Abstract: Code coverage testing of an application (e.g., to determine which blocks of source code are executed during run-time testing) in an operating system is accomplished using instrumented code and a performance analysis profiler. That is, non-executable code statements (e.g., T-SQL in-line comments) are injected into the source code at respective executable statements, and metadata is generated for respective source code elements. The performance analysis profiler monitors the testing of the application, generating trace data. Trace data is combined with metadata to generate code coverage reports for the application's source code, which provide, among other things, an indication of the thoroughness of the test (e.g., number of available application instructions that are actually executed during the test).