| So we use pipeline to build several platforms of our codebase, and prod/nonprod etc. They all produce a unit test report, and the paths to the result files all contain the specific parallel branch they were built in. I was hoping this would be enough for it to properly identify the tests, but it seems that unless the tests have the build specific info in the test name, the graphing and result display of the tests gets very confused and will combine some tests based on race conditions which leads to a very spiky looking graph. I confirmed this by using the jest config in a node project we build the same way. Jest lets you set the pattern used to build the suite name and test name. Once I set this to include the build info, the graph flattened out and remained consistent. In other projects we use other frameworks that you support googletest, etc. I don't know if each of them support the option to configure the test suite/name patterns... Could you please look at how the plugin could properly handle parallel builds for test display and graphing? |
