[BenchMark] Number of Test Case Dismatch

[Cypher-Bruce]

Describe your question

When benchmarking online or locally for step 19 Test VideoService#reviewVideo(AuthInfo, String), my passcnt is 3. There is only 1 wrong answer debug info in the console. So the actual number of test case is 4? However, both the benchmark_small.json and OJ tell me number of test case is 35.

I try to add this code at the beginning of my reviewVideo function:

debug ++;
log.debug("VideoServiceImpl::reviewVideo: debug {}", debug);

where debug is a static variable for the impl class outside any function.

And the output in the console confirms there are only 4 test cases:

[hezean]

除了ser里的几个测试点，我们还测了review之前post的视频，这部分测试是并发的，你的debug变量如果是普通的int/long会没办法正确统计，因为race condition了，需要用atomic int来统计，参考benchmark service的写法

而且如果前面post video错了导致posted video记录少了，会影响到这个step的结果

[Cypher-Bruce]

Thank you for your reply.