use qWaitForWindowExposed instead of qWaitForWindowActive

Forwarded: no
Last-Update: 2019-04-21

The qWaitForWindowActive() function returns false when the test is
run inside Xvfb and there is no window manager.

Gbp-Pq: Name wait_for_window_exposed.patch

Make sure that tests run with just compiled versions of tools

Origin: Debian
Last-Update: 2020-09-08

Last-Update: 2020-09-08
Gbp-Pq: Name fix_test_remove_qlibraryinfo.patch

Disable all tests that fails with xvfb

Origin: Debian
Forwarded: not-needed - it is a limitation so far inside Debian
Last-Update: 2019-03-14

So far some tests need real OpenGL support

Gbp-Pq: Name disableopengltests.patch

QQuickItem: avoid emitting signals during destruction

Origin: upstream, https://code.qt.io/cgit/qt/qtdeclarative.git/commit/?id=74873324bdf33997
Last-Update: 2023-02-26

If a QQuickItem is in the QQuickItem destructor, then it is both unsafe
and unnecessary to emit property change notifications. Connected code
can no longer rely on the state of the emitting object - if it was
originally a subclass of QQuickItem, then those subclass destructors
will already have run. And the QQuickItem destructor will also have
partially run, leaving the object in an undefined state.

Add a flag that we set to true at the top of ~QQuickItem, and don't emit
visibleChildrenChanged, parentChanged, visibleChanged, and
childrenChanged for items that are partially destroyed already.

[ChangeLog][Qt Quick][QQuickItem] QQuickItem no longer emits change
notifications for the parent, children, visible, and visibleChildren
properties while it is being destroyed.

Gbp-Pq: Name qquickitem_no_signals_on_destruction.patch

fix sweep step for tainted QObject JavaScript wrappers

Origin: upstream, https://code.qt.io/cgit/qt/qtdeclarative.git/commit/?id=e6b2f88d892dcf39
Last-Update: 2022-02-13

Currently, whenever the garbage collector runs, it will destroy all
valid tainted wrappers.

Only null or undefined wrappers will be preserved in the
m_multiplyWrappedQObjects map.

It seems like "!" was overlooked in
3b5d37ce3841c4bfdf1c629d33f0e33b881b47fb. Prior to that change, it
was "!it.value()->markBit()", so calling erase() in the then branch
did make sense. But with "!it.value().isNullOrUndefined()", erase()
will be called for every valid wrapper, which is the opposite what we
want.

Gbp-Pq: Name fix_sweep_step.patch

Support ' in styled text and document it

Origin: https://invent.kde.org/qt/qt/qtdeclarative/-/merge_requests/3/diffs
Forwarded: not-needed
Applied-Upstream: https://invent.kde.org/qt/qt/qtdeclarative/-/commit/0dda47d9f1a22567ad8f1266be2f0cd8a9226c7f

This ensures that some translations really look ok.

Gbp-Pq: Name support_apos_in_styled_text.patch

make tst_qmldiskcache::regenerateAfterChange() pass on big endian systems

Origin: upstream, https://code.qt.io/cgit/qt/qtdeclarative.git/commit/?id=dab0d62b655ce9a4
Last-Update: 2021-02-11

Gbp-Pq: Name tst_qmldiskcache_big_endian.patch

build fixes for GCC 11

Origin: upstream, https://code.qt.io/cgit/qt/qtdeclarative.git/commit/?id=eb6525f126f680f9
Last-Update: 2022-11-05

Gbp-Pq: Name gcc_11.patch

qtdeclarative-opensource-src (5.15.8+dfsg-3) unstable; urgency=medium

  * Backport upstream patch to fix crash when destructing QQuickItem
    (closes: #983597).

[dgit import unpatched qtdeclarative-opensource-src 5.15.8+dfsg-3]

Import qtdeclarative-opensource-src_5.15.8+dfsg-3.debian.tar.xz

[dgit import tarball qtdeclarative-opensource-src 5.15.8+dfsg-3 qtdeclarative-opensource-src_5.15.8+dfsg-3.debian.tar.xz]

Import qtdeclarative-opensource-src_5.15.8+dfsg.orig.tar.xz

[dgit import orig qtdeclarative-opensource-src_5.15.8+dfsg.orig.tar.xz]