The jbyte type is a signed char, which also promotes to int in variadic
argument functions.
Extend the test case to make sure that we don't get any warnings for
the most relevant parameter types.
Change-Id: I7811e1eebdbc989ab5989eb1a2c502acd0540bc7
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Now that QtJniTypes::Objects are no longer primitive types that are the
same as a jobject, using those types in registered native functions
breaks. JNI will call those function with a jobject on the function
pointer, and lacking any type safety, the call to the registered
function will proceed with a wrong type of object on the stack.
To fix that, register the native function via a proxy that is a variadic
argument function, and unpack the variadic arguments into a list of
typed arguments, using the types we know the user-code function wants.
Then call the function with a tuple of those types using std::apply,
which gives us type safety and implicit conversion for free.
Add a test that exercises this.
Change-Id: I9f980e55d3d13f8fc16c410dc0d17dbdc200cb47
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Instead of having a type that doesn't behave like a QJniObject, which
includes not holding a proper reference on the Java object, make the
QtJniTypes::Object type a QJniObject subclass that can be specialized
via CRTP to provide type-specific constructor and static functions.
QJniObject doesn't have a virtual destructor, but we subclass it only to
add a typed interface, without adding any additional data members.
Add versions of the static functions from QJniObjects to the
QtJniTypes::Object so that they can be called without explicitly
specifying the type or class name. This includes a constructor and named
constructors.
Constructing such objects means constructing a Java object of the class
the object type represents, as per the Q_DECLARE_JNI_CLASS declaration.
This is not without ambiguity, as constructing a type with a jobject
parameter can mean that a type wrapping an existing jobject should be
created, or that a Java object should be created with the provided
jobject as the parameter to the constructor (e.g. a copy constructor).
This ambiguity is for now inevitable; we need to be able to implicitly
convert jobject to such types. However, named constructors are provided
so that client code can avoid the ambiguity.
To prevent unnecessary default constructed QJniObjects that are then
replaced immediately with a properly constructed object, add a protected
QJniObject constructor that creates an uninitialized object (e.g. with
the d-pointer being nullptr), which we can then assign the constructed
jobject to using the available assignment operator. Add the special
copy and move constructor and assignment operators as explicit members
for clarity, even though the can all be defaulted.
Such QJniObject subclasses can then be transparently passed as arguments
into JNI call functions that expect a jobject representation, with the
QtJniTypes::Traits specialization from the type declaration providing the
correct signature.
QJniObject's API includes a lot of legacy overloads: with variadic
arguments, a explicit signature string, and jclass/jmethodID parameters
that are completely unused within Qt itself. In addition the explicit
"Object" member functions to explicitly call the version that returns a
jobject (and then a QJniObject). All this call-side complexity is taken
care of by the compile-time signature generation, implicit class type,
and template argument deduction. Overloads taking a jclass or jmethod
are not used anywhere in Qt, which is perhaps an indicator that they,
while nice to have, are too hard to use even for ourselves.
For the modern QtJniTypes class instantiations, remove all the overhead
and reduce the API to the small set of functions that are used all over
the place, and that don't require an explicit signature, or class/method
lookup.
This is a source incompatible change, as now QJniTypes::Object is no
longer a primitive type, and no longer binary equivalent to jobject.
However, this is acceptable as the API has so far been undocumented,
and is only used internally in Qt (and changes to adapt are largely
already merged).
Change-Id: I6d14c09c8165652095f30511f04dc17217245bf5
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
If we construct the QJniObject from a jobject, then we know the jclass,
but not the class's name. If className is called while the stored name
is empty, get the name of the jclass and updated the stored value.
Change-Id: Ic3332a6da2dac1eb6842f90da1b9264398a43155
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
currentCompatProperty should point to the compat property that's
currently being evaluated. As soon as we start evaluating a new compat
property, it's invalid by definition. Temporarily disable it then.
Pick-to: 6.6 6.5 6.2
Fixes: QTBUG-109465
Change-Id: I7baba9350ebf488370a63a71f0f8dbd7516bf578
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Fail the test when an unexpected warning message about a field, class,
or method not being found is generated from a JNI exception. Fix the
failure when trying to look-up a boolean field that is no longer
available, and ignore the one expected message from looking up an
unavailable class explicitly.
Change-Id: Ic5e4c003c64272f06a6d4da028e232abba75c4e4
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Support this both for parameters and return types, in all template
functions for calling methods or accessing fields.
To manage the life-time of the temporary objects, create a local stack
frame in the JVM around the calling function. Popping that implicilty
releases all local refernces, so that we don't have to worry about
doing so explicilty. Adding a local reference to the temporary jstring
objects is then enough for the object to live as long as it's needed.
The LocalFrame RAII-like type needs a QJniEnvironment to push and pop
the frame in the JVM, so we don't need a local QJniEnvironment anymore.
Reduce code duplication a bit as a drive-by, and add test coverage.
Change-Id: I801a1b006eea5af02f57d8bb7cca089508dadd1d
Reviewed-by: Tinja Paavoseppä <tinja.paavoseppa@qt.io>
Add compile-time testing to make sure that we can declare a JNI
class String that maps to java/lang/String.
Change-Id: I2b68b2b46112e56b279f3fcddc3d71847a005924
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Tinja Paavoseppä <tinja.paavoseppa@qt.io>
d026fad3d9 added converting constructors
for QPointer. This however made converting _assignments_ ambiguous,
introducing a regression for users coming from Qt < 6.6.
This code:
QPointer<Base> base;
QPointer<Derived> derived;
base = derived;
used to convert `derived` to `Derived *` (using the implicit conversion
operator from `QPointer<Derived>` to `Derived *`), and then the
assignment operator for `QPointer<Base>` that took a `Base *`.
The introduction of the conversion constructor in 6.6 makes it possible
to convert `QPointer<Derived>` to `QPointer<Base>`, and then fall back
to the compiler-generated assignment operator for `QPointer<Base>`.
The result is that the code above is now ambiguous and stops compiling.
Fix this by adding a converting assignment operator for QPointer.
I'm only adding the const-lvalue overload because the implementation
requires going through the private QWeakPointer::assign helper. We
cannot copy-assign or move-assign the inner QWeakPointer, as those
assignments require lock()ing the QWeakPointer and that's not possible
on a QObject-tracking QWeakPointer (but cf. QTBUG-117483).
Assigning from a rvalue QPointer would mean calling assign() on
the internal QWeakPointer _and_ clear the incoming QPointer,
and that's strictly worse than the lvalue overload (where we just call
assign()).
Change-Id: I33fb2a22b3d5110284d78e3d7c6cc79a5b73b67b
Pick-to: 6.6 6.6.0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Android APIs use integer constants like enum values, so they are mapped
to one of the integeral types (jint, jshort, jlong etc) on the C++ side.
Enable C++ code to declare an equivalent enum (scoped or unscoped), and
to use that enum as a type in JNI calls by treating it as the underlying
type in the signature() generator.
Add a helper type trait that maps enums to their underlying type and
other integral types to themselves (we can't use std::underlying_type_t
on a non-enum type). Add tests.
Note: Java Enums are special classes with fields; this change does not
add any special support for those.
Change-Id: Iec430a1553152dcf7a24209aaebbeceb1c6e38a8
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Juha Vuolle <juha.vuolle@qt.io>
Equivalent to get/setStaticField.
Add a test, and tighten up the surrounding test code a bit.
Change-Id: Ic0993c5d6223f4de271cb01baf727459b5167f94
Reviewed-by: Tinja Paavoseppä <tinja.paavoseppa@qt.io>
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Template functions don't permit partial specialization, e.g. we cannot
specialize typeSignature() to return an array signature for any
std::vector or QList type. We need to do that for better array support,
so move those functions as static members into a template class, which
then can be specialized.
Since submodules are both calling and specializing typeSignature and
className as template functions, keep and use those until the porting is
complete.
Change-Id: I74ec957fc41f78046cd9d0f803d8cc9d1e56672b
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Tinja Paavoseppä <tinja.paavoseppa@qt.io>
The type lives in the QtJniTypes namespace, which is where types end up
that are declared through the Q_DECLARE_JNI_CLASS/TYPE macros. Having a
String type in that namespace prevents us from declaring the Java String
class as a QtJniTypes type, which is silly.
Perhaps this type becomes obsolete at some point with std::string being
a constexpr type in C++23, but until then we need it. It has no ABI, so
renaming it us safe.
Until submodules are ported, leave a compatibility alias String type,
which also prevents us from declaring a String JNI class in tests until
the alias is removed in a later commit.
Change-Id: I489a40a9b9e94e6495cf54548238438e9220d5c1
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Tinja Paavoseppä <tinja.paavoseppa@qt.io>
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
The QVariant(QMetaType) constructor is a major anti-pattern: unlike
*every* other QVariant's constructor, it doesn't build a QVariant
holding the QMetaType object, but a QVariant of the specified type.
Introduce a named constructor for this use case instead.
In principle, this should lead to a deprecation of the QMetaType
constructor... except that it's used everywhere, so I can't do it at
this time.
Drive-by, improve the documentation of the QVariant(QMetaType)
constructor (since it's basically c&p for the new fromMetaType
function).
[ChangeLog][QtCore][QVariant] Added the QVariant::fromMetaType named
constructor, that builds a QVariant of a given QMetaType.
Change-Id: I4a499526bd0fe98eed0c1a3e91bcfc21efa9e352
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
That we have two macros to declare a C++ type to represent a Java class
is confusing. The TYPE macro as of now allows us to declare array types,
but with QJniArray we won't need that anymore, and can just use Class[]
as the type instead. Changing that will be a follow-up commit; for now,
get rid of TYPE-usages to declare regular classes.
Change-Id: Iea0a9548772ca701148442412cf6ad567583213f
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Petri Virkkunen <petri.virkkunen@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
The bindable property tests were not using the QTestPrivate helpers, so
add a new test which uses them.
The new tests revealed a binding loop for the interval property.
Fix it in a usual way by explicitly removing the binding and using
{set}ValueBypassingBindings() in the setter.
Task-number: QTBUG-116346
Pick-to: 6.6 6.5
Change-Id: If94f57938da449a68e3527aead5ebd55ba410adb
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
The 6.5 versions of the overload not taking a context/receiver object
were constrained by requiring a functor to be free function or lambda.
207aae5560 removed that constraint, which
might be source incomaptible if wrapper functions in user code forward
the constraint using Expression SFINAE. Those wrappers would no longer
be removed from the overload set based on the same criteria as the
function they wrap.
We can't constrain the new functions based on the same predicate as
before, as after the simplification we have only one overload with, and
one without context object. But we can still remove overloads for
incompatible functors.
Add the respective scenario to the QPermission test as a compile-time
test.
Found during 6.6 header review.
Pick-to: 6.6
Change-Id: Id21391b4a6b78a29de2f8fa04374f4262e5fafa7
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
...to set m_o to nullptr and prevent an existing QSignalBlocker from
touching the QObject, which it was created for.
Add documentation and implement an autotest.
Change-Id: Ic18e80af5a57df1928f9d36aa0ab7ad79b6525fd
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
QWidget already handles this, but it might be useful for non-Widget
object hierarchies as well, such as in Qt Quick.
The flag is opt in, and as QWidget already handles these events by
itself (without checking any flags), we assert that we don't end up
in this code path, instead of enabling it for QWidget. The latter
would mean refactoring the QWidget code, with possible regressions.
Docs and header comments have been updated to reflect that this
event is not widget specific. (This is an issue with other events
as well, that are documented to say "widget", since they came
from a time when there was only QWidget, but nowadays apply to
e.g. QWindow as well. That's something for another fix though).
Change-Id: Ib71962131d6011c17dcce8c01bd8adcdaa58d798
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Due to the pre-existing overload of invokeMethod that accepts a void-pointer there is
concern that the new overloads would not be preferred.
Here we add a test for this to verify all supported platforms work.
Change-Id: Ie5ac7bf16643599006ac57e0145feb6aace3fa87
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This was missing for a while, and there is nothing fundamentally
missing for it to work.
With the more recent work around slot objects and invokeMethod in
general, it is a good time to add support for this.
In this patch, when connecting to a functor, it automatically deduces
the overload to call based on the arguments passed to invokeMethod.
Sharing code with QObject::connect could be done, but they have a
key difference that makes it harder:
With signal emissions we throw away trailing arguments that are not
used: i.e. `signal(int, int)` can be connected to `slot(int)` or
`slot()`. With invokeMethod that's not a thing. So we will need a way
to toggle that behavior during resolution.
[ChangeLog][QtCore][QMetaObject] Added support for passing parameters
to the overload of QMetaObject::invokeMethod that takes a functor. These
new overloads must have the return-value passed through qReturnArg().
Change-Id: If4fcbb75515b19e72fab80115c109efa37e6626e
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The was introduced with the rewrite of QMetaObject::invokeMethod() in
commit 0f76e55bc4, because we have an
optimization for zero timers to avoid creating a temporary
QSingleShotTimer object. The old implementation did attempt to normalize
the target slot name, but did so because it looked metamethods up using
QMetaObject::indexOfMethod:
int idx = meta->indexOfMethod(sig.constData());
if (idx < 0) {
QByteArray norm =
QMetaObject::normalizedSignature(sig.constData());
idx = meta->indexOfMethod(norm.constData());
}
The new implementation does not use this method so it didn't need to
attempt to normalize.
I am fixing this only in QTimer and not in QMetaObject::invokeMethodImpl
(even though it is trivial to do so) because I don't believe spaces in a
pure string to invokeMethod were ever expected to work:
QMetaObject::invokeMethod(obj, "slotName ", Qt::QueuedConnection);
The Q_ARG and Q_RETURN_ARG (for code not recompiled) still does
normalization inside QMetaType::fromName().
Fixes: QTBUG-116060
Pick-to: 6.5 6.6
Change-Id: I964c2b1e6b834feb9710fffd177cac60c83ef413
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Unix dispatcher is not used and - as such - redundant on WASM.
Change-Id: Ia8789ef783b06ce9cfba2ce9d67159db2355b594
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Observers should only be registered when _reading_ the property.
Otherwise we get binding loops.
Pick-to: 6.6 6.5
Change-Id: I974f6ea444fa7a5d333ed79eea6f34e3d757d169
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
The loop was iterating over a temporary container, so it couldn't have
changed it. Store the container in a const auto variable and use
ranged-for.
In files where Q_FOREACH isn't used any more, remove
"#undef QT_NO_FOREACH".
Task-number: QTBUG-115839
Change-Id: I402df5fa48f4287f3cc989ddae1524da43999049
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The density of Q_FOREACH uses in this and some other modules is still
extremely high, too high for anyone to tackle in a short amount of
time. Even if they're not concentrated in just a few TUs, we need to
make progress on a global QT_NO_FOREACH default, so grab the nettle
and stick to our strategy:
Mark the whole of Qt with QT_NO_FOREACH, to prevent new uses from
creeping in, and whitelist the affected TUs by #undef'ing
QT_NO_FOREACH locally, at the top of each file. For TUs that are part
of a larger executable, this requires these files to be compiled
separately, so add them to NO_PCH_SOURCES (which implies
NO_UNITY_BUILD_SOURCES, too).
In tst_qglobal.cpp and tst_qcollections.cpp change the comment on the
#undef QT_NO_FOREACH to indicate that these actually test the macro.
Task-number: QTBUG-115839
Change-Id: Iecc444eb7d43d7e4d037f6e155abe0e14a00a5d6
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
It has always returned dash-joined forms of the locale names, and
callers who need an underscore-joined form have been obliged to
replace('-', '_') before using them. Given that everything it adds to
the list comes from QLocaleId methods that accept a separator, it's
trivial to let it offer the same choice to its callers and save them
this hassle.
Amended code in QTranslater and QMimeType to save them that hassle.
[ChangeLog][CoreLib][QLocale] QLocale::uiLanguages() now lets the
caller choose what separator to use between the tags that make up each
locale-identifier in the list returned.
Change-Id: I91fcd0b988d9a64e0e9ad9e851f6cb8c1be8ae50
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Everyone must have this by now. This test was 1193 ms of CMake time.
Since this was a PUBLIC feature, I've left it around with a constant
condition.
Change-Id: Ifbf974a4d10745b099b1fffd177754538bbff245
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Commit ca54b741d6 used the internal
has_operator_equal (and commit 01d94760d8
copied that for has_operator_less_than) instead of using the recursive
expander that was being used here. That assumed that the contained type
in std::optional would always be the last final check, which is an
incorrect assumption.
Fixes: QTBUG-115646
Pick-to: 6.6 6.5
Change-Id: Ifbf974a4d10745b099b1fffd177702934bec27ff
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
There is no QT_CONFIG_thread.
Thanks to Alexey Edelev for spotting why I couldn't run the test.
Pick-to: 6.6 6.5
Change-Id: I11c99d9b1ff8fed67b118028b76fba8ee6db3c42
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
The test was relying on the fact that, having written 1 byte to both
writeEnd1 and writeEnd2 (and ensured those bytes were written with
waitForBytesWritten()), both read ends would be activated by the next
event loop. It turns out that this was an unreliable assumption, because
the processing of that 1 byte on the second socket may not have happened
yet. So firm up by waiting that both read ends are readable before even
creating the QSocketNotifiers we will read on.
I'm not entirely sure what this test is attempting to test. Its
documentation says it's testing a QAbstractSocket condition, but the
read ends aren't QAbstractSocket (this test should have been in
tst_QAbstractSocket if so). It may be testing the condition that caused
that QAbstractSocket behavior, but that wouldn't be a good test.
Drive-by remove redundant flush()-after-waitForBytesWritten() calls.
Fixes: QTBUG-115154
Pick-to: 6.5 6.6
Change-Id: I61b74deaf2514644a24efffd17708f8071f707ed
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
The test hasn't failed on Windows for a long time, but regularly fails
on macOS in CI, so replace the entry accordingly.
Task-number: QTBUG-115154
Pick-to: 6.6
Change-Id: Ib89d15cb9edafad5dd71f6e3f830d03aaeb16331
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Add the boilerplate standalone test prelude to each test, so that they
can be opened with an IDE without the qt-cmake-standalone-test script,
but directly with qt-cmake or cmake.
Boilerplate was added using the following scripts:
https://git.qt.io/alcroito/cmake_refactor
Manual adjustments were made where the code was inserted in the wrong
location.
Task-number: QTBUG-93020
Change-Id: I28b6d3815c5f43d2c33ea65764f6f3f8f129eaf3
Reviewed-by: Amir Masoud Abdol <amir.abdol@qt.io>
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
`QVariant::convert` may lead to crash or produce garbage data when
attempting to convert a gadget between a pointer type and a value type,
for example from a variant holding a QLocale gadget to a QLocale*
pointer and vice versa. Similarly, `QVariant::view` may crash under the
same conditions.
The reason is that conversion is implemented through copy construction
assuming that both source and target types are either both pointers or
both values. If converting from pointer to value type, the result is
crash during destruction of the QVariant. If converting from value to
pointer type, the result is a QVariant holding a pointer to garbage
data (and possibly crash if pointer is dereferenced).
Similarly, if attempting to convert a pointer to a QObject derived type
to its value type, the system crashes, with a slightly different failure
mode. During `QVariant::convert`, a temporary `QVariant` of the target
type is created. Since objects that can not be copy constructed are
invalid for `QVariant`, the temporary is left empty without constructing
the target value. Then, when attempting to convert from a pointer type
to a value type, the temporary's destructor is incorrectly called on the
owned object. Since the owned object is never constructed, this leads to
a crash.
The proposed fix is to return false from `QMetaType::view`,
`QMetaType::canView`, `QMetaType::convert`, and `QMetaType::canConvert`
if the target type is of different 'pointedness' than the source type.
After this fix, converting and viewing gadgets and QObjects behaves the
same way as primitive types and core types, which already returned false
when converting between value type and pointer type.
Fixes: QTBUG-114797
Pick-to: 6.5 6.6
Change-Id: If5ad764a60f2f3c912070198073b28999d995f17
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Under some circumstances, MSVC seems to complain about SlotArgumentCount
being unused
qobject.h(210): warning C4189: 'SlotArgumentCount': local variable is
initialized but not referenced
note: see reference to function template instantiation
'QMetaObject::Connection QObject::connect<void(__cdecl QAction::* )(bool),
main::<lambda_1>>(const QAction *,Func1,
const QtPrivate::ContextTypeForFunctor<main::<lambda_1>,void>::ContextType *,
Func2 &&,Qt::ConnectionType)' being compiled
This is nonsense, as SlotArgumentCount is used in the next line, to
construct the list of signal arguments, but the workaround to declare
the variable as [[maybe_unused]] is trivial.
Add a connect statement to the test case that creates such a connection.
This does not produce any warning with or without the attribute (and if
it did, the build would fail for CI configuratinos setting -Werror).
Pick-to: 6.6
Fixes: QTBUG-114781
Change-Id: I4ee6f7d57c2836ef3dd9741d037d48181af2cdec
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This is in line with how QMetaType handles QMetaTypeInterface*. You can
retrieve a const pointer to it.
Pick-to: 6.6
Task-number: QTBUG-113690
Change-Id: Iaf3c10603dc6049a5553987c90006807867abc0d
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
QObjectPrivate::getPropertyAdaptorSlotObject called
connectionsForSignal.
Calling this function is only safe after it has been ensured beforehand that the vector has size > signalIndex. As getPropertyAdaptorSlotObject
is not supposed to modify the vector, it does not resize the vector and it could consequently end up with an out-of-bounds read.
To avoid that issue, we instead first check if the vector can
potentially contain an entry for the signal. If not, we simply return
nullptr, and avoid the call to connectionsForSignal.
The issue and its fix can be verified by running the modified
tst_qproperty test with ASAN enabled. The test is modified in the
following way:
- We first create a signal connection to a dummy slot. Otherwise,
connections.loadRelaxed() would return a nullptr, and the problematic
code would never be reached.
- We add enough signals to ensure that the fooChanged signal will
actually be out of reach (which means >= 8 signals, as the initial
capacity of the vector is 8)
Running the test without ASAN will most likely not result in a failure,
as then the out-of-bounds read will simply read garbage, and the most
likely result is that the cast below will fail.
Pick-to: 6.6 6.5
Change-Id: I18a3c4f52769c2b6491a685abb84f6fcfb44e4d8
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
If we have a rvalue reference to an unshared QVariant, we can avoid
potentially expensive copies, and use move semantics instead.
[ChangeLog][QtCore][QVariant] Added rvalue QVariant overloads of qvariant_cast<T>() and QVariant::value<T>().
[ChangeLog][Potentially Source-Incompatible Changes][QVariant] It is no
longer possible to take the address of a specialization of
qvariant_cast; consider using a lambda function instead.
Change-Id: Ifc74991eadcc31387b755c45484224a3200bb0ba
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
QDeferredDeleteEvent has the loopLevel field, which is a sum of
scope and loop levels found at posting. In sendPostedEvents however,
it is impossible to only use this information to find delete events
posted before the outermost loop (which should be handled by any loop)
based solely on this information, as the scope level essentialy removes
the information on loop level.
Break the loopLevel in two, storing both loop and scope levels in
QDeferredDeleteEvent, so that we can check whether an event was posted
before the outermost event loop (for which we need to compare only the
loop level).
QDeferredDeleteEvent was also made private as it should - it is an
implementation detail that wasn't hidden properly.
Change-Id: I0a607a0bd3a2deb5024acad67f740dbf4338574c
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
Extract Method fromStdVariantImpl() to share the otherwise
more-or-less identical implementation between the two overloads.
Don't use a constrained template version of fromStdVariantImpl() as
fromStdVariant(), because the constraint would have to be very complex
to continue allowing subclasses of std::variant to be passed.
As a drive-by, mark the valueless_by_exception() path Q_UNLIKELY.
[ChangeLog][QtCore][QVariant] Added overload of fromStdVariant()
taking rvalue std::variant<>s.
Fixes: QTBUG-114134
Change-Id: Ia1c7ae93ab421e6689dc9f2d8d0c2295b23cbbf6
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
When passing an rvalue-reference to QVariant, there is no reason to make
a copy if the type is moveable. Moreover, we know that the pointer which
we construct from the object passed to fromValue non-null. We make use
of both facts by parametrizing custom_construct on
non-nullness and availability of a move-ctor, and then dispatching to
the suitable template.
We need to keep the const T& overload, as otherwise code which
explicitly specializes fromValue and passes a const lvalue to it would
stop to compile.
[ChangeLog][QtCore][QVariant] Added fromValue() overload taking rvalues.
Change-Id: I44fb757d516ef364fe7967bc103b3f98278b4919
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
[ChangeLog][QtCore][QVariant] Implemented in-place construction for
QVariant. The constructor taking std::in_place_type<Type> constructs
an object of type Type directly inside QVariant's storage, without any
further copy or move operations. QVariant::emplace() does the same
when replacing the content of an existing QVariant and tries to reuse
previously-allocated memory.
Fixes: QTBUG-112187
Change-Id: I16614ad701fa3bb583976ed2001bb312f119a51f
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
It's not needed, but makes the point for having a mutable lambda in the
first place.
Change-Id: I483862d6aee90bb62d4b5363c56a80bb05e14df7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This places the first through third parameters on the exact positions
that they will be used to perform the operations in the switch, saving
the compiler from generating a few instructions to move data around. All
ABIs Qt supports that pass any function parameters in registers at all
pass at least 4.
We keep the return type as void (instead of returning bool, for the
Compare case) so the compiler can apply tail-call optimizations for
those two typical cases.
PMF case: https://gcc.godbolt.org/z/9oP5boKfj
Function case: https://gcc.godbolt.org/z/e9vEzd5dj
Functor case: https://gcc.godbolt.org/z/s8Ejjra7P
Change-Id: I3e3bfef633af4130a03afffd175d3e3009c56323
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Amends 3bf5b5f894, after which free
functions and std::bind could no longer be used as callables in
QMetaMethod::invokeMethod.
For free functions to work we need to decay to function pointers when
choosing what type QtPrivate::Callable aliases.
And std::bind has operator() overloads and the return type cannot be
deduced. So simplify the definition of the ZeroArgFunctor - we know
the function prototype if we know the return type.
Add testcase for calling std::bind and free function, and remove the
now unneeded helpers for functor argument and return type deduction.
Change-Id: I54aac5cb6d660267e6b2f5ab05d583e8826cdf9a
Reviewed-by: Zoltan Gera <zoltan.gera@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The QWeakPointer conversion SMFs cannot actually be used for QObject
payloads, as, for unknown reasons (some comment about vtable this
author doesn't understand), conversion goes through QSharedPointer,
the creation of which throws the checkQObjectShared() warning and
yields a nullptr.
We need to continue to use the QWeakPointer(T*, bool) constructor the
QPointer(T*) ctor also uses.
It's high time we dissociated QPointer from QWeakPointer...
Amends 5f28d367d9.
Fixes: QTBUG-112464
Change-Id: I2f93843af3daf02323d77a4259eaa3745d8de3a8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We have quite a few Qt API that assumes this, so making this change
helps transitioning them to QDeadlineTimer.
[ChangeLog][Important Behavior Changes] QDeadlineTimer will now
interpret negative millisecond remaining times as "forever", instead of
only the value -1. This brings the API closer in line with other API
like QMutex. This change does not apply to the nanosecond counts in the
API, nor to the API based on std::chrono.
Change-Id: I6f518d59e63249ddbf43fffd175a3e5bead564ae
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Commit b498e1ae3a removed the last
distinction. And since there was no distinction, the code that was
previously under a conditional for CoarseTimer must work for precise
too.
Change-Id: I6f518d59e63249ddbf43fffd175a3eddbd41611a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
After the recent changes we only have a single implementation of
QSlotObjectBase, which can handle free functions, member functions,
functors, and lambdas. Rename it to callable, and explicitly hide
the static implementation function so that it doesn't become a symbol
of static libraries using Qt.
Also rename makeSlotObject to makeCallableObject, and polish coding
style and comments in the qobjectdefs_impl header a bit.
Change-Id: Id19107cedfe9c624f807cd8089beb80e9eb99f50
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
[ChangeLog][QtCore][QDebug] Added pretty formatting of C++ <chrono>
durations.
[ChangeLog][QtTest] Added pretty formatting of C++ <chrono> durations
for QCOMPARE expressions.
Change-Id: I3b169860d8bd41e9be6bfffd1757cc087ba957fa
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Add helper that allows us to determine the argument list and return type
of a functor. This triggers a compile time error if the functor has
operator()() overloads (we only support zero-argument call operators, but
there might be const/noexcept variations). Use that helper to declare a
ZeroArgFunctor type which also declares a ReturnType and Arguments alias.
Add a Callable alias that now combines FunctionPointer and ZeroArgFunctor
into a single type that we can then use to merge the specializations of
QMetaObject::invokeMethod.
[ChangeLog][Potentially source-incompatible changes] Using a functor
with several operator() overloads in QMetaObject::invokeMethod now causes
a compile time error. Qt would previously ignore const and noexcept
overloads and always call the mutable version on a copy of the functor.
Change-Id: I3eb62c1128014b729575540deab615469290daeb
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Amends 642f799fc6 to avoid unnecessary
copies in between the calls to the QFunctorSlotObject construcotr. We
can't use a univeral reference in the QFunctorSlotObject constructor
as the call is not deduced. So provide two overloads for lvalue and
rvalue references instead.
The compile check in the test now no longer fails as we delay the
storage until one level later, but that's acceptable.
Change-Id: Ide95b4a73c70f6f47698dd1e95138aa5d48ee95d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Remove duplicate test and outdated comment about move-only functors,
and include return value in move-only functor test.
Change-Id: I58dffe0ccf3ec12e7e05e2c9588303da4a7e75ff
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Amend 3c6e9dcc623c9d7281a81174bb3a696e030f30a7 by making sure that
we explicitly move move-only functors into the slot object in the
respective tests, and that failing to do so doesn't compile.
Also add test coverage for mutable lambdas, which work as they do
with connected functors: the connection stores a copy, and calls
don't modify the original functor.
Change-Id: I637e6f407133e2f8f72109b3fe5369a11d19da93
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This reverts commit 9958edba41, which
incorrectly tested a move-only functor without actually moving the
functor.
Change-Id: I3707f9f8e5055102f7edfb3e1cb9750978356dd7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QWeakPointer can do the same, so there's no reason to not allow it for
QPointer.
[ChangeLog][QtCore][QPointer] QPointer<T> can now be (move- and
copy-)constructed from QPointer<X>.
Fixes: QTBUG-112464
Change-Id: I77cf5d39974bf2b3ec849b4afc33e286e864821e
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
In Qt >= 6.1, < 6.5, a trivially constructible type would have the
NeedsDestruction flag set, but it's dtor pointer would have been null.
In Qt 6.5, the meaning of the NeedsDestruction flag was changed to be
more aligned with what the name suggests, and thus would only be set for
non-trivially destructible types. For QMetaType this was fine, but
QVariant has a check for acceptable metatypes which attempts to verify
whether a QMetaType is usable for QVariant. The check assumes the
semantics of Qt 6.5, and thus fails for metatypes created by older Qt
versions.
To fix this issue, we increment the QMetaType revision field, and only
check the metatype's destruction support if the revision is high enough.
In theory, that allows passing unsuitable metatypes from older Qt
versions to QVariant; however, such code would have been broken in prior
Qt releases already (which didn't attempt the check), and no code that
used to work in any released Qt version will break (as we simply skip a
check that was passing before).
Fixes: QTBUG-113227
Pick-to: 6.5
Change-Id: I12e02bd97d2c410ea1a36efb0ce2389f21d50a30
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Amends 207aae5560 to make it easy to
create human-friendly error messages. Since the functor-accepting member
functions are not removed from the API, the first compile error will be
that there is no suitable overload of the makeSlotObject helper, which.
With the assert helper, the first error message is easier to understand.
Change-Id: I4878ec35a44ddfa5dc9d9e358d81c3fd40389c0c
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Move-only functors must never be passed by value, so fix the
QFunctorSlotObject constructor accordingly.
This then requires adjustments to the various QMetaMethod::invokeMethod
overloads, as those must also perfectly forwad the functor type.
Enable the previously failing test case for move-only functors.
Change-Id: I9c544fd3ddbc5e1da3ca193236291a9f83d86211
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Amend 207aae5560, as code checker
complained that we std::move'd a potential lvalue. This warning was
valid if the public API did not accept the functor parameter by value.
Fix this by consistently std::forward'ing the parameters through the
call stack, and add a compile-time test. Writing that test revealed that
the helper API didn't work with free functions, so fix that as well. It
also revealed that QFunctorSlotObject couldn't work with a const
functor, which is also fixed by this change.
We cannot support move-only functors with that change, as it requires
a change to QFunctorSlotObject that breaks the QMetaObject test.
Change-Id: Iafd747baf4cb0213ecedb391ed46b4595388182b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This avoids constructing an object just to copy (later: move) it into a
QVariant.
ChangeLog will be in a follow-up change adding emplace support.
Task-number: QTBUG-112187
Change-Id: I444e580c7d8927d41b3d21d5a521e7c475119e4c
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Functions in Qt that take a callback need to support callables with or
without context objects, and member functions of an object. The
implementation of those overloads follows a pattern that ultimately
results in a QSlotObjectBase implementation being created and
passed to an implementation helper that takes care of the logic.
Factor that common pattern into a new helper template in QtPrivate
that returns a suitable QSlotObjectBase after checking that the
functor is compatible with the specified argument types.
Use that new helper template in the implementation of
QCoreApplication::requestPermission and QHostInfo::lookupHost.
The only disadvantage of centralizing this logic is that we cannot print
a more detailed error message indicating which argument types the
caller expects. However, that information is visible from the detailed
compiler errors anyway.
Change-Id: I24cf0b2442217857b96ffc4d2d6c997c4fae34e0
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The expected states are:
- nothing sets theMainThread before main()
- theMainThread is reset when the last QObject (the QCoreApplication in
the test) is destroyed
The GUI version of this test appears to leak a lot of QObjects. By the
time this function runs, theMainThread's QThreadData still has a
refcount of 66 on Linux/XCB. The Windows non-GUI version also
failed. Neither situation was investigated to see why objects are
getting leaked.
Pick-to: 6.5
Change-Id: Idd5e1bb52be047d7b4fffffd17507d9e6ef08743
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Amend 4d90c4e74a by clarifying why
moving the QSingleShotObject to the receiver's thread is a good
idea. Move that logic into a separate function and use that also
for the string-based connection.
Optimize the implementation by delaying the timer creation until
after we have moved the QSingleShotTimer object to the target
thread, using a queued metacall if needed to create the timer.
This avoids the creation of the timer in the wrong thread and
then the recreation of the timer in the new thread when QObject
handles QEvent::ThreadChange.
To avoid that the timer is created when it has already expired in
real time, use a deadline timer and fire timeout immediately when
it has expired by the time we get the metacall.
Since the timerId might now not be initialized in the constructor,
initialize it to -1.
Augment the crossThreadSingleShotToFunctor test function by
deliberately not starting the thread until the deadline expired.
[ChangeLog][Core][QTimer] Single-shot timers with a string-based
connection are now started in the receiver's thread, not in the
thread that creates the timer.
Task-number: QTBUG-112162
Change-Id: I3fc94c34c21d9f644da41a2e4c2da479980b8580
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This removes all uses of QDeadlineTimer::t2 member in the .cpp (so it
gets marked [[maybe_unused]]) and greatly simplifies the code.
Change-Id: Ieec322d73c1e40ad95c8fffd17465bd50c1113ea
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
You can't partially specialize a template function, so these
specializations for steady_clock only worked if the Duration parameter
was nanoseconds. This could have been solved with function overloads
instead, but I find the if constexpr code simpler to read.
Pick-to: 6.5
Change-Id: Ieec322d73c1e40ad95c8fffd17468bd73fc2fe24
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
This matches the work that was done for QElapsedTimer. The
QDeadlineTimer::t2 member is now always 0.
This also removes the last distinction of timer types. Originally I had
intended to use CLOCK_MONOTONIC_COARSE on Linux[1], but that created
more problems than was worth, so I abandoned the idea in 2016.
[1] https://codereview.qt-project.org/c/qt/qtbase/+/159933
Change-Id: Ieec322d73c1e40ad95c8fffd17468b313798ef79
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
This commit deletes the direct, low-level functionality that
QElapsedTimer has carried since it was introduced. Everything now uses
only std::chrono::steady_clock and std::chrono::nanoseconds.
QDeadlineTimer temporarily still uses qt_gettime(), which is moved to
qcore_unix.cpp.
Task-number: QTBUG-110059
Change-Id: Ieec322d73c1e40ad95c8fffd174641a469b1eee5
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Avoid capturing the same property multiple times in a binding by
storing them in the BindingEvaluationState. We store them in a
QVarLengthArray array, as the number of properties involved in a binding
is expected to be rather low, so a linear scan is fine.
Avoiding double capture is a good idea in general, as we would otherwise
needlessly reevaluate bindings multiple times, and also needlessly
allocate memory for further observers, instead of using a binding's
inline observer array.
Even more importantantly, our notification code makes assumptions that
notify will visit bindings only exactly once. Not upholding that
invariant leads to memory corruption and subsequent crashes, as
observers allocated by the binding would get freed, even though we would
still access them later.
Fixes: QTBUG-112822
Pick-to: 6.5 6.2
Change-Id: Icdc1f43fe554df6fa69e881872b2c429d5fa0bbc
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
When a context object is provided, then callers expect that the functor
or slot is executed in the thread of the context object. And if the
context object has been destroyed by the time the permission response
is received, the functor shouldn't be called at all.
To implement this, we either have to plumb the call back through a
signal/slot connection and benefit from QObject's infrastructure. This
is not practical here, as we don't have an "engine QObject" that would
emit a signal.
Instead, we can create a QMetaCallEvent explicitly, following what we do
in e.g. QHostInfo, and using a temporary QObject that handles the event
to then call the functor.
Add test coverage.
Pick-to: 6.5
Change-Id: Id878e45b304857304165ab4a7c6aae76fbee46ce
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This will be helpful in a number of places, in particular in order to
support enums of different sizes in QML. We record the type as string in
the JSON output and as QMetaTypeInterface in the generated C++.
Task-number: QTBUG-112180
Change-Id: I943fac67f8b25b013d3860301416cdd293c0c69e
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Amends 4d90c4e74a, after which the test
became flaky. We need to wait for the functor to be called before
quitting the thread, otherwise we have no guarnatee that any of the
queued metacall events have been processed by the thread. Since
QThread::quit is thread-safe, we can just call it from within the
functor. This guarantees that at least one of the single-shot timers
is processed before we quit.
And since QTimer::singleShot has special code paths for 0-ms timers
(going through an explicitly queued QMetaObject::invokeMethod call
rather than through an actual QSingleShotTimer object), we need to run
the test logic with different timeouts to cover both code paths.
Task-number: QTBUG-112162
Pick-to: 6.5 6.2 5.15
Change-Id: Ide1e7b4b74dcbda72144a0d73ef5f64b0694ddbc
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Currently, Qt assumes that enums always have int as their underlying
type (both in QMetaEnum::keyToValue and in the QML engine). This change
makes it possible to to retrieve the underlying type from an enum's
metaype - or rather, a metatype of an integral type with the same size
and signedness. The use cases aobve don't really rely on the exact same
type. In most cases, we wouldn't even need the signedness, however that
is already available anyway, and it will come in handy once QML supports
bigint, and we need to decide whether we should return
While it would be possible for individual users of this function to
manually query the size and signedness, having a function returning a
metatype offers additional convenience - especially in QML, where the
conversion APIs generally operate on metatypes.
Task-number: QTBUG-27451
Task-number: QTBUG-84055
Task-number: QTBUG-112180
Change-Id: Icf733b42df0ea64017d69f4d94cb7c855d9e3201
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
A null QString used to make a QVariant containing it null in Qt 5, but
this is no longer the case. Add a test that get_if works as expected.
Task-number: QTBUG-111598
Change-Id: I0f3511e1b33f4a9d67755269455680feda22ddca
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This second patch of the series implements get().
Unlike other get() implementations in Qt, don't use my trick with the
constrained friend free function here. Instead, provide the four
overloads manually, like mandated by the standard library for
std::variant (and, indeed, tuple), such that these functions can also
be used on subclasses of QVariant.
[ChangeLog][QtCore][QVariant] Implemented the type-based std::variant
access protocol (get<T>()/get_if<T>()) to allow easier access to the
contained element than the previous solution of casting data(), as
well as to allow generic code to treat QVariant and std::variant the
same. The holds_alternative<T>() function is not provided, since it's
the same as get_if<T> != nullptr. The index-based variant access
functions (get<I>()/get_if<I>()) are also not provided, because,
unlike std::variant, QVariant does not have a bounded number of
alternative types, and QMetaType IDs are not (all) compile-time
constants.
Fixes: QTBUG-111598
Change-Id: Id7bc41f7d91761b3483ec5604f1a4685c8079844
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
QVariant supports non-default-constructible payloads, in principle
(QTBUG-105140). And fromValue() works with such types, but value()
insists on providing a wide contract and therefore accidentally
requires default-constructible.
We can now invent other "Qt-ish" API like optional::value_or() or a
value() that returns optional<T>, but we should first get the
interface in that generic code must use, and which at the same time is
the most versatile, because it gives write access to the element
stored in the variant: [variant.get], consisting of get_if(), get(),
and holds_alternative(). The latter is the same as get_if() !=
nullptr, so we won't provide it.
This first patch implements get_if(), adds test for it.
As a Hidden Friend supposed to be called with explicit template
arguments, we run into the problem that wg21.link/P0846 solved for
C++20. Add the usual work-around, and check it works.
The ChangeLog will be on the last patch.
Task-number: QTBUG-111598
Change-Id: I23f57ea2de3946944810c5552c68a7a3060a44f2
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
I must have broken this in the 6.5 work I did for QMetaType and
QVariant, but I haven't searched which commit exactly did it. Our
QVariant tests are old and thus only checked the type ID, which meant
that they caused the registration by the act of asking for the ID in the
first place; this commit adds a couple of explicit checks for the type
registered by name before the ID.
Fixes: QTBUG-112205
Pick-to: 6.5 6.5.0
Change-Id: Idd5e1bb52be047d7b4fffffd174f1b14d90fd7a3
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
For single-shot timers, at least. QSingleShotTimer had either an
optimization or the only way to make new-style slot invocations by
storing the QSlotObject pointer and calling it directly. Instead of
doing that, let's just actually connect and let QObject handle the
actual delivery.
[ChangeLog][QtCore][QTimer] Fixed a bug that caused slots connected to
single-slot timers using the new-style connection mechanism to be
delivered in the wrong thread.
Fixes: QTBUG-112162
Pick-to: 5.15 6.2 6.5
Change-Id: Idd5e1bb52be047d7b4fffffd174eadb227ab65ee
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Ahmad Samir <a.samirh78@gmail.com>
All the other overloads are implemented using the new one.
Windows change relies on the pre-check in the code review making sure it
compiles.
[ChangeLog][QtCore][QThread] Added sleep(std::chrono::nanoseconds)
overload.
Task-number: QTBUG-110059
Change-Id: I9a4f4bf09041788ec9275093b6b8d0386521e286
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
250ca8d5f8 changed the qConvertToNumber
to have a std::optional return instead of a boolean out-argument.
In doing so a code path that was supposed to report a failure (string
could not be converted to an integer) accidentally starting reporting
success (and converting the string to 0).
The problem is that the `ok` check from QString::toLongLong was
accidentally dropped in the refactoring; previously the function set
`ok` to false and returned 0, now the function just returns 0.
Instead, amend that return to return nullopt (because the conversion has
failed).
Change-Id: Iaedef5463f3ec500a97bd4c9bbddf977f66df61a
Pick-to: 6.5 6.5.0
Fixes: QTBUG-111867
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
In most QObjectPrivate-subclasses, the Q_DECLARE_PUBLIC macro is used in
the private segment of the class declaration. In that case, the q_ptr
becomes a private member of the private class, and then the
QObjectPrivate::connect function can no longer be used, as it needs to
access the d_ptr.
Fix this by declaring QObjectPrivate, and the static-assert-helper, as
friends of the class using the Q_DECLARE_PUBLIC macro.
Adapt the QObject test by moving the Q_DECLARE_PUBLIC macro into
the private section of the test-private, and add a compile test.
Pick-to: 6.5 6.5.0
Change-Id: Ifc04be3b305221e138b1e08bb3a3838d871f4fcb
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Axel Spoerl <axel.spoerl@qt.io>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
QPropertyAlias might be deprecated, but using it should still not fail
to compile.
Pick-to: 6.5
Fixes: QTBUG-111735
Change-Id: I486cddb424b60cd3e5c539e26afca3726e29bb09
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Both QTimer's and QObjectPrivate's destructors print a warning if the
current object lives on another thread and has an active timer:
QWARN : tst_QTimer::moveToThread() QObject::killTimer: Timers cannot be stopped from another thread
QWARN : tst_QTimer::moveToThread() QObject::~QObject: Timers cannot be stopped from another thread
This timer is used to ask the thread to quit, which in turn allows us to
destroy this QObject without a cross-thread warning. Because it's
already fired once and done its duty, we can make sure it's not active
by simply making it single-shot.
Pick-to: 6.4 6.5
Change-Id: Ieec322d73c1e40ad95c8fffd17465067b27c044b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
We so far refetched the first observer after evaluating bindings, as
binding evaluating might change the list of observers.
However, that approach did not take into account that the 'this' pointer
might no longer be valid after binding evaluation: In case of a
QObjectBindableProperty (or a QObjectCompatProperty), binding evaluation
might cause a reallocation of the binding storage, and consequently the
invalidation of the QPropertyBindingData.
Fix this by refetching the QPropertyBindingData from the storage (if a
storage has been provided, which is always the case for the affected
classes).
Fixes: QTBUG-111268
Pick-to: 6.5 6.4 6.2
Change-Id: Ie7e143a0bbb18f1c3f88a81dd9b31e6af463584f
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The code for the QMetaMethod queuedConnectionTypes overload relied on
QMetaMethod::parameterMetaType and QMetaType::flags() to detect whether
we're dealing with a pointer (and then use the VoidStar metatype instead).
However, if the type was incomplete when the slot was defined, and the
type was not registered when connect was called, we would not find a
metatype for the argument.
However, in that case we might still be able to handle the method, by
checking whether the type name of the arguments ends with a "*".
This patch does that, fixing a regression from 5.15.
Pick-to: 6.5 6.4 6.2
Fixes: QTBUG-108537
Change-Id: I54cc48a3343444480ab9094fe1ebaaa5aa75cee0
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
A RAII wrapper around Qt::{begin,end}PropertyUpdateGroup().
[ChangeLog][QtCore][QScopedPropertyUpdateGroup] New RAII class
wrapping Qt::beginPropertyUpdateGroup() and
Qt::endPropertyUpdateGroup().
Fixes: QTBUG-110710
Change-Id: If2619e9584dd9d57985d63e3babca75421499ab9
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
A QVariant(QString) was not convertible to an enum not registered with
Q_ENUM() which worked fine in Qt5.
The same problem exists for QVariant(enum) to QString.
Fix it by not bailing out when no metatype for the enum was found and
try to convert it to a qlonglong instead (which is then correctly
converted to the enum type).
Fixes: QTBUG-109744
Pick-to: 6.5 6.4
Change-Id: Ie7bb016a860455b69508f0f46b36474c9c294f3a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This fixes a use-after-free in QPropertyDelayedNotifications::notify.
Before this patch, evaluateBindings or a notify from a property index
might have caused the originalBindingData to become reallocated.
However, at that point, we've already restored the original bindingData
in evaluateBindings, so we won't track updates, and thus won't adjust
originalBindingStatus, which will then point to already freed data.
To remedy this, we no longer do the notification with data fetched from
originalBindingData, but instead use the information we have in the
proxyData.
We also need to enure that referenced bindings do not get deleted; for
that we keep the PendingBindingObserverList alive for the whole duration
of the endPropertyUpdateGroup.
As we now have the PendingBindingObserverList, we use it for the
notification logic, and only notify change handlers in
QPropertyDelayedNotifications::notify. That will allow a follow-up
cleanup of QPropertyObserverPointer::notify, and aligns the logic for
grouped updates with the logic for "nornal", non-grouped updates.
Amends f1b1773d0a.
Task-number: QTBUG-110899
Pick-to: 6.5 6.4 6.2
Change-Id: Iae826e620d9614b7df39d86d8a28c48c8a5c4881
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
That use case for the class never materialized, and it was actually
meant for internal use. However, we put it into the public header, so
we cannot remove it (and while undocumented, someone actually used it at
some point, compare e4d62651c2).
Mark it as deprecated instead so that it can be finally be removed in Qt
7.
Change-Id: I058c5831a44610121fbec6eaddebd8b33d4a16c9
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Marc Mutz
Volker Hilsheimer
Ulf Hermann
Giuseppe D'Angelo
Ahmad Samir
Ivan Solovev
Axel Spoerl
Tor Arne Vestbø
Mårten Nordheim
Thiago Macieira
Mikolaj Boc
Edward Welbourne
Alexandru Croitor
Jøger Hansegård
Fabian Kosmale
Friedemann Kleint
Christian Ehrlicher