This isn't exactly a challenge, but we started seeing some new failures in the old-tck after moving up to MyFaces RC3 in Open Liberty. They were all inexplicable HTMLUnit failures with no clear indication of what had gone wrong. When I went to debug the problem, I noticed that the old-tck is running with a ~9 year old version of HTMLUnit, so this is an effort to modernize that aspect of these tests. The only changes to classes are ones necessitated by the removal of methods from HTMLUnit in the intervening years.
In the process of making this update, all of the new failures went away for us, so this is a change we need to have included in a forthcoming 4.0.1 version of the TCK in order to achieve compatibility.
This update will also make it easier to someday port these tests to the more modern portion of the TCK, since the HTMLUnit update part of the work would be complete. All that would remain is moving from JavaTest to JUnit.
This isn't exactly a challenge, but we started seeing some new failures in the old-tck after moving up to MyFaces RC3 in Open Liberty. They were all inexplicable HTMLUnit failures with no clear indication of what had gone wrong. When I went to debug the problem, I noticed that the old-tck is running with a ~9 year old version of HTMLUnit, so this is an effort to modernize that aspect of these tests. The only changes to classes are ones necessitated by the removal of methods from HTMLUnit in the intervening years.
In the process of making this update, all of the new failures went away for us, so this is a change we need to have included in a forthcoming 4.0.1 version of the TCK in order to achieve compatibility.
This update will also make it easier to someday port these tests to the more modern portion of the TCK, since the HTMLUnit update part of the work would be complete. All that would remain is moving from JavaTest to JUnit.