![]() ![]() This can be solved by waiting for the synthesized touch event to be dispatched to the page (then we know it has been processed by APZ), then calling promiseApzFlushedRepaints() and finally the scrollTo(). I am trying to adjust the Firefox smooth scroll settings using about:config in. In such a situation, the relative delta from the scrollTo(0, 0) will be combined with the scroll from that last animation sample (often (0, 1)), and result in the scroll offset ending up at (0, 1). When this happens, the test can get to the scrollTo(0, 0) while there is still a repaint pending from the last time the animation was sampled prior to being cancelled. (It was calling promiseApzFlushedRepaints(), but the notification that sends to the compositor can get there before the synthesized touch events are processed, and therefore before any repaints have been enqueued.) The issue seems to be that while the test synthesized a touch-tap to interrupt the animation, it wasn't correctly waiting for the consequences to be propagated to the main thread. I did another Try push from the autoland branch, and this time the failures are readily reproducible (they happened on the first try). Michigan Attorney General Dana Nessel, a Democrat and Michigans first openly gay top statewide official, called the bill an. It would probably be nicer to add a new test API specifically for the purpose of cancelling ongoing animations.) (Having to synthesize a touch-tap to interrupt ongoing scrolling animations feels a bit hackish. With these changes, the test seems to be green on Try. I worked around this by running the tests with apz.allow_double_tap_zooming=false. This then caused another complication, which is that in the case of instant scrolling test cases, two touch-taps could be sent in quick enough succession that they'd be gesture-detected as a double-tap and trigger double-tap-zooming, which also interferes with subsequent test cases. Plexiglass General Rough Cutting Roof Rafters and General Frame Cutting - Plunge Cutting Smooth Scroll and Circular Cutting Masonite - Plywood. The touch-down interrupts all in-progress animations, since a touch is supposed to "hold" the page in place. I looked around for what would do that, and settled on sending a synthesized touch-tap event. We really want to do something that cancels any in-progress animation to properly clean up the current testcase. You will need to change the values of the bottom two rows to modify scrolling sensitivity. The remaining Mac failures were caused by the fact that even if the relative scroll update from window.scrollTo(0, 0) is handled correctly, if the smooth scroll animation is still in progress, we're still going to get residual scrolling from it that interferes with the next test case.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |