Churn is a measure of change in defect reporting between two sequential Coverity Analysis Feature releases (for example, 2017.07 and 2018.01).
For most programming languages, churn can be discovered through a two-step process:
- Using the two Coverity Analysis releases to separately analyze the same code base.
- Using Coverity Connect to triage the resulting CIDs into the following classifications: False Positive and Bug. Churn does not apply to the CIDs that have Pending or Intentional classifications.
Churn formula:
(New False Positives + Lost True Positives)/Total Number of CIDsNote: Coverity expects the churn between contiguous Coverity Analysis releases to be less
than 5% for the following programming languages:
- C/C++
- C#
- Java
- JavaScript
- New False Positives (FPs): The number of CIDs that are reported by the more recent version of Coverity Analysis and that developers marked with an FP classification in Coverity Connect.
- Lost True Positives (TPs): The number of CIDs that are reported by the earlier version of Coverity Analysis and that developers marked with the Bug classification in Coverity Connect.
- Total Number of CIDs: The number of CIDs produced by the separate analyses. This total includes CIDs that are common to each version of Coverity Analysis and any unique CIDs from either version. Common CIDs are the result of defect merging in Coverity Connect.