Consider about:blank as a legal "empty" location in test #1991

[fedejeanne]

This approach needs to be discussed, don't merge it yet!

(Fixes #1991)

[github-actions]

Test Results

   539 files  ±0     539 suites  ±0   32m 10s ⏱️ + 3m 13s
 4 330 tests ±0   4 320 ✅ ±0    9 💤 ±0  1 ❌ ±0 
16 579 runs  ±0  16 477 ✅ ±0  101 💤 ±0  1 ❌ ±0 

For more details on these failures, see this check.

Results for commit 8c352e4. ± Comparison against base commit 0f21a6e.

[fedejeanne]

@amartya4256 I'm not sure this is the correct approach but I already tried using your PR #1973 as a fix and it didn't work.

Do you happen to know if Edge does the same as IE (see footnote 1 in the code) and also triggers an "empty" location even after canceling the event? I am tempted to say "it shouldn't" since otherwise this test would consistently fail, but it only fails intermittently so I think the problem might be timing + asynchronous executions.

WDYT?

[fedejeanne]

Since the issue is that now the asynchronous execution of some methods makes canceling calls impossible, this PR is wrong. We should either revert the decision of doing things asynchronously (which requires major changes) or accept it and remove tests accordingly.