We use the FxCop-runner to perform code-analysis for our project. We currently have quite a large number of errors and warnings that we hope to deal with in the coming months and have set the error and warning limits accordingly to keep te build in the green.
However, since a few days the runner consistently fails, because it reports an incorrect number of errors and warnings. When viewing the details of the report, it reports the correct numbers again which are within the set limits.
The build log contains inconsistencies regarding the number of errors.
For instance, the build log of the first failure - a few days ago - reports the following:
[10:43:21]: Errors limit reached: found 3445 errors, limit 3100
[10:43:21]: Warnings limit reached: found 15308 warnings, limit 11000
[10:44:54]: Total: 10964 (+0 -0) Errors: 3048 (+0 -0)
I've compared the fxcop-report with the report from the previous build that still passed and noticed no differences (except for the dates/times).
To my knowledge, there were no changes in the configuration of this buildtask in TeamCity.
Does anyone have an idea why TeamCity suddenly reports these incorrect number and what we can do to fix it?