Tibor17 commented on pull request #493: URL: https://github.com/apache/maven-surefire/pull/493#issuecomment-1075633525
Hey @SammyVimes , I would rather see your help in the reimplementation of `TestSetRunListener` and `StatelessXmlReporter`. The patches don't make sense because they would have short life anyway and it would very consufe the development. These implementations are not in perfect state and adding workarounds would mean that we would not be sure during the refactoring. Anyway, I appreciate your efforts and I would like to invite you into the refactoring. Meanwhile pls close this ticket and write integration test which fails in current master and open a new PR with the test. This would really help! Pls see my notices from JIRA: > It really does not make sense to report any pull requests against these two classes with small workarounds or patches or fixes for three reasons: 1. they are problematic and they exist for many years 2. we have changes or prerequisite for (3) 3. we have to refactor both classes but but... but as you have noticed the implementation relies on ordering of classes and methods, and the report is finished in one point of when test set has finished and not when the Provider has finished and thus we have problem with re-run feature, this issue, natural ordering. We have this in our roadmap and we have to do one thing, rework these two classes and employ the testRunId and RunMode and not to accept patches because these would kill us definitely. We have to smesh down the current implementation and then verify your requirements are met by the new implementation. So this effort you have made is pushing us but unfortunately it won't help, even if you want to help us, but the turn is on our side. So nowadays we should release M5, and then we should rework the implementation and we can do it together with your ITs in a new PR. What would definitely help us is to provide us with integration tests which fail now in master, and you can do it in a PR even if it would have a broken CI, that's not a problem, and after we have got these classes reimplemented, we would make an agreement to adopt these ITs and push the entire work to the master as one complete solution. Currently accepting the patches into an archaic reporter would be problematic and cannot be accepted, otherwise we as maintainers would blow up the problem to bigger and bigger one, and of course this has to be stopped. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@maven.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
