We have a weird situation with one of our repositories. After one of our merge to master SQ analysis has started to show 0% of new code coverege. And only 2 new lines to cover. Although this last push had much more code than those 2 lines of code. Overall code coverage is 80% for this project. Our New Code period is 365 days and our first SQ analysis was 23rd of October 2020.
Must-share information (formatted with Markdown):
- which versions are you using (SonarQube, Scanner, Plugin, and any relevant extension)
SonarQube Community Version 8.2 (build 32929)
- what are you trying to achieve
Proper measurement of new code coverage.
- what have you tried so far to achieve this
I have made master analysis on few different configurations of New Code Period:
Last 365 days
Specific version 23th of October
Specific version 6th of November (which was the next analysis)
In all configurations the results were exactly the same.
Additionally I created a new branch based on version from 23th of October and run the analysis, next I push force master branch to this newly created branch and run the analysis again.
The 2nd analysis was set up to be compared with the 1st analysis on that branch (specific version of new code period).
The result was completly different. It shows 95% new code coverage on 45 new lines to cover which seams to be a valid result.