Edge: Ignore multiple gotFocus events if we were the one who caused them

[sratz]

Follow-up on #1849 / #1848.

[github-actions]

Test Results

  200 files  ±0    200 suites  ±0   27m 21s ⏱️ + 1m 40s
4 742 tests +2  4 718 ✅ +1   24 💤 +1  0 ❌ ±0 
6 880 runs  +6  6 712 ✅ +1  168 💤 +5  0 ❌ ±0 

Results for commit 3344006. ± Comparison against base commit f20619c.

[HeikoKlare]

The change looks sound to me. We had already shortly discussed it in #1849 and just didn't do it without knowing about a concrete case that needs it to keep complexity and risk as low as possible.
I also checked again that ignoreGotFocus is only accessed from the UI thread, such that no memory consistency mechanisms need to be applied.

Just a nit question: shouldn't the test succeed on every browser (not only Edge), so that removing the assumption may be reasonable?

[sratz]

Just a nit question: shouldn't the test succeed on every browser (not only Edge), so that removing the assumption may be reasonable?

In principle, yes. But timing is slightly different. Also, for some reason on Linux/WebKit I am receiving 2 Events for each Browser.setFocus().

Not sure what exactly is going on there, haven't looked into this further. It's probably fine and the general assumption that "1 SWT.FocusIn event per Browser.setFocus()" just doesn't hold generally.

So only for the Edge case we know that should be the case and can attribute a mismatch to a regression for this particular issue.

I propose to proceed with having the test flagged Edge-only, WDYT?