Issue metadata
Sign in to add a comment
|
[EventTiming] Fix tests event-timing-crossiframe.html and event-timing-observer-manual.html |
||||||||||||||||||||
Issue descriptionCurrently the simulated clicks on the child frame's button are incorrectly captured by the main frame's event-timing observer and buffer. However, manual clicks on the child frame's button is correctly captured by child frame's observer and buffer. This behavior causes event-timing-cross-iframe.html to timeout. We need to fix it. Parent bug: crbug/823744
,
Apr 11 2018
,
Apr 30 2018
,
Apr 30 2018
This probably just needs a test fix like the one in https://chromium-review.googlesource.com/c/chromium/src/+/1006528/8/third_party/WebKit/LayoutTests/external/wpt_automation/fullscreen/auto-click.js
,
Apr 30 2018
Though I can't seem to find this test in codesearch?
,
May 2 2018
The CL is still working on progress: https://chromium-review.googlesource.com/c/chromium/src/+/971361 I haven't have time to take a closer look at crbug/820617. But from the fix you gave me it looks like the bug is irrelevant? My current problem is that neither webtestdriver or gpuBenchmarking API could correctly perform a click on an iframe's element. For webtestdriver, it's just not supported yet: https://cs.chromium.org/chromium/src/third_party/WebKit/LayoutTests/resources/testdriver.js?rcl=7270258e2271da0a2fc622c2e8dd9fa9e059f49c&l=62; for gpuBenchmarking, it's as described here in this bug.
,
May 2 2018
,
Aug 24
This bug is listed as "stale" according to the CSM triage guidelines. Max, any update on what the timeline is looking like for fixing this?
,
Aug 24
,
Aug 24
@nzolghadr, do you know whether we can simulate a click on iframe elements yet?
,
Aug 31
The NextAction date has arrived: 2018-08-31
,
Aug 31
Are you using gpu benchmarking or (test_driver) click? There was another bug opened for that as well. Ria was that assigned to you or someone else? I don't seem to be able to find it.
,
Aug 31
I am currently using gpu-benchmarking. I can switch to test_driver if it support it. From this line it seems like clicking in iframe is not supported by test driver yet: https://cs.chromium.org/chromium/src/third_party/WebKit/LayoutTests/resources/testdriver.js?rcl=7270258e2271da0a2fc622c2e8dd9fa9e059f49c&l=62
,
Sep 4
test_driver underneath goes to gpu benchmarking when running the tests on the try bots. Although it comes after that check. So it is worth giving it a try. But you cannot have that submitted to wpt repo as that is Chrome specific.
,
Sep 4
the other bug is crbug/877961, but that one is more viz hit-testing specific because it relies on renderers to submit compositor-frame/ hit-test data to receive clicks
,
Nov 29
,
Nov 29
+Nicolas, I am looking at https://chromium-review.googlesource.com/c/chromium/src/+/1348855. According to #14, we shouldn't have moved the tests to wpt as test_driver is chrome specific. But one point I am not clear is why test_driver is in wpt folder if that's the case. https://cs.chromium.org/chromium/src/third_party/blink/web_tests/external/wpt/resources/testdriver.js
,
Nov 29
It should be fine to use test_driver in WPT, but not call Chrome's gpu benchmarking directly. I don't know what the plan is there for other browsers but I imagine they should each implement their click versions within the test_driver wrapper.
,
Nov 29
I see. Thanks for explanation.
,
Nov 29
test_driver is that standard API that should work for every vendor. It will call into ChromeDriver and then Chrome Dev tools protocol and something similar for Firefox (call into Gecko driver and whatnot). However, for the time being it doesn't do that on our try bots and will go through gpu benchmarking directly. But that is just a hack to get the tests running on the try bots. In the meantime, whatever you see on test_driver APIs you can use in wpt tests.
,
Nov 29
Thanks for explanations. So, we are safe for moving the tests to wpt. For whether we can perform click on iframe's buttons, I will dig into it at some point.
,
Dec 5
|
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by maxlg@chromium.org
, Apr 11 2018