Remove the old revision of the interface, this is not
required with Qt 6 anymore, as everything is being
recompiled anyway.
Change-Id: I66070c4dc6b5e2a6d22f5a9ebea7688ed38333fe
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
And remove the old manual registration code for those operators.
Add some special handling for long/ulong, as these types could be
streamed as a QVariant so far, but are not directly streamable
through QDataStream.
[ChangeLog][QtCore][QMetaType] The QMetaType::registerStreamOperators()
and QMetaType::registerDebugStreamOperator() methods have been
removed. The streaming operators for a type are now automatically
registered together with the type registration. This implies that the
operators should be visible wherever the type is visible and being used.
[ChangeLog][Behavior Incompatible Changes] Because the QDataStream and
QDebug serialization operators are automatically registered with
QMetaType, the declarations of those functions must be present at any
point where the type is used with QMetaType and QVariant.
Change-Id: I4a0732651b20319af4a8397ff90b848ca4580d99
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
The string representation of std::pair<T1,T2> is now always
"std::pair<T1,T2>". This is in line with how we translate QPair,
avoiding typename mismatches that would previoulsy occur, because the
full name of pair on libc++ was "std::__1::pair".
Fixes: QTBUG-84924
Change-Id: Ia6c044a7327d69e4b4f4a31496c6b2408d85ebb9
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It was marked internal anyway. Use the constructor taking a
QMetaType instead.
Change-Id: I15b9cd0911aac063a0f0fe0352fa2c84b7f7c691
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
isNull() would forward to the contained type and check that type's
isNull() method for some of the builtin types. Remove that behavior
and only return true in isNull(), if the variant is invalid, doesn't
contain data or contains a null pointer.
In addition, implement more consistent behavior when constructing
a QVariant using the internal API taking a copy from a void *.
isNull() should return true in both cases. This mainly changes behavior
for some corner cases and when using our internal API.
[ChangeLog][Important Behavior Changes] QVariant::isNull()
no longer returns true when the variant contains an object of some
type with an isNull() method, that returns true for the object;
QVariant::isNull() now only returns true when the variant contains
no object or a null pointer.
Change-Id: I3125041c4f8f8618a04aa375aa0a56b19c02dcf5
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Comparing two variants will not try to convert the types
of the variant anymore. Exceptions are when both types are
numeric types or one type is numeric and the other one a
QString. The exceptions are there to keep compatibility with
C++ and to not completely break QSettings (which needs automatic
conversions from QString to numeric types).
[ChangeLog][Important Behavior Changes] Comparing two
variants in Qt 6 will not try attempt any type conversions before
comparing the variants anymore. Instead variants of different type
will not compare equal, with two exceptions: If both types are numeric
types they will get compared according to C++ type promotion rules. If
one type is a QString and the other type a numeric type, a conversion
from the string to the numeric tpye will be attempted.
Fixes: QTBUG-84636
Change-Id: I0cdd0b7259a525a41679fb6761f1e37e1d5b257f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Remove the compare method in the QVariant::Handler struct. Rely
on the generic support provided by QMetaType instead.
[ChangeLog][Important Behavior Changes][QVariant] QVariant will now use builtin support in
QMetaType to compare its content. This implies a behavioral change
for some graphical types like QPixmap, QImage and QIcon that will
never compare equal in Qt 6 (as they do not have a comparison
operator).
Change-Id: I30a6e7116c89124d11ed9052537cecc23f78116e
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
mkspecs/features/qt.prf adds a dependency on the system threading
library if the Qt Core thread feature is enabled. Because qt.prf is
loaded by any public or internal Qt project, it's essentially a public
dependency for any Qt consumer.
To mimic that in CMake, we check if the thread feature is enabled, and
and set the Threads::Threads library as a dependency of Qt6::Platform,
which is a public target used by all Qt modules and plugins and Qt
consumers.
We also need to create a Qt6Dependencies.cmake file so we
find_package(Threads) every time find_package(Qt6) is called.
For the .prl files to be usable, we have to filter out some
CMake implementation specific directory separator tokens
'CMAKE_DIRECTORY_ID_SEP' aka '::@', which are added because we call
target_link_libraries() with a target created in a different scope
(I think).
As a result of this change, we shouldn't have to hardcode
Threads::Threads in other projects, because it's now a global public
dependency.
Task-number: QTBUG-85801
Task-number: QTBUG-85877
Change-Id: Ib5d662c43b28e63f7da49d3bd77d0ad751220b31
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
... and QMultiMap as std::multimap.
Just use the implementation from the STL; we can't really claim that
our code is much better than STL's, or does things any differently
(de facto they're both red-black trees).
Decouple QMultiMap from QMap, by making it NOT inherit from
QMap any longer. This completes the deprecation started in 5.15:
QMap now does not store duplicated keys any more.
Something to establish is where to put the
QExplictlySharedDataPointer replcement that is in there as an
ad-hoc solution. There's a number of patches in-flight by Marc
that try to introduce the same (or very similar) functionality.
Miscellanea changes to the Q(Multi)Map code itself:
* consistently use size_type instead of int;
* pass iterators by value;
* drop QT_STRICT_ITERATORS;
* iterators implictly convert to const_iterators, and APIs
take const_iterators;
* iterators are just bidirectional and not random access;
* added noexcept where it makes sense;
* "inline" dropped (churn);
* qMapLessThanKey dropped (undocumented, 0 hits in Qt, 1 hit in KDE);
* operator== on Q(Multi)Map requires operator== on the key type
(we're checking for equality, not equivalence!).
Very few breakages occur in qtbase.
[ChangeLog][Potentially Source-Incompatible Changes] QMap does not
support multiple equivalent keys any more. Any related functionality
has been removed from QMap, following the deprecation that happened
in Qt 5.15. Use QMultiMap for this use case.
[ChangeLog][Potentially Source-Incompatible Changes] QMap and
QMultiMap iterators random-access API have been removed. Note that
the iterators have always been just bidirectional; moving
an iterator by N positions can still be achieved using std::next
or std::advance, at the same cost as before (O(N)).
[ChangeLog][Potentially Source-Incompatible Changes] QMultiMap does
not inherit from QMap any more. Amongst other things, this means
that iterators on a QMultiMap now belong to the QMultiMap class
(and not to the QMap class); new Java iterators have been added.
Change-Id: I5a0fe9b020f92c21b37065a1defff783b5d2b7a9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We can end up in a situation where a (soon to be destroyed) observer is
owned by a binding which is about to be deleted. If in that situation
the binding is destroyed first, we end up with a dangling pointer
and ensuing memory corruption. Instead, we now first transfer the
ownership of the observer and only destroy the binding afterwards.
Fixes: QTBUG-85824
Change-Id: I721c0319281ada981ae7896bd2e02e9a0cc901b8
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We already have the information in the QMetaTypeInterface, and provide
functions to access sizeof. Adding alignof support seems natural, and
should make it easier to handle over-aligned types.
This should also be helpful in QVariant.
Change-Id: I166be76f4b7d2d2e524a3a1e513bd2f361e887c1
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
For types that don't have an operator==(), always trigger the binding
and the changed notification.
Task-number: QTBUG-85578
Change-Id: I41374f6d13c88106f4de83864e82172f3a248150
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Smart pointers like QSharedPointer<T> do have a value_type, but their
equality does not depend on T being comparable. Therefore, instead of
simply checking for T::value_type, test for a few other container
requirements.
This also required to add an additional check for std::optional, as that
one has an unconstrained operator== on MSVC.
Change-Id: Iefd048f7aa360f4713ecd79f80acd7dae72ee18c
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Given we optimize for the case where the new value is of the
same type of the one already stored in the variant, enable move
assignment for that case.
As a drive-by, avoid a path to detach() for data() if we know
we're detached.
Change-Id: I9abbdc10637ce77ebb747b49d83e1ef914d997bb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
And mark some methods as inline.
Performance is critical for our new property system. Compiling
it in one unit makes it possible for the compiler to do a much
better job at inlining and generating optimized code.
Improves performance of binding evaluations by another 20%.
Change-Id: I5a2aa93c74d2b68418b0a9d2e34d8199bb71e3ad
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Avoid any QVariant or type dependent code in the cpp files.
Instead, let the binding wrapper determine if the value
has changed and return true/false accordingly.
This required also some reworking of the guard mechanism
for notified properties, where the guard function wrapper
now calls first the binding evaluation function and then
passes the result to the guard.
Change-Id: I350d07a508ccc0c5db7054a0efa4f270b6a78ec3
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
There's no point in returning a usually empty error when
evaluating bindings, adding overhead to the regular code
path.
Instead, the error can be set on the currently evaluating
binding if required. This streamlines the functor used to
wrap the binding and should thus expand to less code and
execute faster in the regular case.
To achieve this, expose a pointer to the currently evaluating
binding in the private API (as QtQml needs it to be able to
report errors).
The error case now requires one additional TLS lookup, but
we don't really care about performance in that case anyway.
Change-Id: Iecb450e765244930a41d813fcf8eb4013957a6a3
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Remove location(). The method would always return an empty value. If you need the location,
the binding itself has it.
Remove setDescription() and require that the description gets passed
in the constructor. Never create a d pointer if type is NoError, so we
can quickly check for it inline.
Change-Id: I7eb8a94786281069d6ea2d82567c09aa50c52ef6
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
And generate a few more test projects that were missing.
Change-Id: I5df51106549aa5ae09bc3c42360e14b143719547
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Use pro2cmake with '--api-version 2' to force regenerate
projects to use the new prefixed qt_foo APIs.
Change-Id: I055c4837860319e93aaa6b09d646dda4fc2a4069
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
This is in preparation for regenerating them with the new qt_foo
prefixed APIs.
Change-Id: Iff34932d642b1c0186ee39f952adf3ad367fd602
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
This removes the fully manual registration of comparison operators in
QMetaType and replaces it with an automatic registration through
Q_DECLARE_METATYPE().
[ChangeLog][QMetaType] The QMetaType::registerComparator() and
QMetaType::registerEqualsComparator() have been removed.
Q_DECLARE_METATYPE() now automatically registers any
operator==() and/or operator<() for a type visible where
it is used on that type, as part of declaring its meta-type.
Change-Id: I3df451b652b735c093533838bf32f3cc785439f8
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Containers often define an operator==() or operator<() which is very useful
for generic code. But those operators can usually not be instantiated if
the template argument doesn't implement the operator.
This sometimes leads to the compiler trying all possible template expansions
and implicit conversions for the type, giving extremely long error
messages. The traits support can be used to safely constrain those
operators.
Being able to safely detect this will also allow us to fold the comparison
support that is currently a large cludge for user types directly into
QMetaType.
Change-Id: Ib84afb5348c3eb0be5161d6ba9d5fe237709c65f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This change only happens to files touched
by the commit to add missing ; to Q_UNUSED.
Task-number: QTBUG-82978
Change-Id: I10e6993a2bb3952cf9a262708b8573550e0dbe63
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Applied to headers only. Source file to be changed separately.
Omitted statemachine for now to avoid conflicts.
Omitted qmetatype.h for now - to be handled later.
Task-number: QTBUG-84469
Change-Id: I317376037a62467c313467d92955ad0b7473aa97
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Adding support for a static notifier within QProperty itself - through a
QProperty "sister" class - is more efficient in terms of memory
consumption and run-time performance.
The MemberChangeHandler permanently takes up at least three pointers,
while the notified properties only cost one pointer in the binding.
Change-Id: Ia1a8c2b66f1f3c2fe13ae0ad9f12cdb6bdcc35ef
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
A guard callback is a predicate which takes the new value set by
setValue or computed as the result of a binding expression. If it
returns false, the value is discarded and the old value is kept.
Note that due to lazyness, when setting a binding, we still notify
everyone as the binding is only evaluated on demand, and the guard can
thus only run when someone actually queries the value.
Note further that a guard is allowed to modify the value that is passed
to it (e.g. to clamp it to a certain range).
Task-number: QTBUG-85032
Change-Id: I3551e4357fe5780fb75da80bf8be208ec152dc2a
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Check at compile time whether the static callback takes an argument
(which has to be of the same time as the type of the property). If so,
retrieve the old value and pass it to the callback.
Change-Id: Ib1c4c9e05b826b6be492b03f66fa72ad015963ee
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
[ChangeLog][QtCore][QByteArray] Remove method overloads taking
QString as argument, all of which were equivalent to passing the
toUtf8() of the string instead.
Change-Id: I9251733a9b3711153b2faddbbc907672a7cba190
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
And name the main class QList. That's also the one we document.
This gives less porting pain for our users, and a lot less churn
in our API, as we use QList in Qt 5 in 95% of our API.
In addition, it gives more consistent naming with QStringList and
QByteArrayList and disambiguates QList vs QVector(2|3|4)D.
Fixes: QTBUG-84468
Change-Id: I3cba9d1d3179969d8bf9320b31be2230d021d1a9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
There is no reason for keep using our macro now that we have C++17.
The macro itself is left in for the moment being, as well as its
detection logic, because it's needed for C code (not everything
supports C11 yet). A few more cleanups will arrive in the next few
patches.
Note that this is a mere search/replace; some places were using
double braces to work around the presence of commas in a macro, no
attempt has been done to fix those.
tst_qglobal had just some minor changes to keep testing the macro.
Change-Id: I1c1c397d9f3e63db3338842bf350c9069ea57639
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Property flags should be compile time booleans, not something to
be determined at runtime.
We've been using this to dynamically disable some properties in QWidget
based classes dependent on the state of a different property, but this
should better get implemented on top of our widgets.
Change-Id: I6296e8761303ecdf24d9e842142e8596304c015d
Reviewed-by: Simon Hausmann <hausmann@gmail.com>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Make QPair an alias for std::pair, and qMakePair just a forwarder
towards std::make_pair.
Why? Fundamentally to ditch a bunch of NIH code; gain for free
structured bindings, std::tuple and std::reference_wrapper
compatibility, and so on.
Breakages:
* Some that code manually forward declares QPair.
We don't care about it (<QContainerFwd> is the proper way).
* Some code that overloads on std::pair and QPair. Luckily
it's mostly centralized: debug, metatypes, testing macros.
Just remove the QPair overload.
* Usages of qMakePair forcing the template type parameters.
There are a handful of these in qtbase, but only one was actually
broken.
* std::pair is NOT (and will never likely be) trivially copiable.
This is agreed to be a mistake done by practically all implementations
in C++11, can can't be fixed without breaking ABI.
Some code using QPair assuming it's trivially copiable may break;
exactly one occurrence was in qtbase.
* QMetaType logic extracts the type names in two different ways,
one by looking at the source code string (e.g. extracted by moc)
and one via some ad-hoc reflection in C++. We need to make
"QPair" (as spelled in the source code) be the same as "std::pair"
(gathered via reflection, which will see through the alias)
when compared. The way it's already done e.g. for QList is
by actually replacing the moc-extracted name with the name
of the actual type used in C++; do the same here.
On libc++, std::pair is actually in an inline namespace --
i.e. std::__1::pair; the reflection will extract and store
"std::__1::pair" so we need an ad-hoc fix to QMetaType.
[ChangeLog][QtCore][QPair] QPair is now an alias to std::pair,
and does not exist as a class in Qt any more. This may break
code such as functions overloaded for both QPair and std::pair.
Usually, the overload taking a QPair can be safely discarded,
leaving only the one taking a std::pair. QPair API has not changed,
and qMakePair is still available for compatibility (although
new code is encouraged to use std::pair and std::make_pair
directly instead).
Change-Id: I7725c751bf23946cde577b1406e86a336c0a3dcf
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Macros and the await helper function from qfunctions_winrt(_p).h are
needed in other Qt modules which use UWP APIs on desktop windows.
Task-number: QTBUG-84434
Change-Id: Ice09c11436ad151c17bdccd2c7defadd08c13925
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The code was reading from uninitialized memory when creating a
variant. Fixing that reveals that the test semantics
are broken: when dealing with datatypes without a registered
operator==, QVariant resorts to memcmp, so the two objects
would've actually compared equal. Amend that.
Change-Id: I36bad7ee6a45154d5d534b7dd8b618cc0a900126
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The static observer can live in a union with the inline observers. We
only need to take care of calling the ctors and dtors manually then.
In order for any observers to be called in the presence of a static
observer, the static observer has to be called after the other
observers.
Change-Id: I2f56fa64f3fe6fcd7f06cc403929362da7d86f65
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Create a QMetaTypeInterface for void. This allows us differentiate
QMetaType::Unknown from QMetaType::Void. In addition, this will enable
the usage of QMetaMethod::metaReturnType in QMetaMethod::returnType,
and will facilitate using metaReturnType in declarative, which needs to
distinguish between Unknown and Void.
Change-Id: I83296b49587f3deb7ec73e25a33f0d8c98cf8da0
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This does the analog of 46f407126e for the
methods we care about (signals, slots, Q_INVOKABLEs). In addition to the
actual QMetaType, we store an array with offsets so that we later can do
a mapping from methodIndex to metatype.
The newly added QMetaMethod::{return,parameter}MetaType methods can then
be used to retrieve the metatypes.
This does however require that all involved types are complete. This is
unfortunately not a feasible requirement. Thus, we only populate the
metatype array on a best effort basis. For any incomplete type, we store
QMetaType::Unknown. Then, when accessing the metatype, we fall back to
the old string based code base if it's Unknown.
Squashes "moc: support incomplete types" and "Fix compile failures
after QMetaMethod change"
Fixes: QTBUG-82932
Change-Id: I6b7a587cc364b7cad0c158d6de54e8a204289ad4
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
A common pattern in Qt Quick will be QProperty members that are
connected to a callback that needs to perform something when the value
changes, for example emitting a compatibility signal or marking scene
graph node data dirty.
To make such a pattern more efficient, a new QNotifiedProperty type is
introduced that offers the same API as QProperty<T>, with two changes:
(1) The template instantiation not only takes the property type as
parameter but also a callback pointer-to-member.
(2) Since that member itself cannot be called without an instance
and to avoid storing an instance pointer permanently, the API for
setBinding and setValue are adjusted to also take the instance
pointer. For the former it gets stored in the binding, for the
latter it is used to invoke the callback after setting the new
value.
Change-Id: I85cc1d1d1c0472164c4ae87808cfdc0d0b1475e1
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
A property alias is the equivalent of the "alias" keyword in QML. It
provides the same API as QProperty, but redirects any access to the
QProperty it was initialized with. When the original property is
destroyed the binding becomes invalid and ignores any further acccess.
Task-number: QTBUG-84370
Change-Id: I0aef8d50e73a2aa9e7703d51194d4c5480573578
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Previously, only the first observer would get notified. Also, make sure
that the notifiers are always retained when switching between bindings
and values.
Change-Id: I9c25c0f2e288dac3a335b68e618f7ddeb44be25a
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
