JingGe commented on pull request #18333: URL: https://github.com/apache/flink/pull/18333#issuecomment-1012166559
> I don't think it's a good idea to break IntelliJ support for this test. Not being able to run the tests in the IDE is very frustrating, and will almost certainly cause people to repeatedly ask why that is. We need to figure out how to make it work in both cases (as it does currently). > > A first step to debugging this would be to just print all the locations which are being checked, once for IntelliJ and once for Maven, and seeing how they differ; I'm pretty sure it's the import options that just need to be adjusted. Me too, It would be great if it could simply work. But I don't think we have any promise to support IntelliJ with any price. It is clear that the import options need to be adjusted. The question is how and how hard. The debug info of ArchUnit is not clean and hard to check and there is almost no javadoc in the ArchUnit codebase. Current ArchUnit rules do not touch the most test classes, therefore is relative simple to handle. BTW, there is also some implicit issue with the current status, since some test classes are considered as production code. What if all the violations found by IntelliJ have been resolved? We can only find them by running the `TestCodeArchitectureTest` via IntelliJ. There is no impact on the `ArchitectureTest`. If anyone later found any new violations for the ITCase, that means they are working on it on purpose, I would recommend to fix them asap. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
