The covariance unit tests (based on the optimization of circle paramters) can be a bit flaky due to our expectations on the covariance matrices and the random numbers generated during the test. As these unit tests were developed to validate our understanding of the covariance calculation (and not to prove the functionality of any calibration-related code), we should disable them (at least temporarily) so they do not continue randomly failing on other valid pull requests
The covariance unit tests (based on the optimization of circle paramters) can be a bit flaky due to our expectations on the covariance matrices and the random numbers generated during the test. As these unit tests were developed to validate our understanding of the covariance calculation (and not to prove the functionality of any calibration-related code), we should disable them (at least temporarily) so they do not continue randomly failing on other valid pull requests