> Then that would be a problem. We'll have to figure out how to make > them fail benignly if not run on a specific build (i.e. don't use > internal rendering APIs). I'd hold off on putting these back for now...
I see. Ok, I'll think about a way to fix this. Perhaps before running each test we could check that the implementation class we want exists and is visible to us and if not simply print a warning? In the meantime, I'll go ahead and push it without the tests. Thank you, Denis. ----- Original Message ----- > > ...jim > > On 2/4/2011 6:02 AM, Denis Lila wrote: > >> Do these also pass closed JDK? So far I don't know of any way to > >> identify a test as "Open Rasterizer only"... > > > > I don't think so. When I run it with closed source java it > > complains about no PathConsumer2D. > > > > ----- Original Message ----- > > > >> > >> ...jim > >> > >> On 2/3/2011 5:59 PM, Denis Lila wrote: > >>> Hi Jim. > >>> > >>>> The code changes look good to go. > >>> > >>> Nice! Thanks. > >>> > >>>> I'm guessing there is no regression test for this as it is > >>>> entirely > >>>> performance. If this fixes any quality or correctness bugs it > >>>> would > >>>> be > >>>> nice to develop an automated test case to go back with it, but if > >>>> it > >>>> is > >>>> just performance then they don't usually have tests... > >>> > >>> This changeset would be entirely performance and I don't have a > >>> test > >>> for it. I do however have a few regression tests that I would like > >>> to > >>> push. They're already in icedtea: > >>> http://icedtea.classpath.org/hg/icedtea6/file/12df222ab029/patches/rendering-engine-tests.patch > >>> One of them doesn't have a sun bugid, so we should probably just > >>> skip it. > >>> I'm not sure how we would include them, since they all test bugs > >>> that > >>> have already been fixed. > >>> > >>> Regards, > >>> Denis.
