Hello everyone, As part of the refactoring process, I suggest to completely remove cobertura and sonar from the framework. My proposal is based on the following:
- The startup logic is more complex because of the existence of legacy classes (Instrumenter, InstrumenterWorker, etc ...). - No one (AFAIK) is actively using cobertura or sonar, and the targets in build.xml are actually broken - The way cobertura is integrated with ofbiz is poor and ugly - Before integrating cobertura, ofbiz first needs a better testing framework that allows for TDD and red-green-refactor. Otherwise, this whole issue with test coverage is a moot point - Too much complexity and legacy code in build.xml, common.xml, ivy.xml, macros.xml and others. It's just really ugly All the code that I saw for cobertura is just ugly and broken. Now it's perfectly fine to reintroduce cobertura cleanly in the future, but I would not use the existing code anyway, I would just wipe it all out and start fresh. I'm not sure whether we need to vote on this? Appreciate your feedback. Taher Alkhateeb