Intermittent test failure: `179 - trace.CircularBufferTest.Simulation (Failed)`

[maxgolov]

This test is occasionally failing in CI.

Example run with the failure:
https://github.com/open-telemetry/opentelemetry-cpp/runs/2778747475

[ThomsonTan]

With PR #796, more information on the test failure was included in the output as below, but seems the difference of the arrays is in the ... part, sigh. But it does show that there is a race condition in the circular buffer.

/home/runner/work/opentelemetry-cpp/opentelemetry-cpp/sdk/test/common/circular_buffer_test.cc:158: Failure
Expected equality of these values:
  producer_numbers
    Which is: { 20462, 39215, 229725, 291755, 674915, 718234, 909998, 1150462, 1530501, 1569515, 1671718, 1872383, 1932766, 2076752, 2142344, 2164891, 2220685, 2292977, 2321066, 2329286, 2347442, 2400544, 2525088, 2695210, 2754107, 2782193, 2984922, 2988760, 3005971, 3321013, 3433521, 3595424, ... }
  consumer_numbers
    Which is: { 20462, 39215, 229725, 291755, 674915, 718234, 909998, 1150462, 1530501, 1569515, 1671718, 1872383, 1932766, 2076752, 2142344, 2164891, 2220685, 2292977, 2321066, 2329286, 2347442, 2400544, 2525088, 2695210, 2754107, 2782193, 2984922, 2988760, 3005971, 3321013, 3433521, 3595424, ... }
[  FAILED  ] CircularBufferTest.Simulation (924 ms)

[lalitb]

but seems the difference of the arrays is in the ... part

Murphy's law - the most important debug data never get's logged :)

It seems googletest PrintValue function for the container only prints the first 32 elements. To print all the elements, we need to define the printTo function for container.