This allows V4 and other garbage collectors to pass through our
custom stack during the mark-phase without triggering access
violations.
Change-Id: Icafcf4df3537c628c641fe694bb9fe2016519a83
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Update the Qt 4 documentation of QSessionManager to reflect the changes
in Qt 5
Task-number: QTBUG-33528
Change-Id: I74286134155efc2781f9f6fc505fb6cf736d814e
Reviewed-by: Martin Smith <martin.smith@digia.com>
QSpiAccessibleBridge uses a D-BUS connection, which in turn uses socket
notifiers and timers internally. Neither of these can be used before
a event-dispatcher is in place, so we need to defer creation of the
accessibility interface until later. We assume that clients will
only call QXcbIntegration::accessibility() when an event-dispatcher
is set up, but to be extra safe we do an assert, so that failures of
this pre-condition will trigger at the place they are caused -- not
as failures to register socket notifiers and timers in the D-BUS code.
Change-Id: I4f9d8362a3f285c3da9045d1ff6b8e7b04570488
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
The size is queried from the window before create() is
called. Therefore Quick apps were not covering the entire screen as
they should. This is now fixed by properly returning the fullscreen
geometry from QEglFSWindow::geometry() even when create() has not yet
been called.
Pre-5.2 this was working because the QEglFSWindow constructor changed
the window state. This is not feasible anymore with the recent
compositing changes since the second, third, etc. windows are not
fullscreen.
Task-number: QTBUG-33620
Change-Id: If66d35f6f4768485d4e68c5ad825fe7a9d948a64
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
The QPlatformIntegration::guiThreadEventDispatcher() function acted as an
accessor to event dispatchers created in the constructor of each platform
plugin, but the logic and semantics of event-dispatcher handling in Qt
itself (QCoreApplication/QGuiApplication) still assumed both ownership
and control over the event dispatcher, such as when to create one, which
one to create, and when to delete it. This conflicted with the explicit
calls in the platform plugins to QGuiApplication::setEventDispatcher(),
as well as left a possibility that the event-dispatcher created by
the platform plugin would never be deleted, as none of the platform
plugins actually took full ownership of the dispatcher and deleted it
in its destructor.
The integration function has now been renamed back to its old name,
createEventDispatcher(), and acts as a factory function, leaving
the logic and lifetime of event dispatcher to QtCoreApplication.
The only platform left with creating the event-dispatcher in the
constructor is QNX, where other parts of the platform relies on
having an event-dispatcher before their initialization. We then
need to manually take care of the ownership transfer, so that the
event-dispatcher is still destroyed at some point.
Change-Id: I113db97d2545ebda39ebdefa865e488d2ce9368b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
QCoreApplication sets up the event dispatcher in the constructor, through
createEventDispatcher(), which is overridden in QGuiApplication to take
the platform-integration into account. The problem is that the platform
integration is created before the event dispatcher, so when we create the
QXcbConnection we are not guaranteed to have an event dispatcher yet.
This is not an issue when using xcb_poll_for_queued_event() in a thread,
but for the fallback case we connect to the event-dispatcher's awake()
and aboutToBlock() signals. To ensure that we do this only when we have
an event dispatcher we post a queued method invocation, that will be
processed as one of the first events once there is an event-dispatcher.
Change-Id: I623011af447b585884b84c7559737f134aab83e8
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
bab29dd76 changed scroll event handling to update
the keyboard modifier state during scroll event streams
and brought back the bug where Qt Creator would suddenly
change the font size if you pressed the command key
while scrolling.
Use momemtumPhase to detect scroll event streams.
momemtumPahse will be NSEventPhaseNone as long as there
are fingers on the trackpad. After liftoff it will
have a different value, and at this point we freeze
the keyboard modifier state until the scroll event
stream has ended.
This approach does not interfere with normal mouse
scroll wheels since momentumPhase is PahseNone for all
such scroll events.
Change-Id: I668ade8e794a97491487c9465366a316a7d6a9c8
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
QCocoaWindow keeps track of popup windows in order
to close them when appropriate, for example on mouse
clicks outside the popup. In this case the mouse click
will be used to close the popup, and will not be
send to the QCocoaWindow.
Clear m_activePopupWindow when the popup is closed
to fix a bug where the first mouse click after
closing a popup would have no effect.
Change-Id: I7a3e850e4fbf4686540d5d89dac6d7b613ed10a4
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
This is necessary for QMacNativeWidget, where there exists a QApplication, but
is never executed. This directly translates in calls to the event dispatcher's
processEvents() function, whose calls we keep track. If no calls have been made,
we always allow timer and posted events to be processed.
Change-Id: Ia0062ee8c59a2572082f520a2eb85ed44a9856a7
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
In particular, this is needed when a QWidget has WA_TranslucentBackground set
and nobody is painting anything behind it (except maybe some native view).
Change-Id: Ib1f0714f85fa7eeced527617ecd09bb2ed6ddfc9
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Even if there is not a new window, the notifications from the old one
are not needed.
Change-Id: I9c1858d25e49379ca4737e23beec06623e91b69c
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Task-number: QTBUG-26722
Change-Id: I579111b5d34f8e3cdc6bb016d9c0e42ec3ffb8c9
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Stephen Chu <stephen@ju-ju.com>
On Mac the QT_NO_PRINTER build was not implemented.
Task-number: QTBUG-33565
Change-Id: I118472f9400aa0a0d0e192ae39a11ea38a66f340
Reviewed-by: John Layt <jlayt@kde.org>
Check only bit 8 to decide if the mouse event is generated from
a touch or pen event.
Task-number: QTBUG-33460
Change-Id: I83b23267b5de6df5e0e6b7113ecf377dd7e86c84
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
We tended to ignore the original application delegate a bit too often.
Change-Id: I0844c8658d128e4fbb9a6fc5000025f55e5293c2
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
This patch simplifies the implementation of touch events to use a
views superview for calculating global touch coordinates rather than
the screen. This removes the need for taking orientation
into account, and will also play better along in a mixed environment.
This will also fix touch events reported for inverted orientations.
Change-Id: I0c8fd8745a1f65f0f4a97447a5676a38165ed032
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Unfortunately we returned the column description when the AT client
asked for the row description....
Change-Id: I46bc0edb4fd0f7cc6d98d7d6e0d8ca6f77553a26
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Before this patch, we would only rotate if no autoRotationMask was set.
This was a temporary way to lock orientation from code until a better
API for this was in place.
But this causes problems for applications that both wants to auto rotate
but at the same time sets a mask to get QScreen::orientation
updates. So remove this heuristic before application code starts to
depend on it.
Change-Id: Idb54abd471b33afd866322738f4860c57bc9dcf7
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
This patch aims to implement the session management for Windows.
Based on the Qt 4 QApplication windows specific code
Task-number: QTBUG-28228
Task-number: QTBUG-33032
Change-Id: I7059298ad9661aebd51e77a03f55a7a04f461479
Reviewed-by: Kai Koehne <kai.koehne@digia.com>
Similarly to how it's done in xcb. And add support for eglwindow since
the WId will soon once again cease to be an EGLNativeWindowType.
Change-Id: I0e3b86a21179439821550c9423f0e747ccae5897
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
On the BlackBerry platform when an application is minimized an
alternate cover window is shown. This patch allows Qt to render
into those windows.
A new window type "Qt::CoverWindow" has been introduced.
Change-Id: If5e321f5cb00f6d3ed7a5dc0af425655d7a9253a
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
This fixes the build breakage caused by Change I50b33d05 when attempting
to compile on Linux with session management enabled but libSM and/or
libICE not present.
Change-Id: I127f32e7041deb2ff968eae8458fee7d4e95d7a5
Reviewed-by: Samuel Gaist <samuel.gaist@edeltech.ch>
Reviewed-by: Bernd Weimer <bweimer@blackberry.com>
Reviewed-by: aavit <eirik.aavitsland@digia.com>
This change makes use of the private QJNI api that now exists in QtCore.
Change-Id: Id03664942d6c6cbfec028a3b52729057153a8466
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
This avoids the assertion in iaccessible2.cpp(510)
Change-Id: I1a4c007ffcbcda70f0e37ef3cf55a303683b58c1
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Various global changes, primarily preprocessor flow, to support the
WinRT platform.
Change-Id: I3fa9cf91d5fb24019362e88fcf205e31b4f810b5
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
To be set to false by eglfs and similar platforms where no real window
management is provided.
When this capability is not set, some of the changes done for
QTBUG-26903 will be disabled, allowing dialogs to show up at their
QDialog-chosen place on platforms where no WM is present and so
windows are never repositioned by the system.
Change-Id: If1dac3cd11f2a856913a51277431fe3ec644b719
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Allow widget apps with popups and multiple top-level widgets to
function on the eglfs platform. GL and Quick2 apps are not affected.
Instead of trying to create a native window and EGL surface for each
window, do it only for the window that is created first. This first
window is forced to fullscreen as usual. Later windows however are
treated differently: These will not have a native window, surface or
context, and keep their normal size. All the textures belonging to the
raster windows are then rendered in one step, using a stacking order
maintained based on visibility changes and window activation.
Note that this will only help apps that create a main window first and
have everything else inside that window or on top of it as menus,
dialogs, popups, etc.
Change-Id: Ia435458ba81bf3c35cc8f61bcb2d2a50cf17f0e3
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
This reverts commit 04325bdd26.
This change breaks qtdeclarative on Mac.
Change-Id: I77b121cc6b283cf7498e93e4b914d2fb68808ab6
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
There is no point in having QAccessible2, so merge it with the normal
QAccessible. The header will be removed in a subsequent commit as it is
still needed by declarative at this point in time.
Change-Id: I1fc47d484d482f25387eba827bc5a373536b7a8b
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Make sure that we are not reading wrong PropertyNotify events.
From icccm:
Requestors may receive a property of type INCR in response to any
target that results in selection data. This indicates that the owner
will send the actual data incrementally.
The selection requestor:
- Waits for the SelectionNotify event.
- Loops:
+ Retrieving data using GetProperty with the delete argument True.
+ Waiting for a PropertyNotify with the state argument NewValue.
- Waits until the property named by the PropertyNotify event is zero-length.
- Deletes the zero-length property.
The issue with the current approach was that after receiving INCR we
simply fetched the first PropertyNotify event from the queue, where
timestamp indicates that it was generated before INCR, which is not
what need. We need PropertyNotify events with a timestamp older than
the one on INCR.
Task-number: QTBUG-32045
Change-Id: I3b9a006f7d346f67c51e1a296d67a9dc89efadb1
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Implement INCR property mechanism according to the icccm specification.
Change-Id: Ic8f85b71cab825d70ee1b61f29acd09fa4c3e642
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
There was a disagreement between the a11y plugin and QTreeView
whether the horizontal header should have been exposed or not.
When the header was hidden, this resulted in that we sent an event
with a child id that was wrong, or in worst case higher than
QAI::childrenCount(). This was the reason we got the warning
output as described in the task.
With this commit, we consistently *expose* hidden headers both for
QTreeView and QTableView, but ensure that their state().invisible is
set to true instead.
This makes it consistent with how hidden cells are exposed.
This also fixes a bug in QTableViewPrivate::accessibleTable2Index
where we always added 1 to the index, which was spotted while
writing the test.
Task-number: QTBUG-33247
Change-Id: Ifd1f83d56296dd071424fdb81fce7628bc24fe0a
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
The backtrace in QTBUG-33012 indicates that the
screen list is getting out of sync on plugging/unplugging
an external monitor. This happens on 10.9 only.
Check if we are about to out of bounds access the
screen list; call updateScreens() which will add the
new screen.
Task-number: QTBUG-33012
Change-Id: Id1acc3baca9e60e7f2a83188d66cdf9d322a582c
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Replace all xlib calls with xcb equivalents, leaving only the
absolutely required xlib calls. Handle WM_DELETE_WINDOW so that
closing the window exits the app as expected. Finally, introduce
EGLFS_X11_FULLSCREEN to enable requesting a fullscreen native window.
Change-Id: I8c46ae832d38549ec7d673592f400a4f34bf4314
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
With XCB a client cannot retrieve the information on which X screen it
is running. This information is only available when opening the xcb
connection. As this is done by Qt no application is able to retrieve
this information. By exporting the x11 screen we can provide this
information again in QtX11Extras.
Change-Id: I50f1d3e803dc7e3afac0e5c7f1648ccda4502e7c
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
While at it, I hope that qplatformtheme API's are stable
by now :)
Change-Id: I83bf3118c91608710bb19380458e8f55cb3d427b
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Clang 3.3 found a variable assigned to itself, probably to cause
GCC to shut up about a variable set-but-unused. So simply stop
assigning the value at all to this variable.
qxcbxsettings.cpp:155:16: error: explicitly assigning a variable of type 'uint' (aka 'unsigned int') to itself [-Werror,-Wself-assign]
Change-Id: I74ae8f724e87c1b3f6b9d358e13d6a440ff4a3e1
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
The responsibility of sendWindowSystemEvents() is to process events from
the window system. Historially that logic was part of the QPA/QWS event
dispatcher, which naturally also sent posted events. Through refactoring,
the code at some point ended up in in the QWindowSystemInterface class,
still with the posting of events in place.
This resulted in QPA event dispatchers adopting a pattern of just calling
sendWindowSystemEvents(), as that would cover both posted and window system
events. Other event dispatchers would call sendWindowSystemEvents(), and
then use a base-class implementation from QtCore for processing events,
resulting in two calls to QCoreApplication::sendPostedEvents() per
iteration of processEvents(). This breaks the contract that processEvents
will only process posted events that has been queued up until then.
We fix this entanglement by removing the sendPostedEvents() call from
QWindowSystemInterface::sendWindowSystemEvents() and move it to the
respective event dispatchers. For some EDs it means an explicit call
to sendPostedEvents, while others were already doing sendPostedEvents
though a separate source (GLib), or using a base-class (UNIX/BB), and
did not need an extra call.
We still keep the ordering of the original sendWindowSystemEvents()
function of first sending posted events, and then processing any
window system events.
Task-number: QTBUG-33485
Change-Id: I8b069e76cea1f37875e72a034c11d09bf3fe166a
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
QPlatformBackingStore::endPaint does not take a QRegion parameter.
6ce6b8a378 set the API, but the
platform implementations were not all synced up since then. There
was anyway no point in overriding beginPaint and endPaint on
platforms which don't need to do anything there. This fixes
clang warnings of the form
QXcbBackingStore::endPaint hides overloaded virtual function
Change-Id: Id6cd0fc2c831a34576ac2c73eeb0d5741d26e622
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Center gravity doesn't mean center the window, it only affects the
method of converting between internal window bounds and decorated bounds.
So wanting to have each dialog centered w.r.t. its transient parent
is not a reason for using center gravity. Instead it caused the
bug that when you resize a QMessageBox by clicking the Show Details
button, it jumps downwards on the screen.
Task-number: QTBUG-32473
Change-Id: I3fabf3caa1e4d10fd4f7508e297f73efe5cc51b6
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
This reverts commit 59fd36bb914cae248d3c5100dd734d6d90a58dca,
which caused other drawing problems.
Change-Id: I3bb75fd3ca1cd21ffbb9ef5474266f4cd615a64a
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Our previous event loop integration had two unfortunate flaws:
1. We would call qt_user_main() from a timer, after returning from
didFinishLaunchingWithOptions. This had the effect of showing the
iOS application window long before the Qt application UI had been
set up, resulting in a 1-2 second flash of black/pink between the
launch image disappearing and the actual application showing.
2. We spun a nested event loop, where our implementation of the
different event loop modes did not perfectly match the Apple
implementation. This resulted in scrolling being busted in
some cases such as when showing the virtual keyboard for
Emoji characters.
These two issues have now been solved by calling the user's main()
from didFinishLaunchingWithOptions. Normally this would not work, as
the user's main would call QApplication::exec() at the end of their
main(), which would block and we would never return back from the
didFinishLaunchingWithOptions callback, resulting in no UI on screen.
We work around this by longjmp'ing out of QApplication::exec(), back
into didFinishLaunchingWithOptions, so that it can return. Again,
this would normally not work, as the call stack where QApplication
and friends would live would get smashed as the application
continued executing. We work around this by allocating a block
of stack space at the start of main(), which we then redirect the
stack pointer to before calling the user's main. This results in
the whole stack of the user's main() and below being preserved, even
if we longjmp out of the call stack (which then restores the
stack pointer).
This approach should work fine together with garbage-collection as
well, since the mark-and-sweep phase will walk the stack from the
stack pointer to the stack base, including sections of the stack
that were part of qt_user_main() and live in the reserved area.
One case where GC will fail though is if it happens as part of the
qt_user_main() call, where the GC will not mark anything in the
'real' callstack below UIApplicationMain(), but this is not
expected to happen.
The size of the reserved stack can be controlled through the
Info.plist key 'QtRunLoopIntegrationStackSize', as well as the
'QtRunLoopIntegrationDisableSeparateStack' key to disable the
separate stack approach completely. This will fall back to the
old approach. The amount of stack space used by the user's
main can be determined by enabling a special debugging mode,
using the 'QtRunLoopIntegrationDebugStackUsage' key.
Change-Id: I2af7a6cfe1a006a80fd220ed83d8a66d4c45b523
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Added QXcbSessionManager to the Xcb plugin.
QXcbSessionManager inherits from QPlatformSessionManager, it's a port of
QSessionManager as it is in Qt 4.8.
Minor changes also in QPlatformSessionManager and QGuiApplication to
hook it up.
Task-number: QTBUG-28228
Task-number: QTBUG-30011
Task-number: QTBUG-33033
Change-Id: I50b33d05a1e32c5278dea339f693713acc870a70
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Mesa does not like eglSwapInterval calls without a current context.
Change-Id: I7ec2d4311586cf74da0461bc951a0e5d9399c35b
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
When trying to create more than one window, stop with a helpful error
message since this is not yet supported. Also, return a fake WId for
desktop windows.
Change-Id: I9859b62b1d4f6b6142982d2e5a90afc1fc3c6a28
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
It is not really fatal but fixing it gets rid of a Mesa warning.
Change-Id: I3045b2691e7457541d6524c3e3ff8a1882ca460b
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Move scriptRequiresOpenType() body right into QFontEngine::supportsScript(),
thus centralizing use of this performance cheat.
Change-Id: I5f494b086f8f900b631c491f41e9cb800002c0f6
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Implement the foreignWindowCreated and foreignWindow closed signals, emitted when foreign windows are created and closed, respectively.
Change-Id: I72dd5380e6061f191eb8362fda5dd8fb8e9ed06b
Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com>
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Implemented nativeResourceForScreen api to return QObject* which can be used to connect foreignWindowCreated, foreignWindowClosed signal in QQnxScreen.
Usecase is to connect signal in custom QML component as below
QObject * obs = interface->nativeResourceForScreen("QObject*", screen);
connect(obs, SIGNAL(foreignWindowCreated(void*)), d, SLOT(newForeignWindowCreated(void*)));
Change-Id: I512c3b6d188a2e90ef7b8e89c413ca420a29dd9b
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Fixes qxcbsystemtraytracker.cpp:125:134:
error: 'connection' was not declared in this scope
Change-Id: If881aa9466ea94d5392da9f177e4b79e044710b7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Windows will only be exposed and hence rendered when they are not
minimized. This will save useless computations and hence battery.
Change-Id: I83166cc6c3d89e878106c998a35890dd7788ed8a
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Use the UpdateLayeredWindow() functions only for windows with alpha.
Task-number: QTBUG-33025
Change-Id: I64b0c28ee0997cd3d09dc76babe105ed474c6835
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
We recommend using pixel sizes for predictable results. For those who
use point sizes in their UIs, we will now give them results that look
similar to what we do on iOS. The default font is changed to give
the same size as before this change.
Task-number: QTBUG-32096
Change-Id: Ia25506ba721a39d31340f3df8bc14129e507af14
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Reviewed-by: BogDan Vatra <bogdan@kde.org>
IFileDialog::close() only works from callbacks. Try to find
the dialog window and send it a WM_CLOSE in addition.
Change-Id: Id0f89f8781564e19e4763d43a71df55d5299fb35
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Instead of using a define to rename the user's main() function during
compilation, we leave the user code alone, and inject our wrapper one
step earlier in the process, at the application entry point 'start'.
This entry point is provided by crt1.o, which is normally linked into
the application automatically. The start() function sets up some state
and then calls main(), but we change the start() function to instead
call our main wrapper.
Instead of shipping our own crt1 binary/sources, we make a copy of
the appropriate crt1.o at build time, and modify its symbol table in
place. This is unproblematic as long as we keep the same length for
the wrapper function name, as the symbol names are just entries in
the global string table of the object file.
The result is that for the regular Qt use-case the user won't see
any changes to their main function, and we have more control over
the startup sequence. For the hybrid use-case, we no longer rely
on the fragile solution of having our back-up 'main' symbol in
a single translation unit, which would break eg with --load_all,
and we don't need to provide a dummy 'qt_user_main' symbol.
OSX 10.8 and iOS 6.0 introduced a new load command called LC_MAIN,
which places the state setup in the shared dyld, and then just
calls main() directly. Once we bump the minimum deployment target
to iOS 6.0 we can start using this loader instead of LC_UNIXTHREAD,
but for now we force the classic loader using the -no_new_main flag.
There's also a bug in the ld64 linker provided by the current Xcode
toolchains that results in the -e linker flag (to set the entry
point) having no effect, but hopefully this bug has been fixed
(or Apple has switched to the LLVM lld linker) by the time we
bump our deployment target.
Change-Id: Ie0ba869c13ddc5277dc95c539aebaeb60e949dc2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
When the QIOSApplicationState object owned by the platform integration
was deleted we would deallocate QIOSApplicationStateListener, but would
then get a callback on the main queue later on where we would reference
the now invalid 'this' variable.
By moving the dispatch_async call to QIOSApplicationStateListener and
using 'self' we ensure that the listener is retained for as long as the
block is valid. This opens us up for receiving application state callbacks
after QCoreApplication has been deleted, so we need to guard against
that.
Change-Id: I2ac14d28d72fd79764e12b6657234b54d846cb79
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
You are not supposed to call NSOpenGLContext -setView: for a view that
has not yet called drawRect. Doing this would result in a invalid
drawable error.
Similar to 4.8 commit cd2a51a66f
Change-Id: Ibb2300a8c6fe52f786f813987e93d4a3dc145366
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Change-Id: Ic78160d27b2c768054feefe6ec3fd4aaf8280c83
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
On wince with freetype engine can't render text for non-latin text.
because wince doesn't have truetype font FONTSIGNATURE api. so match
function in qwindowsfontdatabase_ft.cpp was always failed and render
incorrect text.
this patch has 3 changes.
1. extract font unicode signature using GetFontData function
2. append font fallback data from registry
(see. http://msdn.microsoft.com/en-us/library/ms901076.aspx)
3. wince's default font path is windows. correct fontdir
Task-number: QTBUG-31974
Change-Id: If969df353492141669eeab33119f3506602871b3
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Whilst having the objectName set for each engine is
somewhat handy when debugging, deriving from QObject just for that
is a wasting of memory in all other cases.
This also broke the font engine abstraction by allowing qobject_cast()
to access some private data; the only sane way to distinguish
engines is querying their Type value.
Change-Id: Ib1d195692859eb39089f6d8d9016cb8f9dcc0400
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
The dialog thread can outlive the platform dialog helper if
the helper is destroyed. In that case, IFileDialog::Show()
returns since the parent window is destroyed and then tried
to emit signals on the destroyed helper class instance.
Pass a shared pointer to the native dialog instead of a pointer
to the helper class to the dialog.
Task-number: QTBUG-32494
Task-number: QTBUG-30513
Change-Id: I7c2e769460270a26d886fdefee93ea59c2a17196
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Remove global variable and duplicated lookup in the old code
(map.contains() followed by map.value()).
Change-Id: Id68c34bf38c6706db69dcb8422c3b1ea718aa064
Y# issue or contains a behavior change that is relevant to others,
Reviewed-by: David Faure <david.faure@kdab.com>
Documentation for SHFILEINFO says iIcon is an index, so lets
initialize defaultFolderIIcon with -1 so it doesn't clash
with some icon that might exist at index 0.
Change-Id: Ic16538ee62e5433f3cdcceee19eb5d8d18d55c1e
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Added code using image lists to windows theme.
Initial-patch-by: Max Desyatov <max.desyatov@gmail.com>
Task-number: QTBUG-4970
Change-Id: I6e82f4edec60e456d0b1759bb1771955edb94491
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
The excessive padding has serious performance implications throughout
Qt and need to go. The original reason for having them was to fix
issues with GL text rendering and issues with unspecified OpenType
fonts. There have been several other changes to the glyph cache
since then and neither of the issues can be reproduced today.
Change-Id: I6de67598a079c296daf48be07cc5c2d67768c384
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
We're currently adding a lot of transparent pixels to the cache,
wasting both memory and cpu cycles while drawing. AlphaMapBoundingBox
was introduced to return the exact same bounds as the alphaMapForGlyph
function so we should only rely on this instead of adding arbitrary
padding and margins all over the place.
Windows still has an arbitrary +4 in the its drawGDIGlyph() which
means batching will not work on windows, but at least now
other platforms do not need to suffer.
Change-Id: I714903fa195004400c09c3bf6570e46179775f09
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
There was a error: narrowing conversion of 'Myns::XEMBED_VERSION'
from 'unsigned int' to 'long int' inside { } [-Werror=narrowing]
Change-Id: I1e155870b0b632bfe5ffa9deac69c4f151f5aa23
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
KWin supports the XSync extension. Since the check got added the XSync
implementation in KWin got reworked (release KDE SC 4.8).
Change-Id: I15c37fe4dd5501ca107b90a07ae8d70295ffaeda
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Those should be in window coordinates (or rather, its content view)
not view coordinates.
Task-number: QTBUG-32826
Change-Id: I52dddeccf17b359163ad477ce4299b934633b4fa
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Otherwise the view will miss its own frame change notifications.
And we must unregister from all the notifications during dealloc.
This would also reveal a bug where we would expose an NSView before
its super view is visible, leading to those "invalid drawable" warnings
when using QQuickViews. Therefore, we add this extra check in
QCocoaWindow::exposeWindow().
Task-number: QTBUG-32826
Change-Id: I69018cb6f199b242768d114b2aa34c7f2d243196
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Native widgets, which have a NSView but not a
NSWindow, must be created in the hidden state to
prevent Cocoa from selecting them for event delivery.
Change-Id: I741e52729047ad4e03959f2244abe5b14b5df46b
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
In the case of a window that is not yet shown, or destroyed, there
will not be a native handle, and this would crash.
Change-Id: Ib8c1a4490d82654d04073dac240eaa1275cbc1b0
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
If you click outside a popup window, it only closes the popup; any
other widget that was under the cursor at that time should not get
the event. The bug was about the popup list on a combobox, but this
patch assumes that this rule is universal; can't think of any
exceptions at this time. (E.g. a tooltip is not a popup) Clicking
on the application menubar still does not close the popup though.
Task-number: QTBUG-33241
Change-Id: I2444b7cfd40cf75ff7b70e3fecfeceb2fd4623bf
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
GCC complains about this class having virtuals and no virtual destructor.
C:/Qt/qt5/qtbase/src/plugins/platforms/windows/qwindowsdialoghelpers.cpp:750:20: warning: deleting object of polymorphic class type 'QWindowsNativeFileDialogEventHandler' which has non-virtual destructor might cause undefined behaviour [-Wdelete-non-virtual-dtor]
Change-Id: I7494fce4bb71ed65954226259fa65fdc507ca403
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Kai Koehne <kai.koehne@digia.com>
Fix compilation with e.g. mingw-builds 4.8.1-rev4 package. The Mingw-W64
headers define SHSTOCKICONINFO only for vista and newer.
Task-number: QTBUG-33225
Change-Id: I30a62c642ae017c7eafb99b1efb06578fd61a12e
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
"Make QGuiApplication::exec() run within NSApplicationMain()"
"Make Qt process native and timer events on Cocoa applications"
"Cocoa: Fix QFontDialog, QColorDialog auto-tests"
This reverts commits
1e14762b8de4b2a0b4badf7944e7d7
Change-Id: I80b65b5ee0297b090f807bd420664233dfc44f7b
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
The new Cocoa event dispatcher made apparent some deficiencies in the
way the dialog helpers were being hidden. In particular, we would not stop
a dialog helper's modal loop when closing the dialog, resulting in the
auto-tests hanging. Also, since the QApplication event loop is runnig with
[NSApp run] in the stack, the previous workarounds are no longer needed.
Task-number: QTBUG-24321
Change-Id: Ifba713c286638d78a699c319a15683d09714f06f
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Since we don't spin any QEventLoop, we need to make sure events posted
by the window system interface are processed.
Change-Id: I0e7b541d86a7a14f2ecd3702e7cbdbdee9920859
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
We follow the same pattern as for iOS and Windows ports, making
sure the user's main() runs in a platform friendly environment. In
this particular case, it means calling the user's main() during the
call of NSApplicationMain(), and calling the user's main() function
(renamed to qMain() as in Windows) after receiving
NSApplicationDidFinishLaunchingNotification. In practice, this means
that NSApp is running when qMain() is called, and therefore when
QCoreApplication::exec() is called.
For those command-line utilities running on QGuiApplication, or any
deriving class, and that do not provide a bundle, we override the main
bundle's dictionary and get NSApplicationMain() to run as usual.
Added cocoa/cocoamain "subdir" to build libqtcocoamain.a (together with
cocoa/cocoaplugin -- plugins/platforms/cocoa is made a subdirs project).
This library is linked against all GUI Qt apps and provides the actual
main() function. It also catches the launching NSApplication notifications,
and calls the user's qMain() function. Note that this will happen in the
same cases when the user's application will run with the Cocoa QPA plugin.
Launch related code in QCocoaApplicationDelegate is moved to libqtcocoamain,
QNSApplication is removed (but sendEvent: redirection still there), and
code in QCocoaEventDispatcher dealing with calling [NSApp run] and related
has been removed since it's become unreachable.
ChangeLog: [Qt for Mac] Make QGuiApplication::exec() run within NSApplicationMain()
Change-Id: I790e5138c29aac2e0215a9147d0148fece40ca22
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Maliit upstream has its own IC which is more advanced than this one, as well as
being supported.
Change-Id: I4b820fc7f39468e53c113699c4c43c75d490f65b
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Jan Arne Petersen <jan.petersen@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Reviewed-by: Pekka Vuorela <pvuorela@iki.fi>
The application state is tied to the QtActivity lifecycle.
Mapping of states between Android and Qt is as follows:
onResume --> ApplicationActive
onPause --> ApplicationInactive
onStop --> ApplicationSuspended
Change-Id: Iefef08d6c7a7fde28fba1f4886882458cda6a0c0
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Starting with Windows Vista, we can control the compositing window
manager using the DWM APIs. This allows us to make truly transparent
OpenGL windows.
We also have to detect and listen if compositing is enabled, otherwise
enabling transparency will cause glitches all over the place.
This (partially) fixes
Task-number: QTBUG-28214
Change-Id: I0fe1ec7adec8181b788c32de03c59142731d9e7f
Done-with: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
If m_window is invalid, devicePixelRatio returns 0, which leads to
adverse effects in other parts of the code. For example,
qquickshadereffectsource.cpp will get stuck in an infinite loop
trying to multiply 0 by 2.
Task-number: QTBUG-32975
Change-Id: Ie3db86f1f459df018ebce67bcb4226f6cffe854e
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Prevent windows from being recreated due to screen changes which
can cause GL widgets to stop updating among other things.
EnumMonitors returns only one temporary monitor named "WinDisc"
when locked. Do not remove monitors when that happens.
Task-number: QTBUG-33062
Change-Id: Ia2247bb04b3e10f99f594245f84238b5f9044f70
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Now that it lives in QPlatformSupport, will be fleshed out more, and
might be used on OSX at some point in time. Still iOS specific, as
none of the iOS API usages have been ifdef'ed.
Change-Id: Ib7fde6403ef2dfef175a6f306a85d58027569a30
Reviewed-by: Ian Dean <ian@mediator-software.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
We resize the render-buffer based on the CALayer of the UIView that's
backing the QPlatformWindow, so the logic in defaultFramebufferObject()
to determine if a resize is needed should be based on the relationship
between the render buffer-and the CALayer, not the render-buffer and
the QPlatformWindow.
There is still an issue of the QPlatformWindow and its UIView/CALayer
not being in sync, but that's a separate issue.
Change-Id: I84f617d07ec64fea0d027473e9720523eeae0c7a
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
The XCB backend runs a thread which gets events out of the XCB event
queue and feeds it to the main thread via another queue. This queue is
protected by a mutex.
However, when the event thread exits, it cleans up after itself and
frees all remaining entries in the queue. This code messed with the
event queue without acquiring the needed mutex and left behind a list
full of stale pointers.
Fix this and protect the freeing with the correct mutex and clear the event
queue afterwards.
Change-Id: Ie49cf6241b76be86d8cebbc931f7226a3f6a14e5
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Since QPlatformTheme covers all dialogs whereas QPlatformDialogHelper is
really only for the native dialogs then the SnapToDefaultButton hint is
moved as it has relevance for all dialogs
Task-number: QTBUG-32631
Change-Id: I1dce0bb4abcd4cfd39c4a199a33fc7078176ab4b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
On BB10, NAVIGATOR_WINDOW_INACTIVE is called before
NAVIGATOR_WINDOW_INVISIBLE, already triggering handleWindowGroupDeactivated()
Change-Id: I7d82c0220fe8dc8e87bfa2b31af6085c7d1d6cee
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Bernd Weimer <bweimer@blackberry.com>
Reviewed-by: Nicolas Arnaud-Cormos <nicolas@kdab.com>
On Playbook, rendering should happen when the application is thumbnailed,
therefore we need to send a window activated event to resume rendering once
the thumbnail is restored from the minimized state.
Change-Id: I0fa5da483dc101e25f718e52859a66edfe5c66c7
Reviewed-by: Bernd Weimer <bweimer@blackberry.com>
Reviewed-by: Fabian Bumberger <fbumberger@rim.com>
Reviewed-by: Nicolas Arnaud-Cormos <nicolas@kdab.com>
We retain each modal NSWindow as long as its modal session is running,
and we should release it every time that modal session ends.
Task-number: QTBUG-32728
Change-Id: Ia30c9c2d15be1350e7150a0d3c2f530a2fe4f38b
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
This enables both modes for TalkBack,
explore-by-touch and the normal swiping mode.
It is partially inspired by the BarGraphView example
of the Google/Android Eyes-Free project.
Note that for any accessibility to work you'll need
a device with api level 16 at least.
Using reflection we should be able to dynamically pick up
the classes if we have the high enough api level.
Change-Id: I11b93bead451483782a1711434d45c8f9a35996f
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
The API is broken and available in individual functions.
Don't make it more complicated than necessary to implement
the interfaces for new widgets.
Change-Id: Ie408c369ef05b2b8e7ac666b25153d090fcf3aae
Reviewed-by: Jan Arve Sæther <jan-arve.saether@digia.com>
Since we await giving focus to a focus object until a press
release, it also makes sense to await activating a window until
a press release, since they both have to do with focus. By doing
so, the input panel now stays open if the user selects a line edit
in one window when a line edit in another window still has focus. We
also avoid activating a window in case of a touch cancel (e.g as
a result of the user flicking or triggering a gesture).
Change-Id: Ic00c4be69c257fceb10ce2d5a81cb490ea93710f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Layered (translucent/non-opaque) windows do not receive WM_PAINT,
expose events need to be generated. Improve
6800728d09 to handle transient
children as well.
Task-number: QTBUG-17548
Change-Id: Id113604512692dfbea1f2b10d0db3068213cf599
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
When suspending the app we would destroy the surface twice,
causing libEGL to output a warning on the second attempt.
We would also destroy the surface before all references were
released which is not 100% nice. We don't need to call the
super class implementation at all, since we are managing
the EGL surface ourselves.
Change-Id: Ie1ab2ea8561d0018b5f16ac8cdf3296313a0a92c
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
It is not guaranteed that buffers are cleared when they
are allocated. So don't use a buffer for the root window,
and clear buffers for the actual content.
Change-Id: I42939baec03fa05968c83bbf2739ab8d1d70c8be
Reviewed-by: Matt Hoosier <matt.hoosier@garmin.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
In Android 4.3, we will get a new native window pointer for the
same surface (the old surface has not been destroyed), whereas
before we would get the same pointer, thus hitting the
"sameNativeWindow" branch. This revealed some bugs in the
surfaceChanged code path when the old surface had not been
destroyed. This path is now taken both when there's no old surface
(the app has been suspended in the mean time) and when the
orientation changes.
To handle the second case, we need to make sure:
1. We update the static pointer
2. We update the pointers in the platform windows
3. We don't add a second reference to the static data for
windows
4. We schedule an update of the window size
Task-number: QTBUG-32878
Change-Id: I47257615f9ba820315fc98d7a804e52223f430bf
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
XI2 sends events for individual touch points, but QTouchEvent
sends all of them with a stationary state if they didn't change.
If a touch pressed event is received, and the next XI2 event
is about a different touch point, we wouldn't update the state
of the previously pressed touch point.
Change-Id: I1ebcbea1cea54872064ef7710e2aac7b0b41cd70
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
After change 4dbf574b7a touch worked
only if XI2_DEBUG is turned on. That is because it creates the
QTouchDevice and calls QWindowSystemInterface::registerTouchDevice,
which must be done at startup so we can receive events.
Change-Id: I9446d72bc702fbd819bf26bcdc2a3d657180f642
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
With a _NET_WM_PING message, a window manager can check if an
application still works or if it froze. The WM sends such a message to
the application window and the application sends the message back to the
root window.
Since change Ia04268b0 / commit d8090022f "Move the X11 system tray code
from widgets into XCB-plugin.", Qt selects StructureNotifyMask on the
root window. Due to this, we now also receive replies to _NET_WM_PING
and treat them like a _NET_WM_PING request.
This caused an endless loop as soon as any _NET_WM_PING was seen where
Qt would send a _NET_WM_PING to itself and handle it again and again.
Fix this by ignoring _NET_WM_PING messages that are sent to the root
window. According to EWMH, such messages can only be replies to
_NET_WM_PING and GTK does this, too.
Task-number: QTBUG-32957
Change-Id: I1b0aa682f99b17d633baacc14b18b3adca7a1aba
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Optimize the parser of the X11 compose tables. Parsing these was
responsible for over 90% of the startup time in Qt 5.1.
Change-Id: Ifddc3f30828791e51a755f92791c26ffe43a9cd3
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
To avoid using modifier keys when releasing a two-finger
mouse wheel flick on a trackpad, we listen to event
phases. But for a normal mouse, a separate phase
NSEventPhaseNone is given. Ensure that we still send
modifier keys when that state is reported.
Task-number: QTBUG-32098
Change-Id: Ib840dd661b7842ae49127e5a8d42e3666ae2da4e
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Change QNSView convertFromEvent: toWindowPoint: andScreenPoint:
arguments from QPoint to QPointF.
QWindowSystemInterface event handlers already use QPointF and
QPointF is needed for tablet support
(https://codereview.qt-project.org/#change,62740).
Change-Id: I6e5b2f19777fe24ae34e9e85028772e4be0a94f6
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
A QWindow should keep its title after removing and restoring the
frame, so set it if we are setting window flags that include a frame.
Task-number: QTBUG-32978
Change-Id: I0fe1b651eac05a210b06ec4f7f6fb78f2536834d
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
This patch fixes the issue when the escape '\' character was
interpreted as a compose value.
On the whole en_US.UTF-8 Compose file it only affects the following
sequences:
<dead_diaeresis> <space> : "\"" quotedbl # REVERSE SOLIDUS
<Multi_key> <slash> <slash> : "\\" backslash # REVERSE SOLIDUS
<Multi_key> <slash> <less> : "\\" backslash # REVERSE SOLIDUS
<Multi_key> <less> <slash> : "\\" backslash # REVERSE SOLIDUS
By coincidence the last 3 sequences were working even before this fix.
Task-number: QTBUG-32972
Change-Id: Ia4fc2156d982cf2918d0d1be6ee07706cfbfd091
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
On Linux (XCB), resizing a dialog shifts its position. The fix corrigates
the geometry of the dialog to the right values.
Task-number: QTBUG-32473
Change-Id: I6d38539a3ebc3b95eacc7f13a76f83fc9e4d821c
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
We take the path of least resistence and keep the old API code for 10.6 while
updating the code for 10.7 and newer. This means we have some code duplication.
It also means that we only compile the 10.6 code for QCoreWlanEngine when the
deploymen target is 10.6.
The 10.6 version file should be removed once we drop support for Snow Leopard.
Change-Id: If4702b155bcdb7522800bf99a4dd37d4efed803a
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Reviewed-by: Peter Hartmann <phartmann@blackberry.com>
Reviewed-by: Lorn Potter <lorn.potter@jollamobile.com>
Windows Server 2003 is based on Windows XP and should use
QWindowsXpFileDialogHelper as it does not support the CLSID-based
IFileDialog interfaces that are available from Windows Vista onwards.
Change-Id: Idd973f9ec4c98d1f2fb7e835de64532edeccfc72
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
This doesn't actually fix the source of the problem, but its harmless.
This was reproduced with tst_qcolumnview and voiceover enabled
Task-number: QTBUG-32440
Change-Id: Iad27884e1ca9194f911271c16908ef358e4b1875
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
The started & ended phases are required for implementing correctly
behaving transient scrollbars (ie. they become and stay visible when
touching the pad with two fingers).
Change-Id: I718d991ba6fd7e949cf9790f3bae285000fce576
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
QEglFSWindow::winId() method was changed to return EGLNativeWindowType
EGL window handle instead of static window number as it recommends in
documentation.
QPlatformWindow documentation reads: "The platform
specific window handle can be retrieved by the winId function." and
also for winId() method itself: "Reimplement in subclasses to return
a handle to the native window".
Task-number: QTBUG-32564
Change-Id: I634c5b4d966b6aebde72518a2c39717d1b39af08
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
So far only the xcb plugin is able to handle xcb errors and there is no
way for an application to get access to the errors even if the app is
using xcb directly.
This means Qt is filtering out all xcb errors which are relevant to the
app and at the same time the xcb plugin is getting spammed with errors
it did not cause and which are logged with a qWarning.
By passing the error event through the native event filter an app can
filter out all errors it knows to have caused.
Change-Id: I158deee2e1c71630f2b1d77174f1091532851b3d
Reviewed-by: David Faure (KDE) <faure@kde.org>
The printer should not be released after changing a property on it as it
is still needed by QPrinter elsewhere. It is released as appropriate
automatically already.
Task-number: QTBUG-32831
Change-Id: Idb2d98b25b62f343015a0a0fb3c9a0d506546132
Reviewed-by: Liang Qi <liang.qi@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
This issue was introduced when porting key handling code from Qt4 to Qt5.
To conform to the implementation of QKeyEvent::modifiers() we have to invert
modifier state logic before sending them as QKeyEvents.
Task-number: QTBUG-31332
Change-Id: I3bb41169f8ab2a4b0a13a224bb461d2792d3a65f
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
The functionality to get the current x11 timestamp is also needed by
applications which need to interact with the X11 directly. With XCB
it is not possible to inspect the event queue and by that an
application is not able to retrieve the current timestamp using the
property change pattern and waiting for the matching event.
Change-Id: Ie7ba78ecbe509ed3a902c702266917f65bf5ad07
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
Reviewed-by: David Faure (KDE) <faure@kde.org>
Add support to request EglDisplay resource for a window. This is
supported in other platform plugins, like kms and xcb and is required by
wayland-egl based QtWayland compositors.
Change-Id: If3f8c625682915c9671e820974e810763c3c8b66
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
Add support for a get_egl_context NativeResourceForContextFunction to
the eglfs platform plugin.
Change-Id: I155952797f340dd00ab9864852add4b44d841042
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
Change-Id: I27cbcd8c59bdc34493931f341341cc25b4aba9e7
Reviewed-by: Kurt Pattyn <pattyn.kurt@gmail.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Similar to what the EGLFS plugin does, to avoid having to explicitly
specify which device to use for keyboard and mouse.
Change-Id: I00502be8c767487490f3fd36fb06d3eb56548ada
Reviewed-by: Pier Luigi Fiorini <pierluigi.fiorini@gmail.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
Prevents crashing when some menu or similar is shown, although the
visual result might not be ideal.
Task-number: QTBUG-29729
Change-Id: Ia840b3ec17f5ef30ee58150bd2f807ca5e72cc12
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
Some systems may have multiple DRM devices attached to a single seat and we are
not guaranteed to pick up the primary GPU. With this flag we can control how
DRM devices are detected and consider only the primary GPU.
This is very useful for the kms plugin or QtCompositors running on kms in order
to use the right DRM device.
Change-Id: I8b91e78f148b25aaa4e40724e39e0ed0918ca100
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
Change-Id: Icd7a192701958673fe216f40ddab710f5f63a8b8
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Reviewed-by: Andy Nichols <andy.nichols@digia.com>
2af0a778f464980594c36098e4a8ba4448edfd29 the fix for QTBUG-27410
caused this Designer bug. Doing the automatic level escalation only
for "special" windows and avoiding it for plain Qt::Window type
windows is one way of fixing the Designer problem.
Task-number: QTBUG-31779
Change-Id: I1da5454f31111f36480fac3b53be6d5f0ce40047
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Since updating the xkb_state throght core events is more
tricky (opposed to the xcb-xkb code path) where we have
to use xkb_state_update_key, we need to make sure that our
local state is in sync with the X server's state.
The local state was getting out of sync if key was pressed
down in a Qt application and released in other X client (by
changing the focus to another window with a mouse, for
example).
Task-number: QTBUG-32660
Change-Id: I662bf5aad3ab0e8591109994e746d85ff61ad6ef
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
* changeFont is a delegate method of NSFontManager.
* exec() will not be called in Qt Quick.
Task-number: QTBUG-32450
Change-Id: I05e43ce84d28ff32c88d75bd6ee60653de6ca4f0
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Just moved applyCursor() and defaultCursor() to a #ifndef block.
Change-Id: I14c21aa509395fb1bd72d389cfc46f0f34ab7649
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Change the checking for created EGLNativeWindowType so that zero
is a valid value. This is the case e.g, with BeagleBoard, where
widget application cannot be run without this change.
Change-Id: I36c30091e1a5a0598ae3822d0be8dc4362779c0b
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
We don't support native widgets in Android, so we can get into a mess
when people set this widget attribute, since the FB compositor
will assume that all widgets have their own backing store. This adds
a capability flag to the QPlatformIntegration which allows the plugin
to disable the WA_NativeWindow feature.
Task-number: QTBUG-32685
Change-Id: Ic200487da4a297f71ab594cf7c90d1e1d53bacd3
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
We would resize the backing store without resizing the viewport,
which would cause all subsequent blits of the backing store to the
screen to look broken.
Task-number: QTBUG-32146
Change-Id: I65bae051b7cfbbc61fc285e4baa74685d5639569
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
The reference color dialog helper change is
eeffcfbd89.
Change-Id: I35b505cce133adeed91df7c95b21f7c74c55cb53
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
The window should react to the wheel event (e.g. scroll content)
but without becoming focused; this is the X11 convention.
Task-number: QTBUG-32517
Change-Id: I7e12425e5a6e1549b7f23dc318612a436c24d14b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
There is no signal for it in gtk font selection. We should emit this
signal just before accept() was emitted.
Change-Id: Ie31d96b789436607b134c84dd77a4b9be9e9a550
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
The QWindow::resizeEvent documentation states that resizeEvent
is invoked after the windowing system has acknowledged a
setGeometry() or resize() request.
The Cocoa plugin however did set the platform window geometry
immediately so that the qnsview's updateGeometry returned too
early.
Task-number: QTBUG-32706
Change-Id: I1f359ab368833d174ab6740f4467b0848c290f13
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
- Add system tray tracker class to XCB plugin and
provide functionality via invokable slots
of the native interface.
- Remove XLib-dependency of widgets/utils.
- Reintroduce tracking of tray window destruction
and recreation, which was removed in the XLib-code
when porting it from Qt 4 to Qt 5.
This paves the way for implementing the tray icon
completely in terms of QPlatformSystemTrayIcon at some point
later.
Change-Id: Ia04268b0e2919c05874a3e9548930535332897c7
Reviewed-by: Alberto Mardegan <mardy@users.sourceforge.net>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
On the Android Emulator, the shaders will be compiled by a desktop
GL driver, since the GL driver in the emulator is just a thin
wrapper. The GL driver does not necessarily support the precision
qualifiers, which can cause applications to break. We detect this
at runtime in the platform plugin and set a workaround flag to
Task-number: QTBUG-32557
Change-Id: Ied00cfe8e804d1f7862697dd379a14f3bed3d980
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Several QOpenGLContext methods fails incorrectly on QWindows from EGL or
MinimalEGL. This is happens because they are incorrectly marked as raster
surfaces instead of OpenGL surfaces.
Change-Id: Ic9b3859915a9049fce442216b01dce89521fa5ee
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
If some of the X11 requests fail, QXcbXSettings::QXcbXSettings() prints a
warning and returns. These error paths all caused memory leaks.
Change-Id: Idfecf03dd412c35552c3bbbebdda9c039aeadc13
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Laszlo Papp <lpapp@kde.org>
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Tor Arne Vestbø
Samuel Gaist
Friedemann Kleint
Jan Arve Saether
Laszlo Agocs
Morten Johan Sørvig
Gabriel de Dietrich
Shawn Rutledge
Liang Qi
Vitalii Shastun
Yoann Lopes
Richard Moe Gustavsen
Giuseppe D'Angelo
Sergio Ahumada
Bernd Weimer
Teo Mrnjavac
Christian Strømme
Andrew Knight
Gunnar Sletta
Paul Olav Tvete
Frederik Gladhorn
Gatis Paeglis
Fabian Bumberger
Martin Gräßlin
Kai Koehne
Thiago Macieira
Marc Mutz
Konstantin Ritt
Chris Meyer
MyoungSeok Song
darius
BogDan Vatra
Eskil Abrahamsen Blomfeldt
Sérgio Martins
Oswald Buddenhagen
hjk
Robin Burchell
Robin Burchell
Daiwei Li
Uli Schlachter
Andy Shaw
Rafael Roquetto
Andreas Holzammer
Stephen Kelly
Jocelyn Turcotte
Ryohei Ikegami
Lars Knoll
J-P Nurmi
Balazs Domjan
Simon Hausmann
Jonathan Liu
Alexey Chernov
Jan Arne Petersen
Samuel Rødal
Pier Luigi Fiorini
Jorgen Lind
Samuli Piippo
Nils Jeisecke
Allan Sandfeld Jensen