Otherwise doing deleteLater() on an ancestor of the QQuickWindow will
cause double deletion for that ancestor. This is typical when using
WA_DeleteOnClose since that uses deleteLater().
Task-number: QTBUG-40920
Change-Id: I9146fc0854f77c42cad2eda23b3cdaa7a9e22080
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
And how to use use two textures in the same material.
Change-Id: Idcc618ed359422c2a104eeed105b7c4f5086ee4e
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
This was an oversight the the implementation of
setOpacity. QSGOpacityNode already does the right thing.
[ChangeLog][QtQuick][Item] When Item.opacity is set to a value outside
the range of 0 to 1, it will be clamped.
Change-Id: Ib7f23c5d8860c76e009a25e5ab2c36db20301faf
Reviewed-by: Raine Mäkeläinen <raine.makelainen@gmail.com>
Reviewed-by: Alan Alpert (Personal) <416365416c@gmail.com>
prevModel points to d->instanceModel and is being
used further down in the code. So reset the pointer
to 0 after freeing the old instanceModel.
Change-Id: I8854ae3e09c8b2fe50ad311f3dbc7b8ed26805e1
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
With gcc 4.6.3 armhf (defualt for Ubuntu 12.04) some optimization bug
results in QML2 v4 compiler hang during type inference.
Task-number: QTBUG-40364
Change-Id: Iea1a8be3b5a7d9410304110d89dae3735339cb72
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>
Implements the suggestion from Simon Hausmann (codereview 89716
from 08-05 14:46) to use QQmlScriptString rather then the previous
used MetaObject-manipulation.
This also introduces comparison operators for QQmlScriptString
to be able to determinate if a QQmlScriptString changed what
is needed cause there is otherwise no way to access (all) the
needed details within QQmlScriptStringPrivate.
Change-Id: I198479eac8fd37cbdd98a99aacdd8eebf7b75d21
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
In the reported bug, it can happen that we try to access the compile-time resolved
QObject property of an object that is referenced by id. The binding that uses this is
triggered when the property changes but _also_ when the id referenced object gets either
created or deleted. The first time the binding is evaluated is very early on, when the
id referenced object is not created yet, so the binding evaluation fails. However the
dependency is set up, and so later then the id referenced object is created and the id
property is set on the context, the notification triggers and the binding is re-evaluated.
During that binding evaluation a QObject property access happens by index on an object that
doesn't have its VME meta-object set up yet. Therefore the property access fails and a
crash occurs or the Q_ASSERT(property) assertion fails.
The fix is to set register the id named object in the context _after_ the VME meta-object is
setup.
Task-number: QTBUG-40018
Change-Id: Ic2d7b4a0c49635efe68e93f2f6c316eb65f0c309
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
In case they try to remove themselves from our list of transforms when
that list has already been destroyed after ~QQuickItem() has run.
Task-number: QTBUG-40877
Change-Id: Ie57f5dd1e8b791846f08629183974c771553c4bf
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Alan Alpert <aalpert@blackberry.com>
Avoid dangling pointers when nested incubators are being used.
Task-number: QTBUG-40437
Change-Id: I73922d2f373b2efbc00983305cdea9e8d60f0c41
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Using new to do a reinterpret_cast is rather confusing.
Also protect the list model case against self-assignment
which could lead to memory corruption.
Change-Id: I10b81644d0004d4a50a5a74e5b8c58e27cbe6934
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
It is impossible to use in a shader effect otherwise.
Change-Id: I489e6ae43cdc3fe8f54efa6d86ead4d21831b84d
Task-number: QTBUG-40809
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Convenience for cleaning up texture providers. Private class for now,
we can maybe make it public in 5.5.
Change-Id: I3489fde2a268bc988634a16f87ee558b9ac5dba2
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
Allow the implementation of visit to control whether the children should be
processed afterwards or not. This is modelled after the JS AST visitor in
Qml.
Change-Id: I5b1adf5e1aefd8899bf444e993ddfbf2c5c125c9
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
This way they will at least render.
[ChangeLog][QtQuick] Images exceeding GL_MAX_TEXTURE_SIZE will be
downscaled to fit so they will still show.
Change-Id: I169ecac768036812b8e14265ec1a0a8902655666
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
This basically renames QQuickShaderEffectTexture to QSGDefaultLayer and
introduces QSGLayer as interface to be used. QQuickShaderEffectTexture is
generic for the scene graph and has no QtQuick dependencies. The interface
separation allows scene graph backends to customize layers.
Change-Id: I9a7f37addaa4b80a34ff9a1456b0cb9b16d4e9f3
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
And don't use "file watchers" for resource file paths,
otherwise it generates the following warning:
QInotifyFileSystemWatcherEngine::addPaths: inotify_add_watch failed: No
such file or directory
Change-Id: I6b75c9195fb2b2ba7b3e0bb7d146fc5cd343927e
Task-number: QTBUG-40307
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Get rid of incomplete attachment warnings on Android.
Change-Id: I7da546861b1814677b0c8ae110416f69bbbb7edc
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
When a cached compilation unit depends on an import that is implemented as a
Qml C++ plugin and that plugin in turn loads an asynchronous component that is
cached, we would get a crash in the typeloader and a failing assertion
(ASSERT: "d->m_mainThreadWaiting == false" in file qml/ftw/qqmlthread.cpp, line 300)
This is because we did not implement the asynchronous loading within the loader
thread for cached compilation units. This is simply done using a posted event,
using the same mechanism used for other async load methods.
Change-Id: Iefce67ab634ce26122c348dcdfc8e66b00fec671
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
We can simply use a standard image node and do the dirty marking also
directly in the texture (both live in the same thread).
Change-Id: I7345c702613782adec3c6a36a7705a2a54ed82ec
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
This makes following QML-code proper working:
ParentChange {
x: 0
Component.onCompleted: x = 10
}
where x is a QQmlScriptString.
Before this patch an error-message would be thrown that the
bool/int/string/etc cannot be converted to a QQmlScriptString.
With the patch primitive types including null and undefined are
proper converted to a QQmlScriptString. The patch ignores (as
in not implements) function/binding assignment.
Unfortunately since commit aa25ad8d5f its not possible any
longer to instanciate QQmlScriptString what means there is
otherwise no (easy) way to inject a QQmlScriptString from
within Javascript.
Change-Id: I18aac6a6e9a57f3b7d0a2d66cdab2be6c3c153c5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Change ee616b3905 accidentally
removed the member variable used to refcount the attribute
names. This resulted in the names being deleted and the
stored const char *'s to become invalid. Add it back.
Change-Id: Ie33f75cd76085283a5ee685602e023bb3c42c896
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
By peeling the first iteration off of a loop and putting it in front of
the loop, type inference can deduce more type information for esp. loop
induction variables. To prevent increasing the code size too much, only
the inner-most loops are peeled.
This gives a 10% speed-up on crypto.js.
Change-Id: I57f9611695bc8defc0bff84e440b8a20b2c8a34e
Reviewed-by: Fawzi Mohamed <fawzi.mohamed@digia.com>
When removing edges, the control-flow graph changes, so some immediate
dominators might need to be updated. This change collects all candidates
and processes them in a single batch.
Change-Id: I928f42232427a84bcb9658e314dadd0bd021b12f
Reviewed-by: Fawzi Mohamed <fawzi.mohamed@digia.com>
-url inherited from the url variable set in qtbase/doc/global
Change-Id: I94a0da7aa98af1fdd6140168fe9feb27b11c0bdb
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@digia.com>
Reviewed-by: Jerome Pasion <jerome.pasion@digia.com>
Fix QQmlContext::setContextObject to reevaluate bindings as the
documentation says that it should.
Task-number: QTBUG-40798
Change-Id: Ifbd97c7a07a5432f4948937da863370b05705206
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Without this information the app fails to find the plugin
required to load the QtQuick.LocalStorage module on Android.
Task-number: QTBUG-37203
Change-Id: I245012e78ab8d1fe3205e08939f4baec58c20759
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Merge QV4::CompiledData::QmlUnit into QV4::CompiledData::Unit. For pure JS
units it means a slight increase of memory usage by a few bytes, but overall it
makes the code a lot simpler.
Change-Id: Ib48927749720b056f004aac0fe22cb8ec729e3f6
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
if there's an error during loading, it's much easier to spot if this fails
rather than crashes.
Change-Id: If0106729e18768819da8a877959a095398af8547
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
We only assign to the variable once, so the value may not change in time.
Change-Id: I54ac7faad60d3984bfde4606aaf7e30a2340ba3e
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
This also reintroduces the change from commit 14b4747b17.
Due to some bizarre merge issue that change has disappeared from qquickwidget.cpp
and its history. It is there in the main log, though.
The new fix depends on this older one, so the combined one fixes the following
three issues:
1. Polish and sync cannot be skipped when we are rendering for the first time.
2. Recreating the fbo twice when the widget is shown. Calling
createFramebufferObject() is not necessary in this case since createContext()
will trigger this anyhow due to scenegraphInitialized().
3. Avoid recreating the fbo when the size is the same as before. Some platforms
are keen on sending resize events with the same size. These should be ignored.
What's worse, some platforms (cocoa) generate a resize on exitting (Cmd-Q)
and not ignoring the resize at that stage is dangerous since the scenegraph
is already invalidated.
Task-number: QTBUG-40710
Task-number: QTBUG-40505
Change-Id: I2e897acc68fa68233563a1db63ffb6c5d0baad73
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Missing else: two (nearly identical) conversions would get generated.
Change-Id: I745120f81d42bf28fbce7ab6a62da909a8e14458
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
VFP3-d16 is the minimum implementation available on ARMv7, so these
registers are also always available. The big added bonus is that they
are callee saved, so using them will result in less loads/stores for
doubles.
Change-Id: I0cab3fe24d8677315b4c7d9449d8619cbf9a2919
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
This is not used yet by any platform/abi we support, because we do not
define any callee-saved FP registers. Yet.
Change-Id: I5857a452456175398c5e9681ff33800b9431b9da
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Instead of saving all possibly used ones. Note that floating point
registers are not saved yet, as we don't support callee-saved FP
registers yet.
Change-Id: I1db2ba2513f7b466c64ec103eda0c464269247b0
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
... with the regular (non-FP) registers that need to be saved. This
patch shouldn't change any of the JIT generated code, because all
regular callee saved registers are passed in.
Change-Id: Id11b8f37f06d80e8015ac6f0d0ccefdfa3342cbe
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Laszlo Agocs
Simon Hausmann
Gunnar Sletta
Gunnar Sletta
Lars Knoll
Tasuku Suzuki
Pelle Johnsen
Sebastian Sauer
Albert Astals Cid
Mitch Curtis
Jani Heikkinen
Gatis Paeglis
Alan Alpert
Erik Verbruggen
Topi Reinio
Alex Montgomery
Venu
Robin Burchell