Hi Glenn,

Thanks for posting!

Most likely this is due to a difference in the way that the tests are being counted. Many areas of NCrunch (including the corner spinner) will tend to count a fixture as a test, since it has an independent pass or fail status. This means that if you have a fixture that contains one test, this is effectively two tests as far as some aggregation systems are concerned.

There are also situations where tests can be queued for execution twice. This can happen if you have a test currently being executed, and a change is made to the codebase. The change will result in a new execution run for the same test being queued, while the test is still executing with an older version of the code. When NCrunch claims that a test is 'queued', this is inclusive of tests that are already under execution. This can be quite noticeable if you have large tasks in the processing queue consisting of many tests.

There is also a known issue (due for fix in the upcoming 2.8 release) that can sometimes result in tests being queued more than they should be. This can result in a marginal increase in the number of tests queued under certain circumstances (i.e. during the first build/test run of a session).


Cheers,

Remco