Laszlo Agocs [Fri, 17 Feb 2012 23:01:32 +0000 (01:01 +0200)]
Fix wayland plugin and qwindow-compositor compilation
Apparently some qtbase change triggers the need to include a bunch of
headers which were not properly included before. The extra mouse
button support in the platform plugin simply did not compile.
Change-Id: I7d518021fa81f2d5232277f0bce26e48effbeae3
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Mon, 20 Feb 2012 07:42:49 +0000 (09:42 +0200)]
Send mouse release generated from TouchCancel to the proper window
Change-Id: I3d0ab79eb261de4d22c6aadf0672f5c84ff3533e
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Tue, 14 Feb 2012 15:33:29 +0000 (16:33 +0100)]
Compositor memory leak fixes
Change-Id: Ia0510e00ffc925cde98ec5345a8129d1979c478c
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 14 Feb 2012 18:30:43 +0000 (19:30 +0100)]
Fix qwindow-compositor memory leaks
Change-Id: I85aa62c9deae308612ed2b13468e7f1e48f82acf
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Rick Stockton [Wed, 15 Feb 2012 02:42:37 +0000 (18:42 -0800)]
qtwayland: Add support for 16 mouse buttons
Update expands the mouse button assignment code block, which formerly
handled only values 272, 273, and 274. New version handles all mouse
mouse buttons shown in kernel's input.h header (decimal values 272
through 285).
Value 286 appears to be the first button in the range for joystick
devices, so I did not include any other kernel-enumerated values.
(Other values might be possible- but I don't know of a mouse which
is currently being sold with more than 15 non-wheel buttons. The
mouse with 15 buttons is the Razer Naga, and I don't own it.)
Task-number: QTBUG-22759
Change-Id: I77f16eb89e24d7d6410199badff28c0d48f7df93
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jørgen Lind [Wed, 15 Feb 2012 12:28:10 +0000 (13:28 +0100)]
Expose in a virtual function that surfaces are being destroyed
Change-Id: I979e7d9ca7152d4db6c3e7adae8353935a05d6c5
Reviewed-by: Andy Nichols <andy.nichols@nokia.com>
Samuel Rødal [Tue, 14 Feb 2012 14:22:21 +0000 (15:22 +0100)]
Pass the correct window on to QWindowSystemInterface.
Sometimes, the window's coordinate system is not bounded by
the coordinates of the QScreen (such as if the window uses a
landscape window orientation on a portrait screen).
QGuiApplication::topLevelAt() will fail to find this window, but since
we have the window already at the wayland plugin level we just use that.
We already did this for mouse events, so it's natural to do it here as
well.
Change-Id: Id086dd940aa1cf4eb4480117d243ab8a76da6cca
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Samuel Rødal [Mon, 13 Feb 2012 19:39:22 +0000 (20:39 +0100)]
Added convenience properties in WaylandSurface for window orientation.
The new windowRotation property is especially useful for the compositor
to know how much the surface needs to be rotated in order to be
displayed correctly.
Change-Id: I81ee17880ceafd465c39b4439dee978f2cb80924
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Samuel Rødal [Mon, 13 Feb 2012 19:38:44 +0000 (20:38 +0100)]
Prevent occasional crash in qwaylandinputdevice.cpp
Change-Id: I86559add9925dad6aef530433dd92066f1708757
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Jørgen Lind [Mon, 13 Feb 2012 13:19:22 +0000 (14:19 +0100)]
Removing Wayland::SurfacePrivate
And moving Wayland::SurfaceBuffer into its own file.
Also remove the extra functions in Wayland::Surface which just relayed
down to extended surface. The pattern is that WaylandSurface is the
entry point so we check there is the Wayland::Surface has the extension
we'r looking for, then calling it directly from WaylandSurface.
Last, the data of an extension is suppose to be members of the
extension, so I moved the properties to be members of ExtendedSurface
Change-Id: I5877671a17ee873b7644094f695fcd3e4677964a
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Sun, 12 Feb 2012 12:28:04 +0000 (14:28 +0200)]
Fix surface wrapper build when QT_COMPOSITOR_WAYLAND_GL is not set
Change-Id: Ic03d08ec936059344a1e092329a9def0b4ceb9ee
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Pier Luigi Fiorini [Sat, 11 Feb 2012 21:12:38 +0000 (22:12 +0100)]
Fix build of compositors by always including QOpenGLContext.
QOpenGLContext is always required even when QT_COMPOSITOR_WAYLAND_GL is
not defined. Moving the include outside the ifdef makes compositors
build without problems.
Change-Id: I4b8903b06a5f58ad8fd2953dbde1476877683cb4
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Jørgen Lind [Mon, 30 Jan 2012 08:53:25 +0000 (09:53 +0100)]
Surface::isInverted Dont use xor and move the ret variable
outside the ifdef
Change-Id: I6f9048cd7d5992ef46dcc0e9d05e86f2f7ca4033
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Gunnar Sletta [Fri, 10 Feb 2012 14:43:59 +0000 (15:43 +0100)]
Added window flags to the surface-extension protocol
Change-Id: I1b53c5f0b0f908490dcd0ab30e46ab681f885954
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Thu, 9 Feb 2012 13:36:28 +0000 (15:36 +0200)]
Add TouchCancel support.
It maps to the protocol's touch_cancel pretty well. :)
Change-Id: Id417aac768106c2f6618b3e7ecb51d6929977c37
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Andy Nichols [Thu, 9 Feb 2012 18:08:30 +0000 (19:08 +0100)]
Remove hack that causes deadlock when waiting in eglSwapBuffers
This hack was added to work around the problem of calling
wl_display_iterate in two separate threads, but now that we are
are enforceing that wl_display_iterate can only be called in one thread
it now causes a deadlock when we are waiting in eglSwapBuffers for the
gui thread to actually call wl_display_iterate.
Change-Id: Id3153b184c191494838fc363380f7cae621ac64e
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 7 Feb 2012 14:37:11 +0000 (16:37 +0200)]
Fix data device handling.
Clearing the device list on resource destroy is wrong: it cleans away
data devices for all clients. Calling free when removing stale devices
for a client is also wrong.
Change-Id: I7bfcc928762dc4ca7dbf5abeebcd2489956e6828
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 7 Feb 2012 09:39:49 +0000 (11:39 +0200)]
Add mouse event synthesizing to the touch extension protocol.
The compositor can now be configured to tell the clients to generate
mouse events when receiving touch events. In touchscreen-only systems
this can be used to reduce the number of Wayland events.
Change-Id: I231a15cd4ed463ee81c510c082a270efa255a1f3
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Fri, 3 Feb 2012 08:30:40 +0000 (10:30 +0200)]
Set pointer focus to null when surface is destroyed.
We must do something otherwise the next inputDevice->mouseFocus() call
will dereference an already deleted instance.
Change-Id: I3ac92a37715f63c6dc43191968a8da51befbccde
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Lasse Holmstedt [Thu, 2 Feb 2012 14:54:05 +0000 (15:54 +0100)]
Don't try to carry over qPrintable() pointers
The pointer is out of scope by the time it's sent through wayland.
Change-Id: I00e9f050dfa47bf078921600d2bf63502f0b4897
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Ian Monroe <ian.monroe@nokia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@nokia.com>
Laszlo Agocs [Thu, 2 Feb 2012 16:03:10 +0000 (18:03 +0200)]
Fix mouse handling in qml-compositor.
The pointer focus must be set in pressed. We cannot rely on
the implicit mouse tracking behavior of QWindow for QML stuff
because the WaylandSurfaceItems will not get mouse moves outside
of a press-release sequence.
Change-Id: Ic2cf2a7497cac7397bcfa26440dd05cded6182dc
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Samuel Rødal [Wed, 25 Jan 2012 13:37:22 +0000 (14:37 +0100)]
Added reporting of window and content orientation to surface extension.
Change-Id: I6e182c048282f5edd30f49be19dcc0f020679b85
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Mon, 30 Jan 2012 14:16:22 +0000 (16:16 +0200)]
Revert "Do not attach null buffer."
This reverts commit
42ed12656016a96b966b7b5e52f0b684ad65c227.
Weston has been fixed so attaching null buffer on hide is now a valid
operation.
Change-Id: I0fc6af7f87c97e07075da03ed1b54d2e3c1d45fd
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Sat, 28 Jan 2012 18:56:28 +0000 (20:56 +0200)]
Do not attach null buffer.
Apparently Weston does not like this and causes instant crashes when
hiding a top-level widget. The QtCompositor-side behavior is dubious
too although at least that does not crash. Nevertheless attaching a
null buffer seems unnecessary.
Change-Id: I1d7fe3e0b05e5d77773b661a71bc37311968eded
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Sat, 28 Jan 2012 18:47:11 +0000 (20:47 +0200)]
Null out the frame callback ptr properly.
Otherwise shm based apps (widgets) will crash randomly.
Change-Id: I5345ca9419a6702170f93fd817bbab4d9ef2bc59
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Mon, 30 Jan 2012 07:57:04 +0000 (09:57 +0200)]
Fix compilation for qwidget and qwindow compositor examples
Change-Id: If720f7f4797e319ab03d02b6ecb9c88cb71356c2
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jason McDonald [Tue, 24 Jan 2012 05:35:27 +0000 (15:35 +1000)]
Remove "All rights reserved" line from license headers.
As in the past, to avoid rewriting various autotests that contain
line-number information, an extra blank line has been inserted at the
end of the license text to ensure that this commit does not change the
total number of lines in the license header.
Change-Id: I08d8673382f0230d5c5a22a25efaf5c45cef539a
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Laszlo Agocs [Fri, 27 Jan 2012 11:02:32 +0000 (13:02 +0200)]
Fix qt_compositor.pri to make header exporting work.
Change-Id: I5c54c65f57805372cba101dd62f34818580a90aa
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Fri, 27 Jan 2012 11:03:53 +0000 (13:03 +0200)]
Fix compilation
Change-Id: Ia3e62ed30467bb52555c353a30440ebcc3529d0b
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Thu, 26 Jan 2012 14:28:36 +0000 (15:28 +0100)]
Add possibillity of overriding invertedY flag for wl_surfaces
Change-Id: I0f4c2696ae5c7071275af326c924dc1eaed6d5b6
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Andy Nichols <andy.nichols@nokia.com>
Jørgen Lind [Thu, 26 Jan 2012 14:02:09 +0000 (15:02 +0100)]
Make it possible not to have queing of buffers on the serverside
Change-Id: Ic7d3221321e2d7ecc69e56b7ad7a23322e65995d
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Andy Nichols <andy.nichols@nokia.com>
Laszlo Agocs [Thu, 26 Jan 2012 13:19:42 +0000 (14:19 +0100)]
Revert "Do not generate mouse event on the first motion event."
This was a misunderstanding. The default (and the only) behavior with
QWindows is to have mouse tracking enabled. QML or QApplication can
then filter out the button-less moves, if needed.
This reverts commit
bd0681daf6fa91afee16bba619f17b7bd3bd5460.
Change-Id: I2f641a2826e1578f103d1e29939b661cbae4fcc2
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Sanity-Review: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Wed, 25 Jan 2012 18:36:19 +0000 (19:36 +0100)]
Do not generate mouse event on the first motion event.
This event comes before the button event and thus must not result in a
Qt mouse event. Instead, the button event will generate the
MouseButtonPressed with the position sent in this motion.
Change-Id: I97887d4d5fcd54575d6f4e9523a04569424c0bc2
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Samuel Rødal [Wed, 25 Jan 2012 08:45:40 +0000 (09:45 +0100)]
Added missing include to fix build.
Fixes build after change in qtbase requiring Q_DECLARE_METATYPE(T*) to
be fully defined.
Change-Id: I741d91eb0e7e480b5a35a4616b32a90043f01a63
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Samuel Rødal [Fri, 13 Jan 2012 12:58:53 +0000 (13:58 +0100)]
Adapted to new orientation API in qtbase.
Change-Id: I455d06a364a89b6558ab00b15d9541ea397adcd5
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Jason McDonald [Fri, 20 Jan 2012 05:26:51 +0000 (15:26 +1000)]
Update obsolete contact address.
Replace Nokia contact email address with Qt Project website.
Change-Id: Ib67c393df41c539deae0ef71e0acc13029ceb46d
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Laszlo Agocs [Fri, 20 Jan 2012 13:22:23 +0000 (15:22 +0200)]
Fix duplicated QOpenGLFunctions usage
Change-Id: I3c90b705c0ef07e824768975fe1889e394acd82b
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Gunnar Sletta [Wed, 18 Jan 2012 17:16:28 +0000 (18:16 +0100)]
pick the right gl version
Change-Id: I1bee7028f9cf53383bb76678f1446b86cc2d007f
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 17 Jan 2012 15:30:42 +0000 (17:30 +0200)]
Offer the retained selection when no data source is available anymore
When retained selection is enabled, clipboard data will be available
even after the client, that offered the data, exits.
Change-Id: I8158eba29b14092ce150245410b62912011051fd
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 17 Jan 2012 13:32:32 +0000 (15:32 +0200)]
Implement selection offers from compositor to clients.
It is a hack but works beautifully. It allows the compositor to
participate in copy-paste which becomes mandatory when there is UI
running in the compositor process.
Change-Id: I1993d8705a26159eff0c9947244b66e954b9f460
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jørgen Lind [Tue, 17 Jan 2012 09:46:56 +0000 (10:46 +0100)]
Update Wayland sha
Change-Id: Ic248b08863e661259ad5f96c35b1135d9de594f8
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Jørgen Lind [Tue, 17 Jan 2012 09:45:07 +0000 (10:45 +0100)]
Remove declarations of functions not longer implemented
Change-Id: Ia258fe47d59e3c5a844b939245ad50e2a4f1a6f1
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Laszlo Agocs [Tue, 17 Jan 2012 13:50:14 +0000 (15:50 +0200)]
Remove unnecessary client list generation in Compositor
Change-Id: Ia240a19a14098bdc76374b9537d501927a6a1a74
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Tue, 17 Jan 2012 14:26:05 +0000 (15:26 +0100)]
Null pointer check
managedClient() is allowed to return 0.
Change-Id: I67f4ad9cc9bf40a829c2a26895529cabf5a716c0
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Andy Nichols <andy.nichols@nokia.com>
Jørgen Lind [Tue, 17 Jan 2012 09:44:01 +0000 (10:44 +0100)]
Make sure that subSurface extension handling in client is initialized
to zero, so that if the compositor doesn't support the extension we know
about it.
Change-Id: Ia39a832d08fe4e2a2d94e18b84f2783206283ae4
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@nokia.com>
Laszlo Agocs [Mon, 16 Jan 2012 07:39:29 +0000 (09:39 +0200)]
Avoid using an extra event for each touch report.
Instead of using touch_frame the number of non-stationary points (that
is, the number of touch events sent) is included in the touch events
themselves.
Change-Id: Ied29f05d7e00676d47ee0b91006cafb4569951c4
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jørgen Lind [Mon, 16 Jan 2012 07:14:06 +0000 (08:14 +0100)]
Fix license in extensions.
The scanner adds the commenting syntax.
Change-Id: Ia480dab42302395e2f562e5bfc9015ef470ddc08
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Laszlo Agocs [Mon, 16 Jan 2012 06:51:18 +0000 (08:51 +0200)]
Do not send stationary points
Change-Id: Id534965df614bcbff3bfe5f8fd5484c370222196
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Sat, 14 Jan 2012 13:31:58 +0000 (15:31 +0200)]
Send the contents of the rawScreenPositions() list too.
Change-Id: If43c4e07ae919b77050aaf5dcffea3b8d48ac460
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Sat, 14 Jan 2012 12:58:35 +0000 (14:58 +0200)]
Rename touch extension interface and fix event sending
Follow the naming convention of other extensions.
The events are now posted only to the client to which the focus
surface belongs, the check was previously missing.
Change-Id: Ie43b6e887b64c7bed1377babb368129753a3cd0f
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jason McDonald [Mon, 16 Jan 2012 04:33:00 +0000 (14:33 +1000)]
Update license headers.
All code in this module outside of src/plugins/platforms/wayland should
carry a copy of the BSD license header. The master copy of this header
lives at qtbase/header.BSD.
Change-Id: Icd56f24f4a3d309db4c1c3aaef25618680c07f6e
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Fri, 13 Jan 2012 15:37:49 +0000 (17:37 +0200)]
Introduce a touch protocol extension.
Using this opt-in protocol extension all data from QTouchEvent will
properly be delivered to the clients.
By default compositors will still use the standard and limited
protocol for touch events to remain compatible with non-Qt
clients. However by calling enableTouchExtension() in the compositor,
touch events will be sent through the new protocol and thus Qt clients
will receive proper QTouchEvents. Non-Qt clients will get no touch
data in this mode.
The contents of the rawScreenPositions() list are not yet delivered as
the size implications remain to be checked.
Change-Id: I981b9c4d44ea808a95b770bf948320abd47fc036
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Robin Burchell [Fri, 13 Jan 2012 08:40:40 +0000 (10:40 +0200)]
Remove the MDI compositor.
This was broken by the recent input rewrites, and isn't really the 'correct'
future-proof path anyway. Wayland is more advocating client-side decorations
rather than compositor-side decorations, and this won't work in that world.
Change-Id: I187084d7eea785eda43e8658cfd2120205eea211
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Fri, 13 Jan 2012 12:18:02 +0000 (14:18 +0200)]
Fix handleScreenOrientationChange call in wayland plug-in.
Change-Id: Ib8abd146c2bd30b59c6aba0aa9a28f70133b25e9
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Fri, 13 Jan 2012 12:13:50 +0000 (14:13 +0200)]
Do not use the removed primary flag for touch points.
Change-Id: I5ac4bf15b89a7d352c9971e6ce4882d49e686838
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Robin Burchell [Thu, 12 Jan 2012 18:41:32 +0000 (20:41 +0200)]
Use QOpenGLFunctions instead of presuming GL methods exist.
They may not always exist, especially on desktop GL, so this is incorrect and
won't build in some cases.
Credit to Samuel Rødal for setting me on the path :)
Change-Id: I89296c5ed67f9388cceedb08cd68a7b5560be5e9
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Andy Nichols [Thu, 12 Jan 2012 18:24:28 +0000 (19:24 +0100)]
Replaced missed instance of Wayland::Surface->clientHandle()
This was if EGL_WL_request_client_buffer_format was defined.
Change-Id: Ic96be10a981b79d0d2f62bc165b57130398c385a
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@nokia.com>
Kent Hansen [Thu, 12 Jan 2012 06:34:41 +0000 (07:34 +0100)]
Fix linking error
Missing "QT += declarative" caused the Q_DECLARATIVE_EXPORT macro to
be defined incorrectly, so the linker would complain about missing
symbol qdeclarativeelement_destructor.
Change-Id: I9638377ddf9478b452362f5ee8c86c0e7b844fc0
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Wed, 11 Jan 2012 14:18:21 +0000 (16:18 +0200)]
Make touch events working in the qml example compositor.
Change-Id: I03318359d13caf7287671bcc9e01da91acb928dd
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Wed, 11 Jan 2012 14:01:30 +0000 (16:01 +0200)]
Make touch event sending more robust in the InputDevice.
At least it won't crash and proper warnings are shown when the
slightly obscure pointer focus is not used properly in the compositor.
Change-Id: I2c970d197176b588f38a750e3cc805aae4d6c049
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Wed, 11 Jan 2012 13:50:51 +0000 (15:50 +0200)]
Don't crash when sending touch events.
Change-Id: I62bf4913338e6daa3245400b0e2c7004907cb29a
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jørgen Lind [Wed, 11 Jan 2012 08:35:39 +0000 (09:35 +0100)]
Move event handling into WaylandInput api
Qt only gives us 1 input device as of now, ie. the mouse/keyboard and
touch events. However, at some point in the future, this will change, so
that the events will have a device id. This will ie on x map to the same
device, but on evdev this can be different devices.
Also this is part of what is needed to implement grabbing
Change-Id: Ice049502d6f0f53fd06142d4dedde05806d60120
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Jørgen Lind [Thu, 5 Jan 2012 12:13:29 +0000 (13:13 +0100)]
Fix so that not all mouse events end up being key events as well
in the Window compositor
Change-Id: I9d88e15ea7ac810a18f6a5b962109d27d2d883a7
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 10 Jan 2012 11:28:32 +0000 (12:28 +0100)]
Add global x and y coordinates to our mouse events
Qt applications don't need global x and y, but the protocol dictates it,
so someone else might need it. We still keep the api to just use the
local coordinates
Change-Id: I4b34df935a6692a8d72946c16603d3d600514161
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 10 Jan 2012 11:07:48 +0000 (12:07 +0100)]
Remove unused variables from wlcompositor and wlsurface
and add the popupinterface with updated sha1
Change-Id: I4c35039451edc0175be4f570b6de1036a147d86f
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Sanity-Review: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 10 Jan 2012 07:14:36 +0000 (08:14 +0100)]
Handle changeCursor in QWindowCompositor
Change-Id: I8ef3b9b36aad00177170f98bcc5c832f884bab04
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 10 Jan 2012 07:13:55 +0000 (08:13 +0100)]
Handle mouse focus in the QWindowCompositor
Change-Id: Ibb994eb8b9e2c69aecdee7daf704475d4d84b183
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 10 Jan 2012 07:09:58 +0000 (08:09 +0100)]
Expose the mouse focus functions from Wayland::Compositor
For Compositors to handle mouse focus handling, ie enter/leave. This
calls down into wayland-server which keeps track of which surface holds
the focus. Trivial implementation in compositor would be to call
this function for all mouseMove events. I believe overhead should not be
significant.
Change-Id: I8fdd5115cc512562c1a8d8b90c28f8ef2d5b53c9
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 10 Jan 2012 07:06:45 +0000 (08:06 +0100)]
Cleanup Wayland::Surface's client handle logic
This is code which is either fairly old or where the api has changed.
Anyway, Surfaces allways have a client handle in the resource, so no
need to check it or store it.
Change-Id: If0be4bed177d42f168b5384a820b8e681e6d9539
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Mon, 9 Jan 2012 17:53:22 +0000 (18:53 +0100)]
Create a ShellSurface class
so that we can actually do something usefull in the callbacks
Change-Id: Ie8e85757bf5c90caa40db34df1b8bc642ba09962
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Mon, 9 Jan 2012 18:02:16 +0000 (19:02 +0100)]
Dont keep track of the focus surfaces
waylands-server does this for us
Change-Id: Ib9980d6023b9b80a9823162d1685ae64da5b61e0
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Paul Olav Tvete [Tue, 10 Jan 2012 12:40:31 +0000 (13:40 +0100)]
Direct rendering fix for previous fix.
In the direct rendering case we disown the buffer, so we end up in
yet another strange state. This code definitely needs refactoring...
Change-Id: Ie5495f9cb1e993e195ba9ecbe19f74425600efa9
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jørgen Lind [Sat, 7 Jan 2012 19:28:25 +0000 (20:28 +0100)]
Make compositor handle subsurfaces with shm buffers.
So we'r lying for now for shm buffers, and assuming they
have allways been displayed.
Also, this patch fixes so that if you attach a buffer, and the last
buffer has not gotten any damages, then it will be released.
Change-Id: Id1b9789aa8e4dcff75147f6c96220039defc0d8f
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@nokia.com>
Jørgen Lind [Mon, 9 Jan 2012 07:58:08 +0000 (08:58 +0100)]
When changing the cursor, call damage on it
so the image is recreated on the server side. If we don't do this any
the QImage/QPixmap on the serverside will not change its cache key
Change-Id: Ie085038e467a4f9676a32bdbecbd27e7cca4c0ff
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Mon, 9 Jan 2012 07:56:35 +0000 (08:56 +0100)]
Add an interface to call damage on a wl_buffer
Change-Id: Id3814572be31fc2e7812fea33e8ed21c53f2d6f1
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Tue, 10 Jan 2012 08:47:03 +0000 (10:47 +0200)]
Remove unnecessary state keeping in input device
Change-Id: I6325573d27b099a21fa5ac5e141764d5172b2453
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 10 Jan 2012 08:33:55 +0000 (10:33 +0200)]
Add a function to the compositor api to send full touch events
A simple sendTouchPointEvent(id, x, y, ...) type of function will not
be sufficient in the future due to the amount of data in a touch
event; therefore an additional sendFullTouchEvent is introduced. This
function takes a QTouchEvent and posts a series of down, motion, up,
frame events as needed. In the future it may be changed so that it
maps to a protocol extension instead of the standard events.
As an example qwindow-compositor is updated to use this new function.
Change-Id: I39d3df1c6d4868364440f59789d01fc5b7b80dac
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Tue, 10 Jan 2012 08:29:38 +0000 (10:29 +0200)]
Add an option to qwindow-compositor for disabling random positioning
When passing -stickytopleft all surfaces will be shown at position 0,0
instead of picking a random one. This is useful for apps with one
single large window in combination with -nofullscreen.
Change-Id: I0834c43981a2d2c7b9796e273ac0eb0cfd738132
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Mon, 9 Jan 2012 16:24:45 +0000 (17:24 +0100)]
Workaround for race condition when compositor is animating
If the compositor is animating, and rendering in another thread, we
may get a frameFinished call for a buffer that has not been displayed
yet. This buffer would then never get released, and the client would
hang. This is just a workaround: the real solution would fix the
race condition between markSurfaceAsDirty() and emit damaged().
Also fix misplaced paranthesis.
Change-Id: Iadecc77a6c6e3999a9f07815db0d0f810d92c1d3
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Mon, 9 Jan 2012 13:12:49 +0000 (14:12 +0100)]
Make the queueing work.
Unconditionally dequeue in frameFinished, instead of checking for
whether a texture has been created.
This removes the frameFinishedTooEarly hack, which was not only ugly,
but also wrong. We have to rely on the compositor calling frameFinished()
for every buffer anyway.
Change-Id: Id284b4578ff5f0813b2e8a12a5cc43293cf79bf8
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Mon, 9 Jan 2012 11:57:47 +0000 (13:57 +0200)]
Remove unused clientAdded signal in compositor.
This signal is not used by anything anymore and so there is no need to
spend time on the contains() call.
Change-Id: I003370e66048a54d1c9e2ec384ad740506d113a2
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Laszlo Agocs [Mon, 9 Jan 2012 09:19:56 +0000 (11:19 +0200)]
Use QOpenGLFunctions in qwindow-compositor.
To keep it building with desktop OpenGL.
Change-Id: I82685926461fc3249ed7a53e741b653273fe6f5c
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Laszlo Agocs [Mon, 9 Jan 2012 09:21:21 +0000 (11:21 +0200)]
Remove logging of key presses in qwindow-compositor.
Change-Id: I505b262687bb32853a69c4dccb79a05f68b1dced
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Robin Burchell [Thu, 5 Jan 2012 00:19:52 +0000 (01:19 +0100)]
Have a crack at fixing setGeometry calls.
Commit
a5ce91c023aeb433f1b5b792291a88c9a5683955 removed setGeometry, so let's
make an effort at still functioning.
Change-Id: Iaae877a2312712ba63aa94a8ae3a3939554c5592
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jason McDonald [Thu, 5 Jan 2012 07:06:27 +0000 (17:06 +1000)]
Update copyright year in license headers.
Change-Id: Idfa1e4479b9ea0c793dcbf8570354d657a1db8cb
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Samuel Rødal [Thu, 5 Jan 2012 14:36:45 +0000 (15:36 +0100)]
Corrected wayland_sha1.txt
Change
d9ea435f0179fe2b5fe0 updated wayland_sha1.txt with a SHA-1 that
was too old, touch_down had not gotten the focus surface as a parameter
yet.
Change-Id: I1979238261758ef0b145997d5cd11c4c8953f1c5
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Andy Nichols <andy.nichols@nokia.com>
Paul Olav Tvete [Wed, 4 Jan 2012 13:51:29 +0000 (14:51 +0100)]
Fix direct rendering after buffer queue change
Change-Id: Ia2ff6f539a8e59ca00ae8506b9367c277ae50304
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Wed, 4 Jan 2012 12:27:32 +0000 (13:27 +0100)]
Queue buffers instead of discarding them
Queue up buffers if the client renders faster than the compositor can display
Change-Id: I1449f60ae35a813c7ef88a15e50fb330e9668231
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Fri, 16 Dec 2011 10:13:59 +0000 (11:13 +0100)]
Add isYInverted property with notification
Change-Id: I6a835a2be4513fa6c24f172503669bc3c95fad0b
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Tue, 3 Jan 2012 08:58:43 +0000 (09:58 +0100)]
Clear dirty set before sending frame callback.
Previously, markSurfaceAsDirty() would have no effect when triggered
by the frame callback function.
Change-Id: Ife7c40757f1cba10e8cdbfab3d514a1dcdac76db
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Paul Olav Tvete [Thu, 5 Jan 2012 13:24:09 +0000 (14:24 +0100)]
Actually return the size from the size() function
Change-Id: I885b1c39f934f2b806843cd9b813a75ea9d8954f
Sanity-Review: Qt Sanity Bot <qt_sanity_bot@ovi.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Robin Burchell [Thu, 5 Jan 2012 06:33:58 +0000 (07:33 +0100)]
Remove QSize from surfaceMapped signal.
Commit
6307ca5e22d81da9a39289c2cca9165fea81b3a5 removed this from the signal, so
we must follow suit.
Change-Id: I4073cfb7e9f6c722c1ac097e3d939767868027cb
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Robin Burchell [Thu, 5 Jan 2012 00:09:28 +0000 (01:09 +0100)]
QSGTextureProvider is now public API in QtQuick.
Commit
42d6acd0a3ea5fa5cf7ab69f1f8f2206559107ea in the QtDeclarative repo moved
QSGTextureProvider to be public API, so we should probably follow if we want to
keep building :)
Change-Id: I1514c0eef9d3d2c2660b38f0bd352c0f885641ba
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Jørgen Lind [Wed, 4 Jan 2012 09:26:20 +0000 (10:26 +0100)]
Remove the geometry from the wayland surface
and add pos and size properties instead. The pos is a PointF while the
size is a integer based Size since pos can be transformed, while the
size reffers to the pixel size.
Change-Id: I5d84aa6661405cb0df356b787246d0d73ad0c503
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Wed, 4 Jan 2012 09:22:49 +0000 (10:22 +0100)]
Fixup waylandsurface item after QSize was removed
from the mapped signal
Change-Id: Ib05fad9393d33a965e50ea3459fa83b10cd6d567
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Wed, 4 Jan 2012 07:16:27 +0000 (08:16 +0100)]
Make it possible to drag windows around
while holding the windows key down.
Change-Id: I1b5cd7fb39b5ec0ead6a4c312b4bb653a0191a00
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Wed, 4 Jan 2012 07:33:28 +0000 (08:33 +0100)]
Implement subsurfaces extension
This allows you to have subwindows in the compositor.
We tried to experiment with composing subsurfaces client side, but the
architecture did not feel very lean. This however, requires the
compositor to compose each surface before drawing the surface. The
example compositors render the subsurfaces into the wl_surfaces texture.
This might not be a good idea.
Change-Id: I6e186b62d7b490de7f4e6c6f22fcf6c1e0a70df3
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen Lind [Tue, 3 Jan 2012 12:47:33 +0000 (13:47 +0100)]
Add mouse move events to the window compositor
Change-Id: Iacdfbd55feb5db4d38eac6e2729684052a46afa9
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen [Tue, 3 Jan 2012 11:36:38 +0000 (12:36 +0100)]
Add a .gitignore
Change-Id: I3e5e42425825c61a76197c91c8bd2faf7bc64231
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Jørgen [Tue, 3 Jan 2012 11:30:58 +0000 (12:30 +0100)]
Dont have subdirectories in the extensions folder
This folder is just for the xml files
Change-Id: I20820bac6ed46cb0267b20afddd129ddf64141c5
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>