Building a static library Qt configuration will now require a
minimum CMake version of 3.20.
Qt builders and packagers can still opt out of the mentioned minimum
required version by configuring Qt with QT_FORCE_MIN_CMAKE_VERSION.
Such a Qt configuration is /NOT SUPPORTED/.
To facilitate these changes, the minimum version check has been moved
to happen after the BUILD_SHARED_LIBS option is computed by either
QtAutoDetect.cmake or set by a user provided cmake toolchain file.
Introduce a new QT_MIN_SUPPORTED_CMAKE_VERSION_FOR_STATIC_QT variable
in .cmake.conf to mark the minimum version for a static Qt build.
Pick-to: 6.2
Task-number: QTBUG-95018
Change-Id: Idc1875729f26a7c635b6bd26ac0c1be973917c13
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Craig Scott <craig.scott@qt.io>
In order to filter out file names based on the user's settings,
QFileSystemModel used to have a loop that tested if a given file
name matched one of the filters. The problem is that each filter
(a wildcard) was converted to a QRegularExpression _inside_ the loop.
This causes a quadratic behavior (number of files * number of filters).
Instead, build the regexps once when the filters are set (or the case
sensitivity is changed, as that affects the filtering), and simply
_use_ them in the loop.
Simplify and correct some related code as a drive by.
Done-with: Jean-Michaël Celerier
Fixes: QTBUG-95383
Pick-to: 6.2 6.1 5.15
Change-Id: I6bc336364c145bb05793a8f867545d7715d35832
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
The port from hand-rolled function storage to std::function
inadvertently removed the helper variable of static storage duration
whose dtor would unregister the conversion function. This caused
QTBUG-94831, where the cleanup of conversion functions attempts to
call code (via std::function) from a library that has already been
unloaded.
Restore the 5.15 behavior by adding a static-storage-duration scope
guard to unregister the conversion and view functions from Qt upon
library unload (when static objects are destroyed). Unlike 5.15, only
install the scope guard upon successful registration, ensuring that
only the DLL which successfully registered its conversion function
unregisters it again.
Amends 0e4ae4fbf8.
Add some strategic std::move()s as a drive-by.
Pick-to: 6.2
Task-number: QTBUG-94831
Change-Id: I391ca667420cf0d98a166676b9bb363d6e190306
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead of contains(), 1-2x operator[](), and remove(), equalling 3-4
separate lookups, use find() + erase(), which does just one
lookup.
Since our erase() function is C++11-compliant these days and takes
const_iterator instead of (mutable) iterator, we can use the const
find() overload to delay a detach (attempt) until we actually erase().
Change-Id: I8e67a48e221e548528049fa093ab7ef2f1802f7e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Move the documentation of the format and precision parameters to
QLocale::toString(double, char, int), reference it from various
QString methods (instead of repeating there and referencing one of
those from QLocale). Add brief first lines for various documentation
comments.
Mention the special handling of negative precision in the moved
documentation. Mention QLocale::FloatingPointShortest, add its type to
\sa lines of methods it affects. Change a comment on some code
implementing its special treatment to make clear that it does apply to
'e' and 'f' formats, not only to 'g', even though it has no overt
special handling in that code; and update docs to report the
undocumented behavior the comment previously described.
Document how infinity and NaN are represented. Be somewhat more
consistent about single-quoting the format names where referred to and
applying \c to character constants.
Make clear what things are different between different methods using
these parameters. Reference QString::number() from QByteArray's
relevant methods, since they share its quirks.
In the process, rename the format and precision parameters of relevant
functions so they're consistently named, replacing a mixture of
abbreviated forms.
Pick-to: 6.2
Change-Id: I077521b30346000e4b4f6968a8e750e934f72937
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Because we're using a thread-specific global static, we need to recreate
the QCollator type every time setDefault() is called. Since one is
stored per-thread, we simply increment a generational counter and check
that the next time that QString::localeAwareCompare is called in each
thread.
The issue was introduced on f6425da424
(5.12) even though we were supposed to be using QCollator since commit
5279134935 (5.3).
Fixes: QTBUG-95050
Pick-to: 6.2
Change-Id: I7e0b82c2d2fe464082d8fffd1696a8c9f74e5b30
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
USE_COMPARESTRINGEX is not defined anywhere. Its definition was removed
by commit 45b0f1be68 ("Remove winrt").
Change-Id: I7e0b82c2d2fe464082d8fffd1696a931fa301986
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This follows up on commit 98666c8afc,
which did the same for QString. If someone wants to get formatting
suitable to an unsigned value, they can cast the value to that
unsigned type and the correct overload shall pick it up.
[ChangeLog][Important Behavior Changes] QByteArray's formatting of
negative whole numbers to bases other than ten now, like QString's
(since Qt 6.0), formats the absolute value and prepends a minus sign.
Task-number: QTBUG-53706
Pick-to: 6.2
Change-Id: I91fee23d25ac0d5d5bcfcbeccbac1386627c004a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use _qt_internal_set_up_static_runtime_library() to set correct MSVC
runtime library for targets created by qt_add_executable(),
qt_add_library(), qt_add_plugin() and qt_add_big_resources(), to bring
convenience to user projects.
Pick-to: 6.2
Change-Id: Ia49f1e90989c464820824be3e1eef3df1351cdcf
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This amends 3224c6d7d1 to account for the
case when the dockwidget is already floating.
Task-number: QTBUG-70137
Pick-to: 6.2 6.1 5.15
Change-Id: If8b345565b11b44beb3fb4b697cfe812c29c6396
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
One is a bad application or library in this case, but nonetheless
we should handle this more gracefully then just crashing due to
the QRhi already having been destroyed. Mainly because in Qt 5 one
could get away with the same: releasing OpenGL objects underneath,
for example, a QSGPlainTexture with no (or wrong) GL context did
not generate any user visible fatal errors. So we should not crash
in Qt 6 either with these code bases.
In debug builds or when QT_RHI_LEAK_CHECK is set, one will get the
unreleased resources warning printed in Qt 6, which is a step
forward compared to Qt 5. So there is still some indication that
something is badly designed, even if the application survives.
Task-number: QTBUG-95394
Pick-to: 6.2
Change-Id: I944f4f425ff126e7363a82aff926b280ccf1dfc3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
The value returned from shouldHaveNonClientDpiScaling()
controls two related behaviors:
1) Should Qt call user32dll.enableNonClientDpiScaling()
2) Should Qt code treat NonClientAreaScaling as enabled.
Commit c35643db updated shouldHaveNonClientDpiScaling()
to account for the fact that PerMonitorV2 always enables
NonCLientAreaScaling, with the intent to disable 1)
However this also disables 2), which was not intended.
Instead, make shouldHaveNonClientDpiScaling() always
return true when PerMonitorV2 is enabled, and then also
omit calling the user32dll API in this case.
Change-Id: I1d06f36a3d06becc667351fadcb00ab28af6ec4b
Pick-to: 6.2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
We were using CGDisplay callbacks to determine when a screen reconfiguration
had happened, and when it had propagated to changes in NSScreen.screens,
so that we could update our QScreen view of the world.
Unfortunately the CGDisplay callbacks were not deterministic enough to
use as a signal for when a reconfigure had completed.
Since we can't rely on NSApplicationDidChangeScreenParametersNotification
either (it comes in too late), we're now resorting to updating our QScreens
at every chance we get:
- On every CGDisplay reconfiguration ending
- On QCocoaWindow::windowDidChangeScreen() as a result of AppKit
moving the window.
- On NSApplicationDidChangeScreenParametersNotification
- On QCocoaScreen::get() as a last resort
Since the result of these updates are only reflected as QScreen property
updates or QGuiApplication signals if a change actually occurred, it should
be safe to update early and often.
Task-number: QTBUG-77656
Fixes: QTBUG-80193
Pick-to: 6.2 6.1 5.15
Change-Id: I98334a66767736d94ad2fcb169e65f0d8bc71a30
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Currently number(double) creates a QByteArray-instance and calls
setNum(double).
setNum(double) creates a QByteArray-instance and copy-assigns it to
itself.
By making setNum(double) call number(double) we can cut one creation
from numbers(double)'s path, and making setNum no worse than it was.
The other pairs of setNum/number are left alone as they don't have this
issue.
Change-Id: Ib611ebc78db64f74244a56cb8e384e42bc7d1758
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Also change the timer default value to use QDeadlineTimer::Forever.
Pick-to: 6.2
Change-Id: Ia6d0101872a5d01b04e146cd9b2f90315cb8eb2e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The actual buffering is performed either in the inner socket (Unix),
or in the pipeReader (Windows). To avoid unnecessary copying of data
between buffers, enforce unbuffered mode for all implementations,
which slightly improves read performance.
Change-Id: Idb2a09f35823b9945c71231710290be12d9930f7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
For subclasses that implement alternative buffering (such as
QLocalSocket), it would be useful to bypass the QIODevice's internal
buffer, even if the QIODeviceBase::Unbuffered flag is not set when
opened.
By setting the readBufferChunkSize member to 0 in their constructors,
these classes now unconditionally switch the base class to unbuffered
mode, while still reporting buffering with openMode() function.
Change-Id: I351bc57ac0ccb45c81f8c6be15f1745131aa26ba
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
It would have been nice to ensure that a device is registered already in
mouseEnteredImpl(); but in that context, NSEvent.deviceID is always 0,
and we can't find out anything else about the device.
QWindowSystemInterface::handleEnterEvent() doesn't currently take a
QPointingDevice either.
In handleMouseEvent() and scrollWheel(), deviceID seems unique for each
trackpad or Magic Mouse, but 0 for any plain USB mouse. There, the first
mouse that the user interacts with becomes primaryPointingDevice():
its deviceID is assigned to systemID (except if deviceID == 0, we use
1 instead, to avoid the auto-incrementing device ID assignment in the
QInputDevicePrivate ctor.) When scrolling occurs, we update the
capabilities to have PixelScroll if theEvent.hasPreciseScrollingDeltas.
So over time, QInputDevice::devices() should build up to a complete
list, with capabilities() also distinguishing plain mice from those that
have the PixelScroll capability. And in the common case that the user
has only one Apple pointing device, it becomes primaryPointingDevice().
Pick-to: 6.2
Task-number: QTBUG-46412
Task-number: QTBUG-63363
Task-number: QTBUG-72167
Change-Id: Id9771b4dfd765e49023bd57d42a2aa4d0635a3b2
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The key part of FailureReportingMode::QtTest is that it will result
in a test failure (via QTest::qVerify); not that it's using the
QtTest logging machinery.
Pick-to: 6.2
Change-Id: I18472fba173ce1323748f1f87301a4defa87020e
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
- avoid calls to private and virtual functions, if the device is not
open;
- avoid repetitive checks in loops;
- add missing checks in readLine() overloads;
- remove check against unsuccessful resize().
Change-Id: I973d5931163b25db1c09c7c3b66f29ea90bb1b29
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
This signal is emitted by the QIODevice itself, so we don't have to
forward it from the internal socket.
Pick-to: 6.1 6.2
Change-Id: I85745f36d7a27d92f339a9184de3b6e5d46f6f34
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Remove the QString/QStringView/QLatin1String/const char* overloads
from the API, but not the ABI.
As a drive-by, replace a use of QStringView::left() by truncate(), as
suggested by a comment.
[ChangeLog][QtCore][QUuid] The from-string constructor and the
fromString() function now take QAnyStringView (was: overload set with
a subset of QString, QByteArray, const char*, QLatin1String,
QStringView each).
Change-Id: If7fa26cfbef9280480c78b669d9f5f14118995ed
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Apart from a more fitting, minimal, API, QDuplicateTracker also
transparently uses C++17 pmr::monotonic_buffer_resource to avoid, or
at least reduce, memory allocations.
Change-Id: I155f5518190c5f8f6d21fbec3fcecd6bcc7ff852
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
From the API review session, a potential deadlock behavior might
occur when using QFuture's synchronous APIs on the UI thread. Also
the fact that this api currently have an implementation only for
Android. For those reasons we thought this API could be postponed
until Qt 6.3, when the QFuture concern is addressed and other
platforms other than Android are implemented as well.
Pick-to: 6.2
Change-Id: I1aef025488c24791da85d15fb57367d3e5e681be
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The comment indicates how to use QT_BUILD_REMOVED_API, which includes
a #include of qglobal.h; but we have a tool, run somewhere in
configure, that grumbles about this #include (unaware, I suspect, that
it's in a comment), saying we should #include <QtCore/qglobal.h>; so
change the comment to say that. Presumably this is our recommended way
to do includes, so we should follow it when suggesting how to write
code.
This amends commit 0c8b98774c.
Change-Id: I4683ad6f1c0eedec3eaa02ac40c5d017059a49b0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Formatting using scientific notation with extra precision wasn't tested
Change-Id: I7a89a0f3d6468515604e43e52fc366dedf3c39ea
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
According to Microsoft Docs [1], DWM composition is always
enabled and can't be disabled since Windows 8. Now that
we are cleaning up pre-Windows 10 code, this apparently
needs to be removed as well.
[1] https://docs.microsoft.com/en-us/windows/win32/api/dwmapi/nf-dwmapi-dwmiscompositionenabled
Task-number: QTBUG-84432
Change-Id: I64dc049e0741600c8d0ae4db0e9e3bc98a211339
Reviewed-by: André de la Rocha <andre.rocha@qt.io>
The format was renamed in 27db9e458c.
Pick-to: 6.2
Change-Id: I53975c7467d8768dc9dc9ac2d89c42eefa12e22f
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
For historical reasons a few of the subtests are skipped when
running with anything but the plain text logger to stdout.
To ensure we have as broad test coverage as possible for the
expected output of the various loggers we run these tests in
stdout-mode.
Pick-to: 6.2
Change-Id: I856905d1543afe89710533657a55bd599c0305fd
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
If Vulkan headers are present on the system when qtbase is configured,
QtGui and QtOpenGL should be compiled with Vulkan support.
If a user project uses a Qt built with Vulkan support, but their system
is missing Vulkan headers, the project configuration needs to succeed.
The project will get compilation errors if it uses Vulkan headers, but
that's intended.
This use case was broken when fixing Vulkan to be found when building
Qt for Android.
Fix the regression with a combination of things
1) Mark the WrapVulkanHeaders package as optional (already the case)
2) Use the include directories directly when compiling Gui and OpenGL
3) Propagate WrapVulkanHeaders::WrapVulkanHeaders link requirement to
consumers only if the target exists. It won't exist if Vulkan
include dirs are not found
This also requires some changes in pri and prl file generation.
For prl file generation, we don't want to link to the
WrapVulkanHeaders target, so we filter out all dependencies that
use TARGET_NAME_IF_EXISTS for anything that calls
__qt_internal_walk_libs which includes qt_collect_libs.
For pri files, we make sure to generate a uses=vulkan/nolink clause
by inspecting a new _qt_is_nolink_target property on the target.
We also don't add include dirs to the pri file if the new
_qt_skip_include_dir_for_pri property is set.
This is intended for Vulkan, because there is separate qmake logic to
try and find the include dirs when configuring a user project.
As a drive-by, fix nolink handling for WrapOpenSSLHeaders.
Amends bb25536a3d
Amends 7b9904849f
Pick-to: 6.2
Fixes: QTBUG-95391
Change-Id: I21e2f4be5c386f9e40033e4691f4786a91ba0e2d
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
This fixes building Qt with Vulkan support on macOS which aside from
the regular Vulkan headers also needs the location of the MoltenVK
headers.
This assumes the usage of the LunarG Vulkan SDK for macOS.
Pick-to: 6.2
Fixes: QTBUG-80576
Change-Id: I030b2fdeaafdd7bf5ebe6fb4c9ea860917e84056
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The JUnit reporter was initially named xunit, but the naming was inaccurate
and the reporter was renamed in 27db9e458c.
The corresponding test has now been renamed as well, and as an added bonus
we only run it for that reporter.
Pick-to: 6.2
Change-Id: I59cb7d949514cdf46a0199a53a7a3e39f833207c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
This patch introduces some test improvements to check the calls of
different methods on an empty default-constructed string.
Apart from that, many other tests are added to extend code coverage.
Task-number: QTBUG-91736
Pick-to: 6.2 6.1
Change-Id: If86ef3d8611a678798b1bcc60a1a4f5598fd2179
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This patch fixed two bugs in indexOf/lastIndexOf:
1. The lastIndexOf(char, qsizetype) overload was crashing with an empty
QByteArray. It was unconditionally calling lastIndexOfCharHelper()
which assumes that this QBA is not empty. An explicit check for
the empty case is added.
2. The indexOf(QByteArray, qsizetype) overload was behaving incorrectly
while searching for an empty QByteArray. In this case it
unconditionally returned its second parameter (from). However, from
can be negative, or even exceed the size of this QByteArray. This
patch handles this cases properly.
As a drive-by: this patch adjusts the QByteArray::indexOf(char, qsizetype)
and QByteArray::lastIndexOf(char, qsizetype) overloads to match with the
QByteArrayView implementation. This is done to have similar code paths
in both cases and avoid tricky bugs in future.
Ideally we had to adjust the QByteArrayView implementation, but it's
fully inline, so can't be changed without breaking BC.
Task-number: QTBUG-91736
Pick-to: 6.2 6.1
Change-Id: Iaef2fdc5b99cce6aa342cca2d17544a1ad7ca677
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
There are few slots whose lambdas are empty; most will at least
capture [this]. But there are a few in Qt examples that do, e.g. []{
qApp->quit(); }. Logging is also an example. So go the extra mile and
optimize for empty functors by inheriting from them as opposed to
storing them in a member variable.
Change-Id: I3904f10db5ebe904ba889d29c08569edd804df3b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The QString itself can be compiled without QRegularExpression, but
the tests do not check if they are supported or not.
This patch fixes the issue by introducing the proper #ifdef guards.
Task-number: QTBUG-91736
Pick-to: 6.2 6.1
Change-Id: I797691f78a34d4f78a86af99c78bf06e26e846d1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The vsnprintf we use in QTest doesn't not have a portable %ls:
It accepts wchar_t, so it's UTF-32 on Linux and UTF-16 on Windows
Change-Id: I9ebda1e92b6e8e4dbbb79c6f2e35a833c587a089
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In commit 9bad096c09 I gave
qFuzzyIsNull(qfloat16) a more forgiving threshold, consistent with
qFloatCompare(qfloat16, qfloat16)'s fractional threshold. The selftest
failed to catch two of the tests failing, so fix one of them to use
different values, and remove one.
Updated test expection for txt and deleted for other formats, as
they're skipped (in tst_selftests.cpp) for this test. Refined the
generator script to know about this test only being tested for txt.
Pick-to: 6.2
Change-Id: I109547cf92178bb9f5ff0b06e0b3bb40c881b41b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Building a shared library Qt configuration already required a minimum
CMake version of 3.16, because we depend on CMake's Autogen.json files
which are used for metatype.json file generation.
If a lower version was used, a FATAL_ERROR was issued in
qt6_extract_metatypes.
This change is essentially moving the check to happen earlier in the
qtbase configure process.
User projects will now /also/ officially require a minimum of 3.16
(up from 3.14).
As a consequence, the min/max version policy range that is set in
the public QtFooConfig.cmake files is changed
from 3.14..3.19
to 3.16..3.20
The upper bound is raised because building and using Qt works fine
with all CMake 3.20 policies set to NEW.
[ChangeLog][CMake] Building Qt as shared libraries now requires
CMake version 3.16 or later. Building user projects with CMake using
that Qt installation also requires a CMake version of 3.16 or later.
Pick-to: 6.2
Task-number: QTBUG-95018
Change-Id: I77d2829370f1dfc90b4071bebc8a3ade654e59e6
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Craig Scott <craig.scott@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
