qt/qtdeclarative.git, branch 6.7.2

Update dependencies on '6.7.2' in qt/qtdeclarative

2024-06-07T21:28:01+00:00

Change-Id: I4561ca428f7204827c7a704275cac56a08f34a27
Reviewed-by: Qt Submodule Update Bot

Update dependencies on '6.7.2' in qt/qtdeclarative

2024-06-06T03:48:28+00:00

Change-Id: I153ee6983f78023984a01175b4bf9d49f462296d
Reviewed-by: Qt Submodule Update Bot

Update dependencies on '6.7' in qt/qtdeclarative

2024-06-03T03:46:09+00:00

Change-Id: I32c52a84a17577ca949bbe6f66d7cce45a712f9d
Reviewed-by: Qt Submodule Update Bot

Update dependencies on '6.7' in qt/qtdeclarative

2024-05-31T03:40:36+00:00

Change-Id: Iefd020971414bc6e970bdc911ebab53bfb52a838
Reviewed-by: Qt Submodule Update Bot

DelegateModel: Fix crash when model changes during model reset

2024-05-30T23:25:53+00:00

There are two issues here at play that interact to a crash. First is using
sender() inside the slot of a single shot connection. This doesn't work
reliably as inside doActivate the connection is removed before invoking
the slot. However QObject::sender() will look through all connections to
find one with the same sender as the currentSender otherwise it returns
nullptr. Which means as long as there are other connections to the same
object sender() would still work. This brings us to the second issue,
if the model changes between modelAboutToBeReset and modelReset, down
the line the signal from the old model would still be handled and
since there were no other connections to the old model anymore crash.

Change-Id: I71af46ec691e0c130cf367ac9fa04308c0826481
Reviewed-by: Ulf Hermann 
(cherry picked from commit 1c2ff06654bf33dffbceaadcf777134a9121d5e4)
Reviewed-by: Qt Cherry-pick Bot

Don't update transform if it hasn't changed

2024-05-30T23:25:52+00:00

Updating the transform marks it as dirty and causes rebatching.
Since we were doing this even when nothing had changed, we were
essentially rebatching every time there was any update to the
shape. We should leave it to the node to decide if it has
been updated.

This adds two tests that verify that rebatching still happens
when changing the color/gradient of the shape.

Change-Id: Iefb7a21a89ac2c31053e473cf689fa06e34a40d5
Reviewed-by: Eirik Aavitsland 
(cherry picked from commit e3d0233bbe3e130c72e3957c45007097c019530d)
Reviewed-by: Qt Cherry-pick Bot

QtQml: Document and uphold precondition of metaTypeFromJS()

2024-05-30T07:55:41+00:00

The value needs to be a default-constructed instance. Otherwise a number
of branches in this method produce unwanted effects, such as appending
to an already existing array rather than creating a new one.

Amends commit 1b89c1edcae68351632c2755e5408410c2ff98e3.

Pick-to: 6.5
Fixes: QTBUG-125429
Change-Id: If175a02b3a794573abc03df206fbddd41f2855b4
Reviewed-by: Fabian Kosmale 
(cherry picked from commit c16a3e5adb59f6da5e39e51ca15a4d5324d68d1c)
Reviewed-by: Sami Shalayel

Doc: Capitalize Singleton in code snippet

2024-05-29T19:25:00+00:00

Writing singleton with a small 's' leads to a compile error.

Fixes: QTBUG-125127
Change-Id: I7edc4512aceea34bcf2a0884c6b6e48ab3069f19
Reviewed-by: Fabian Kosmale 
(cherry picked from commit f4fc10172e78a66de15ed67fe65ac5e03c64950c)
Reviewed-by: Qt Cherry-pick Bot

quick_test_main_with_setup: only activate windows where supported

2024-05-29T19:24:59+00:00

Otherwise we get warnings before initTestCase, meaning not even
failOnWarning can catch them:

"qWaitForWindowActive was called on a platform that doesn't support
window activation. This means there is an error in the test and it
should either check for the WindowActivation platform capability before
calling qWaitForWindowActivate, use qWaitForWindowExposed instead, or
skip the test. Falling back to qWaitForWindowExposed."

Task-number: QTBUG-124230
Pick-to: 6.5
Change-Id: I8111928917b94d51ac9b94f470f5837a98f44a71
Reviewed-by: Doris Verria 
(cherry picked from commit 0a458847e9d8dffadfbb8de349e2d26e4ce34bdb)
Reviewed-by: Qt Cherry-pick Bot

graph example: Do not use GLSL functions as uniform names

2024-05-29T15:03:20+00:00

Using textureSize in the uniform block leads to renaming it
to _textureSize when transpiling to GLSL. This is not
reflected in the reflection metadata for some reason.
For OpenGL this matters because the QRhi backend won't
be correctly able to set up the uniforms when the
names do not match. To overcome this, use a name
in the shader that does not conflict with a built-in
GLSL function. Strictly speaking the name (textureSize)
is somewhat incorrect anyway. Use texCoordScale to
express its meaning better.

Pick-to: 6.6 6.5 6.2
Change-Id: I863e98c6d56cc46fad895763b9b5106a40e2952c
Reviewed-by: Andy Nichols 
(cherry picked from commit 3d8c86976844bc9507b0ff677188eee64e49b732)
Reviewed-by: Qt Cherry-pick Bot