Fix flaky test on Windows for scalar.Retry

[victornoel]

The test scalar.RetryTest.runsScalarMultipleTimesWithWait fails only sometimes: https://github.com/yegor256/cactoos/pull/1491/checks?check_run_id=1336830111

We've seen it only on Windows for now and only once.

Let's make the test more robust, either with a simple fix if possible or replacing it with something different that test the same functionality.

[victornoel]

@0crat in

[victornoel]

@0crat status

[0crat]

@0crat status (here)

@victornoel This is what I know about this job in C63314D6Z, as in §32:

• The job #1497 is in scope for 7mon
• The role is DEV
• The job is not assigned to anyone
• The budget is not set yet
• These users are banned and won't be assigned:
  • @victornoel/z: This user reported the ticket
• Job footprint (restricted area)

[pnatashap]

also faced with this error and add an iteration number to the error message. The error message was

Expected: a value less than or equal to <2024-02-20T15:21:39.556203100Z>
     but: <2024-02-20T15:21:39.556203700Z> was greater than <2024-02-20T15:21:39.556203100Z>

So the difference is 0.0000006 sec = 0.00006 milli sec
I think that is just system accuracy. May be good to compare time up to 0.1 millisec and it should be good