571c407ea6
Currently the delivery of pointer events to childMouseEventFilters differs depending on if the event is a mouse, touch, or synthesized mouse event. If case of a mouse event, it will be sent to all the filters up the parent chain, even if one of the filters along the way returns true. If it's a touch event, propagation will stop as soon as a filter returns true. What does it mean that childMouseEventFilter returns true? According to tst_QQuickWindow::testChildMouseEventFilter(), if a childMouseEventFilter returns true, the event should be stopped from being sent to the receiver, and only the receiver. It should not be stopped from propagating to other childMouseEventFilters up the parent chain. This is explicitly tested with data row "r1 rejects and filters". Therefore, in order to make testChildMouseEventFilter() pass, not only for mouse events, but also for touch and synthesized mouse events, this patch will make the following changes: 1) Remove the early 'return' statement after a touch event was filtered by a childMouseEventFilter. This will make sure that the touch event will continue to propagate to parent childMouseEventFilters, equal to how it works for mouse and synthesized mouse events. 2) For both touch-, and synthesized mouse events, we deliver a (localized) copy of the original touch event to childMouseEventFilter(). The filter can then choose to accept or ignore this copy. But as it stood, we would never sync back the accept state from the copy to the original event. The result was that the original event would continue to propagate, regardless of accepted state set by the filter. This patch will therefore sync the accepted state from the copy back to the original event, when the event is filtered. This will make sure that if a filter e.g ignores the event, the receiver will not receive the event (since it was filtered), and the event will propagate to the parent (since it was ignored). Which is equal to how it works for mouse events. 3) For both touch and synthesized mouse events, we used to always set an exclusive grab on the affected event points if a childMouseEventFilter filtered an event. This is different from mouse event delivery, where we only set a grab when the event is also accepted. And the latter is also (most likely) the correct thing to do; If the event is ignored, it means that the filter says (on behalf of the receiver) that it doesn't want the event. And it doesn't make sense then (AFAICS) to still grab the event points. This patch will therefore, equal to mouse event delivery, ensure that we only give a filter an exclusive grab on the touch points when the event was actually accepted. With these changes applied, we then also change the tst_qquickwindow::testChildMouseEventFilter() to run three times, once for mouse event, touch events, and synthesized mouse events, to verify that they're all aligned. Fixes: QTBUG-115953 Pick-to: 6.7 6.6 6.5 Change-Id: I8b5b1faadc907e804b7e21c667888db7cfe28872 Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io> |
||
|---|---|---|
| .. | ||
| auto | ||
| baseline | ||
| benchmarks | ||
| global | ||
| libfuzzer/qml | ||
| manual | ||
| system | ||
| testapplications | ||
| CMakeLists.txt | ||
| README | ||
README
This directory contains autotests and benchmarks based on QTestlib. In order
to run the autotests reliably, you need to configure a desktop to match the
test environment that these tests are written for.
Linux X11:
* The user must be logged in to an active desktop; you can't run the
autotests without a valid DISPLAY that allows X11 connections.
* The tests are run against a KDE3 or KDE4 desktop.
* Window manager uses "click to focus", and not "focus follows mouse". Many
tests move the mouse cursor around and expect this to not affect focus
and activation.
* Disable "click to activate", i.e., when a window is opened, the window
manager should automatically activate it (give it input focus) and not
wait for the user to click the window.