1678303 - testing/mochitest/chrome/test_sanityEventUtils.xhtml fails with --headless, passes without

Status: RESOLVED FIXED

(Testing :: Mochitest, defect)

Description

[Andrey Bienkowski]

Headless

./mach test --headless testing/mochitest/chrome/test_sanityEventUtils.xhtml

Overall Summary
===============

mochitest-chrome
~~~~~~~~~~~~~~~~
Ran 2 checks (1 subtests, 1 tests)
Expected results: 1
Unexpected results: 1
  subtest: 1 (1 fail)

Unexpected Results
------------------
testing/mochitest/chrome/test_sanityEventUtils.xhtml
  FAIL uncaught exception - NS_ERROR_XPC_GS_RETURNED_FAILURE: ServiceManager::GetService returned failure code: at synthesizeDrop@chrome://mochikit/content/tests/SimpleTest/EventUtils.js:2753:56
test@chrome://mochitests/content/chrome/testing/mochitest/chrome/test_sanityEventUtils.xhtml:109:16

simpletestOnerror@chrome://mochikit/content/tests/SimpleTest/SimpleTest.js:2112:18

Not headless

./mach test testing/mochitest/chrome/test_sanityEventUtils.xhtml

Overall Summary
===============

mochitest-chrome
~~~~~~~~~~~~~~~~
Ran 19 checks (18 subtests, 1 tests)
Expected results: 19
Unexpected results: 0
OK

Comment 1

[Andrey Bienkowski]

Based on the above stack trace this is the failing line:

https://searchfox.org/mozilla-central/rev/7b40f0b246ad0b54975b1525811f2ad599b95f33/testing/mochitest/tests/SimpleTest/EventUtils.js#2742-2744

  var ds = _EU_Cc["@mozilla.org/widget/dragservice;1"].getService(
    _EU_Ci.nsIDragService
  );

Comment 2

[Andrey Bienkowski]

What I found out so far

1. The headless mode uses a separate "widget backend" located in widget/headless.

2. There are 3 backends that have nsDragService.h

   widget/cocoa/nsDragService.h
   widget/gtk/nsDragService.h
   widget/windows/nsDragService.h

3. There is currently no widget/headless/nsDragService.h

Comment 3

[Andrey Bienkowski]

Attachment: Bug 1678303 - Mark testing/mochitest/chrome/test_sanityEventUtils.xhtml as failing in the headless mode r=ahal

Comment 4

[Andrey Bienkowski]

I added a patch to mark this test as failing in the headless mode. After it lands (if it does) I should be removed as an assignee because I currently have no plans of implementing nsDragService for the headless mode.

Comment 5

[Andrew Halberstadt [:ahal]]

I think we can let the bug resolve itself after this lands since this test doesn't make sense to run in headless mode anyway.

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/bd6a5134912f
Mark testing/mochitest/chrome/test_sanityEventUtils.xhtml as failing in the headless mode r=ahal DONTBUILD

Comment 7

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/bd6a5134912f