Hi Joseph, I'm pretty sure this would have happened to anybody that migrated the code. I agree that we don't need an additional IT since it doesn't break Maven itself, plugins/extensions that rely on this mechanism should validate it. And Eclipse already added a test for it, so we should be good.
thanks, Robert On 21-10-2019 05:55:31, Joseph Walton <[email protected]> wrote: Robert and Stuart; thank you for fixing the bug introduced by my change, with such a clear explanation and without a major reversion. >From your description of this as an edge case, it doesn't seem like this quite deserves a new test in maven-integration-testing, but it may be worth testing with the Polyglot plugin (any others?) when finishing off MNG-5577. On Sat, 19 Oct 2019 at 23:19, Stuart McCulloch wrote: > ... > which is why > I added a detailed description in the original patch (and not just the > annotation change) >
