How do I resolve False Positives or False Negatives with Bug Finder?

When reviewing the results of Bug Finder for a coding rule violation (MISRA, CERT-C...) or a defect, it seems that Bug Finder is raising a false violation/defect (false positive), or is missing a violation/defect (false negative). How to deal with this issue in general?

Answer 1

[MathWorks Support Team]

First, check the Bug Finder documentation to see if there is an example that could explain the results given by Bug Finder, especially if a coding rule is interpreted in a specific way. See also Diagnose Why Coding Standard Violations Do Not Appear as Expected.

If you are not using the latest Polyspace Bug Finder release, it would be worth to test with it. If behavior has disappeared, it has certainly been fixed. 

If you cannot find an explanation, or if you are able to reproduce behavior with latest release, please contact MathWorks Technical Support and send:

• A screenshot showing the section of code where you expect the defect or rule violation, or a piece of code that reproduces the false positive or the false negative (recommended) with the analysis log. It will help Technical Support to understand the options used during analysis.
• The results folder containing the FP or FN, if possible (recommended as well to speed up resolution of the TS request).
• The analysis log, if possible, so that Technical Support can understand the options used during analysis.

As soon as Technical Support is able to reproduce the FP or FN, they will forward the reproduction step to Development. It will then, certainly be fixed in a further release.

Until it would happen, we recommend to annotate the finding with a no action planned status (see documentation here: https://www.mathworks.com/help/bugfinder/ug/annotate-hide-known-acceptable-polyspace-results-web-browser.html).