Add a QSGOpacityNode flag which allows a node to ignore inherited opacity

See merge request !31

[qtdeclarative] Add a QSGOpacityNode flag which allows a node to ignore inherited opacity. Contributes to JB#50336

[qtdeclarative] Load images in an OpenGL texture friendly format. Contributes to JB#49193

See merge request !30

Pre-emptively convert images in the pixmap cache to RGBX8888 or
RGBA8888_Premultipled to avoid potentially doing the conversion in the
render thread where it will directly impact the frame rate.

[qtdeclarative] Dump types reachable through internal dependencies. Contributes to JB#46800

See merge request !29

Do not exclude types that are reachable through dependencies, but only
become registered (exported) after the dumped module is imported.  This
happens e.g. when a private submodule exists as a dependency and it
extends types from the public module. The extended types become
reachable when the private submodule is loaded as a dependency, but only
become registered (exported) after the public module is loaded.

[qtdeclarative] Fix crash due to an interaction between an item layer and anchors. Fixes JB#44681

See merge request !28

Evaluating anchors in QQuickItem::componentComplete() will change size
and position bindings and the right code reacting to those changes could
create a layer after QQuickItemPrivate::completeCreate is set to true
which means QQuickItemLayer::classBegin() would not be called, but
QQuickItemLayer::componentComplete() would be meaning
QQuickItemLayer::activate() would be called twice and it would be added
as an item change listener twice. If the layer was then ever deactivated
it would only be removed as a listener once and changes to the parent
size would update the layers size when it wasn't expecting it.

Add API to get more information for each line in a QML Text element

See merge request !27

Previously there was no way to know what area is occupied by each line
in a QML Text element.

This commit adds new API to expose implicitWidth and isLast
on QQuickTextLine for use in the lineLaidOut signal.
It also adds improved documentation to the lineLaidOut signal and
an example usage of the new API to the text layout example.

An example use case of the new API is eg. to allow embedding
timestamps and indicators within a text paragraph, to enable
creating more efficient layouts.

[ChangeLog][QtQuick][Text] Added new API that exposes implicitWidth,
and isLast on the QQuickTextLine for use in the lineLaidOut signal.
This allows the user to layout other items relative to the lines
of text.

Fixes: QTBUG-78277
Change-Id: Ibc754db17c78efb01468106aba32e30d70d2f4df

[qtdeclarative] Package qmlimportscanner with devel-tools. JB#46477

See merge request !26

Without the qmlimportscanner tool available, qmlplugindump fails
(silently!) determining module dependencies, dumping unnecessarily
components provided by those.

This commit exists to ensure a changelog entry exists for CI.

[qtdeclarative] Ask rendering thread to release resources when window is closed.

See merge request !25

Dpi scaled configuration

See merge request !24

On runtime the active dpi is compared to "General/BaseDPI" on
QtQuick2.conf file to multiply configuration items value marked as
scalable. That is done by having configuration item name appended with
"Scaled", e.g.
[QuickFlickable]
FlickThresholdScaled=15

On MouseArea the old, not anymore used, configuration for press and
hold timeout was removed.

This commit can be squashed with earlier configuration support commit
when rebased to newer Qt releasese:
"Allow QtQuick2 parameters to be configured".

[qtdeclarative] Fix crash in ModelNodeMetaObject::emitDirectNotifies. Fixes JB#41922

See merge request !20

If we call get() on a model in a worker thread, we may end up creating a
ModelNodeMetaObject (aka cacheObject). Subsequent mutation of properties
may make us end up in emitDirectNotifies(). However since we can't have
bindings in there, we should shortcut/suppress the notify emission, which
we can do by checking ddata->context via qmlEngine(). The previous code
crashed when qmlEngine() return a null pointer but
QQmlEnginePrivate::get(const QQmlEngine *) would attempt to dereference
the parameter.

Started-by: Slava Monich<slava.monich@jolla.com>
Change-Id: I880619c686436c053692faafa5dba2c96c2ace96
Reviewed-by: Robin Burchell <robin.burchell@crimson.no>
Reviewed-by: Slava Monich <slava.monich@jolla.com>

[qtdeclarative] Keep 5.6.3 version id. Contributes to JB#41362

See merge request !18

Only bug fixes so far and 5.6.4 cmake rules demand qtbase 5.6.4.

This reverts commit ad66f10f.

Merge latest 5.6 changes

See merge request !17

Hopefully fixing the few seen ListModel related crashes. Not including
too many changes.

The generated code for jump-on-strict-not-equal-undefined used the
same logic (but with inverted conditions) as the equal case. For
equality, one can jump to else if the value parts are not the same.
So, for not-equal, if the value parts are the same, it would jump
to the else block if they are the same. Meaning, an encoded int
value of 0 (which is strict-not-equal to undefined) would end up
being evaluated as equal.

Task-number: QTBUG-66832
Change-Id: Id27bb44eccbf39608ae8cebab634c8bcd4c8adfc
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

Assigning to a group property inside a property value source or
interceptor as part of an "on assignment" is perfectly valid. That is
because while "color" is a value type property, the on assignment means
we're actually setting easing.type (in the example and test) on the
property value source, not the color, and that one is a QObject. The
same goes for interceptors.

 Conflicts:
	src/qml/compiler/qqmlpropertyvalidator.cpp
	src/qml/qml/qqmlvmemetaobject_p.h
	tests/auto/qml/qqmllanguage/tst_qqmllanguage.cpp

Change-Id: I505a658977a578894d6dfb00bf5c65b41e42b12f
Task-number: QTBUG-56600
Reviewed-by: Michael Brasser <michael.brasser@live.com>
(cherry picked from commit 2659c308792967322564b5088e0e21bb371e0283)

This is a regression introduced with commit
4876ea6a. Where we previously always
returned the same JS object, we would afterwards return a new JS object
for every invocation, which breaks reference comparison. As we store the
JS wrapper for the list element in the QQmlData->jsWrapper we can avoid
repeated allocations. In order for that wrapper to keep working after
modifications (insertion, etc.) to the list model, we have to replace
the static element index with a reference to the node model meta-object,
which also has an element index that however is kept up-to-date by the
list model itself.

 Conflicts:
	src/qml/types/qqmllistmodel_p_p.h

Change-Id: I4368de6b6d86687fe96fbf73bd60b80b69d7b058
Task-number: QTBUG-52017
Reviewed-by: Michael Brasser <michael.brasser@live.com>
(cherry picked from commit 44a89492b49f23a975377795dbb7a48916cb5081)
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>

This is a combination of 2 commits that went into 5.9. They cannot be
cherry-picked however, because some c++11 constructs were used. Hence
this combo commit, which is a squashed version of those two commits that
have the c++11 bits replaced by Good Old boilerplate code.

Original commit 1 (e29ffa17):

Fix use-after-free when removing elements from a ListModel

Detaching delegate instances from model items is done after the
destruction of said model items. The problem is that after the model
item is destroyed, it will emit a change/destroyed signal. As the
delegate is still referencing the item, this will result in a
use-after-free. To provent that, the items are kept around until after
everyone (notably the delegate model) has been notified of the removal.

[ChangeLog][Qt][Qml] Fix possible use-after-free when removing items from a ListModel through JavaScript.

Original commit 2 (163c5157):

Fix use-after-free when clear()ing all elements from a ListModel

Same problem as the problem with remove(), so now clear will call into
remove to do the correct thing.

[ChangeLog][Qt][Qml] Fix possible use-after-free when clearing all items from a ListModel through JavaScript.

Task-number: QTBUG-63383
Change-Id: I9a6bdf65da63b33833f18c80e74ad7bb93409627
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

This is a regression introduced with commit
3cc589c9, which in turn fixed a leak
with QV4::QObjectWrapper objects. Unfortunately the allocate() call into
the persistent (weak) value storage in the list model introduced a leak
of the weak value itself. This is fixed by replacing the free standing
weak value allocation with the use of the existing jsWrapper weak value
in the declarative data (QQmlData). That weak value is freed property in
the destroy() method of the QV4::QObjectWRapper. The extra QQmlData
allocation is hidden behind a unified allocation, similar to what we do
in void QQmlType::create(QObject **, void **, size_t) const.

Task-number: QTBUG-66189
Change-Id: I5351e3e484542709a6b210e84aa19b14d28e11ad
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
(cherry picked from commit 22d43f74e264626d0c28654c42c91839f9de45b5)
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>

Merge 5.6.3

See merge request !15

When the layout changes, we mark all rows as changed but do not track
where the individual rows get moved.

The only reason why one would want to track the moves is to persist
the current item selection across a layout change. But even the previous
code did not achieve that. I'll create a follow up patch to this one
that also implements this behavior as seen in Qt Widget item views.

Note that removing this code brings a tremendous performance win
on larger models. The repeated calls to _q_itemsMoved triggered O(n^2)
behavior in the number of top items in the model. Even with "only"
tens of thousands of items in the model, a layout change became very
costly and took seconds on a beefy modern desktop machine.

Calling _q_itemsMoved in a loop is bad because it:

- leads to O(N^2) behavior within QQmlChangeSet when merging the small
  moves into the item view's current change set
- potentially triggers tons of binding/property updates when the cached
  model indices are updated in _q_itemsMoved

Removing this slow path, I did not yet find a behavior change to the
previous code. Instead, it just does it all much faster.

Change-Id: I67fa99a1c5d8e05d17497d29391da9458bd9bdd0
Task-number: QTBUG-51638
Task-number: QTBUG-45674
Task-number: QTBUG-53677
Reviewed-by: Daniel Vrátil <daniel.vratil@kdab.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
(cherry picked from 84f61dd2)
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>

Change-Id: Ic878796354a42cb5b6ba29b272c603ce762cf232

[qtdeclarative] Add an environment variable to disable use of BGRA textures. Contributes to JB#39372

See merge request !14

[qtdeclarative] Add an environment variable to disable use of BGRA textures. Contributes to JB#39372

Some hardware while supporting BGRA textures has extremely slow
upload times.  Add an environment variable which will force
the use of RGBA textures instead.

Also try and claw back some of the time lost converting from BGRA
to RGBA by avoiding a coversion to an intermediate
RGBA32_Premultiplied image from non 32-bit formats.

If debugging is disabled, enableProfiler() is an inline empty function.

Task-number: QTBUG-63118
Change-Id: I28fa7b7d3390482e7b0cc6419c5a1f47a5b9916f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

We have to make sure all profiler calls to one adapter are done from the
same thread.

It turns out that all the calls from QQmlTypeLoader are done from the
type loader thread. By using a separate adapter for that, we avoid any
extra locking.

Task-number: QTBUG-62987
Change-Id: Ice400b1c3b7bd920d855ceb1ba0d46417f50348d
(cherry picked from commit 4578a927)
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>