1550112 - OSX Mojave (10.14) - widget/tests/test_native_mouse_mac.xul fails in CI on 10.14

Status: RESOLVED DUPLICATE of bug 1137575

(Core :: Widget: Cocoa, defect, P5)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

I am working on getting all our tests running on OSX 10.14 in CI in preparation to upgrade from 10.10.

I found an error in widget/tests/test_native_mouse_mac.xul:
https://treeherder.mozilla.org/#/jobs?repo=try&searchStr=mochitest-chrome%2Cosx&revision=aa6f565ce1649da28a111af74858bca0a9541ab9&selectedJob=243068492

and from the log file:
15:34:39 INFO - TEST-PASS | widget/tests/test_native_mouse_mac.xul | 5 | wrong Y coord for event {"type":"mouseover","target":"box","screenX":170,"screenY":120}
15:34:39 INFO - TEST-PASS | widget/tests/test_native_mouse_mac.xul | 5 | wrong target for event {"type":"mouseover","target":"box","screenX":170,"screenY":120}
15:34:39 INFO - TEST-PASS | widget/tests/test_native_mouse_mac.xul | 6 | wrong X coord for event {"type":"mousemove","target":"box","screenX":170,"screenY":120}
15:34:39 INFO - TEST-PASS | widget/tests/test_native_mouse_mac.xul | 6 | wrong Y coord for event {"type":"mousemove","target":"box","screenX":170,"screenY":120}
15:34:39 INFO - TEST-PASS | widget/tests/test_native_mouse_mac.xul | 6 | wrong target for event {"type":"mousemove","target":"box","screenX":170,"screenY":120}
15:34:39 INFO - Buffered messages finished
15:34:39 INFO - TEST-UNEXPECTED-FAIL | widget/tests/test_native_mouse_mac.xul | uncaught exception - NS_ERROR_FAILURE: Component returned failure code: 0x80004005 (NS_ERROR_FAILURE) [nsIDOMWindowUtils.sendNativeMouseEvent] at testMouse@chrome://mochitests/content/chrome/widget/tests/native_mouse_mac_window.xul:132:13
15:34:39 INFO - runNextTest@chrome://mochitests/content/chrome/widget/tests/native_mouse_mac_window.xul:761:9
15:34:39 INFO - callbackFunc@chrome://mochitests/content/chrome/widget/tests/native_mouse_mac_window.xul:121:9

here is the line of code we fail on:
https://searchfox.org/mozilla-central/source/widget/tests/native_mouse_mac_window.xul#132

and a screenshot for this failure:
https://taskcluster-artifacts.net/FDqwDPUEQqGPOkiC1rKA2w/0/public/test_info/mozilla-test-fail-screenshot_KVmigT.png

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

:spohl, I see you as the triage owner for cocoa widgets, could you help me find a solution to fixing this test as it looks to be osx specific and disabling it on osx doesn't seem like a good idea!

Comment 2

[Stephen A Pohl [:spohl]]

Do I need to do anything in particular to get this test to run on 10.14? I built a local copy of Nightly on 10.14, using the 10.11 SDK. When I run the tests under widget/tests, I get the following output for native_mouse_mac_window.xul and the test doesn't actually seem to run:

1:54.75 TEST_START: widget/tests/test_native_mouse_mac.xul
1:54.83 INFO must wait for load
1:54.83 INFO must wait for focus
1:55.19 GECKO(38549) MEMORY STAT | vsize 8758MB | residentFast 527MB | heapAllocated 166MB
1:55.30 TEST_END: Test OK. Subtests passed 0/1. Unexpected 0

Comment 3

[Joel Maher ( :jmaher ) (UTC -8)]

this is what we currently have as the test is skipped on osx 10.10 (i.e. never ran anywhere in CI for the last year):
23:08:56 INFO - TEST-START | widget/tests/test_native_mouse_mac.xul
23:08:56 INFO - TEST-SKIP | widget/tests/test_native_mouse_mac.xul | took 0ms

as you are not getting focus, I am not sure why that is the case- I haven't seen that error in working on the tests in CI to run- maybe there are permissions settings we have differently in CI? I personally don't have access to an OSX machine, so I need to rely on pushes to try in order to test theories and get data.

So you are seeing the test run to some degree, although it appears to do nothing.

When this testcase was disable in bug 1137575 4 years ago, it was having different failures. Possibly we should treat this as a new test given the history?

Comment 4

[Stephen A Pohl [:spohl]]

(In reply to Joel Maher ( :jmaher ) (UTC-4) from comment #3)

When this testcase was disable in bug 1137575 4 years ago, it was having different failures. Possibly we should treat this as a new test given the history?

It looks like bug 1137575 comment 12 describes the same failure that we're seeing now and this test may never have successfully run on 10.10.2 or above. Treating this as a new test (or a test that needed to be rewritten since bug 1137575) seems like the right approach.

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

The priority flag is not set for this bug.
:spohl, could you have a look please?

For more information, please visit auto_nag documentation.

Comment 6

[Stephen A Pohl [:spohl]]

This was disabled in bug 1137575.