The applicationWillResignActive documentation describes that the app
should "use this method to pause ongoing tasks, disable timers, and
throttle down OpenGL ES frame rates". Similarly, the more modern
sceneWillResignActive describes that the app should "use this method
to quiet your interface and prepare it to stop interacting with the
user. Specifically, pause ongoing tasks, disable timers, and decrease
frame rates or stop updating your interface altogether."
In practice the application enters the inactive (as opposed to the
backgrounded) state when the user brings up the app switcher, or
enters split-view mode, as described by the "Adopting Multitasking
Enhancements on iPad" documentation.
We do propagate the inactive state as Qt::ApplicationInactive, but
it was up to the application itself to react to the state.
To adhere to the documented behavior, we now skip update request
delivery if the app is not active. Once it becomes active again
we re-try the update request delivery.
Relying on the exposed state of window was not an option, as the
window is still technically exposed, even if the app itself is
inactive.
This hopefully fixes a GPU crash observed (only) on A10 iPads
(iPad 7) with iOS 18, when the user entered split-view mode,
possibly due to changes in the system Metal-based OpenGL driver:
void IOGPUScheduler::signalHardwareError(eRestartRequest, int32_t)
void IOGPUScheduler::hardware_error_interrupt(IOInterruptEventSource *, int)
void IOGPUCommandQueue::retireCommandBuffer(IOGPUEventFence *)
Deny submissions/ignore app[openglwindow] with 2 GPURestarts in 31 submissions.
Execution of the command buffer was aborted due to an error during execution
Caused GPU Timeout Error (00000002:kIOGPUCommandBufferCallbackErrorTimeout)
GLDRendererMetal command buffer completion error
Error Domain=MTLCommandBufferErrorDomain Code=2
Execution of the command buffer was aborted due to an error during execution
Ignored (for causing prior/excessive GPU errors)
Terminating due to blacklisting by kernel driver
The back-trace always pointed to the display-link callback,
so the assumption is that the crash was triggered by the app
rendering when it shouldn't, during the split-view transition.
Fixes: QTBUG-132314
Pick-to: 6.5 5.15
Change-Id: Idf1f692daa9d437ee69f9436706777b220fbfbf4
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
(cherry picked from commit c812190a921660c748be3324c3a34245b24354e9)
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 1b993a787d7e1e2ed7bde8102847424db676d2c7)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
This is more efficient, because it maps to a single assembler
instruction on most ISAs, and also avoids a Coverity warning about the
loop not terminating and thus 1 << i overflowing, if type == {}.
The clone class, QSizePolicy, is not affected, since it already uses
qCountTrailingZeroBits().
Amends 6c322a917a.
Coverity-Id: 474336
Pick-to: 6.5
Change-Id: Idf62d045d8ac3f0cb15196dd8664ac37d92a9cc2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit d71379d447e87843c5a5d91857c4ff253ad4b847)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 4ed56eaafcc6270855b305d08d551f6a48c9667b)
std::slist doesn't seem to even exist on Google, while std::forward_list
exists in the standard since C++11.
Change-Id: I9869c716fb631bc743e656a3b8bbb8ca35c9a1ed
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
(cherry picked from commit 9e5dae9b8c3d43cfd86719e429f35b032d14c66c)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 21a0818903bca4a31e58abbdf58cb9c94830d5a3)
A few values have meaningful defaults, for the rest only warn about
implemented metric.
Change-Id: I80c6c8dba6c8d2715daac0463a411bc5117f0bc0
Reviewed-by: David Faure <david.faure@kdab.com>
(cherry picked from commit da1fbf6053c07df76a4d2518103bdfc40d59cfa8)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit a69fdf64104670d970b255d555fbabaa9a81185d)
Commit 353ce5344fbde5a6cecbdd2c131e1cf0f4b7f383 made QMenu::exec()
return immediately (without showing the menu), which meant it never
emitted aboutToHide() and QToolButton didn't reset mouseButtonDown.
Therefore, on the next click on the toolbutton (maybe by then the
menu has actions), it won't even try to show it (because mouseButtonDown
is still true).
Task-number: QTBUG-129108
Change-Id: I3f84b9a35eb444fc33d8516c4be130293c04017d
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit f0508b7ff3d2393521f49dcb1a9e880bf7665a87)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 05016926ae37ae73a9da9375a95f3a4e8234477e)
On Windows, using a URL of "file:///c:/" as a base to be resolved with
"../" should not result in the Windows drive being removed.
[ChangeLog][QtCore][QUrl] Fixed a bug (regression from 6.7) where
resolving a base URL of an absolute file path containing a Windows drive
could result in said drive being removed (e.g., resolving "file:///c:/"
with "../" would result in "file:///").
Fixes: QTBUG-133402
Change-Id: I58286b9c5e5d02363f0efffdb06f983b560340df
Reviewed-by: David Faure <david.faure@kdab.com>
(cherry picked from commit 340c9d88ab353e201f117d64609fa5f7d2fa2b21)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 8c5736e68f9619a6f2f223c7363a820e528d089c)
We set it when manipulating full URLs. If you're creating them from
parts, you may end up with one without a host. We will still fix that up
in QUrl::toString() ("manipulates full URL").
This allows the path normalization code to avoid accidentally creating a
URL with no host/authority and with a path that starts with double
slash.
[ChangeLog][QtCore][QUrl] Fixed a bug (regression from 6.7) where
QUrl::resolved() could create invalid URLs when the relative URI being
resolved contained a path with double slashes (e.g., combining
"scheme:a" with "..//b.txt")
Fixes: QTBUG-133403
Change-Id: I3fe9d5fbd2efcaa66d66fffdc010e5a84066b641
Reviewed-by: David Faure <david.faure@kdab.com>
(cherry picked from commit 7d05f5ed7d3472028e28a09eeda175bb1b1eeb00)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 86597d289116789463e905d3fdcdb3850f4b5918)
AllowUncPaths (and the whole enum) is private API that is always enabled
only on Windows (build-time decision). So, remove it and change the
unittests to match reality. Now DefaultNormalization on Windows implies
handling UNC paths.
This was suggested in code review by Thiago; since this is a hot code
path, the goal is letting the compiler keep the flags parameter to
qt_normalizePathSegments() in a register, by keeping the
PathNormalization enum as small as possible.
Drive-by change: since those lines in the unittest are changed anyway,
take the chance and use u""_s syntax.
Change-Id: I3dcf30d888a0ea9f8898e260e65c5f85655296d5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 4f47ee4de46f1fa2df9fef163590807e2fae28d1)
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
... for local URLs too.
For example normalizing "/Users/sam/troll/qt4.0/.." by following the
alogrithm at https://www.ietf.org/rfc/rfc3986.html#section-5.2.4:
- replace "/.." with a "/" in the input buffer, then remove the previous
path segment from the output buffer up to, and including, the
previous "/"
- next iteration will find the input buffer starting with a "/" which
is moved to the output buffer; the final result would be:
"/Users/sam/troll/"
Remove the KeepLocalTrailingSlash enumerator, it's become redundant with
this change.
De-duplicate a row in tst_QDir.
[ChangeLog][Important Behavior Change] Aligned how QDir and QUrl
normalize paths with respect to preserving a trailing slash. That is,
QDir::cleanPath("/b/.") and
QUrl("file:///b/.).toString(QUrl::NormalizePathSegments) will return
"/b/" and "file:///b/" respectively. For more details see:
https://www.ietf.org/rfc/rfc3986.html#section-5.2.4
Task-number: QTBUG-128940
Change-Id: Id5a4bdab14e7312a60b1b54de07e5d9f3e8aa40a
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit ec894d694506fcce75fc9643f08dfb1c79f64c8b)
On some very rare cases the continuation could be executed twice in a
multithreaded scenario. Consider the following example:
auto f = QtConcurrent::run(someFunction);
f.then(someOtherFunction);
Assuming that QtConcurrent::run() is executed on Thread B, and the
continuation is attached in Thread A, the following call stack is
possible:
1. Thread B: someFunction() is finished
2. Thread B: QFutureInterface::reportFinished() is called
3. Thread B: QFutureInterfaceBase::reportFinished() is called and
finished.
4. Thread A: then() is called. It checks that the future is already
finished, so the attached continuation is executed
immediately.
5. Thread B: QFutureInterfaceBase::runContinuation() is executed.
At this point it sees that a continuation exists, and
executes it again.
Step 5 will lead to a crash, because we'll try to access a moved-from
QPromise.
To fix the issue, introduce a flag that is explicitly used to test if
the continuation was already executed or not.
The pre-existing code already uses a concept of a continuation state,
but it cannot be easily extended without convering it from an enum
to flags, because the canceled continuation should still be called
in order to actually execute the onCanceled() handler.
Wrapping QFIBP::ContinuationState into QFlags would increase its size
from 1 to 4 bytes, and manually treating the enum as flags will result
in a more complicated code. So, I simply picked the approach of adding
an explicit flag for this case.
Writing a unit-test is not really possible, but I verified that the
reproducer from the linked Jira ticket does not crash anymore.
Amends dfaca09e85 that introduced the
continuations, so picking to all active Qt 6 branches.
Fixes: QTBUG-118032
Pick-to: 6.5
Change-Id: I3e07722495e38367e8e4be2eded34140e22b0053
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
(cherry picked from commit 04b5d2b94f96c73a13973f6a57cefbf07d2e850b)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 5b55e8d1f066945dc36deda6285aebc3fd315e00)
This amends commit 353ce5344fbde5a6cecbdd2c131e1cf0f4b7f383 so that
the behavior doesn't depend on the widget style.
If QStyle::PM_MenuPanelWidth returns 2, then size will be (2x2),
which is hardly a valid menu to show - don't show that either.
Task-number: QTBUG-129108
Change-Id: I2daa05932ea00971880747daeffdfa20a8be3f22
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
(cherry picked from commit 9dfd8410b07b0c484443ad6355004ace5f9b2b8b)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit bbc46bbc7a15728264574c39e13b6be4c9229b54)
This is hot code, run early in the application's life, usually as a
result of the first qDebug() because QLoggingRegistry wants to find
qtlogging.ini. So let's not use QFileInfo, which allocates memory, just
to do string manipulation. These are always filesystem paths.
See commit d59e640c868f3db2d661970f3d34a22013d49053 for a similar change
in findConfiguration().
Change-Id: I012daf4a3aa2ebc2c5abfffd2ef7ceac1df2626d
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
(cherry picked from commit c74cba1117355a6312b1f0cc815efa4cdea4bbfa)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 3d43223448b4d5c1828501de4ec23edb2edc64d7)
This commit adds a new parameter (--pre-test-adb-command) to
AndroidTestRunner. The new parameter allows to pass an extra adb command
which will be called by AndroidTestRunner after installation and before
running the test.
To set the mentioned argument the new parameter for qt_internal_add_test
was proposed: ANDROID_TESTRUNNER_PRE_TEST_ADB_COMMANDS.
The new parameter is needed especially for multimedia screen capture
tests. ScreenCapture feature needs an acceptation of Security Popup. It
can be automatically accepted with additional adb command.
Fixes: QTBUG-132249
Change-Id: Ib70cd05d60d4594961ca68b554c7aae11cf42240
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit b4c82eba03388aa4f9e5b56f633e0e679a3b0123)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit bc533b0fb2d50a8e124e5e54628ad847733a8de8)
Reviewed-by: Bartlomiej Moskal <bartlomiej.moskal@qt.io>
QStringViews should be passed by value, not cref, so sane ABIs pass it
in registers instead of forcing the variable to the stack.
As a drive-by, s/QVector/QList/ in the return type.
Amends 51e2a64ed9.
Pick-to: 6.5
Change-Id: I6b51dfda87643f6c7ec219ac4a8970537bb6b2ed
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 8113a50d71cb9ea320e69c2b10452972e3310f4d)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 76c50ca1674eecca230d4bff9001fc952cc20b3b)
Amends commit 6708107873 from 6.8, which
removed qloggingregistry.cpp from the bootstrap library build.
Change-Id: I3bb5609707478df771b8fffd3e71eb889eaac53f
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
(cherry picked from commit 696e11ec729452d246e736e54790411b408f9c3c)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 00caa645ff064d9c26e70699b9f14c62172b71c9)
After the commit 28df9a49776a88cb1a8e69348ae19a59b16a5b7e, a regression
occurred in the positioning of cursorHandle and EditPopup.
Previously, these positions were calculated using QtEditText
coordinates, which worked correctly because the QtEditText size matched
the QtWindow size. However, after the mentioned commit, the QtEditText
size no longer reflects the window size. In this case, we need to use
the parent View for calculations.
This adjustment was already made for the single cursorHandle.
This commit also updates the positioning of selection handles and the EditPopup.
Fixes: QTBUG-132589
Change-Id: I861292e363452d487284e3f603fe03a21a334aa4
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit 5bd26fda7a3f0a509a64847b58b916830ebc2d0c)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 941434e81fe073d55607267ea839025282413900)
ReadOnly was used in places where ReadWrite was the correct value.
Change-Id: I26a2f0de55665ac015fe269e8e5f0d23a2f00e5f
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 4dcff4ed6d24e6aa061e145bb9de9ff93351a2b1)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit a8fe95f0eab7facc274327a54ca88d4b139eb385)
On 150% DPI scaling, the Radio/Checkbox indicators were clipped. This
patch adds an extra margin to the left to avoid this clipping.
Fixes: QTBUG-133117
Change-Id: I2e0186cbdad12481922de5ddf00f50daeb94bb37
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
(cherry picked from commit 3d45d22b536cd5fa074f6de4d02b1fd049e6b220)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 9f7437fb858cc69fcd50b32809e9bd83355d0947)
Having the input field length too wide causes the view to scroll
to contain it. Since it is not visible, setting it to 1px width and
height should work.
Further, input field contents where set on the showInputPanel
call. At this point in time we do not necessarily have an updated
focus object. Hence the input contents is now set on the
setFocusObject call.
Ignore input on showInputPanel/hideInputPanel. This reflects
the current design where showInputPanel do not act unless
inputMethodAccepted(), and hideInputPanel do not act if
m_focusObject.
Still hideInputPanel used to set m_usingTextInput to false,
this behavior is not kept.
Fixes: QTBUG-132057
Change-Id: I2d7088def8060df267c5da9e045824decd913faa
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
(cherry picked from commit 7dd94d74e4497647ff62c85460643487baf372bc)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 43eb0dbd38e74158e7eb36a28cb416b357445d6b)
When rendering hinted text on Windows with DirectWrite, the goal was
to keep the rendering and metrics as close to GDI as possible,
minimizing the impact of changing the default font backend.
Therefore, the DWRITE_RENDERING_MODE_GDI_CLASSIC was always preferred
when hinting was on.
However, DWRITE_RENDERING_MODE_GDI_CLASSIC only applies antialiasing
in the horizontal direction, but GDI applies vertical antialiasing
as well. The result is that text will look more aliased with
DirectWrite than it did with GDI, which arguably looks ugly
especially at large sizes and can be perceived as a regression.
Microsoft documentation recommends using symmetric antialiasing
when the pixel size exceeds 16, so this patch enables
DWRITE_RENDERING_MODE_NATURAL_SYMMETRIC for fonts larger than
16px regardless of whether hinting is on.
It's worth noting that for fonts with heavy hinting, such as
Times New Roman, the rendering using
DWRITE_RENDERING_MODE_NATURAL_SYMMETRIC is different from
GDI, although more similar to Freetype. However, the impact
of not having vertical antialiasing is worse and as native apps
are moving towards using DirectWrite, the GDI rendering is not
going to be the definition of a "native" look anymore.
A second thing to note in this patch is that we always pass in false
for the useGdiNatural argument in GetGdiCompatibleGlyphMetrics().
According to the documentation, we should be passing in true for text
rendered with CLEARTYPE_NATURAL_QUALITY. However, doing this causes
wider kerning in certain cases. Since the tighter kerning matches
the layouts we get in native apps, as well as when using Freetype,
I've chosen to pass false for now, to be consistent.
This change also adds a manual test which can be used to switch
between DirectWrite, GDI and Freetype rendering on Windows, so
that it's easy to compare.
[ChangeLog][Windows] Improved hinted text rendering at font sizes
larger than 16px.
Fixes: QTBUG-131946
Change-Id: Iebbe5c7affe7df6266ade6b161c31bde3d2caa84
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
(cherry picked from commit 407a98d94fb05780f30e77463fb2bd535041a044)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit e59362678880cf89a2e7870365a030f0b9c254a7)
Like QGuiApplication does as well. Also add a note that screen and
window DPR can be different.
Change-Id: I1a24cf7f11e945187da9dce4421039350e811997
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
(cherry picked from commit 97049278443860d3dd1b1efd20d1f7dabaf72a35)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 281c45452b93b38d50b4c43a743017c5a1aa7cd1)
Reviewed-by: David Redondo <qt@david-redondo.de>
We look it up with a binary search, so it needs to be sorted.
Pick-to: 6.5
Change-Id: I0aa25e6ff9722372c0a09f2dce9f9c839fedf00e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit 382a7d006b52faa08b84e56d97cc6c216925a43e)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 00ddc174b49ef4ed0e2384a2377df55d9201ad9f)
isActive() may return true if a (transient) parent window
is active (has focus).
Compare against QGuiApplication::focusWindow() instead.
Fixes: QTBUG-129233
Change-Id: I834a5814a5db161bdea2ba3524d1010e0c0e3e6e
Reviewed-by: Even Oscar Andersen <even.oscar.andersen@qt.io>
Reviewed-by: Piotr Wierciński <piotr.wiercinski@qt.io>
Reviewed-by: Lorn Potter <lorn.potter@qt.io>
(cherry picked from commit 31762e2e704673057a292ac59977b501919f2450)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 55148eb94d314187986492214415e0f0472bbb26)
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
- The docs were restructured in v. 13.0 and 14.0, so most headings
changed, some topics were removed and lots of topics were added
- Use the macro \QC
- Fix the broken links in qtbase
Fixes: QTBUG-132516
Change-Id: Ic173a3e9a1c80322162c3feb277098de2a9f1cc6
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
(cherry picked from commit 7419c9f77c0b56cc98d5b44f1fd89da6bca8b148)
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
memcpy() takes void*, not char*, so the cast from QChar* to char* is
not necessary (and potentially harmful).
Amends the start of the public history.
Pick-to: 6.5
Change-Id: I7b7178290bd30d93ed9e64b6b84294d0f7576c80
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
(cherry picked from commit 7206f6a39d298ed9c94bccf7bdc57498fa73ac36)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 0e5e850691250144eb9bd5a3f88c7e2f3264e940)
Amends 73405890b8911dd1a58120665ce63a38281c6ce4. The patch here
does the same for D3D12 as that previous one did for the Vulkan
backend.
Others, like D3D11, are not effected because the loop logic there
does not branch.
Fixes: QTBUG-133405
Change-Id: I88ac7c841dc0c28fc3f6b37c3402315b52de0a66
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
(cherry picked from commit 68e48ac60a9681887999e31f6dff622eb81bdb7d)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 987ff07a85e7d9676e10e1792a1749d290d7ffa5)
Fixes clicking UI elements with a mouse on Android.
8d8cbe87e2 introduced changes to support
mouse buttons other than Qt::LeftButton, but the release always looked
like no buttons changed, nor were they tracked (m_buttons is always
Qt::NoButton). Qt was not notified of mouse up, so nothing was clickable.
Now all mouse events go through sendMouseButtonEvents, and the last seen
button state is tracked for every event. If a mouse up with no buttons
occurs, the last seen set of buttons is used instead, so Qt correctly
handles the event. Also adds the mouse button state information to
mouse move events from Android, so the workaround for delivering
Qt::LeftButton when a window is tracking a move while a button is
pressed has been removed.
Tested on a Samsung A1 with a Bluetooth mouse.
Fixes: QTBUG-132700
Fixes: QTBUG-130297
Change-Id: I241282c2915d7e6cf99db7f0bc1ad2d541349077
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
(cherry picked from commit d908e043984dcfed3aa80e30cd1cafacd13b644d)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 65e9eca63b3f522065018777ac07c5a71c08ff37)
The various base classes that abstract away template-independent parts
of QVLA have protected destructors to prevent slicing. Re-add the
special 5 (using the macro) and a defaulted default constructor.
Pick-to: 6.5
Change-Id: I60f036eafc808862a1abee72fd53023dd99b4cc3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit c10072895d1061e258d23f3cc2a9050de5236a46)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 5ca7ad268a3b2fb68c9505eee888aa62c2aa6189)
If a class honors the RO0 we usually just leave a comment in the class'
body, and don't explicitly redeclare any of the RO5 special member
functions.
In some cases we may need to redeclare (some of) them. The main use-case
so far is to declare a protected destructor for a base class that is
not polymorphic, in order to prevent slicing; the compiler-provided
destructor is always public.
We can easily declare and default such a protected destructor, but that
comes with the problem that now we're violating the RO5 (as far as C++
is concerned, declaring a SMF counts towards the RO5, even if it's
immediately defaulted).
Specifically: by declaring the destructor, the class loses the
compiler-generated move operations, and the copy operations are
generated but deprecated. Clang >= 18 warns about this.
In such a scenario we *must* redeclare all five SMFs... and default them
all. This is boilerplate, therefore I'm adding a macro to streamline it.
Apply the new macro to a couple of cases were manual refactoring has
already been done.
Pick-to: 6.5
Change-Id: I5dc0ff9770621fbec0b1057c164d0623f901e3e9
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit fcb57391f73d36914e10ba964dbd9b01fe6f3af2)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 711b86fa2471cb294d02b2ace89204b22fef0120)
Amends commit 0756cc1eae.
The generic instantiation of this function had a std::nullptr_t *
parameter, but callers had special code to pass a nullptr there because
we never needed a value of a nullptr (it's always a null pointer). For
example, in compare_ptr_helper():
auto lhsFormatter = Internal::pointerToString<QObject>;
auto rhsFormatter = Internal::genericToString<std::nullptr_t>;
return compare_helper(t1 == nullptr, "Compared QObject pointers are not the same",
const_cast<const QObject *>(t1), nullptr,
lhsFormatter, rhsFormatter, actual, expected, file, line);
But in debug mode, some compilers did emit a load from this memory
location, causing a crash. So we just specialize this function to avoid
such.
We had a test for this... except it was never reached because the
earlier QCOMPARE() had already failed. For the test, this amends
commit ae02188233.
Conflict resolution for 6.8: regenerated the expected output for
tst_selftest.
Fixes: QTBUG-133330
Change-Id: I2cd3bb475788431c6a0dfffd28e730e8b613e033
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
(cherry picked from commit e19b633c468123526660b40ae110f46090682c76)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 042e713cb7800f97c10d2d75a717ceb3e41ab29a)
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Definitely a compiler bug because it passes for everything we threw at
the compiler in the CI but fails for a user. At least the workaround is
simple.
Fixes: QTBUG-133032
Change-Id: Ie84590c466cf1d6f01bbfffd3ca1d6c262d77200
Reviewed-by: Irfan Omair <irfan.omair@gmail.com>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 0a5b37b706ee72c7d9042efcde244aa93a5ce9a5)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit d37960b35ab17fddd6b52862f33f07b82d83145c)
Amends commit 19b7f854a2, which removed
the __PIE__ check because of the GCC version check that was alongside.
GCC and Clang define both __PIE__ and __PIC__ for -fPIE, so checking for
__PIC__'s presence isn't enough.
Drive-by amend the no-PIC #error message to also suggest
-DFEATURE_no_direct_extern_access=ON, which is more modern and produces
better code for executables too. Because of that, I've also changed the
reduce_relocations option to the CMake one, as there is no configure-
style equivalent for the modern option.
Pick-to: 6.5
Fixes: QTBUG-133101
Change-Id: I957b264eaee360fd964afffd60617964b26fe934
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 014e77801fa4b99cdb03330ccce1e212ac13a0af)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit d07a849e9853c200556abf277759679d50c645de)
The selection of which set to iterate over and which one to remove
from based on their relative size violates the function's
documentation, which clearly states that items are removed from *this,
and not from `other`, so the result must never contain any elements
from `other`.
Amends 4f2c96eaa8. Instead of reverting
to the gruesome old code with the forced detach-just-to-remove copies,
distinguish four cases:
- if the two sets are shallow copies of each other, then their
intersection is *this
- otherwise, if either set is empty, clear() *this. This is required
for one of the tests that 29017f1395
added to succeed.
- otherwise, if *this is detached, perform the operation in-place,
using removeIf()
- otherwise, create a new set and move-assign to *this to avoid
detaching just to remove something again. In this case, we can
continue to iterate over the smaller set, but we need to keep
picking elements from LHS into the result.
[ChangeLog][QtCore][QSet] Fixed a regression (introduced for Qt 5.2)
in intersect() that caused equivalent elements of `*this` to be
overwritten by elements of `other` if `other.size()` was larger than
`this->size()`.
Not picking to 5.15, as users will have likely adjusted their code to
the buggy behavior, and because removeIf() isn't available there.
Pick-to: 6.5
Fixes: QTBUG-132536
Task-number: QTBUG-106179
Change-Id: Idfa17c3b3589c4eacec27259fc01df6aeaa6c45f
Reviewed-by: Øystein Heskestad <oystein.heskestad@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
(cherry picked from commit 162015e9c6f469951d9212ef655cff16dcace071)
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit eb13efc4a606cdb0c6f785a6319c489ad53626f3)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
This reverts commit 92acc94fa9.
The change broke QSet ordering guarantees: The documentation clearly
states that each item from `other` that isn't already in `*this` is
inserted ("STL insertion behavior"). Swapping *this and other breaks
this.
Independent of STL vs. Qt insertion behavior, making the picking of
elements from containers "random" in the sense that the size of the
container is now important, and not merely LHS vs. RHS, is a bad idea.
[ChangeLog][QtCore][QSet] Fixed a regression in unite() that caused
equivalent elements of `*this` to be overwritten by elements of
`other` if `other.size()` was larger than `this->size()`.
Fixes: QTBUG-132500
Change-Id: Ia636b62325139d618b5467a643ff710716324296
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
(cherry picked from commit 2d1b3028673493cb144060cbec49b1b95f4188d2)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 3dff1a3f7b22bc623b3c326630bfd94d2bd35d08)
- There is no load() in Q*Atomic (unlike std::atomic), so link to
loadAcquire()/loadRelaxed(), depending on context.
- Also remove the template argument, <T>, which prevents qdoc linking.
Amends b62c3a8545 and
329e09a3c3.
Pick-to: 6.5 6.2 5.15
Change-Id: I8c58674ffc7e779fee982555edcd45ed40d4c17d
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
(cherry picked from commit 9854482835071993968a524878b601c1d9327f1e)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 40dc33a9d333cb6befae040d64847e175d6dafa2)
To prevent colored glyphs (like emoji) from being cut off at the edges,
we can simply apply the solution from bbba71748438dc05d344bb9727f10b7f55a5d229
to colored glyphs as well (i.e. to glyphs with QFontEngine::Format_ARGB format).
Fixes: QTBUG-133297
Change-Id: I5578e2d12f956d0635b662b3ecf714aa63474945
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
(cherry picked from commit 7cc0ea0b5a1dc1d9e1a9a8cc8639958e42adba3a)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit d35a1eb1ac44a890268954a148ea3907cb78aea6)
I've been meaning to add this since at least 2012 (cf. commit message
of 9180d9aa8b), this is as good a time
as any.
Detection taken from qcompare.h, except for
- libstdc++: using version macro instead of date one
- Dinkumware, which is taken from the above-mentioned commit (adjusted
for the LIBCPP->CPPLIB typo),
- STLport/SGI, which are taken from stlport's _stlport_version.h,
- RogueWave, which is taken from Apache stdcxx's rw/_config.h, a fork
of RogueWave, contributed to Apache by RogueWave themselves.
It looks like not all STLs provide a version macro, and those that do
vary a lot in granularity, so for the time being, just define these
macros to nothing and not to some form of version.
[ChangeLog][QtCore] Added Q_STL_ macros for stdlib detection (libc++,
libstdc++, MSSTL, Dinkumware, STLport, SGI, RogueWave). If your STL is
lacking, please file a bug report. Note that these macros are not
considered public API just yet.
Picking to all active branches, since not picking is more risky (code
that assumed these macros existed in older branches could silently
change behavior if they don't).
Fixes: QTBUG-132908
Fixes: QTBUG-132909
Pick-to: 6.5 6.2 5.15
Change-Id: I8f956d131292483b7727f11f69b460b12a06b583
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
(cherry picked from commit b6f825a857d0d2266ea89879f26703f86ddf669a)
Reviewed-by: Qt Cherry-pick Bot <cherrypick_bot@qt-project.org>
(cherry picked from commit 719799234136570c1dda6c649e39d08989bb5ec1)
Tor Arne Vestbø
Marc Mutz
Aurélien Brooke
Allan Sandfeld Jensen
David Faure
Thiago Macieira
Ahmad Samir
Ivan Solovev
David Boddie
Bartlomiej Moskal
Friedemann Kleint
Wladimir Leuschner
Even Oscar Andersen
Christian Ehrlicher
Eskil Abrahamsen Blomfeldt
David Redondo
Mårten Nordheim
Morten Sørvig
Assam Boudjelthia
Leena Miettinen
Liang Qi
Laszlo Agocs
Andrew Forrest
Giuseppe D'Angelo
Kai Uwe Broulik
Vladimir Belyavsky