It is important to regularly run tests to keep control of user-visible changes incured by code modifications. The tests that are part of synopsis are not an absolute measure for success or failure, but rather reflect a given state of the system at a particular point in time.
It sometimes happens that a new addition to the code will make a particular test fail, simply because that test now produces a different output compared to what it used to produce before.
This may indicate a true regression, or it may mean that the expected output should be adjusted if the new output is valid and should be the new reference.
The 'official' regression test results are available at all times as a test report.