too, to make the whole window consistent. For that, we need to
make sure the frame event from the previous eglSwapBuffers() actually
arrives.
This patch adds an extra wl_surface.commit(parent), when the window is
being resized, which should guarantee, that the previous eglSwapBuffers
gets its event.
To properly handle an EGL widget in a sub-surface, running in its own
thread, the EGL widget's automatic updates should be paused before
sending the extra wl_surface.commit(parent). A natural place for the
pause would be in the widget's resize hook. However, wl_surface.commit
cannot be called right after resize hooks, because it would commit new,
incomplete surface state. Therefore this patch is not enough for
threaded toytoolkit applications. Luckily those do not exist yet.
When eglSwapInterval() gets implemented, this patch should be reverted.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Add a demo program with:
- a main surface (green)
- a Cairo-image sub-surface (red)
- a raw GLESv2 widget (triangle)
Sub-surface input region is set empty to avoid problems in toytoolkit.
If Cairo links to libGL, then we will end up with also libGLESv2 linked
to subsurfaces program, and both libs getting really used, which leads
to disaster.
Do not build subsurfaces demo, if Cairo links to libGL and cairo-egl is
usable.
The GL rendering loop is not tied to the toytoolkit or the widget, but
runs directly from its own frame callback. Therefore it runs
independent of the rest of the application. This also relies on one of
two things:
- eglSwapInterval(0) is implemented, and therefore eglSwapBuffers never
blocks indefinitely, or
- toytoolkit has a workaround, that guarantees that eglSwapBuffers will
return soon, when we force a repaint on resize.
Otherwise the demo will deadlock.
The code is separated into three sections:
1. The library component, using only EGL, GLESv2, and libwayland-client
APIs, and not aware of any toolkit details of the parent application.
This runs independently until the parent application tells otherwise.
2. The glue code: a toytoolkit application widget, who has its own
rendering machinery.
3. The application written in toytoolkit.
This patch also adds new toytoolkit interfaces:
- widget_get_wl_surface()
- widget_get_last_time()
- widget_input_region_add()
Toytoolkit applications have not had a possibility to change the input
region. The frame widget (decorations) set the input region on its own
when used, otherwise the default input region of everything has been
used. If a window does not have a frame widget, it can now use
widget_input_region_add() to set a custom input region.
These are not window methods, because a widget may lie on a different
wl_surface (sub-surface) than the window.
Changes in v3:
- replace set_commit_mode with set_sync and set_desync
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Add redraw_needed flag to all surfaces, in addition to having one in
window. The window redraw_needed flag is changed to force a redraw of
the whole window, regardless of frame events.
widget_schedule_redraw() now schedules the redraw only for the surface,
where the widget is on. window_schedule_redraw() is equivalent to
scheduling a redraw for all (sub-)surfaces of the window.
We still use only one deferred task for all redraws.
surface_redraw() will skip the redraw, if the window does not force a
redraw and the surface does not need a redraw. It will also skip the
redraw, if the frame callback from the previous redraw has not triggered
yet. When the frame callback later arrives, the redraw task will be
scheduled, if the surface still needs a redraw.
If the window forces a redraw, the redraw is executed even if there is a
pending frame callback. This is for resizing: resizing should trigger a
window repaint, as it really wants to update all surfaces in one go, to
apply possible sub-surface size and position changes. Resizing is the
only thing that makes a window force a redraw.
With this change, subsurfaces demo can avoid repainting the cairo
sub-surface while still animating the GL sub-surface.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
The new application API window_add_subsurface() will create a plain
widget that is on a new sub-surface.
The sub-surface position is taken from the surface's root widget
allocation. This way widget allocations are always in the main surface
(i.e. window) coordinates. However, Cairo drawing coordinates will now
be different to widget coordinates for sub-surfaces. Cairo coordinates
are fixed by applying a translation in widget_cairo_create(), so that
widget drawing code can simply use the widget allocation as before.
Sub-surfaces are hooked up into resize, window flush, redraw, and
find_widget. Window maintains a list of sub-surfaces in top-first order.
Add a client settable default commit mode, and toggle the mode when
resizing to guarantee in-sync updates of a window and its sub-surfaces.
Changes in v3:
- replaced set_commit_mode with set_sync and set_desync
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Increase the maximum number of shm "leaves" to three, and rewrite the
leaf release and pick algorithms. The new algorithms hopefully improve
on buffer re-use while freeing unused buffers.
The goal of the new release algorithm is to always leave one free leaf
with storage allocated, so that the next redraw could start straight on
it.
The new leaf picking algorithm will prefer a free leaf that already has
some storage allocated, instead of just picking the first free leaf that
may need to allocate a new buffer.
Triple-buffering is especially for sub-surfaces, where the compositor
may have one wl_buffer busy on screen, and another wl_buffer busy in the
sub-surface cached state due to the synchronized commit mode. To be
able to forcibly repaint at that situation for e.g. resize, we need a
third buffer.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Resolve a bad frame visible when maximizing toytoolkit programs with the the
maximize button in decorations. Windows now use wl_display.sync requests to
wait for a maximize to finish before drawing again, following suggestions from
http://lists.freedesktop.org/archives/wayland-devel/2013-February/007650.html
Make sure that display_acquire_window_surface() creates the Cairo
surface as necessary. Otherwise surface->toysurface can be NULL.
This fixes weston-screensaver fullscreen mode. Demo mode was not
affected as it uses window decorations, and so the Cairo surface is
created. This regression was introduced by:
commit 0c4445ba57
Author: Pekka Paalanen <ppaalanen@gmail.com>
Date: Wed Feb 13 16:17:23 2013 +0200
window: create Cairo surfaces on demand for redraw
Reported-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Ideally the shell would send an unmaximize event to the client when
we try to move a maximized window, but for now, let's just prevent
moving maximized windows.
https://bugs.freedesktop.org/show_bug.cgi?id=56296
This introduces the function widget_cairo_create().
Instead of directly referencing surface->cairo_surface, use the function
widget_cairo_create(), which will create the cairo_surface as necessary,
and just returns a Cairo drawing context. Also fix window_get_surface()
similarly.
Now we can go through idle_redraw() without always creating Cairo
surfaces and committing them. This will be useful with sub-surfaces,
where repainting one sub-surface does not need to force the repaint of
all surfaces of a window.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Menu and tooltip redraw functions were using the surface size directly.
For consistency, make them use the widget size instead, it is the same.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Widgets should be rendering to a cairo_surface for a particular
wl_surface, just like buffers are per surface.
window_flush() has a change in behaviour: it will now send
wl_shell_surface.set_toplevel also without a cairo_surface to be
attached. This shouldn't change anything in practice.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
So that given a widget, we can access the surface specific data, like
buffers, and input and opaque regions.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
They are per wl_surface state.
The frame widget is always on the main surface, since it can be created
only for the window. That is why frame_resize_handler() can simply
assume that the surface is the main_surface.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Repaint and resizing widget recursions must start from the root widget
of each (sub-)surface, so that buffers and regions get initialized
correctly. Make it easier by moving the widget field from struct window
to struct surface.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
These are surface specifics, since buffers are surface specific.
SURFACE_HINT_RESIZE is moved together to the other SURFACE_* flags, so
that surface_create_surface() would not need two flags arguments.
struct toysurface::prepare vfunc checks for SURFACE_HINT_RESIZE, and
egl_window_surface_create() and shm_surface_create() check for the
non-HINT flags.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Fields 'allocation' and 'server_allocation' are surface specific. Fields
'saved_allocation', 'min_allocation', and 'pending_allocation' are
window specific, and will not be moved.
Field 'toysurface' is naturally surface specific, since it provides the
backing storage for the wl_surface.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Struct window has many fields that are directly related to the
wl_surface, more than to the window as a whole. When we start composing
a window from several wl_surfaces, these fields need to be per
wl_surface, not per window.
Start separating such fields from struct window into struct surface by
moving the wl_surface.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Handle the case when we the compositor somehow migrates from requiring
double buffering into working on single buffering, so we release the
extra shm buffer.
Currently, I do not think this can happen in practice, but in the future
it may happen with sub-surfaces.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Check for errors in the first wl_display_dispatch() call. Otherwise
doing something silly like
$ WAYLAND_SOCKET=999 ./clickdot
will segfault.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Make them explicitly mention EGL, otherwise one can easily think that
"failed to initialize display" refers to Wayland display.
Also explicitly mention falling back to wl_shm. I tested this with a
LD_PRELOAD trick that overrides eglBindAPI and makes it fail.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
All the clients here were missing the global_remove handler. Because
window.c did not have it, weston-desktop-shell and weston-keyboard
segfaulted on compositor exit, as they received some
wl_registry.global_remove events.
Add more or less stub global_remove handlers, so that clients do not
crash on such events. Toytoolkit and all applications would need a lot
more code to properly handle the global object removal.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
After a client has been double-buffering, and then switches to
single-buffering, it should release the 2nd buffer. That never happens
in practice here, so just add a comment and a check in case it ever
occurs in the future.
If we implemented the releasing now, it would be difficult to test.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
In a few cases, we set a motion handler just to be able to set a fixed
cursor. This adds a default cursor helper that can be used in those cases.
In case of the 'transformed' test case, we also avoid a brief flicker
of the pointer cursor, which is set on enter when the move grab is lifted.
Change the boolean parameter 'resize_hint' into a bitmask 'flags'.
Note, that this flags is very different to the other flags used in
creating the toysurface implementations. They do not make sense to mix
one way or the other. Prepare() cannot change the surface type, and
surface constructors do not care for dynamic hint flags.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
When a window's buffer transformation is set, its buffers are
reallocated with the appropriate size (i.e., with width and height
swapped in case of 90 or 270 degree rotation).
Since the opaque region was set in frame_resize_handler(), if a client
created a frameless window setting the toplevel widget as opaque would
have no effect.
This patch fixes this by moving the call wl_surface_set_opaque_region()
to idle_resize(), and changing the latter function to set the whole
window as opaque if its toplevel widget has the opaque flag set.
To reproduce, launch the terminal, open a second window using Ctrl-Shift-N,
go back to the first window, and press Ctrl-D. The terminal's master FD gets
events even after being closed, causing terminal_destroy to be called twice
on the same object.
To fix this, I'm adding a function to stop watching an FD.
Listen for wl_buffer.release events in the shm path, and if a previously
posted buffer is still held by the server, allocate another one. The
maximum of two should be enough, since there is no point for a server to
hold more than one buffer at a time.
Buffer allocation happens as needed instead of window creation time.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
After the toysurface rewrite, windows do not have a valid Cairo surface
outside their repaint cycle, so tooltips are not getting their size
right.
Create a dummy Cairo surface only for querying text extents, so we do
not rely on any window surfaces of parent windows or otherwise.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Implement shm_surface as a sub-class of toysurface, and unify the
toysurface call sites removing most buffer type specific branching.
Do not destroy and create a surface, if the size does not change.
The resizing optimization of shm surfaces is retained, but the pool is
moved from struct window to struct shm_surface, since it does not apply
to egl_window_surface.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
We need more structure to the way we handle the backing storage in
toytoolkit, to make it possible to double-buffer the shm case properly.
The existing buffer handling is very complex with the three
different cases:
- EGLSurface backed Cairo surface with a window associated
- wl_shm backed Cairo surface with a window associated
- wl_shm backed Cairo surface without a window, as used by dnd.c
Introduce the toysurface abstraction, which defines the interface for
the both buffer handling cases that have a window associated. It also
means, that windows will not have a valid Cairo surface outside of their
repaint cycle.
Convert the EGLsurface case into toysurface for starters. For EGL-based
Cairo surfaces, the private data is no longer needed. Destroying
egl_window_surface will trigger the destruction of the cairo_surface_t,
not vice versa. This is possible because display_create_surface() is
shm-only.
The shm cases are left untouched.
As a side-effect, display_acquire_window_surface() and
display_release_window_surface() will no longer use the 'display'
argument. Instead, display will be the one inherited from the window.
Signed-off-by: Pekka Paalanen <ppaalanen@gmail.com>
Pekka Paalanen
Kristian Høgsberg
Giulio Camuffo
Emilio Pozuelo Monfort
MoD
Rafal Mielniczuk
Rob Bradford
Ander Conselvan de Oliveira
Dima Ryazanov
Jan Arne Petersen