Hi @vbortkevic
the language team has looked into this. The mapping logic applied to golancici-lint reports sets an issue type only once per source (the linter name with those reports) and will then stick to it with subsequent analysis (and across branches in some cases). This report source is the identifier mapped to the SonarQube rule key.
This leads to all issues from the same “source”, i.e. linter to be assigned the same issue type, with cross branch and PR mechanisms making changes difficult to predict (and follow).
So if you really need to distinguish different issues from the same linters into code smells and bugs, at this stage a work around could be to append the intended severity to the linter name in the reports.
We may help with an improvement at some point, but we don’t have any ETA. I’ll update you when I have news.
And I would add a final note about this PR analysis scope initial topic with a link to Clean as You Code: How to win at Code Quality without even trying
Best
Sylvain