| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The code assumed that CGImageForProposedRect would return a sensible
sized image, but that can return basically any size (depending on the
NSImage's image representations). In the case of the bug report this is
a 1024x1024 pixmap when requesting a pixmap of size 64x64 and larger.
Make sure that we return a pixmap of the exact requested size.
For this, themeHint must also return sizes in device coordinates.
Task-number: QTBUG-35009
Change-Id: Iaae11023bab6d4122815ca4010aab6967dfb18a0
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Right now /dev/fb0 is hardcoded. This is not ideal. Therefore
QT_QPA_EGLFS_FB is introduced. This environment variable can be set to
a different framebuffer device. Once it is set, eglfs will use the
specific device. This is similar to linuxfb's fb=... plugin parameter.
The actual behavior depends on the board-specific implementations.
For now only iMX6 has real support. It extracts the index from the
device name as bind the EGL display to the corresponding framebuffer
using the vendor-specific fbGetDisplayByIndex(). Other hooks can
follow suit later on.
With this patch eglfs is at least on par with linuxfb, meaning that,
if the board supports it, different apps can run on different screens.
Task-number: QTBUG-36113
Change-Id: Ia3c88bd06e108bc668433e3c5c3fce34a5a0e73d
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
|
|
|
|
|
|
|
| |
On windows (tested on win7) physicalSizeMM is changed after changing
screen resolution. We need to save new size.
Change-Id: I1b38d9ba2af9679812a973737a724c9a9c5395da
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
A parent is required, otherwise the application gets deactivated when
the dialog closes. The same is done in Qt 4.
Task-number: QTBUG-36039
Change-Id: I56688b3259f1d207edd1bec499afd1365ae511dc
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add setContentBorderThickness() to the Cocoa platform
plugin. This functions requests that the platform
plugin draws a gradient in the unified title and toolbar
area and/or the status bar area.
The background gradient is drawn before and under
the Qt backing store content. It is expected that
parts of the backing store will be filled with transparent
pixels to allow the gradient to be visible. To facilitate
this the backing store image is created with an alpha
channel.
Task-number: QTBUG-34411
Change-Id: Iadc5e64ee9b9b42e92fb84a615817fdffd7a8802
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
|
|
|
|
|
|
| |
Change-Id: I6aa3f3bc3c098b96ec0eec369b85a938c552f364
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
This causes all sorts of unintended side effects and
is generally a bad idea.
Task-number: QTBUG-32988
Change-Id: Iebdae34764be4cfd9ced47aa93789871e0455ab3
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The native child windows need to inherit the parent's visual in order
to have a translucent background as well.
Surface with type QSurface::OpenGLSurface should not be forced to use
the parent window's visual - the parent visual for instance, might not
even be GL capable.
Changing WA_TranslucentBackground during runtime is not supported, for
two reasons:
1) Other platform plugins seem not to support it
2) It would require recreating X windows.
Task-number: QTBUG-29625
Change-Id: Ic1474dd2de99069027481c7db6bf865f9b8d616d
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
The function is invoked from the QFileDialog constructor in
save file mode when passing a non-existing file.
Task-number: QTBUG-35703
Change-Id: I5c13efb9b6789404b28739f6ce88ed87b594b6e4
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Processing the object file with ld strips away debug information for
the main() function, resulting in the debugger not being able to
break on specific lines of the function.
It also causes issues when externing sybols in main's object file.
We revert back to the approach of using the strings in-line in the
object file (which is why we keep the name the same length, 'qtmn').
Task-number: QTBUG-35553
Change-Id: I8b0acee36f48ecfefa2e4fd008a842365713d985
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
In qt_xcb_XPixmapFromBitmap(), a new pixmap is generated for the caller.
This pixmap has to be freed after use. However,
createNonStandardCursor() didn't do this and instead leaked the pixmap.
Change-Id: I6fee180ec6508db9e82a5bb028957e7d9f7a4632
Signed-off-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When using a touch screen on a Linux machine, we receive both touch-events
and emulated mouse events from XInput, on top of that we synthesize mouse-
events ourselves for the touch events.
This patch grabs the touch device for touch events whenever it processes
a touch-begin thereby avoiding XInput from synthesizing mouse events.
Task-number: QTBUG-35157
Change-Id: I5849d5841be236d6719cd080af2e9e39eb9cdd84
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
2d05d3bd2815c220474b3c07bf3f2ef7417d3070 was not correct.
On OS X, when user uses CJK input method, only types single punctuation,
it was converted to CJK ones, and not showed in composing text.
Task-number: QTBUG-35700
Change-Id: I919edb3f5165bf943c0d90d06a788a2f335bb1ba
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
|
|
|
|
|
|
|
| |
Task-number: QTBUG-35151
Change-Id: Ie62e50032aaa647a86c4f03b1a3363e5ef6a1bbb
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
|
|
|
|
|
|
|
|
| |
Fix a multiple define when Qt is configured to use both EGL and xcb-Xlib.
Change-Id: I6fdb282f575842711b3b5d377bbdf3bc9909bf0c
Reviewed-by: Christoph Cullmann <cullmann@kde.org>
Reviewed-by: David Faure <david.faure@kdab.com>
|
|
|
|
|
|
| |
Change-Id: I9b9fc6cf000b262277711374e0a2fe119328849e
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On non-Android systems hideCursor should default to false, just
like it is done in eglfs. This is especially useful on udev-less
systems where currently one has to resort to setting
QT_QPA_FB_HIDECURSOR to "0" to enable the mouse cursor. This is
not ideal. Defaulting to showing the cursor unless disabled by
the environment variable or, in absence of that, the lack of a
mouse reported by udev is a better choice.
Change-Id: I7362ac47046179d5eb8ed8b44cf2c36c0fc23432
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
|
|
|
|
|
|
|
| |
move the dpy to the place where it is used inside the egl ifdef guard
fixes compilation on old distros not having egl
Change-Id: I7eebe5305f3a584c0c5da2ea7b9099fdd994249d
Reviewed-by: David Faure <david.faure@kdab.com>
|
|
|
|
|
|
|
|
|
|
| |
export QT_XCB_DEBUG_XINPUT_DEVICES=anything to show detected
input devices at startup
export QT_XCB_DEBUG_XINPUT=anything to log mouse, touch and tablet events
Change-Id: Id14844b68ad376740f82a36aab2c59c84d2017ab
Task-number: QTBUG-35583
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
logicalDpi() already has the logic to check m_forcedDpi, so let's reuse
it. I hope it's not a problem that we send a signal that it changed when
nothing changed.
[ChangeLog][Platform Specific Changes][X11 / XCB]Fixed a bug that caused
Qt applications to think the screen DPI had changed when it had not,
after connecting or disconnecting monitors.
Task-number: QTBUG-32683
Change-Id: I45dd27de5109e65e7599915f11cfdb633a65a67c
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
|
|
|
|
|
|
|
|
|
| |
Then it's possible to embed native cocoa widgets without crashing.
Task-number: QTBUG-35629
Change-Id: Ic212c36178282b4d090c6ce0470012adc8e0c2bb
Reviewed-by: Morten Johan Sørvig <morten.sorvig@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
|
|
|
|
|
| |
Change-Id: Ib7dc8dcbeca7e85d97b8c7fb04d2cf42e5245298
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
|
|
|
|
|
| |
Change-Id: I058b76ff9d40b75eb51f9a6b1b89032ead629aed
Reviewed-by: Liang Qi <liang.qi@digia.com>
|
|
|
|
|
|
|
| |
These are structs in gbm.h.
Change-Id: I448ddc072c3aa924672f802dca42b2183470519e
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
When we get a WM_CHAR message we do not need to add anything to the key
map as these messages already contain the character code in question.
Task-number: QTBUG-35532
Change-Id: If73993cd873b7c616876b9ba52ceba6b27a6a7b1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
Do exactly what windows font database does to determine
writingSystems in case of non-truetype fonts.
Task-number: QTBUG-13585
Change-Id: I4a1a48d74c36e403c31f20847cf80295d89a34bc
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For the linuxfb platform the screen was registered
before its geometry was calculated resulting that
the qpa controller will deal with a screen instance
that has geometry of QRect(0,0,1,1) .
This masks pretty much all the painting and the input.
The fix is to call screen init before registering
the screen instance to the qpa controller.
Change-Id: If84daca6587dc368301b578cdbf7b2a65a56a5b5
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
QMacPasteboardMime is a public class in Qt 4. Qt 5 has
two copies, one public QtMacExtras and an internal
one in the cocoa plugin. This causes a symbol collision
when building statically.
Rename the internal copy to QMacInternalPasteboardMime.
Task-number: QTBUG-35310
Change-Id: I891787b451a0b342ed85aa7196e606bc11623e21
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Window buffers are currently always created with an alpha channel when
the underlying display supports them. This is the correct thing to to
in order to support translucency. However, the cost is that the
composition manager unconditionally performs blending even when the
use-case in the GUI doesn't make use of it. That is rather expensive,
particularly on hardware which doesn't have extremely efficient
dedicated 2D blitting hardware.
This patch adds inspection of the user's requested window format
and--if no blending is desired--sets an additional Screen window
property which permits Screen to skip blending on the window surface.
Change-Id: I83c100a351a00ef0e8bea1b1d1fec10b7218d1de
Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com>
Reviewed-by: Bernd Weimer <bweimer@blackberry.com>
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>
Reviewed-by: Frank Osterfeld <frank.osterfeld@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
|
|
|
|
|
|
|
|
| |
This will tell QtQuick.Dialogs not to use native windows for dialogs.
Change-Id: Ie2e5878b84a9597e1f730d2cb1ebe2f59be6bc75
Reviewed-by: Liang Qi <liang.qi@digia.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
|
|
|
|
|
|
|
|
|
| |
No need to tie the debug bit to OpenGL 3.0+. xcb is correct
in this respect, let's correct the windows plugin too.
Change-Id: I13ea48de067d3fb61575be8f71b97bb547d8eb02
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In layoutSubviews we take the root viewcontroller position into account
when determening the new QWindow geometry, but we were missing this logic
in displayLayer, and would assert if the in-call statusbar was visible.
Since we don't really need the position of the window in displayLayer,
we change the assert to only check the size of the exposed area, which
is independent of the position of the root viewcontroller.
Change-Id: I774b8d9b075518e729f488a789b3a9e584c3f4d3
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
Introduce a hint to QPlatformTheme to control the behavior.
Task-number: QTBUG-35231
Change-Id: Ia28e153a8dd3f1931321a222d8906ca87166ed62
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Added a Enironment variable to make it possible to disable the
automatic installation of EVDEV input handlers.
This is needed if you want to use your own generic plugin instead,
which also uses evdev
Change-Id: I17d47008c10999bf918db62a22a3b6a38d7abb80
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In commit 0c936ca83c40b73f30195a619900f2be98b9e968 we fixed the case
where the window was maximized programmatically, but if the window gets
restored by clicking on the taskbar we don't get to set WithinMaximize
flag, since we immediately get the WM_GETMINMAXINFO request from Windows.
To reproduce the problem, run tests/manual/windowflags, remove frames,
maximize, minimize, and restore by clicking on the taskbar.
Task-number: QTBUG-8361
Change-Id: I8a7cf4fccbb2c3dac5f570848501ac1e8d2c2307
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit a2d3b7c99165cb5c4be6f0dd83a967e1119cf732. That
commit broke -opengl es2 builds on Linux desktops (when GLX is
available).
/usr/include/GL/gl.h:162:17: error: conflicting declaration ‘typedef double GLdouble’
qopengl.h:97:17: error: ‘GLdouble’ has a previous declaration as ‘typedef GLfloat GLdouble’
Change-Id: Id5f48c05803be1a8c03eeffca139b80e9a85a0eb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Brings Windows QPA on par with other platforms.
[ChangeLog][Windows] Don't cover the taskbar when maximizing
frameless windows.
Task-number: QTBUG-8361
Change-Id: Iba35132f697cb7379650a4c883b616c5c2023d4c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
|
|\
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
configure
mkspecs/macx-ios-clang/features/default_post.prf
tests/auto/widgets/widgets/qmenu/tst_qmenu.cpp
Change-Id: Iaba97eed2272bccf54289640b8197d40e22f7bf5
|
| |
| |
| |
| |
| | |
Change-Id: If237f08683290105413dc47923e23a496765bb22
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently there is no way to always hide the statusbar
on iOS other than showing a window in fullscreen. This
patch will check if the statusbar is (initially) hidden
from the Info.plist, and respect that in the application.
SubAttack is an example of an app that (because
of styling issues with MainWindow margins) manually
sets the geometry larger than fullsreen, and calls
showNormal(). In that case we still want the statusbar to be
hidden.
Change-Id: Ia365d14971978360d0b39621ff0f8f82f74b57e2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Going through the platform window failed when the parent wasn't created
yet. We can still get the window state of an yet-to-be-created top level
window.
Change-Id: Iaa61ddc50df037ac0bd2fd0884884c2bfce1dd9a
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
false"
This reverts commit 2eb1e28a90eaf82f4405dc65584021f16415014b,
which makes several of the OpenGL examples crash on startup.
(textures, hellogl_es2)
Change-Id: Id47d2df6f57b2396c333e478b3c062f4cb4becb5
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Qt defaults to creating a QWindow as offscreen surface
if QPlatformIntegration::createPlatformOffscreenSurface
returns 0. Using an offscreen surface is often done
in a background thread, which is problematic, since then
a QIOSWindow will also be created in a background thread.
According to Apple docs, working with a UIView from other
threads than the main thread should not be done. In our
case, we instead hit an assert in QApplication that
checks for the same.
As a quick fix for Qt 5.2, we remove the offending call that
causes the assert, since we anyway will call the same function
lazily when becoming first responder.
Task-number: QTBUG-35378
Change-Id: Id35462f99783a9748c688b163f6497de9bfff73e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The drag and drop event filters do not play nicely with touch events
or virtual keyboards.
Task-number: QTBUG-35348
Change-Id: Id4d079ae72882f48750d394f13e10700d60e4532
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Prevent the cocoa platform plugin from sending expose
events for windows that are off-screen or has invalid
geometry.
Differentiate between initial exposes and geometry
update exposes, in order that e.g. geometry updates
that happen before setVisible() don't trigger expose
events. After the initial expose geometry updates
to and from invalid geometry will trigger obscure
and expose events.
Task-number: QTBUG-35143
Task-number: QTBUG-35091
Change-Id: Ibbff20c69974f098ea7635e9ca38620597840a05
Reviewed-by: Gunnar Sletta <gunnar.sletta@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It turns out we cannot rely on QGuiApplication::focusWindow() to
be non-zero at all times (e.g when pop-ups are closing etc).
So instead use m_focusView.qwindow which points to a
valid QWindow. This window is the same as QGuiApplication::focusWindow
most of the time, except when a focus window closes.
For those cases we get a new call to scrollRootView immediately
after with m_focusView updated to reflect the new focusWindow.
Task-number: QTBUG-35339
Change-Id: Icb3a8d3140af1f1904495a9289c8c26ab79e70f6
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
An application will sometimes crash if the keyboard
is told to hide while the application is about to
quit. This patch will ensure that we set m_qioswindow
(and [UIView qwindow]) to 0 when the window is destroyed.
We also check this pointer before telling QUIView to
resign first responder when closing the keyboard. The
latter will fix the crash.
Task-number: QTBUG-35356
Change-Id: I934088beb7e877c5b33d96225cb215a8ffd4dbb2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
QIOSInputContext controls QUIViews first responder status
based on whether or not the keyboard should be open.
But since QGuiApplication updates focusObject before
focusWindow (when e.g a popup closes), we sometimes ended up
activating the old window upon a call to becomeFirstResponder.
This in turn led the application to hang because of
recursive dependencies in qioscontext when the focus window
changed.
So the solution for now is to avoid activating the window
when the view becomes first responder. This should be
fine since we now activate the window from
QIOSWindow::requestActivateWindow (ref: 6272a816d1)
Task-number: QTBUG-35340
Change-Id: I3068c14fec18d84d4b0b348a043c4c054e366c75
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This patch tells the window system that it should create a mouse enter
event if a window was activated when the mouse was inside. This wasn't
working and was a regression.
Task-number: QTBUG-35109
[ChangeLog][Cocoa] Fix enterEvent not being called on activate.
Change-Id: I4e4662b4a4c58dafa8d0a2c09458ab88f678d243
Reviewed-by: Liang Qi <liang.qi@digia.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If a platform window is created from a QWindow without setting a valid
size or position on the QWindow, the platform window is expected to
apply sane defaults. We use the baseclass initialGeometry() function
for this, similar to other platform plugins.
The default geometry unless otherwise set and/or calculated based on
size hints is that of the screen's available geometry.
An improvement to this is to detect whenever we apply the screen
geometry, and also apply the appropriate window state, but that
needs more testing.
Change-Id: I02b12064ce6d55c04fe0cc2cd1d2816ca1113f40
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
|