I have a problem with some test cases. If I run them from the browser, they run successfully but if I run them from the test runner they fail sometimes but sometimes not. It is not a problem of process because there isn't any fork in the code. By the way, I checked the names of the process for both codes at all time and they are the same.

What is more, if i add something like "Transcript show: 'something'" at the beginning of each method, they always run successfully.

Any idea??