| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
Was never unskipped after being fixed
Task-number: QTBUG-47523
Change-Id: Ibb193b45a73ca627ab3db7db8d1bf7b98f9ee8b4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
When inserting a TableViewColumn after the TableView is created, then
it needs to be inserted into the list of columns at the specified index
so that it responds correctly to things like resizeColumnsToContents()
Fixes: QTBUG-58594
Change-Id: If7ff545306fe8b6616d8e016eb87e565ed40c836
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
|
|
|
|
|
|
|
|
|
| |
Instead of using a timer to terminate the benchmark, let's terminate the
benchmark after we've sent off the first rendered frame. This is a bit
more realistic, although it doesn't change numbers on my machine.
Change-Id: I9a3724ecf781191d2b4dd937a5b74960a19550df
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
|
|
|
|
|
|
|
|
|
| |
Don't measure the time it takes to destruct the QQmlEngine, as that may
be heavy work due to the last sweep of the garbage collector heap and
deleting all engine-owned QObjects.
Change-Id: Ibdd3d92e306d0418ee7080077b8894950878ec43
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
doLayout() updates the dimensions of the scroll view, and therefore can
trigger the scroll bars to be shown or hidden. That, in turn, can
trigger another call to doLayout(), resulting in binding loops. Avoid
those by deferring the call.
The scrollview tests are broken without this change because the binding
loop prevents the scrollbars from being shown or hidden in some cases.
The deferring necessitates a tryVerify(), but fixes the actual tests.
Change-Id: I0dac0ba380240e025237f69c4d357d764a535c3d
Fixes: QTBUG-50605
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add an actionChosen() signal to DefaultDialogWrapper, allowing the
developer to stop further processing of the event by setting the
"accepted" field to false. This is especially useful when some
validation on the dialog contents need to take place before the dialog
can be accepted.
[ChangeLog][Dialogs] Add a signal to the Dialog class to allow the
client to intercept the button presses and optionally prevent further
processing of the event. This allows performing some validation on the
fields before dismissing the dialog.
Task-number: QTBUG-69095
Change-Id: I19bca0bd9fcbafc72d337a5870776a96634ba748
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
|\
| |
| |
| | |
Change-Id: Ie706f27a0c43ec1d1136310fba196d9b41a184ed
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
be4c4877b35ad1b6983cf194e61f0dad1d123af5 fixed Context2D's
createRadialGradient() function, so our usage of it has to be updated
accordingly. This patch results in the same visual result as it was
originally (before the auto test failure).
Task-number: QTBUG-68301
Change-Id: Ibd8aa24e84ceecccf25d7656350738a38cd03e24
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The issue that was causing the tests to fail was most likey fixed by
12d80048. 41669ca6 unblacklisted other tests that were fixed by the
that commit.
Task-number: QTBUG-46393
Change-Id: Ib81e7b1afe0a60961c0b8ad3201802389c7bab01
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\ \
| |/
| |
| | |
Change-Id: I4194dd92ef851407e667bd1c635367a1bae97480
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The test seems to have started failing due to high DPI issues, but
until the root cause has been found, unblock CI.
Task-number: QTBUG-68301
Change-Id: Ia800eb2b568ee0abd35937e436b213321d035290
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\ \
| |/
| |
| | |
Change-Id: I00b66b03717563bb700f8a73c3bc295a55ac8c6a
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When the wheel was used it could cause any binding on value to be broken
since it was updated directly. Going via range.position ensures this is
not the case.
Change-Id: Ic041ea911cee68de34f5c4e184cf8460f5e002da
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: Ibc95b14faaa77f200cf6ea5e05cb227e6a3628ca
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
|\ \
| |/
| |
| | |
Change-Id: I4b1a4ebf258c9d31e537f42c8d0219c4294628bb
|
| |\
| | |
| | |
| | | |
Change-Id: I2d12dd7b271e17662295007f5cf25ff4b3b78e8e
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Calendar was implemented in such a way that pressing on a date selects
that date, rather than releasing or clicking. With the current code,
this presents issues when the month changes as a result of the press.
For example, clicking on a date in an adjacent month will result in the
clicked() signal passing the date under the mouse in the new month,
instead of the originally pressed date.
This patches fixes the issue by storing the pressed date and using it
when emitting the released() and clicked() signals.
Task-number: QTBUG-54129
Change-Id: I0c16293033b77f6ae783b5365d198b4a516af90b
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| | |\
| | | |
| | | |
| | | | |
Change-Id: I556a8404b908c15eaa71b009319c7f27274b40f1
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The qmlbench repo has benchmarks only for a subset of QQC2. They are
part of a larger collection of benchmarks that are run periodically, for
which the results are collected to Grafana. We have limited resources
there, so we don't want to overload it with benchmarks for many QQC1
types. However, for our own benchmarking purposes it is useful to be
able to easily benchmark these types.
Change-Id: I9124791a93f07346dc80f1f08aedd655ed0cfa02
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The "triggered" signal from the ComboBox menu items is handled twice:
once in ComboBox.qml, and once in QQuickMenuItem. The function of the
latter is to update the parent QQuickMenu's selected index.
With the current change we guarantee that the selected index (that is
exposed as "currentIndex" in the public API) is updated only once,
before any activated() signal is emitted. This also allows the handler
of the activated() signal to update currentIndex at will, without
running the risk of having it overwritten again by the ComboBox internal
code.
Task-number: QTBUG-51113
Task-number: QTBUG-43050
Change-Id: Ibfe12752d02eb4c9e2bf9a73ee8b3ab0ca8d7eaa
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
|
|\ \ \
| |/ /
| | |
| | |
| | |
| | |
| | | |
Conflicts:
.qmake.conf
Change-Id: Id56429b283aff90a18576cbd769b96fa1f0781fa
|
| |\ \
| | |/
| | |
| | |
| | |
| | |
| | |
| | | |
Conflicts:
src/dialogs/qquickabstractfiledialog_p.h
tests/auto/controls/data/tst_combobox.qml
Change-Id: I62e54d0a2e89d987e05e8dfad6ae6aac1a32cc72
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
There was a bogus timer firing once after the test had been running 500 ms. The
timer was supposed to be used as a workaround on Mac, but since it had running
set to true, it was actually firing once on all platforms.
The timer was originally added in 6254af5. The test function ended up being
skipped on Mac, so the timer was actually pointless.
Task-number: QTBUG-64664
Reviewed-by: Liang Qi <liang.qi@qt.io>
(cherry picked from commit a749525164b771452599e67e277d2ea6b3866479)
Change-Id: Icebda05c872ba2de690e87bd0a9968670f203ce2
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Task-number: QTBUG-47243
Change-Id: I19d06df099a9da253059d2d0aa4a290b41a17ada
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
QtDeclarative was fixed with 781caafe1fde71b059c0e3a42bda77ce0d7e4c2a
and tests pass now.
Task-number: QTBUG-65521
Change-Id: Ibb88e17a182786f8299a6ab00bd74862aba548ce
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Test was skipped for all the targets when it should only be
skipped on offscreen platform.
Task-number: QTBUG-65307
Change-Id: Id6ef2aefcf885447d4529caecdaf3514e52b3ebd
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Lots of tests fail when qemu crashes with output:
"tb_lock: Assertion `!have_tb_lock' failed." These need to be
fixed with QTBUG-65211.
Task-number: QTBUG-65307
Change-Id: I7d7249d2e7c9a18d10e15a6e79f34dc1ec93e5d6
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
|\ \ \
| |/ /
| | |
| | | |
Change-Id: Ie66ca17fe8ea49b3637accb468e206657346f4ab
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The default handle size is based off the height of text on the system
that the test is run on. By making the size fixed, we remove the
possibility for small differences in compare results.
Task-number: QTBUG-65224
Change-Id: I14061a028360340c6b284fb450bf1252014dc9b2
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| |\ \
| | |/
| | |
| | |
| | |
| | |
| | | |
Conflicts:
.qmake.conf
Change-Id: If27d9c3cbc17765c4180651189c4b8d19b8409da
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
I fixed a similar issue last year on the JavaScript side:
094ad30c94a72784449f43ef06d2172d644ab0fd
At this time the same thing has been happening in C++. See QTBUG-54559
and the patch for a detailed explanation of the issue. In short,
the problem with Calendar is that internally it uses QDate which does
not keep information about time, whereas in the JavaScript world date
is always combined with time. So, QDate(2017-10-15) is valid, but when
during QDate -> JS Date transformation we add time to it (which defaults
to midnight (00:00)), it becomes invalid in time zones where the Daylight
Saving Time -> Standard Time transition takes place at midnight.
To avoid switching the entire QQuickCalendarModel1
to using QDateTime, I modified its date(...) and dateAt(...) methods to
return QDateTime with the time part always set to 12:00.
That transformation required more changes in QQuickRangedDate1, because
Calendar::selectedDate internally is QQuickRangedDate1::selectedDate,
and I also had to fix "selected" property binding in the Base Calendar
Style as it did take into account the time part of the date, which
resulted in wrong behavior with my changes.
Task-number: QTBUG-64068
Change-Id: Ia2f7703ff4e5811ef79438c97739da1d8001a7f5
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The ScrollView is not a control widget, which should be included in the
focus chain when the user presses the Tab key; instead, it's a container
and as such it should not interfere with the navigation.
We also modify TableView, because it derives from ScrollView but we
want to keep its current behavior.
Task-number: QTBUG-64596
Change-Id: Ibd7833603d38171693b2f34c5859e9c4615b8ed4
Reviewed-by: Liang Qi <liang.qi@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
There was a bogus timer firing once after the test had been running 500 ms. The
timer was supposed to be used as a workaround on Mac, but since it had running
set to true, it was actually firing once on all platforms.
The timer was originally added in 6254af5. The test function ended up being
skipped on Mac, so the timer was actually pointless.
Task-number: QTBUG-64664
Change-Id: I5b4803ba49db0ab7e1b4dd632a34d51af59bbf91
Reviewed-by: Liang Qi <liang.qi@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Test was fixed by commit c84f7554894cef0cbeeba8a61f441286e190a797
in qtbase.
Task-number: QTBUG-64523
Change-Id: I86195fe049c61ceab43566d004c8f6e6bfcee041
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
|\ \ \
| |/ /
| | |
| | | |
Change-Id: Iadff14f78c9d4b5aca9cba8a2728c10b7963d8fc
|
| | |
| | |
| | |
| | |
| | |
| | | |
Task-number: QTBUG-64523
Change-Id: Icc369c1933a725321927f46255bb396800a0ac5d
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| |\ \
| | |/
| | |
| | |
| | |
| | |
| | | |
Conflicts:
.qmake.conf
Change-Id: Ia33462a31303ae79f7a6ef26dc6065a0afeef3e1
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I03c222a72490959d55b3077a6282553a4bfcf7d2
Reviewed-by: Jani Heikkinen <jani.heikkinen@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This is exactly the same as the minimum change, when the maximum is
changed, the handle position needs to be re-evaluated.
Task-number: QTBUG-63354
Change-Id: I13a67451190a3bb7915afe8e89501097c1dd3301
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I1fac87fd96898fed487f1621c1d1bdc4c44de1cb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Allowing for testing menu positioning in addition to combo dropdowns.
Change-Id: I0b7ef259a15d8773e82218e1f8a30f15d23e6644
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
[ChangeLog][FileDialog] Added defaultSuffix property
Task-number: QTBUG-39230
Change-Id: I7dc73c332ad62129d7158b6f6f5a614b8582ffbc
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| | |
The x,y params of mouseClick are relative to the item
It makes no sense using the x,y of the item as those parameters
Change-Id: I0cda5053758caec97ad7d1738978bc295ece8ae9
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The tests for boot2qt were disabled with commit
555750221ee3850ba70169a2b6854041b434c592. The tests can
be enabled since qtdeclarative now fallbacks to software
renderer if OpenGL is not supported.
Changes needed:
- Using grabImage does not work correctly on offscreen platform, so
tests relying on it are skipped
- Extra 1ms delay is needed between two mouse clicks to get rid of
test_activeFocusOnPress() flakiness on software quick renderer
Change-Id: Ie985110c171bc395bac0d63cd2fbb94207c63516
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\ \
| |/
| |
| | |
Change-Id: I6fab4a2531f013291329f2277849f3ef251f0eda
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This seems to help get rid of the ApplicationWindows that would
otherwise stick around until the end of the controls tests.
Change-Id: I0f5cc81e07c727070c3f80654d46881c696eebe7
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Liang Qi <liang.qi@qt.io>
|
|\ \
| |/
| |
| |
| |
| |
| | |
Conflicts:
.qmake.conf
Change-Id: Ie6ad3e9490dcf85bd5deb4c6d7004a4aa8b81433
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Increase QtQuick import version to 2.6 to enable the properties of
underlying QML Text type to be usable in the QML Label type too.
Task-number: QTBUG-57895
Change-Id: I7a2b641a56c90ab91b4e6c973278d8943fa6903a
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This is not an automated benchmark, it is meant to be run manually or by
a script.
Change-Id: I72a4c15120ff4a650fb4bd2f90c6db1907dba994
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This enables them to be run with "make benchmark"
Change-Id: I13ecd96087dd6586a47fb2abab269982f04803db
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
|