Fix flaky test in MainLooperLongTaskStrategyTest

[xgouchet]

What does this PR do?

Fix a flaky test in MainLooperLongTaskStrategyTest.

Flakiness arise from the way multithreading can introduce some delays, and some suboptimal assertions:

• Probably a long task would be detected, we used Thread.sleep(duration + 1), which is a bit weird.
  • we adjusted the random duration to always be 1 more than the threshold.
• The assertion used .isCloseTo(duration, offset), which would be the same as .isBetween(duration - offset, duration + offset). This assertion is not perfect as a detected value less than duration is impossible, and wrong.
  • We adjusted it to be .isBetween(duration, duration + 16L)
• The offset itself was set to 10 millisecond, which would prove sometimes to small locally (5% of occurences)
  • We increased it to 16 ms

Note

Although this reduces drastically the flakiness of this test, it can sometimes happen (once every 1000 run or so) that a GC or other stuff will delay the test run, and result in a failing test.

[codecov-commenter]

Codecov Report

Merging #1730 (7e1035c) into develop (9457e96) will decrease coverage by 0.04%.
Report is 6 commits behind head on develop.
The diff coverage is n/a.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #1730      +/-   ##
===========================================
- Coverage    83.41%   83.37%   -0.04%     
===========================================
  Files          465      465              
  Lines        16173    16170       -3     
  Branches      2407     2406       -1     
===========================================
- Hits         13490    13481       -9     
- Misses        2027     2028       +1     
- Partials       656      661       +5     

see 17 files with indirect coverage changes