For some reason, XCB accepts positions as unsigned
integers, even though the X protocol explicitly
allows negative values (windows overlapping the
left/top screen border).
After discussion with Samuel and Laszlo, use a
reinterpret_cast to convert from the signed
representation to the unsigned one.
I also extended the clipping of the extents to
the position. I guess if X can't handle widths
beyond XCOORD_MAX, it won't be happy with
x-coordinates exceeding that limit, either.
Change-Id: I2fa0e61f823b6cd45dad6471eaa55f38bb3c3e52
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Previously the primary sceen number wasn't stripped before adding the
actual screen number, resulting in names such as ":0.0.0".
Change-Id: Ifbcf12dcf1c200fd6bc5e4b95a5d82483957e0d2
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
The XCB plugin requries libxcb >= 1.5. Configure and
config.tests/qpa/xcb now check for this.
Change-Id: I96c688b79bf5b49fd3ecc4ddc12ebdc2d3788790
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Adds missing QT_NO_ACCESSIBILITY checks where required to build
without accessibility support.
Change-Id: Id98ecdcb9b351289b21dc2d382100d0b63857db9
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@nokia.com>
We should only query the position for top levels, otherwise trust the
position given in the event.
Change-Id: Ic29f25983af3e2c2f27eeb527c08069435ac938c
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
In the old implementation, the QPlatformIntegration was asked
for the theme first, so there was no way of overriding that
by a custom plugin. Also, there was a memory leak in case
the platform theme was actually created by a plugin.
QGuiApplication now asks the QPlatformIntegration for a list
of potential theme names first, tries to load them using
the plugin loader and finally invokes a factory method of
QPlatformIntegration in case that fails. The theme is now
owned by QGuiApplication.
In the XCB plugin, the environment variable DESKTOP_SESSION
is queried and appended to the list of themes, making it possible
to load plugins for other session types.
Change-Id: I1a4b4e061815bca16c65b23e591bb7563a3e44e2
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Change 50ca45f059 caused mapped windows to not receive focus by
default. The _NET_WM_USER_TIME spec says that if the user time is set to
0 (or CurrentTime), the window will not be initially focused. Thus, if
the connection time has not yet been set, we skip this part to restore
the old behavior.
Change-Id: I19de3602c78629ad2bc65f5e1976313949c82c4c
Reviewed-by: Jan Arne Petersen <jpetersen@openismus.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
- Introduce cursor() accessor to QPlatformScreen.
- Remove screen member of QPlatformCursor (a
cursor can be shared by multiple screens
of a virtual desktop).
- Add QCursor::pos()/ QCursor::setPos() taking
a QScreen-parameter, use primaryScreen() for
old overloads. QCursor::pos() can then query
the platform cursor for the position and return
the position even if the mouse position is outside
the windows owned by the Qt application.
- Fix tests
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Task-number: QTBUG-22457
Task-number: QTBUG-22565
Task-number: QTBUG-20753
Change-Id: Ia69f37343f95772e934eab1cd806bd54cbdbbe51
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Setting the user time before mapping the window seems to prevent the
crasher from happening. We used to set the user time before mapping in
Qt 4.8 too, so it's probably the right thing to do.
Task-number: QTBUG-24462
Change-Id: Ia670b799bd1ed7a7e6399631d5242e57324918b3
Reviewed-by: Jan Arne Petersen <jpetersen@openismus.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Got rid of unnecessary debug output, and prefixed debug / warnings
with class name to provide more context.
Change-Id: Ia68e85ec0207de4d87d02226bdf6e4cddf464afc
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
This change makes sure that the X11 server really supports the SHAPE
extension before it is used.
Before this, libxcb would disconnect from the server as soon as it was
told to send a SHAPE request to a server which doesn't support SHAPE. Oh
and of course it would do so without any good error message.
Change-Id: I5f9481e488925e74e62f229e9f637ef262020c04
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Screens need to be destroyed in reverse order to ensure the primary
screen stays valid when other screens are destroyed.
Task-number: QTBUG-24300
Change-Id: I9d9d710aa67ec045baa8bf292833ffe7d9eea935
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
Otherwise metacity was crashing when a QWindow was destroyed immediately
after being activated, because metacity was trying to select events
(XSelectInput) for the already destroyed m_netWmUserTimeWindow.
Task-number: QTBUG-24492
Change-Id: Iedbe7bdd6b26110ca8bec6f33525209ae551ffd5
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
In previous commits we took into account bound shaping, but X also
supports input shaping, so make sure it's inside both input and bounding
shaping to consider the position as inside a window
My tests show that when unset Bound Shaping and Input Shaping return the
rectangle of the window itself, so we need to be inside both of the
rectangle sets to consider the position as a dragabble position for the
window
Change-Id: I77973e51168d2a6d274d166f29f32e1bc02c32c3
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Use a _NET_ACTIVE_WINDOW client message instead of xcb_set_input_focus
for activating toplevel windows on xcb.
According to the Extended Window Manager Hints the right way to activate
a top-level window is using _NET_ACTIVE_WINDOW (when it is supported by
the WM). Other approaches like calling xcb_set_input_focus should be
avoided when possible, since the WM cannot intercept them.
Change-Id: I9be4901f56cbcfb563baf73ccd71ff17a9bdc1d2
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
- Give QPlatformDrag a synchronous drag() function returning
the Qt::DropAction
- Move the base functionality for asynchronous event handling
to the platformsupport library as QBasicDrag (extendable base class
handling drag icon and providing new virtuals) and QSimpleDrag
(sample implementation for drag within the Qt application).
- Change the Windows implementation accordingly.
- Change XCB to be based on QBasicDrag.
- Clean up QDragManager.
Change-Id: I654f76f0e55a385ba189bd74f3ceaded6a8fe318
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
tst_qclipboard.cpp still has Q_WS_WIN which must be clear away. After
clean up, auto test will fail under windows: When setMimeData() is
called, dataChanged() signal will be emited twice.
The solution for QTBUG-24184 has partially solved the problem, but it
still there. Make sure emitChanged() only called by QPlatformClipboard
will give our more control for this.
Task-number: QTBUG-24484
Change-Id: I23566c6d3b32828b6865234c311af3635fe9e299
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Even with the more recent libxcb, it's a bit flaky. Just don't list
having this capability until it is more stable.
Change-Id: I1c2b39d757803fc23295944bc9a00e39ff4e1bbf
Reviewed-by: Gunnar Sletta <gunnar.sletta@nokia.com>
Add window flag to support windows which should not get the input
focus.
Sets the input field in the WM_HINTS structure of the window to false
if the WindowDoesNotAcceptFocus flag is set on a window in xcb.
Change-Id: Ifbc10695b83484c17dca0eb13ea826d74f174833
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Otherwise we end up with an error at runtime:
QEGLPlatformContext::makeCurrent: eglError: 3009, this: 0x4b5c58
Change-Id: I1e794f427243e40f0a8f572a936b9fb756e89f8a
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
- Move palette() from deprecated QtWidgets/QGuiPlatformPlugin
to QtGui/QPlatformTheme, Make it return a const * since
QPalette does not have isNull().
- Initialize QGuiApplication::palette() and
QApplication::systemPalette() from it.
- Do not initialize QPalette from QGuiApplication::palette()
unless app_pal is non-null (default to Qt::black if it is 0).
This avoids initialization order crashes/recursions in the
QPA plugin. Streamline initialization function.
- Remove styleName(), systemIconThemeName() and iconSearchPaths()
from QGuiPlatformPlugin and re-add them as
QPlatformTheme::themeHint().
- Remove styleHint() from QGuiPlatformPlugin, add it to
QPlatformTheme::themeHint().
- Add UNIX themes with factory function (Generic,
KDE, Gnome), taking it from Qt 4.8 code (stripping the
KDE 3 code).
- Implement Windows palettes.
Task-number: QTBUG-24204
Change-Id: Ie27ec035df4f84c42deaffc4816b2e53ce705462
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
Change the use of QList<QPair<>> to a simple C array of xcb_keycode_t*s.
This is possible since the number of elements in the container is
statically known.
I have not measured speedup; this is just preventing premature pessimisation.
Change-Id: I8855fc8a4e7ee840d8b7497ec4166074da7d8ea8
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Since a GLuint returning virtual was added to QPlatformOpenGLContext
we need to make sure it's protected by #ifndef QT_NO_OPENGL.
Change-Id: Id2f56ccdccc3863986250ee1b3aa7efccf88ba5c
Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
- Add QPlatformServices as back-end for
QDesktopServices.
- Bring back UNIX/Linux desktop detection in platformsupport
as a generic implementation.
- Add Windows implementation.
Reviewed-by: Morten Johan Sorvig <morten.sorvig@nokia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Anselmo Lacerda S. de Melo <anselmo.melo@openbossa.org>
Change-Id: If94bb65755df4f849edd83c57143ee2c73002137
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Returning pointers to unexpected types for unknown keys is quite
wrong. The clients expect 0 in such a case but what they got (until
now) was whatever was associated with the default constructed enum
value.
Change-Id: Iefd7bf461bfb2c1f4c73f5f9f291aecad60219eb
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Reviewed-by: Jani Uusi-Rantala <jani.uusi-rantala@nokia.com>
It can happen that there is a window covering all the screen but it is shaped
to only take part of the screen. If that happens, besides the condition of
QRect(attr.x,attr.y,attr.width,attr.height).contains(pos)
we also need to query the server for its region rectangles and make
sure the cursor is inside one of those rectangles. If that does not happen
we have to return 0 so the hierarchical xcb_query_tree_children xcb_query_tree_childrenntinues
Change-Id: I67327672e3d854d064190b55a07bd56ade4dfddb
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
If we call any xcb_shm_*() function, libxcb will have to figure out
SHM's major number. However, if that fails because the SHM extension is
not available, xcb will disconnect from the server. Obviously, that's
not what we want and must be avoided.
Change-Id: I8eba5ff9e84e0e2017a5c0d88fda4efd0459bd1e
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
If we call any xcb_sync_*() function, libxcb will have to figure out
SYNC's major number. However, if that fails because the SYNC extension
is not available, xcb will disconnect from the server. Obviously,
that's not what we want and must be avoided.
Change-Id: I9a1032e1cfac074a52bafcb0772304bfd423e770
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
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: I311e001373776812699d6efc045b5f742890c689
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Qt::UnknownOrientation is now Qt::PrimaryOrientation.
Change-Id: I2846a90bd11ecd1416c29b7bd861f180ccb129a1
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
Some older drivers don't fully support glXChooseFBConfig(). As a bonus,
fix some memory leaks here and there.
Task-number: QTBUG-21880
Change-Id: Ie306dee27f616927a6aa55fd71601569b828afcc
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Remove QByteArray-construction and hash lookup in the event
handling; use an enumeration indexing an array instead.
Change-Id: I4d272b32a5ff71c8da58197cf3a0b38c1e61d489
Reviewed-by: Simon Hausmann <simon.hausmann@nokia.com>
Commit e08453f31a added a check which makes
sure qt prints an error message instead of segfaulting when it can't
connect to the X11 server.
However, libxcb will never return NULL from xcb_connect() and thus that
commit only works if XCB_USE_XLIB is defined and used.
Change-Id: I8cc1496494a94f07055a3ac5093ce362dd347c5b
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
The draganddrop examples all crashed here because they were using a
default-constructed QImage() (i.e. one without any content).
I guess this happens here because I don't have any mouse theme set.
To test, one could start a second X server, but without any WM or DE.
The "evil" QImage() came from QGuiApplicationPrivate::getPixmapCursor().
This function seems to just always "return QPixmap();".
This fix is correct because the only caller has another fallback if the
createNonStandardCursor()-fallback didn't work. This caller is
QXcbCursor::createFontCursor().
Change-Id: I7ec7fbcfdf0203e983149b5e73016cc7e85ecf40
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Interface to provide cross-process caching mechanisms in a platform
plugin. Can be used for shared glyph caches and icon caches etc.
Change-Id: If0d89a0a50bbd6eee05daf908448262ff270fc5b
Reviewed-by: Jiang Jiang <jiang.jiang@nokia.com>
The setEventFilter on the platform native interface allows subscribing to
events on the backend by event name.
Change-Id: Ib10077fbc69f0207edbae1177e7bbd18c8d0f9ae
Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
Reviewed-by: Michalina Ziemba <michalina.ziemba@nokia.com>
Set the _MEEGOTOUCH_ORIENTATION_ANGLE property on the window, just like
Qt Components for MeeGo and MeegoTouch itself.
Change-Id: I0b9adf4550593678bbcba89a2d4f1f65c1f4bd20
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
When connecting to an X11 server, the server tells the client what its
maximum allowed request size is. Larger requests are not permitted.
Thus, libxcb verifies that all requests which are sent are smaller than
this size limit and possibly kills the connection (without any good
error message).
Thus, when uploading an image, we could be trying to send more pixel
data than fit into a single request. If this would be the case, the
code will now use multiple requests where each request only sends a
part of the allowed rows.
In case all the data fits into a single request, this commit shouldn't
change any behavior.
Change-Id: I84a4ebfcdb6de7e206015c37e3e33cba3bd309b1
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Be more user-friendly (e.g. when logging into a machine by ssh and
forgetting to export DISPLAY).
Change-Id: I9d07b0af9c5b4841827826053bb27b507801ae61
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
The current way we do it of having the platform or touch plugin send
both mouse and touch events is not ideal. There's no good way to write
an application that works sanely both on a touch-only device and on a
desktop except by restricting yourself to only handling mouse events. If
you try to handle touch events you don't get any events at all on
desktop, and if you try to handle both, you end up getting duplicate
events on touch devices.
Instead, we should get rid of the code in the plugins that automatically
sends mouse events translated from touch events. This change enables
that by making the behaviour fully configurable in QtGui.
Two new application attributes are added to explicitly say whether
unhandled touch events should be sent as synthesized mouse events and
vice versa, and no duplicates are automatically sent as the current
situation. Synthesized mouse events are enabled by default.
We also get rid of the QTouchEvent::TouchPoint::Primary flag, which
was only used to signal that the windowing system automatically
generated mouse events for that touch point. Now we only generate mouse
events from the first touch point in the list.
Change-Id: I8e20f3480407ca8c31b42de0a4d2b319e1346b65
Reviewed-by: Laszlo Agocs <laszlo.p.agocs@nokia.com>
Reviewed-by: Jocelyn Turcotte <jocelyn.turcotte@nokia.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@nokia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
Marc Mutz
Friedemann Kleint
Thiago Macieira
Samuel Rødal
Bradley Smith
Jerome Leclanche
Donald Carr
Lars Knoll
Uli Schlachter
Miikka Heikkinen
Jan Arne Petersen
Albert Astals Cid
Debao Zhang
Alan Alpert
Tor Arne Vestbø
Laszlo Agocs
Albert Astals Cid
Jason McDonald
Simon Hausmann
Eskil Abrahamsen Blomfeldt
Rohan McGovern