We don't need to reload d->threadData, as we've already loaded it into
a register three lines above, so use the already-loaded value.
We don't care whether we might have read a different value the second
time around, because that would mean we had a race condition elsewhere
in the program (e.g. concurrent moveToThread() calls).
Pick-to: 6.3
Change-Id: I29a27ca23302288b5900ac6b45b8690a80e85680
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
We don't need an acquire fence for checking a pointer for nullness,
and we're not dereferencing the pointer later, so use a relaxed load
there.
Pick-to: 6.3
Change-Id: Id84e6fc50100f1bf6a4e33f89424f8b1cbb250cd
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We dereference thisThreadData in the next line, at a point in time
where we haven't, yet, verified that it's this_thread's QThreadData,
so we need an acquire fence.
The alternative would be to re-arrange the code so that dereferencing
the pointer is delayed until after we verified it's this_thread's, but
that doesn't seem readily possible.
Even if it was easy, we'd first need to verify whether there are any
writes into QThreadData objects after they've been constructed, in
which case the acquire fence may be needed even in case it's 'ours'.
So just add the acquire fence.
Pick-to: 6.3 6.2 5.15
Change-Id: I468bc1f971bd87345bfcd6c13b7384bdf09d086a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Several classes in QWidget use QDataStream internally in order
to save and restore state. These QDataStream usages were not
versioned, meaning that if Qt changes the serialization for some
datatype, then the data saved between different Qt versions becomes
incompatible. Note that the save/restore API in question just produce
opaque blobs as QByteArrays -- the user has no control over the
QDataStream objects and thus versions.
Fix by version the usages.
In QHeaderView this has caused a regression because QBitArray *did*
change version between Qt 5 and 6. In general, using QDataStream without
explicit versioning is a mistake, so deploy the same fix elsewhere as
well.
Fixes: QTBUG-99487
Pick-to: 5.15 6.2 6.3
Change-Id: I82bb5c266f4e5dedc0887cbef855dccab1015e29
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: <doctor.whom@gmail.com>
The code is hard to read due to the long line generated.
Task-number: QTBUG-100145
Change-Id: I1d1a7a547db71cca1e710d39df809f079fc2dafe
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
In a QGraphicsTextItem without a width yet, there's no need to do any
layouting. The use case is obviously items with an app-defined size,
not the default where text items adapt to their contents.
Results:
0.065 msecs to create a QGraphicsTextItem with some text (layouted)
0.036 msecs to set everything up in a QGraphicsTextItem with 0 width
QTextEdit was abusing the width 0 to mean "no wrap, width comes from
contents", but since the value -1 means that already in QTextDocument,
QTextEdit now uses a width of -1 for that meaning.
Change-Id: I67ad59c305e5dd34830886e4e6c56dde03c93668
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Use the QLatin1String overload of QString::append().
Pick-to: 6.3
Change-Id: Id8ddfd72199cfb627c2d6648ce3011979f92094e
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Keep it DRY. Just let the compiler figure out the size.
Pick-to: 6.3 6.2
Change-Id: I2bf1c44d4e2060a9398700d16ab98d67c849814c
Reviewed-by: Laszlo Papp <lpapp@kde.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Only qcoreapplication_p.h uses a QReadWriteLock (tanslationMutex), so
include it only there.
This include has caused QAtomicPointer<QReadWriteLockPrivate> to
percolate to the top-5 of template instantiation hogs when compiling
QtWidgets, as reported by Clang -ftime-trace.
This patch improves that, because qcoreapplication_p.h isn't part of
the PCH and is much less frequently included into QtWidgets TUs (68
instead of 142 times).
Pick-to: 6.3 6.2
Change-Id: I2518040d83a04e7cef0645d7f4bf641fb50f49ff
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Manipulating owning containers is costly at runtime, but also at
compile-time, so don't do it in inline functions, do it
out-of-line.
Removes the top-2 entries[1] in a Clang -ftime-trace QtWidget build.
[1] at the time of writing of this patch
Change-Id: I5da5030788b0d976a0ab7875f87ba3b51dbb1231
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The writerCount member variable is only ever accessed under
Private::mutex protection, so we need to lock the mutex here, too.
Pick-to: 6.3 6.2 5.15
Change-Id: I1717e5282eed2ecc14ccdc5090b9fc41174cccc8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
[ChangeLog][QtCore][QString] Fixed result truncation mod INT_MAX in
fromStdSstring(), fromStdU16string(), fromStdU32string(), and
fromStdWstring().
[ChangeLog][QtCore][QAnyStringView] Fixed result truncation mod
INT_MAX in asLatin1StringView().
[ChangeLog][QtCore][QUtf8StringView] Fixed result truncation mod
INT_MAX in toString().
Pick-to: 6.3 6.2
Change-Id: I944c561018a6ab4581ad6b1d71ec1ba1accba4a4
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
... out of QThreadData.
No-one except two functions in qobject.cpp uses the object, and its
creation is constinit, so there's no advantage to expose it to the
world as a QThreadData member.
Remove it from QThreadData, move the class' definition to the unnamed
namespace in qobject.cpp, ensure constinit'ability by letting the
language zero out the members (as opposed to an STL algorithm call),
declare it constinit thread_local static, and adapt the two users
(basically, removing the retrieval of QThreadData::current()).
Almost no effect on Clang, but saves ~400 bytes on optimized GCC 11.2
Linux AMD64 C++20 builds.
Change-Id: I22432d4ec5eb4ab59920656409b21768983fb4db
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Originally, this argument was meant for internal use only. But it
turned out useful for user projects too.
Pick-to: 6.2 6.3
Change-Id: I58af5ca8303fa9166d41a0538a60913d1fba9784
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
If QT_GENERATE_WRAPPER_SCRIPTS_FOR_ALL_HOSTS is ON then we generate
Windows scripts on Unix and vice versa. We always used the host
platforms line endings for generating the scripts. This leads to
Windows line endings in Unix scripts and vice versa.
Explicitly specify the line endings style when generating wrapper
scripts.
Fixes: QTBUG-102747
Pick-to: 6.2 6.3
Change-Id: I1603add46f276a5d91bbf0f103a261cdd84c343b
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
When drawing widgets with graphics effect using repaint manager,
if we do not combine the effect region some child widgets are not
affected by the graphics effect since they are not included in the dirty
region. However always using region bounds breaks the referenced bug use case.
Fixes: QTBUG-102374
Pick-to: 6.3 6.2 5.15
Change-Id: Iaf4eaba34db863500a0600b344e5062f2b36b9bf
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Add the Q_OS_WIN32 guard for the include of QWinEventNotifier.
Change-Id: I7824b2ee236a370c83fd85a2f594a39cf36b36e6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Need to collect all header files that belong to the module to make
them visible in CMake source tree. In future commit they will be used
to consider what headers need to be installed for each platform.
Amends 28b4bd7dbb
Change-Id: I7ab4063350254324869d5c2d0ba0f7e336da96fb
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This broke compilation of qDBusRegisterMetaType<std::vector<MyStruct>>
because std::vector<T> is in fact std::vector<T, std::allocator<T>>.
Pick-to: 6.2
Change-Id: I6a13f5f0476a3faa3a43da54d90d652b4bdd8186
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This allows us to specialize JNI type signature templates for e.g. the
context object, which in Java signatures is "android/content/Context".
Introduce a Q_DECLARE_JNI_TYPE macro that takes care of the plumbing.
The types declared this way live in the QtJniTypes namespace, and
transparently convert from and to jobject. Since jobject is a typedef
to _jobject* we cannot create a subclass. Use a "Object" superclass
that we can provide a QJniObject constructor for so that we don't
require the QJniObject declaration to be able to use the macro.
The APIs in the QNativeInterface namespace doesn't provide source or
binary compatibility guarantees, so we can change the return types.
Change-Id: I4cf9fa734ec9a5550b6fddeb14ef0ffd72663f29
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Since we know at compile time whether the return type is an object type,
we can use 'if constexpr' and auto return type in the call(Static)Method
and get(Static)Field functions to call the object-type methods.
This makes the object-methods conceptually obsolete, but don't declare
them as deprecated as long as they are still used in submodules to avoid
warning floods and build failures in -Werror configurations.
Change-Id: Ic3019ed990a9252eefcb02cdb355f8a6ed6bc2ff
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
QString::utf16() will detach from QStringLiterals in order to ensure
NUL-termination, we don't need the returned pointer NUL-terminated,
because we pass the length explicitly to the u16string ctor, so just
use data().
Task-number: QTBUG-98763
Pick-to: 6.3
Change-Id: If883901a41480f469162ff4ed5bef927a3bf060a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The function dereferences the load()ed pointer, so it needs an acquire
fence.
Pick-to: 6.3 6.2 5.15
Change-Id: Ib951de3f00851d915fec3392cdaba64f4a994300
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Turn off native indicators if we have custom rules, and if we drew
natively, don't draw custom indicators. This amends
ea0e0a8652 which turned off custom drawing
too aggressively, removing custom indicators also if no custom drop
down arrow (which is only relevant for drop down menu buttons) was set.
When then drawing the custom indicator, respect positioning rules in the
style sheet.
Extend baseline test.
Fixes: QTBUG-102866
Pick-to: 6.3 6.2
Change-Id: I5ca353f42e704ec3f6e57677c35118a9cb358b0b
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
We need an acquire fence before we delete the d-pointer. Otherwise,
the reads that the dtor performs (QReadWriteLockPrivate contains many
non-trivial data types such as std::mutex and QVLA), race against
writes performed in other threads. The qWarning() indicates that
QReadWriteLock can not rely on external synchronization to ensure
a happens-before relationship between reads in the dtor and said writes.
While an explicit fence just before the delete would suffice, the guard
return is an extremely unlikely error case, and if we ignore it, then
loadAcquire() is correct, so use that.
Pick-to: 6.3 6.2 5.15
Change-Id: I29773b665a7f864cd6b07a294da326e8b10399b5
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Removes a workaround that was added to force the Windows On-Screen
Keyboard to automatically appear on Windows tablets, since it is
no longer needed in currently supported Windows releases and was
interfering with text input using Input Method Editors.
Also remove logic for programmatically disabling the OSK, since
it is no longer needed here as well, and is already supported in
the UI Automation code elsewhere in the Windows QPA.
Pick-to: 6.2 6.3
Fixes: QTBUG-98003
Change-Id: I6c4781852a5e9f96330d54a24ee3893dcd43a28c
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The first attempt introduced in e9fd1c6aab had an unwanted
side-effect which makes the screen name non-unique and made the
screen manager think multiple different monitors with the
same name is one same monitor. This behavior is of course wrong
and unwanted but it was missed due to there's no unit test
for this property and I was not familiar enough with QPA
stuff. As a result the change was reverted in 851dc581cf.
However, the original change is still a good improvement,
so we bring it back, but fixed the side-effect mentioned above,
by letting the screen manager compare the device name instead
of monitor name.
If multiple monitors have the same name, a numeric suffix
will be appended to the monitor name to let people still be
able to distinguish between them.
Change-Id: Ic03105d0825abbb3d21cb51d1da3ab2cbb4ca913
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Replace the current license disclaimer in files by
a SPDX-License-Identifier.
Files that have to be modified by hand are modified.
License files are organized under LICENSES directory.
Task-number: QTBUG-67283
Change-Id: Id880c92784c40f3bbde861c0d93f58151c18b9f1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
When decoding CF_DIB data through the bmp handler we have to
do some assumptions on where the pixel data starts since there's
no file header to get the offset value from. We have to do this
because theres some optional data after the info header that
needs to be skipped over in some cases.
These optional color mask values are now also written when
putting a CF_DIB into the clipboard for maximum compatibility
with other apps on Windows.
This fixes the issue where pasted dibs would be offset by 3 pixels
on Windows.
Fixes: QTBUG-100351
Pick-to: 6.2 6.3
Change-Id: Icafaf82e0aa3476794b671c638455402a0d5206f
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
LocationChanged event unconditionally triggered invalidateVirtualViewId
call. That call results in TYPE_WINDOW_CONTENT_CHANGED Android event,
which causes a lot of background processing.
That is not correct, because LocationChanged event is generated by
every accessible element, not only the one that has A11Y focus.
This patch checks event->uniqueId(), and processes only events that
come from the focused accessible element.
Done-with: Mike Achtelik <mike.achtelik@gmail.com>
Task-number: QTBUG-102594
Pick-to: 6.3 6.2 5.15
Change-Id: I6b941733c9d215fed5ee5a7aeeb5be234add9ebe
Reviewed-by: Mike Achtelik <mike.achtelik@gmail.com>
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
The item widgets all have an API flaw: getters and setters for the text
alignment deal with int instead of Qt::Alignment.
Deprecate the existing setters and introduce others taking
Qt::Alignment. Store the alignment directly into the item widget (now
that views know how to handle it).
We can't change the getters without cluttering the API, so make that a
Qt 7 change. Users can prepare by forcibly casting the return value to
Qt::Alignment; this is going to work in Qt 6 and 7.
While at it: streamline the handling of Qt::CheckStateRole as well,
avoiding to rely on a pointless Qt::CheckState to int conversion
through QVariant (the setter stores a Qt::CheckState, but the getter
retrieves an int and converts it to a Qt::CheckState).
Task-number: QTBUG-75172
Change-Id: I9f29e818e93cb2dc1d8e042bc320162c2f692112
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
It seems that the size of the corresponding array in the code has a one-off
error. This is now fixed.
The end of the array passed to lower_bound is wrong because the
hard-coded size is size + 1. However, the end is array + size. This is
what lower_bound expects from an array. Or any other algorithm for that
matter expecting the end of a container as an argument.
This can cause issues with something like lower_bound because a
potential "empty" fill is not sorted as lower_bound would expect the
data structure.
It could have been fixed by decreasing the size by one, however it is a more
future-proof solution to avoid hard-coding the size and just use
std::size(array) instead.
Pick-to: 5.15 6.2 6.3
Change-Id: I1d25a5b1a80a3b2634b229e0718108ad5e7808a0
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Also add the beginnings of an autotest for QTextCursor::insertHtml(),
for comparison purposes.
We can see that the block to be inserted is merged with an existing
block by default rather than being inserted as a new one, with both HTML and
Markdown insertions. So now we test for leading and trailing newlines
in the markdown to be inserted, to determine whether we need a new block
into which to insert, and to "hit enter" at the end of the insertion.
QSKIP the toMarkdown() comparisons if GeneralFont is mono. This happens
on Boot2Qt systems in CI.
Task-number: QTBUG-76105
Task-number: QTBUG-94462
Task-number: QTBUG-100515
Task-number: QTBUG-103484
Change-Id: I51a05c6a7cd0be4f2817f4a922f45fa663982293
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
This allows color space conversions that produces values outside the
0.0->1.0 range, which is one of the intended functions of the floating
point image formats.
Change-Id: I63b37b0f6934d4382edafb4709486c785a637c67
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
A class that reimplements a virtual function isn't using the override
keyword, which generates a warning during compilation.
Pick-to: 6.3
Change-Id: Ic39ea24993e031f95ac9e61c3285d3be05fe6c34
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This allows the compiler to deduce the template arguments based on the
provided method parameters, which we can then pass to the methodSignature
and fieldSignature helpers to generate the signature string completely at
compile time.
Since we can't partially specialize template member functions, replace
the specializations for void methods with compile-time-if branches in
the general templates.
This variadic template now prevents implicit conversion from the
LiteralStorage types to const char* signatures, so catch the case where
such a type ends up in the parameter list.
Due to overload resolution rules for constructors, we need to explicitly
disable the constructor if any of the arguments is a string literal type,
as we have to keep the old C-style variadic function working for such
calls.
Add variations that use the variadic templates to the unit tests.
Change-Id: I8734664b38bae932369462330a9a03302254c33c
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
The QList<int> range ctor and other QList<int> methods have percolated
up near the top of the list of Clang -ftime-trace most expensive
template instantiations in PCH libQt6Gui.so builds:
**** Templates that took longest to instantiate:
[...]
5138 ms: QList<int>::QList<const int *, true> (256 times, avg 20 ms)
4327 ms: QtPrivate::QCommonArrayOps<int>::appendIteratorRange<const int *> (256 times, avg 16 ms)
The code in 6.3 is already sufficiently different for this patch to
not be applicable there.
Task-number: QTBUG-97601
Change-Id: I4420c8c90e472ecfd679b414cc4334d2ab55cce3
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
They already nicely factored this error-only code from the normal path
of execution. All that was missing was for them to be marked as COLD,
so the compiler moves them out of the way even further, and optimizes
them for size, not speed.
TEXT size savings: ~400b on GCC 11.2 and ~500b on Clang 10 optimized
C++20 AMD64 Linux builds.
Pick-to: 6.3 6.2 5.15
Change-Id: I64a4123645855b4e34fbb0bc3304d144d7191a0d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
\since will mark the whole enum as new in 6.2
Amends 673ed80fa4
Pick-to: 6.2 6.3
Change-Id: I88bd73d3bf8b0da355e3200b7a9b67fc141a2b93
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
Emscripten no longer prints an error for each getProcAddress
lookup failure and this code can be removed.
Change-Id: I082f420772677196c8eb82c49546825c750377ae
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: David Skoland <david.skoland@qt.io>
This is the correct mapping, and as a bonus the Chrome browser provides
a stack trace on all output from console.warn().
Change-Id: I4a1b95475679d6b54a0690f51c23683514fe7985
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: David Skoland <david.skoland@qt.io>
The fsevents implementation of QFileSystemWatcher on Mac OS X tries to
remove trailing slashes from the paths when calling addPaths(). If the
user tries to watch "/", the path is changed to "".
The fix checks whether the path is longer than 1 before removing
trailing slashes.
Change-Id: Iafb10e449c4f3bd600b02edbe7c549911db05048
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This avoids repeated re-instantiations of unique_ptr with local
deleters, removing that instantiation from the top of the list in
Clang -ftime-trace QtWidgets builds:
**** Templates that took longest to instantiate:
2627 ms: std::__1::unique_ptr<void, free_deleter> (835 times, avg 3 ms)
Amends e297e80fd0.
Pick-to: 6.3
Task-number: QTBUG-97601
Task-number: QTBUG-99039
Change-Id: I1281f6cf9248a3796d9dfdc653f19f5a67dc3bda
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Remove .as_handle() call, which was left over from the previous
EM_ASM-based code; see 0ec75f4b9.
Change-Id: I220304074f4d17e706726dab086c0330edc9ef25
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
