Since the native file dialog doesn't allow changing the name filters
after it's been opened, another way to ensure that we comply with
the provided nameFilters list is to provide them all at once.
[ChangeLog][Android][FileDialog] use all nameFilters at once since
Android file picker can't change nameFilters after being opened.
Pick-to: 6.6
Change-Id: I389feecae190af41b87c7f2070494bd57eec4b69
Reviewed-by: Oliver Eftevaag <oliver.eftevaag@qt.io>
When a QWindow is moved between being a child window and a top level,
it gains a corresponding QNSWindow, which we create as part of
QCocoaWindow::createNSWindow(). NSWindows are not visible by default,
so we need to re-apply the QWindow state to the newly created window,
just like we do for other QWindow properties.
Pick-to: 6.6
Change-Id: I809d12b09eeccea7a4ee4e8444f69db46dbd159e
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Unix dispatcher is not used and - as such - redundant on WASM.
Change-Id: Ia8789ef783b06ce9cfba2ce9d67159db2355b594
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
The [NSWidow windowNumberAtPoint:belowWindowWithWindowNumber] API has issues
with sometimes being out of sync with the window server, resulting in failing
to hit test windows that we know are there.
This has manifested in flakeyness in our tests, for example in tst_QWindow's
testsInputEvents: http://testresults.qt.io/grafana/goto/YNGj7TgIg
A workaround is to call [NSWindow windowNumbersWithOptions:0] to force a sync,
but we might as well use the more modern block based API to iterate our own
windows in Z-order. This API seems to do the required sync on our behalf,
or at least doesn't operate on stale data.
The logic has been otherwise kept as is, including treating non-top-level
windows as candidates for hit testing, which seems strange for a function
named topLevelAt(). This is to be investigated further.
Task-number: QTBUG-108402
Task-number: QTBUG-115945
Pick-to: 6.6 6.5
Change-Id: I5599881c381a0a673d262e4b9585e2c6798c9810
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
During refactoring of the vista style in
46f98147f4 a regression slipped into
subControlRect()'s default branch which lead to a double call of
visualRect(). This breaks nearly all subControlRect() calculations in
rtl mode.
Fixes: QTBUG-116224
Pick-to: 6.6 6.5
Change-Id: Ida722e1b6c6a311b27a9b48d1fe1378b47c21cf4
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Santhosh Kumar <santhosh.kumar.selvaraj@qt.io>
The density of Q_FOREACH uses in this and some other modules is still
extremely high, too high for anyone to tackle in a short amount of
time. Even if they're not concentrated in just a few TUs, we need to
make progress on a global QT_NO_FOREACH default, so grab the nettle
and stick to our strategy:
Mark the whole of Qt with QT_NO_FOREACH, to prevent new uses from
creeping in, and whitelist the affected TUs by #undef'ing
QT_NO_FOREACH locally, at the top of each file. For TUs that are part
of a larger executable, this requires these files to be compiled
separately, so add them to NO_PCH_SOURCES (which implies
NO_UNITY_BUILD_SOURCES, too).
In tst_qglobal.cpp and tst_qcollections.cpp change the comment on the
#undef QT_NO_FOREACH to indicate that these actually test the macro.
Task-number: QTBUG-115839
Change-Id: Iecc444eb7d43d7e4d037f6e155abe0e14a00a5d6
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Accent color role has been renamed according to name rule of other
color roles in QPalette.
Fixes: QTBUG-116107
Pick-to: 6.6
Change-Id: I70ac98a1e97afbdc7ea5f8d79f808c307e170712
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
We already return RSA/DSA/DH/EC, so we can also return generic.
Fixes: QTBUG-115718
Change-Id: I2064c2bbce2df73985609d27a94857a0ee2e6b42
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This reverts commit 566def740e.
The fix had some unwanted side-effects when the
QAccessibleValueChangeEvent carried a value of type string. Only
QComboBox uses QAccessibleValueChangeEvent in such a way.
The consequence of the reverted patch was that it broke QComboBox
so that the screen reader would read aloud the value change event
regardless of the visibility or focus state of the QComboBox. (Thus, if
you e.g. changed the QComboBox::currentIndex on a *hidden* combo box,
the screen reader would still read aloud the event)
This is also the root cause of what is described in QTBUG-93763.
Also, due to the usage of NotificationProcessing_ImportantMostRecent
the screen reader would treat it as such an important event that it
would abort whatever it was currently speaking.
In addition, the reverted change didn't fix any bugs and it failed to
implement the suggested behavior in a correct way (as was described in
QTBUG-75003 - albeit properly described after the change was merged).
QTBUG-75003 has already been reopened due to this, and the change can
therefore be reverted quite risk-free.
Task-number: QTBUG-75003
Task-number: QTBUG-93763
Pick-to: 6.6 6.5 6.2 5.15
Change-Id: Ib91872adc563c31534fe2b30fd9c447bfcca6b40
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
We need to apply properties of the QWindow the underlying "native
window", in this case our <div>, on construction, without waiting
for the user to call one of the QWindow setters.
Pick-to: 6.6
Change-Id: Id422a9424f584e2269ef333e2c7c88a123ecb70b
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Only top level windows should have window manager features,
to match other platforms.
Pick-to: 6.6
Change-Id: I7a0563ef34aeb430d0b1a16633a5626482ccd17d
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
A child window should be possible to place at arbitrary positions
within its parent, even outside the parent's current rect. Once
the parent size is changed, the child might become visible.
The current code also caused issues when the parent did not
have a size yet (0x0) at the time of the child's setGeometry
call, resulting in the child always being placed at y=0.
Pick-to: 6.6
Change-Id: I1534b606ab6eb7d51216d3b305a1b60443c41ec2
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Pass an encoded URI string before parsing them through the Android APIs.
Fixes: QTBUG-114435
Pick-to: 6.6 6.5 6.2 5.15
Change-Id: I65131799fad81bfe7490d663d3b7996c94d37f0b
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
We've known for a long time that this is producing worse code with GCC
because of how we implemented in Q_ASSUME_IMPL(). So bite the bullet and
actually deprecate the macro, replacing all extant Q_ASSUME() with
Q_ASSERT().
The replacement is in C++23. Backporting the support onto Q_ASSUME_IMPL
was previously rejected by reviewers.
[ChangeLog][Deprecation Notice] The Q_ASSUME() macro is deprecated. This
macro has different side-effects depending on the compiler used (GCC
compared to Clang and MSVC), and there are certain conditions under
which GCC is known to produce worse code than if the macro was absent.
To give a hint to the compiler for optimizations, use the C++23
[[assume]] attribute.
Change-Id: I80612a7d275c41f1baf0fffd177a3a4ad819fb2d
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Child windows on Windows cannot have a frame, so checking only for
Qt::FramelessWindowHint is not sufficient. Add an additional check
to see if the window is a native child (has the WS_CHILD flag).
Pick-to: 6.6 6.5
Change-Id: Ib5cbec0f6157da687a5585e12a6c4c6935919538
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
QAndroidPlatformInputContext::focusObjectStopComposing() sends an input
event for each character newly added by the Android virtual keyboard.
It then sends a second input event to notify that the cursor has
advanced to the position after the new character.
The implicit assumption is, that the receiver of the input event does
not change the text.
If e.g. QLineEdit::setText() is called in the QLineEdit::textEdited
slot, the text does change. If the change implies a cursor change,
QLineEdit notifies the platform input context about it.
However, by sending the second input event, QAndroidPlatformContent
returns the cursor back to the position after the last character added
by the virtual keyboard.
This patch joins the composed text and the cursor position into one
single input method event. A new cursor position, set by the receiver
of the input method event, is no longer overridden.
The patch adds test functionality to tst_QLineEdit::setText().
Fixes: QTBUG-115756
Pick-to: 6.6 6.5 6.2
Change-Id: I85ffac5d6bab93ccb144be0f5b8083258a270550
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The density of Q_FOREACH uses is high here, too high for this author
to tackle in a short amount of time. But they're concentrated in just
a few TUs, so pick a different strategy:
Mark the whole plugin with QT_NO_FOREACH, to prevent new uses from
creeping in, and whitelist the affected TUs by #undef'ing
QT_NO_FOREACH locally, at the top of each file. Since the TUs are part
of a larger executable, this requires these files to be compiled
separately, so add them to NO_PCH_SOURCES (which implies
NO_UNITY_BUILD_SOURCES, too).
Task-number: QTBUG-115839
Change-Id: I42c26cb5d95add115e57cc484a0e1a93ed368ddd
Reviewed-by: James McDonnell <jmcdonnell@blackberry.com>
QEventDispatcherGlib: it used the out-parameter timespec to set a
timeout in milliseconds, making timerWait() return milliseconds is more
straightforward. Also timerWait() returns the time rounded to
milliseconds, so the code in QEventDispatcherGlib that rounded up the
timespec::tv_nsec part was no-op, tv_nsec was always 0 IIUC.
In QEventDispatcherGlib, guard against overflow with qt_saturate.
Change-Id: Ie6f78374d00cbe8a6adf7b50ed67c8c86ab4d29d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Description added to the html tags for accessibility in wasm through the
aria-description property
Change-Id: Ic1976d5c4c64ea45298f24b660807adeda6fdb64
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
The density of Q_FOREACH uses isn't very high here, but at this time,
this author just goes through the code-base and white-lists TUs that
still use Q_FOREACH in order to globally enable QT_NO_FOREACH.
Mark the whole plugin with QT_NO_FOREACH, to prevent new uses from
creeping in, and whitelist the affected TU by #undef'ing QT_NO_FOREACH
locally, at the top of each file. Since the TU is part of a larger
executable, this requires the file to be compiled separately, so add
it to NO_PCH_SOURCES (which implies NO_UNITY_BUILD_SOURCES, too).
Task-number: QTBUG-115839
Change-Id: If731d02f65131c94afa8beb51679ed2ff7d2cdaa
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Prevents using the channel after it was destroyed.
Change-Id: I5d65fcde0d451dd8ce2c56f403e0bd1f41510382
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
The density of Q_FOREACH uses is high here, too high for this author
to tackle in a short amount of time. But they're concentrated in just
a few TUs, so pick a different strategy:
Mark the whole plugin with QT_NO_FOREACH, to prevent new uses from
creeping in, and whitelist the affected TUs by #undef'ing
QT_NO_FOREACH locally, at the top of each file. Since the TUs are part
of a larger executable, this requires these files to be compiled
separately, so add them to NO_PCH_SOURCES (which implies
NO_UNITY_BUILD_SOURCES, too). I'm aware that Objective-C/C++ files are
currently probably implicitly NO_PCH_SOURCES, but don't rely on that.
Task-number: QTBUG-115839
Change-Id: If451501f3cb7cc3a182854a94bc5d27c907f6161
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Since we enabled QT_USE_QSTRINGBUILDER by default for all plugins
it doesn't make sense to add this definition explicitly.
Change-Id: I22ad0b1fb567e12a842df4299231a5697ab9216d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Unifies code paths that end up closing and releasing the window.
Pick-to: 6.6
Change-Id: Ie041079a22bbae0912df0406291c8146f02d0cd7
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Same implementation as for macOS.
[ChangeLog][iOS] QWindow::setMask() is now supported for masking the
rendering and touch input of child windows.
Pick-to: 6.6
Change-Id: I2f9429f0f8fa278fdd8edc15c7b242c7c6bc0ff0
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The Android QPA implementation requires a 1:1 link between a platform
window and a platform backing store, to correctly flush a backing
store to the screen. QAndroidPlatformBackingStore has a bool member
m_backingStoreSet, to remember if this link exists. It defaults to
false and is set to true, when setBackingStore() is called in the
constructor. It falsely remains true, when a platform window is
deleted, e.g. because a QWindow has been hidden. When the QWindow is
shown again, a new Android platform window is created. With
m_backingStoreSet still being true, this new platform window will
never be associated with a backing store. As a consequence, it will
never be displayed on the screen.
The 1:1 relationship of an Android platform window and an Android
backing store is neither ideal, nor in line with other QPA layers
(e.g. XCB). Changing the Android QPA implementation is complex and a
short term fix is necessary.
This patch removes the member m_backingStoreSet. Instead of it,
QAndroidPlatformBackingStore::flush() directly checks, if the platform
window corresponding to the QWindow argument is associated to a backing
store. If that is not the case, setBackingStore() is called.
QTBUG-97482 has been fixed with another approach, which this patch
reverts.
The following commits are effectively reverted by this patch:
9a39ad8dfbf91588923ba4ca9e8065dbb072eb28959a8b3967
Fixes: QTBUG-97482
Pick-to: 6.6 6.5 6.2
Change-Id: Ic4344f8df2e954c057dd2705340f11dfd2d4c6fe
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
In b64b0c7947 we bailed out from
QCocoaWindow::recreateWindowIfNeeded() for foreign windows, as
we should not manage any NSWindows on their behalf.
Unfortunately QCocoaWindow::recreateWindowIfNeeded() also took care
of adding the view as a subview to the potential non top level view,
which we do want for foreign views.
Ideally we'd move the reparenting out of recreateWindowIfNeeded()
and into QCocoaWindow::setParent(), but this is a more intrusive
change, so for now just restore the original missing logic.
Pick-to: 6.6 6.5
Change-Id: Ic35ebf94d4adc2f19cedb1cb6a5d0215a1c9c2b4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Christian Strømme <christian.stromme@qt.io>
Implement TCP server to stream trace data. Use the QTRACE_LOCATION
environment variable to configure the server address and port.
The rest of the configuration is done in the client TraceRequest.
The user can configure the compression level, buffer size and
whether the server is buffering when client doesn't have connection.
The server responds with basic response and then begins sending packets
until the client disconnects or the app closes.
Each packet contains stream name, stream data and whether to overwrite
the stream, which is used with metadata streams.
Task-number: QTBUG-105976
Change-Id: I41c708a4c7988666d59f0c6093dd41e8ccd88441
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
Remove old files from the output location before saving new ones.
Pick-to: 6.6 6.5
Change-Id: I75181126c6c920e13951e9a46a6be1c666d457fe
Reviewed-by: Antti Määttä <antti.maatta@qt.io>
Reviewed-by: Janne Koskinen <janne.p.koskinen@qt.io>
QByteArray::assign() re-uses existing unshared capacity(), if any, and
is therefore potentially more efficient than '= QByteArray(~)' (and
never slower).
Task-number: QTBUG-106201
Pick-to: 6.6
Change-Id: I5e5114064c8aaa23892e95c235335632d7c5cd0a
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Implementation for QAccessibleEditableTextInterface is pending for texts
in qml accessibility, adding accessibile role check for text to be
updated when accessibility feature is on in wasm
Change-Id: Ic24ebf6942e50a5952126e312afa633ed4fbe9ce
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Mark the module are Q_FOREACH-free, so it stays that way.
These two instances are risk-free, because the loop is over local
variables that the loop bodies clearly do not touch, so the safety
copy that Q_FOREACH takes is not required here.
Pick-to: 6.6
Change-Id: Ida3c1d51c661d77a59a2ad105e080c3c9b66c53f
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
QAndroidPlatformScreen::addWindow() asserted the existence of a backing
store, when a window is added. But QQuickWindow neither needs nor has a
backing store.
This patch removes the assertion, which was introduced in
dbb072eb28
Task-number: QTBUG-97482
Pick-to: 6.6 6.5 6.2 5.15
Change-Id: Ifb43f83d39e16b972754e86d0ff9633f58dac3f8
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Running a Vulkan-based QRhiWidget (upcoming in 6.7) and QQuickWidget
shows the "...does not support OpenGL-only windows" warning which is
pointless nowadays. Since Qt 6.4 RasterGLSurface is not in use,
whereas the surface type can very well be VulkanSurface too.
Pick-to: 6.6 6.5
Change-Id: I790767e683b2a4cffc99cbc38015aca809cf83c5
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Should be no different from MetalSurface and OpenGLSurface.
Pick-to: 6.6 6.5
Change-Id: I529f2904a43e44376e9f2da2489fac57670d954d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Add an option to disable handling of positional binding so jsonb
operators are not screwed up
[ChangeLog][QtSql][QSqlQuery] Add setEnablePositionalBinding() to be
able to disable positional binding.
Fixes: QTBUG-96636
Change-Id: I428a9d3b10274b97292ab86a74d9b3971d6f10e9
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Use QT_NO_CONTEXTLESS_CONNECT to disable 3-arg connects which are
considered dangerous.
Change-Id: I0ac711491de60e0eeaca9edb60715eafe9da841a
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Instead use composition by using a container member; inheriting from
containers has a couple of issues:
- containers (STL and Qt) don't have virtual destructors, which means
that deleting via a pointer to the base-class would be troublesome
- as it turns out, inheriting from containers, QList in this case,
in an exported derived-class could lead to binary-incompatibility
issues (see linked bug report)
Drive-by change:
- group all private members in one place
- make timerInsert() private, nothing should use it from outside the
class anyway
Change-Id: I69843835d8c854fa4b13e8b4ba3fb69f1484f6a6
Fixes: QTBUG-111959
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Bridge QAccessibleSelectionInterface that was introduced
in commit 9d16d5e224 to
UIA's ISelectionProvider and ISelectionProvider2 interfaces
by extending the existing WindowsUiaSelectionProvider
to make use of the QAccessibleSelectionInterface.
Also make use of that interface to implement handling
for the ISelectionProviderItem interface methods when
an object has a parent that implements the
QAccessibleSelectionInterface.
Sample use from NVDA's Python console [1] with this
commit in place:
1) start NVDA
2) run the interview example (examples\widgets\itemviews\interview\interview.exe)
3) select "Item 1:0", "Item 2:0" and "Item 3:0" by left-clicking
on "Item 1:0" in the left view, then shift+clicking on "Item 3:0".
4) press Numpad_insert+control+z to start the NVDA Python console and
capture snapshot variables
5) query and use the interfaces using NVDA's Python console:
>>> import UIAHandler
>>> selectionpattern2 = focus.parent.UIAElement.GetCurrentPattern(10034).QueryInterface(UIAHandler.IUIAutomationSelectionPattern2)
>>> selectionpattern2.CurrentFirstSelectedItem.CurrentName
'Item 1:0'
>>> selectionpattern2.CurrentLastSelectedItem.CurrentName
'Item 3:0'
>>> selectionpattern2.CurrentItemCount
3
>>> selectionitempattern = focus.UIAElement.GetCurrentPattern(10010).QueryInterface(UIAHandler.IUIAutomationSelectionItemPattern)
>>> selectionitempattern.CurrentIsSelected
1
>>> selectionitempattern.RemoveFromSelection()
0
>>> selectionitempattern.CurrentIsSelected
0
>>> selectionpattern2.CurrentItemCount
2
>>> selectionitempattern.Select()
0
>>> selectionitempattern.CurrentIsSelected
1
(Note that calling selectionitempattern.AddToSelection
in that example would also unselect all other currently
selected entries, because the underlying implementation
has a selection mode of QAbstractItemView::SingleSelection
set in QAccessibleTree::selectRow, which gets called
from QAccessibleTable::select via QAccessibleTableCell::selectCell.)
[1] https://www.nvaccess.org/files/nvda/documentation/developerGuide.html#PythonConsole
Change-Id: I7003bae5bbcfd5c685620bf710781165ed70f106
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
Lorn Potter
Assam Boudjelthia
Tor Arne Vestbø
Mikolaj Boc
Friedemann Kleint
Christian Ehrlicher
Marc Mutz
Santhosh Kumar
Anton Kudryavtsev
Timur Pocheptsov
Jan Arve Sæther
Thiago Macieira
Timothée Keller
Axel Spoerl
Ahmad Samir
Shreya Pattani
Antti Määttä
Alexey Edelev
Thorbjørn Lindeijer
Dennis Oberst
Liang Qi
Laszlo Agocs
Yuhang Zhao
Michael Weghorn