We've been requiring C++17 since Qt 6.0, and our qAsConst use finally
starts to bother us (QTBUG-99313), so time to port away from it
now.
Since qAsConst has exactly the same semantics as std::as_const (down
to rvalue treatment, constexpr'ness and noexcept'ness), there's really
nothing more to it than a global search-and-replace, with manual
unstaging of the actual definition and documentation in dist/,
src/corelib/doc/ and src/corelib/global/.
Task-number: QTBUG-99313
Change-Id: I4c7114444a325ad4e62d0fcbfd347d2bbfb21541
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
I don't know when conversion through sequential and associative
iteratables was added, probably some time in the 5.x. QString and
QByteArray got conversions to sequential iteratables for Qt 6.1 with
commit c9a1102269. Since that was
intentional, I'm just documenting reality.
Fixes: QTBUG-107246
Pick-to: 6.2 6.3 6.4
Change-Id: Id8d5e3999fe94b03acc1fffd171b863b1a0ead68
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
- We don't support HPUX any more, so no need to talk about it
- Clarify that it's the QSharedMemory destructor that releases the
resources on Unix systems
- Explain the System V and POSIX backends and how to detect them
- Add a note about Android not being supported.
- Add a section about using QFile for shared memory
Change-Id: I413ea647c2a5453b8307fffd17174c8083416529
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
If an error is returned from qt_safe_poll inside the event dispatcher
the error is currently printed and ignored.
For most cases this behavior seems to be fine, but when used in critical
systems e.g. automotive or medical, a error might indicate a more severe
problem and the application should be stopped instead.
The system can then decide itself what to do e.g. restarting the
application using a watchdog.
Change-Id: Iaf5abb20bb3941eaeff19d14e41c395c88fa088d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
So we can have Q_ENUM.
Change-Id: If4c23ea3719947d790d4fffd17152a37d0fe551b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The whole point of QPrivateSignal is to forbid anyone but the class
itself from emitting a certain signal. This is a "workaround" introduced
in Qt 5.0, due to the fact that the PMF syntax for connections requires
users to take the address of a signal, and that is only possible if the
signal itself is public. (In fact, signals were protected in Qt 4.)
The Q_OBJECT macro defines the private QPrivateSignal class. A QObject
subclass that wants to "protect" its signal emissions can declare a
signal carrying an argument of type QPrivateSignal:
signals:
void aSignal(int, QPrivateSignal);
If the class itself wants to emit the signal, it can do so:
emit aSignal(42, QPrivateSignal());
But if "someone else" wants to, they can't use the QPrivateSignal type
because it's private.
emit obj.aSignal(42, SomeClass::QPrivateSignal()); // ERROR
Here's a hair in this soup: list initialization. If a braced-init-list
is used, [over.ics.list] simply *ignores* access control. This allows an
"untyped" initializer-list to match QPrivateSignal and perform aggregate
initialization:
emit obj.aSignal(42, {}); // works!
This kind of defeats the whole purpose. Therefore: make QPrivateSignal
not an aggregate and give it an explicit default constructor, disabling
copy-list-initialization for it. This means that using `{}` will fail to
compile (class is no longer an aggregate, a constructor must be
selected, and copy-list-initialization will not select an explicit
constructor).
This isn't a complete fix by any means. There's always the possibility
of using enough template magic to extract QPrivateSignal's type (e.g. as
a local alias) and then create an object of that type; but that sounds
extremely unlikely to be happening "by accident" (whilst it's super-easy
to just type {} as the argument and not realize that you were not
supposed to do so).
[ChangeLog][QtCore][Potentially Source-Incompatible Changes] It is no
longer possible to use `{}` to construct a QPrivateSignal object
(specifically, QPrivateSignal's default constructor is now explicit).
This means that emitting a signal that carries a QPrivateSignal argument
(i.e. a "private signal") cannot any longer be done by using something
like `emit aSignal({})`; instead, such usages must be ported to
`emit aSignal(QPrivateSignal());` or equivalent.
Change-Id: Iac379aee3a8adca5a91d5db906a61bfcd0abc89f
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The typeSignature for a type T[] is always "[" + typeSignature<t>, so we
can just implicitly support arrays of any known type. To prevent support
for multi-dimensional arrays, make sure that the underlying type is not
also an array.
By adding a QJniTypes::isArrayType in addition (that is true for any
type with a signature starting with '['), methods like
QJniObject::callMethod could then return a special QJniArray type that
provides array-specific functionality.
As a drive-by, and since all lines need to be touched to add braces,
replace std::is_same<>::value with std::is_same_v.
Change-Id: Iccadf03cfceb8544381a8f635bb54baeddf46c99
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
I don't know which platforms qpoll.cpp is still used and if in those
there's even a way to increase the file descriptor limit above
FD_SETSIZE's. But this is an easy change and protects against buffer
overruns.
Pick-to: 6.4
Change-Id: I810d70e579eb4e2c8e45fffd1718ca1aac8e6bef
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
None of these users require C++20 constexpr or C++23 noexcept, the
only remaining difference between std::exchange and qExchange.
This leaves a single qExchange() user, in QScopedValueRollback, that
requires the constexpr version, only available from C++20, and thus
remains unported.
Task-number: QTBUG-99313
Change-Id: Iea46f6ed61d6bd8a5b2fd9d9ec4d70c980b443a2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead of using the overly-generic qSwap() monster, use
- qt_ptr_swap() for swapping raw pointers
- member-swap for swapping smart pointers
- std::swap() for swapping scalars
In QtCore, this has proven to give a nice reduction in compile time
for Qt users, cf. b1b0c2970e.
Pick-to: 6.4 6.3 6.2
Task-number: QTBUG-97601
Change-Id: Iad8e6c11ebcc3ff822479c36f5faff88992b1165
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The QSharedMemory backend build system machinery does not currently
support multiple backends, so the choice has to be made at configure
time.
Pick-to: 6.4 6.2
Fixes: QTBUG-106910
Change-Id: I4b814ca1c131a2860467e96cc5a6dd7cd03fc8b7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A recent change fixed headerscheck complaining about missing std::min.
But <limits> was also missing, so included that, too. Added a comment
on <algorithm>, because I, too, thought std::min was in <utility>,
instead.
Change-Id: Ib22e78349c79673f7013a2ed26cd64b06e159ed0
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
std::min is defined in header <algorithm>. It seems that the
latest macOS/iOS clang fails without it.
Change-Id: I3d1cfad27244580ea975db870d993cce394b4b2f
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
qExchange is one of the few remaining functionalities that have not been
moved out of qglobal. Given that std::exchange exists in the standard, we
can simply move to it everywhere...
...if it weren't for the fact that std::exchange is only constexpr in
C++20, and only has its noexceptness specified in (most likely) C++23.
Still, we want to move to the existing std functionality where
possible, to allow the removal of qglobal includes in lieu of something
more fine-grained in the future.
So leave any constexpr calls[1] alone for now (and observe that none of
our current usages cares about the conditional noexceptness), but
replace everything else.
[1] QScopedValueRollback' ctor and QExplicitlySharedDataPointerV2::take
Task-number: QTBUG-99313
Change-Id: I599cb9846cf319c7ffd3457130938347a75aad25
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
To indicate success of a conversion, the public API has previously only
supported registering member functions of the form To (From::*)(bool *).
When adding custom converters for types that cannot be modified, this is
usually not a possibility.
As an alternative, this patch adds support for std::optional in the
UnaryFunction overload of QMetaType::registerConverter. If the returned
optional has no value, the conversion is considered failed.
Task-number: QTBUG-92902
Change-Id: Ibac52d2cb9b5a2457081b4bebb0def1f03e3c55d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
When it was added, along with the type it describes, it wasn't added
to the QDoc-only fake version of Type's declaration, or to the \enum's
list of \value entries. There being no clear reason for those
omissions, I'm presuming to guess they were just an oversight.
This amends commit c7ce1bc05c which was
included in Qt 6.0.0.
Pick-to: 6.2 6.4
Change-Id: I1bf9fe0f0a31219a888666550a197e479d8eadc1
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The real Type has several "administrative" members beside LastCoreType
and LastGuiType, but only these two appear in the qdoc-only fake
version of Type, only to be \omitvalue'd out in the docs. Save the
perpetual "need" to keep updating them with each new addition (they
weren't even in sync with the real versions anyway) by removing them.
Change-Id: If7c5da87655a2da5c7f29f394c9dd9921ff0c1a7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The new function has an advantage of not requring EM_JS.
Change-Id: Ib9ad0e6b59cfe2e6864697a14b5cfdb39f62af2d
Reviewed-by: David Skoland <david.skoland@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
...by making QMetaTypeId2 contain all expected members.
While we don't allow QMetaType from non-const references, we want to
get to the static_assert telling us as much, instead of running into
more or less incomprehensible error messages in the depths of
qmetatype.h.
Pick-to: 6.4
Task-number: QTBUG-106672
Change-Id: Ica9b13fee95eda97cafab2cccdc5249dfc3dcdf2
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Amends 2d0c31e7d9. We were using
MetaTypeDecay in qTryMetaTypeInterfaceForType; but that is not used by
moc when complete types are enforced. Change qt_metaTypeArray to also
use qTryMetaTypeInterfaceForType, so that the code path for "force
complete types"[0] and the normal one do not diverge.
[0] Most easily enabled by using one of the QML type registration
macros.
Fixes: QTBUG-106672
Pick-to: 6.4 6.4.0
Change-Id: I9bf14873d1d0c4127a676643f7e8eb77f6e42dc8
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
This was only needed for bootstrapped builds, but QSharedMemory is no
longer part of it since commit 75082c9f20.
Change-Id: If4c23ea3719947d790d4fffd171522c0d5f9aafb
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
__has_include(<chrono>) is always true, because C++11 chrono include
is required since 6.0.
Pick-to: 6.4 6.3 6.2
Change-Id: I50cb92571bf4f1f86e2f3f2b5f486dd3c3f30f4a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit 71b4d4f150 is likely the source of
the issue. It fixed a race on disconnection, but kept the call to
disconnectNotify() (which is user code) inside the locked section. My
analysis is that by construction the sender object can't be undergoing
concurrent deletion anyway at this point. All call sites
(QObject::disconnect or the signal-slot activations but before the slot
is activated) imply that the user code that reached here cannot itself
be racing the deletion.
There may be one race condition left: if the same signal was connected
earlier to a slot via queued connection and that slot deletes the sender
asynchronously. A synchronous deletion is handled by doActivate(), so
the single-shot connection is never activated in the first place, but an
asynchronous deletion could race past that check and delete the sender
while QObjectPrivate::removeConnection is running. However, I'd call
this a mistake in user code.
[ChangeLog][QtCore][QObject] Fixed a regression from 6.3 that caused
QObject::isSignalConnected() to deadlock if called from inside
disconnectNotify().
Fixes: QTBUG-106025
Pick-to: 5.15 6.2 6.3 6.4
Change-Id: Ic6547f8247454b47baa8fffd170fe0bdb62cfcaf
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
Refactor QWinRegistryKey to make it more modern and more
usable from outside.
Adjust the test for QWinRegistryKey to test the new functions,
merged with the original test.
Will port raw registry accessing code in QtBase to use this
class in follow-up commits. This change is the first step.
The long term goal is to port QSettings registry code to
this class instead of using raw Win32 APIs, however, there's
much more registry code in QSettings and migrate them to this
class needs a large refactor, so jsut leave it for now. Will
fix it in some future commit.
Change-Id: Iada2adb41b4d58e1b658ff6870a4b31ace479d43
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This overload was only needed to avoid breaking compilation of
qtdeclarative. After declarative had been migrated to the new API, it
can be safely removed.
Pick-to: 6.4 6.3 6.2
Change-Id: Ib540b924178b6a69e3bfd89b7c324fbba79c75ff
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Using [NSAutoreleasePool showPools] to debug auto release pools
should now show the call site that allocated QMacAutoReleasePool,
instead of the QMacAutoReleasePool constructor, i.e.:
################ POOL 0x7f844c80c050
0x600003644190 ^-- allocated in function: QCoreApplicationPrivate::init()
0x600003a41080 __NSArrayM
0x60000345cca0 __NSCFString
0x600003a410e0 NSPathStore2
0x600002141680 NSPathStore2
0x6000036441b0 __NSSingleObjectArrayI
Change-Id: I52d22503c1d3de5a9dbae9939569d0836cc1a840
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Not clearing the target time makes the subsequently created timers not
get their time ticks, as no actual timeout will be registered as the
dispatcher wrongly assumes there will be a tick fired, which there isn't.
Since there are no updates, the WASM compositor enters an infinite
window update loop where it requests an update on a window, and the
window requests an update on it, getting the update request back due to
an animation running forever.
Pick-to: 6.4 6.4.0
Fixes: QTBUG-105347
Fixes: QTBUG-102004
Fixes: QTBUG-104518
Fixes: QTBUG-106153
Change-Id: I14b8dd08df81852e28e8527545c8530e0656990d
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
[ChangeLog][QtCore][Meta Object] QMetaMethod::invoke(),
QMetaObject::invokeMethod(), and QMetaObject::newInstance() are no
longer limited to 10 arguments.
[ChangeLog][QtCore][Meta Object] The use of the Q_ARG macro is no longer
necessary when using QMetaMethod::invoke(), QMetaObject::invokeMethod(),
and QMetaObject::newInstance(). Types may now be passed
directly. Similarly, Q_RETURN_ARG can be replaced by the free function
qReturnArg().
[ChangeLog][Potentially Source-Incompatible Changes]
QMetaMethod::invoke(), QMetaObject::invokeMethod(), and
QMetaObject::newInstance() no longer support passing forward-declared
types in the argument list (it was possible to pass them by
const-ref). From Qt 6.5 onwards, all types in the argument list must be
fully defined.
[ChangeLog][Potentially Source-Incompatible Changes] Attempting to use
the internal types QArgument, QReturnArgument, QGenericArgument, or
QGenericReturnArgument directly with QMetaMethod::invoke(),
QMetaObject::invokeMethod() or QMetaObject::newInstance() may fail to
compile. Those are internal types that were never meant to be used
directly and will be removed in Qt 7. If really necessary, ensure all
arguments passed to those functions are directly using those classes and
not mixed with Q_ARG and Q_RETURN_ARG. Implementations of bindings to
other languages should contact the Qt development mailing list to
discuss options.
Change-Id: I36b24183fbd041179f2ffffd1701e3e8e47e0fba
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The doc of QMetaEnum::valueToKey() says to use ::valueToKeys() instead
for flag types.
Pick-to: 6.4
Change-Id: I48e5ba47324137f2ce2710f1d876e93e7c562e9f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This includes QDBusReply, QProperty, and QStringBuilder expressions.
The new constructor subsumes the QStringBuilder case without requiring
jumping though hoops to delay the definition of the ctor the way we
had to for the explicit QStringBuilder constructor, so remove the
explicit QStringBuilder one again.
[ChangeLog][QtCore][QAnyStringView] Can now be constructed from
anything that implicitly converts to either QString or QByteArray.
Fixes: QTBUG-105389
Change-Id: I0e584dd3e20d591381609a3329ef47cec7356ecc
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
And include qcore_mac_p.h where needed.
Task-number: QTBUG-99313
Change-Id: Idb1b005f1b5938e8cf329ae06ffaf0d249874db2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
[ChangeLog][QtCore][Meta Objects] The QMetaObject::invokeMethod() taking
a method name by string, QMetaObject::newInstance(), and
QMetaMethod::invoke() now support more than 10 arguments.
[ChangeLog][QtCore][Meta Objects] The use of the Q_ARG and Q_RETURN_ARG
macros is now optional with QMetaObject::invokeMethod(),
QMetaObject::newInstance(), and QMetaMethod::invoke(): the type name
will be obtained from the C++ type (the same as QMetaType). The function
qReturnArg() can be used in place of the Q_RETURN_ARG macro. The macros
are still useful in rare conditions where the type was typedef'ed from
its original name.
Change-Id: I36b24183fbd041179f2ffffd17022a2b48c7639b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Since Qt 6.0, QMetaType stores the name obtained from the C++ compiler,
which means we know a type like Qt::Alignment by its proper, full name
of QFlags<Qt::AlignmentFlag>. However, the meta object records only the
bare name of the enumeration, not the full flags.
Pick-to: 6.2 6.3 6.4
Fixes: QTBUG-105932
Fixes: QTBUG-96185
Change-Id: Ic6547f8247454b47baa8fffd170eab977e306377
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Type registration isn't necessary any more, unless you're trying to look
up a name back to ID or QMetaType object. It hasn't been since 6.0.
Drive-by update the example not to use deprecated API.
Drive-by remove the paragraph about requirements that aren't accurate
any more.
Pick-to: 6.4
Task-number: QTBUG-104858
Change-Id: Ic6547f8247454b47baa8fffd170eecb66719fa65
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Unifies our approach to calling CoInitializeEx and CoUninitialize,
removing a lot of boilerplate in the process, and also fixes a few
bugs where we would incorrectly balance our calls to CoInitializeEx
and CoUninitialize.
The optimistic approach of qfilesystemengine_win.cpp of calling
CoCreateInstance without initializing the COM library explicitly
has been removed, as calling CoInitializeEx should be a noop in
the situation where it's already been loaded.
Change-Id: I9e2ec101678c2ebb9946504b5e8034e58f1bb56a
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
But not before ::fromValue, since that registers automatically.
Take the opportunity to add the \relates to the two streaming operators
to QDataStream, so the documentation shows up *somewhere* at all.
Pick-to: 6.4
Task-number: QTBUG-105469
Change-Id: Ic6547f8247454b47baa8fffd170eb8ffc31feb5d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
For those, moc does know their type ID, and yet they may be still
forward-declared in the C++ side, so the meta object may have recorded a
null pointer in the metatype array.
Fixes: QTBUG-105832
Change-Id: Ic6547f8247454b47baa8fffd170dae07c0813dc7
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
For now just includes qfunctions_winrt_p.h
Once submodules have been moved over to qfunctions_win_p.h we can move
non-WinRT specific content from qfunctions_winrt_p.h to _win_p.h.
Change-Id: I467bb4991c67a8769b60b9cf9f26aa553c439b92
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Marc Mutz
Thiago Macieira
Dominik Holland
Giuseppe D'Angelo
Volker Hilsheimer
Tor Arne Vestbø
Kai Köhne
Fabian Kosmale
hjk
Arno Rehn
Edward Welbourne
Mikolaj Boc
Konrad Kujawa
Mårten Nordheim
Yuhang Zhao
Eirik Aavitsland
Mike Trahearn
Sona Kurazyan