If we are logging to the console, we should _not_ log to slog2, like
the other log backends.
Change-Id: I6d16704c0f0923f54b3dc030d374af9805c2af06
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This fixes various problems that occur because the current egl context
assumes OpenGL ES 2.0 and does not support newer versions of ES.
Task-number: QTBUG-64306
Change-Id: I81466ba5cf028b47ca5a2ebcdc702167aff655a2
Reviewed-by: James McDonnell <jmcdonnell@blackberry.com>
Blending can fail if one of the input buffers is destroyed in the middle of the
streaming operation. In that case, turn streaming off so blending can be
attempted again after disabling the input.
Change-Id: Id5c82b22a2ed0858ab20902395830efa0e64177d
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Flushing the QPA event queue is problematic since we can then end up
delivering several events to Qt (or the app) at the same time on the
call stack. This again can easily leave objects in an inconsistent
state if they receive callbacks from subsequent events while being
occupied processing the first.
This is also what happens in the listed report. A QMenu shows
a sub menu when the mouse enters a menu item. The show leads to
QPA flushing events, which in some cases also includes flushing
another pending move event. The move event is delivered to the
same QMenu, which will clear a private variable (currentAction).
When the show returns, the state of QMenu has unexpectedly
changed, which causes a crash to happen since currentAction is
null.
This patch will fix the root cause of the problem by
stopping QCocoaWindow from flushing user input events when
the call location does a flush to deliver geometry events.
Task-number: QTBUG-66093
Change-Id: Id277550b0a080ad98c81e8c30dc7098dc73723d5
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
[ChangeLog][Third-Party Code] Sqlite was updated to version 3.22.0
Change-Id: I341c1cabfdd43fac45406c00c33c193ef4bcf402
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
On Android, we load the application library, and its dependencies (Qt),
on Android's main thread (thread 0), and then spin up a secondary
thread (thread 1), that we call main() on.
If any QObject is constructed during loading of the application library
or any of Qt's libraries, via static initializers or constructor
functions, we will set QCoreApplicationPrivate::theMainThread to
thread 0, which will confuse Qt later on when it's being run on
thread 1, and will result in a warning during QCoreApplication
construction:
QApplication was not created in the main() thread
This situation can easily lead to a crash as well.
Unfortunately logging via qDebug/qCDebug and friends will trigger
this too, as they internally use QObject.
Fixing the root cause of this is under investigation, but for now
we will partially revert fa2a653b3b for Android. The effect
is that any qCDebug with a "qt.*" category before qApp construction
will turn into a no-op, like it was before fa2a653b3b.
This patch does not cover the case of a regular qDebug, or a qCDebug
with a non-Qt category. Those will still produce the same symptom,
as before fa2a653b3b.
Task-number: QTBUG-65863
Change-Id: I95675731d233244530d0a2a1c82a9578d5599775
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 538b1b5076)
QFontEngine::cloneWithSize() is used by QRawFont internally when switching
a raw-font from one size to another using setPixelSize. For CoreText, we
use a subclass of QCoreTextFontEngine to keep track of the QByteArray data
of a raw-font, but failed to overload cloneWithSize, so we would lose the
data whenever setPixelSize was called, resulting in missing text rendering
in QtWebKit. We now retain the data as we should.
Task-number: QTBUG-65923
Change-Id: I7d4186a3c32a61d48d1e9388e43f2792e8e46081
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
The target position is passed in physical native pixels, so call
QPlatformWindow::mapToGlobal() instead of QWindow::mapToGlobal().
The latter operates on logical pixels.
Task-number: QTBUG-55251
Change-Id: I789128a0a345d4113fced82ed1b215fe14044634
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
With object_parallel_to_source the workaround of making copies of the
files and using these is no longer needed.
Debug2.h and .cpp were added to the repository by mistake and should not
have been there in the first place.
Task-number: QTBUG-66059
Change-Id: Ib9dbd15be1dee1cb5190762fe06bad56dd40dd47
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Some compilers will assume src and buffer are different and only
vectorize the unaliased case and take a slow path when they are equal.
In our case they are as often equal, so we need to manually unalias the
variables to make sure both cases are fully optimized.
Change-Id: I6ec86171dd179844facdf45376253c55980d9e36
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The target position is passed in physical native pixels, so call
QPlatformScreen::availableGeometry() and QPlatformWindow::mapToGlobal()
instead of QScreen::availableSize() and QWindow::mapToGlobal(). The
latter two operate on logical pixels.
Task-number: QTBUG-55251
Change-Id: I281f47baee727bc0f4738fd6d6cdf12c9f462b0f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
All of these had already no effect.
Cleaned up a bit of code around old Qt 4 #ifdefs and
documentation mentioning Carbon.
Change-Id: Id2858c56a5785f82c47e20e1e760743fb4d08189
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
It looked inverted for some reason. A single line looks
better than before.
Change-Id: Icb214b44ddcc6e9e57fe12e7b4c3f512a5f66452
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
While the QPA theme provides support for dock widgets
title font we never use it.
This is a very similar solution to what we do in
QMdiSubWindow, setting the platform theme font at
few specific points.
This patch also fixes the dock window title font on
macOS, as queried in QCoreTextFontDatabase.
Change-Id: Ie264e4e83e0d3d1e8f78bd378359f3063cc1d525
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Not only they should not appear (icons in menus is a special
thing on macOS), they don't match the look on the window frame
buttons. So, it's better to hide them completely.
Change-Id: I8814ed3e128480abff4f53fdd40a0bb13f1796ef
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
There are a couple issues. Because the actual class for
the buttons returned by +[NSWindow standardWindowButton:
forStyleMask:] is private, we can't fully configure them
as we could with HITheme. Therefore, we don't get the
mouse hovered state and the zoom button will show the
fullscreen arrow instead of the '+' icon.
Hopefully, QMdiArea will go the way of the dodo soon
enough. It is not a thing on macOS and it has been
replaced by tabbed windows UI on most desktop apps.
Change-Id: Ia581f72611ad0224f42657afbc4d9f94bf5e5a3a
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
This is currently only used in QMacStyle since,
on macOS, the spin-box buttons are always beside
the line-edit. Hence, there is no need to tamper
with the line-edit frame.
Change-Id: I8ccfbab547fccae89a8ddeaeb7005cb17d5951e6
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Bonus changes in this edition:
* The focus frame shows only around the spin-box's
line-edit, as it should. This requires getting
access to the spin-box's line-edit, which we do
by setting a property on the former. Notice that
the stepper doesn't get focus. On this point,
macOS is inconsistent (partly because there's no
such thing as an NSSpinBox, and partly because
spin-box buttons are a separate control), so we
copy the same focus behavior as NSDatePicker.
* We clean some QFocusFrame related dead code in
QMacStyle::event().
Change-Id: I204ca2093d9db343c78f0420811dda35c463bbcd
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Regression after 114f795221, which incorrectly assumed that checking
qt_logging_to_console() was a safe behavior change to include.
Unfortunately that broke debug output in Qt Creator, as Creator doesn't
seem to run applications with a controlling TTY unless it itself was
started with one, and doesn't set QT_LOGGING_TO_CONSOLE to force log
console output either.
Task-number: QTBUG-66153
Change-Id: I775cb06953122a3f20c16bfe21e19de2d1aa5b52
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This way, it's lossless.
This commit is a cherry-pick of ab1e507574,
which was reverted.
[ChangeLog][QtCore][QVariant] Conversions of QDateTime to strings now
contain the millisecond components.
Change-Id: I5e421e32396d44e4b39efffd150b744e40fff3a1
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
Call the base class implementations to avoid returning an unmapped
values for non-embedded windows.
Task-number: QTBUG-55251
Change-Id: Ib05fd530498dd4d72d3d4ef37caf4e2f0ebcd2e4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Automatically printing the relationship between the view, platform
window, and window, makes it easier to track each object across
events.
Change-Id: I9fbfaa5c304849ed99dba3b5cd8e7449105d0307
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
AtSpiAdaptor::componentInterface was not checking the coordinate type in
GetAccessibleAtPoint.
Now, if the coordinate type is ATSPI_COORD_TYPE_WINDOW the coordinates of the
window will be added to the specified coordinates.
Change-Id: Iba571109a8da300f4141d616a94ef4ac87918f98
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
If the accessible description of a QLineEdit is requested and none has been
set, the placeholder text will be provided if one is available.
Change-Id: I4c0dad1d06fd3d8a6e00c963402d380c59bd7f05
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Passing on the thread ID is confusing, as it's not really what the
function does. The QNX code path can resolve the thread ID by itself.
Change-Id: I5f0d54621058576cdcf3707d36a11762fe2383c8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The include is not needed and breaks build that do not have process
support.
Change-Id: I3951c24c950dd556a3b26744d8994709e294d397
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The actual implementation for Android will come later.
Change-Id: Ia68fcb03ea5d769d302ec6f77e5666292b567ffa
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
Commit 8f52ad9fe0 ("ucstricmp: compare
null and empty strings equal") made sure empties and nulls would compare
equally, but may have broken the null vs non-empty comparison (which was
not tested). The commit message also said that it expected all callers
to handle null before calling into those functions, but that's not the
case for QStringView created from a null QString: the incoming "a"
pointer was null.
So just remove the checks for null pointers and rely on the size checks
doing the right thing.
[ChangeLog][QtCore][QString] Fixed a regression from 5.9 that caused
comparing default-constructed QStrings to be sorted after non-empty
strings.
Task-number: QTBUG-65939
Change-Id: I56b444f9d6274221a3b7fffd150c83ad46c599b6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
But they are movable.
qxmlstream_p.h:654:32: error: ‘void* realloc(void*, size_t)’ moving an object of non-trivially copyable type ‘struct QXmlStreamPrivateTagStack::Tag’; use ‘new’ and ‘delete’ instead [-Werror=class-memaccess]
Change-Id: I41d006aac5bc48529845fffd150e8115eb852034
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is similar to commit 342bb5b03a.
From GCC 8:
qarraydataops.h:84:17: error: ‘void* memcpy(void*, const void*, size_t)’ writing to an object of type ‘class QStringRef’ with no trivial copy-assignment; use copy-assignment or copy-initialization instead [-Werror=class-memaccess]
[etc.]
Change-Id: I41d006aac5bc48529845fffd150e817e64973bec
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
From GCC 8:
error: ‘void* memcpy(void*, const void*, size_t)’ copying an object of type ‘QJsonPrivate::offset’ {aka ‘class QSpecialInteger<QLittleEndianStorageType<unsigned int> >’} with ‘private’ member ‘QSpecialInteger<QLittleEndianStorageType<unsigned int> >::val’
from an array of ‘const value_type’ {aka ‘const class QJsonPrivate::Value’}; use assignment or copy-initialization instead [-Werror=class-memaccess]
Both types are standard layout and have the same initial sequence (one uint
member), so this is a valid copy. The only difference between the two is that
QSpecialInteger has a private member, whereas in the bitfield it's public.
Change-Id: I41d006aac5bc48529845fffd150e80585fd24db7
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It can be used by custom widgets or for example by
the Breeze style from KDE, which allows to drag windows
by some widgets.
It's important on X11 because _NET_WM_MOVERESIZE requests
induced by touch sequences require support from Qt.
Task-number: QTBUG-58044
Change-Id: I31c37534555a9050cf361cad85bdef13c2808572
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
