Currently, we test on hardcoded versions. It would be better to have a test matrix that covers the set of supported versions.
Pre-requisite is to work on factorizing some of the test logics. We want to have the tests and expected as part of dd-trace-*** repository. Then we want to trigger the analysis in prof-correctness (possibly along with building the images, TBD).
Originally posted by @r1viollet in #4 (comment)