There is no point in having them separate and this way the plugin can be
optional.
Fixes: QTBUG-89804
Change-Id: Ic7de35f6ee7abde4840841e17d21c2b709f6db7d
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
When using refs as loop variables, the clang compiler complains (with
default settings). This prevents that. Note that QJsonValueRef is used
"behind the scenes", which makes this iteration method correct.
Change-Id: I5a5f58ca8ad3887bce2009231cbae5a57c107697
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Move all the code from tools/shared into src/qmlcompiler and build a
static library from it so that we can re-use it in external tools.
Change-Id: I7c8d8e59063dc7c711f4072f103a01095e6f5997
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Tolerate dependencies without version and don't write them into the
output anymore. No one should read them as the same info has to be
available from the associated qmldir files.
Change-Id: I5e4cd93c83cb5e874a8a6e1467461d081b1a089c
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
We don't need to register the Type enum for both QtQml and QtQuick.
QtQml is enough. Removing this makes the whole manual value type
registration obsolete. Furthermore, we want QEasingCurve as QML_FOREIGN
as we have several classes with properties of that type. To keep it nice
and tidy, we make the uppercase-named enum holder class a separate type.
Unfortunately, the Type enums differ in one entry:
QEasingCurve::BezierSpline is called Easing.Bezier in QML. Therefore, we
need to keep the custom enum around. We can change all users in
qtdeclarative to use the name from QEasingCurve, though.
Change-Id: Ibbc78d8bbf8938e2a8722f8c09833a0c73394c3d
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
In many places we carry major and minor versions or revisions that are
loosely coupled to minor versions. As the Qt minor version resets now,
we need to handle these things more systematically. In particular, we
need to add a "major" part to revisions.
QTypeRevision can express the current major/minor pairs more efficiently
and can also be used to add a major version to revisions. This change
does not change the semantics, yet, but only replaces the types.
Change-Id: Ie58ba8114d7e4c6427f0f28716deee71995c0d24
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Qt::staticMetaObject should be used instead
Change-Id: I1e02884f0b1f07e9c51d065f11bdc51c793b7173
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Before we would always merge everything up to QObject. That would
duplicate entries between types that inherit from the same base class.
Also, when the composite type doesn't specify a module (because we just
created a component from a plain QML file), use the module URI from our
version info. As we're dumping the component we can assume it belongs to
the module we're dumping.
Change-Id: Icf9a58cfe1165f557ebbf7309251e98a0782dc33
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Previously the sorting depended on qHash(QQmlType), which hashes a
pointer to QQmlTypePrivate. That was not very useful.
Also, the defaultCompositeTypes were only populated when dumping
builtins, and only used when dumping regular types. Therefore, whenever
they were used, they were empty. We don't have to bother with that.
Change-Id: Ibfbfc7e14bee900d617bab58cbe39d05f5fe91c3
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
If not given a "--strict" argument, qmlplugindump will ignore the given
major version and only check the URI as prefix, not as exact match. This
is in line with what qmlplugindump did in version 5.13 and before.
Fixes: QTBUG-78745
Change-Id: Ib2803d47e6cb60c797fec8dfaa628278307d4cd4
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
We do not put the objects we encounter in defaultReachaleNames. Thus,
the check in prototypeNameForComposite will fail most of the time.
Instead, we now use the same check as in collectReachableMetaObjects
(that is, does the type belong to the module).
Then, we once again only collect the metaobjects belonging to the module
in objectsToMerge, and only dump the relevant properties from there.
Fixes: QTBUG-78744
Change-Id: I78cae1e56b6f59e277198f9bcd0d826c4e0978a7
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Now that our own types are not registered with qmlRegisterType<T>() anymore,
we can use QQmlType::module to decide whether a type should be included
in the qmltypes file.
We include types with no module, to support existing user code which
still calls qmlRegisterType().
Change-Id: Iee0a132bf5b3ddcdfc222ab4b2d37a1f05b1851d
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Singleton object instances were previously managed by
QQmlType::SingletonInstanceInfo and kept in a shared storage. This caused
concurrency problems when instantiating singleton instances from different
QML engines in different threads.
This patch moves the singleton house-keeping infrastructure to
QQmlEnginePrivate and makes SingletonInstanceInfo immutable.
Singleton objects are stored in a QHash with QQmlType as the key because
the qml type id might be 0 for composite singletons. The public API of
QQmlType is extended to provide more information about singleton types so that
access to SingletonInstanceInfo is not needed.
All internal accesses of singleton objects must now take the same code path via
QQmlEnginePrivate::singletonInstance<T>() which simplifies overall usage of
singletons and scatters less implementation details throughout the code base.
Task-number: QTBUG-75007
Change-Id: I13c5fd21cac2eb7291f2cbcf2c2b504f0f51a07c
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Just like with 'normal' types, we don't want to include composite types
that were actually part of depending imports. So we need to remove them
from the list.
This fixes an issue with QtQuick.Controls style plugins, were so
far types from QtQuick.Controls 2 imports were re-dumped.
Change-Id: Ib099ae9cfba35218cab852d4d260369c2a652727
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Fixes -Wclazy-unused-non-trivial warnings, in preparation for using
gcc/clang's warn_unused attribute in all containers.
Maintainers please check if some unused variable isn't hidding an
actual bug.
Change-Id: I83c5a11aa2328db3dce4c6c402295d86ef297c83
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Change 8725f9873d fixed qmlplugindump to always use the current
Qt minor version for the Qt Quick and Qt Qml dependencies.
Anyhow, this is unnecessary; for the dependencies resolution only the
major type is important. Hard-coding the latest import might lead to
people unnecessarily updating the plugins.qmltypes.
Change-Id: Icc969517f268b775492f2bcc901827de64123bae
Reviewed-by: Marco Benelli <marco.benelli@qt.io>
It's not safe to use printf and memory allocations
in a signal handler.
Task-number: QTBUG-70460
Change-Id: I2a4caf302e17992471575bd89faa505a8db4a7e3
Reviewed-by: Marco Benelli <marco.benelli@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Extension classes do augment the primary C++ class with additional
properties that should be part of the QML type. Anyhow, they should
not be visible in the API.
The old logic tried to emulate this by making the extension
class the actual C++ type, and 'inheriting' from the normal
class by setting it in the prototype. Anyhow, this failed for
e.g. Qt3D, where there are numerous types sharing the same
extension class.
Instead, this patch fixes the issue
a) hiding the extension class metainformation itself, and
b) printing properties of the extension class as part of the
main type.
In the QMetaObject collection phase, we now traverse the class
hierarchy using QQmlType::baseMetaObject instead of
QQmlType::metaObject.
In the generation phase, we explicitly resolve the extended type,
and dump additional properties and "DefaultProperty" information
into the main type description.
Note that the ExtendedType sets the DefaultProperty independent
of the version. Changing this would require either revisioning
the defaultProperty, or (again) splitting up the types, which
however brings other problems.
Task-number: QTBUG-70294
Change-Id: I324469152398a605fae9e5a35b619845444ac4e8
Reviewed-by: Marco Benelli <marco.benelli@qt.io>
Do not print the full path passed to -dependencies in the file header.
This avoids local paths to show up in the plugin.qmltypes file, and
therefore in the Qt checkout.
Task-number: QTBUG-70264
Change-Id: I7870d7d4e30e57684c62f4e353a6bf6d72dc5faf
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
QtQuick / QML minor version is nowadays the same as the Qt minor
version.
Change-Id: I4d7ef0b101320ee2179746c95093b62e063adba2
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
From now on we prefer nullptr instead of 0 to clarify cases where
we are assigning or testing a pointer rather than a numeric zero.
Also, replaced cases where 0 was passed as Qt::KeyboardModifiers
with Qt::NoModifier (clang-tidy replaced them with nullptr, which
waas wrong, so it was just as well to make the tests more readable
rather than to revert those lines).
Change-Id: I4735d35e4d9f42db5216862ce091429eadc6e65d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Skipping paths that ends with "private" is not enough.
Task-number: QTBUG-47027
Change-Id: I25af518b76f594c268db6b77dbafd343f2f57ee8
Reviewed-by: Teemu Holappa <teemu.holappa@qt.io>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Make sure any QQmlType stored in the meta type data
has a refcount of 1. This should now make it possible
to clean out unused types by iterating over the
list of types and removing those that have a refcount
of 1.
Some care is still needed for C++ registered types, that
will need to get one more refcount, so we don't accidentally
remove them.
Task-number: QTBUG-61536
Change-Id: Id2a18dae5ddcb815f34013f5fde1f05d2d9d0214
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
QQmlType is now refcounted, and we need to use it by
value, to control it's lifetime properly. This is
required, so we can clean up the QQmlMetaTypeData
cache on engine destruction and with trimComponentCache()
Task-number: QTBUG-61536
Change-Id: If86391c86ea20a646ded7c9925d8f743f628fb91
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Report to the user the stderr of qmlimportscanner, if needed.
Task-number: QTBUG-58930
Change-Id: Ib8ddf12cb3d18c936eb151f12991900a32175576
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Remove "qrc:/qt-project.org/imports" from the paths list passed to
qmlimportscanner.
qmlimportscanner will error out on a path that does not exist.
Since the internal qrc:/qt-project.org/imports path only exists if
the plugins are compiled in statically into qmlimportscanner, we
have to remove this path to stop it from erroring out.
Task-number: QTBUG-58930
Change-Id: I39525f2b6ff29c02b38f5f701342026dea871701
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Some modules (ie QtChart) require a QApplication instead of the
QGuiApplication that qmlplugindump use by default.
This option let the user specify if a QApplication is needed.
Task-number: QTBUG-58264
Change-Id: I10e68d0d3eab2afad315b57fef9ce92c42e70124
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
Having only the possibility to redirect the stdout of
qmlplugindump to a file is both unconvenient and unreliable
since some plugins might write content (such as logs) to
stdout, which pollute the output.
Change-Id: I8b1d482d7674945e6145d59aea839c54600e7784
Reviewed-by: Marco Benelli <marco.benelli@qt.io>
Wrap all global variable in an anonymous namespace in order to avoid
compiler warnings on the uninizialized ones.
Change-Id: Ifabb790d2d76b6056d706580b9e4c0f12f5119f0
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
Ulf Hermann
David Skoland
Olivier Goffart
Fabian Kosmale
Qt Forward Merge Bot
Richard Weickelt
Kai Koehne
Sergio Martins
Liang Qi
Shawn Rutledge
Lars Knoll
Marco Benelli
Simon Hausmann
Mitch Curtis
Peter Varga
Jacques GUILLOU