Hi Jason, On Mon, Jun 8, 2009 at 4:35 PM, Jason Daly<jd...@ist.ucf.edu> wrote: > And your responses show that you didn't grasp the point of mine.
Um.... it was mostly a rant... > You seem to still be under the impression that I expect you to continue > supporting CMake 2.4.5, no matter what. Let me be clear here: > > *If you don't want to support CMake 2.4.5 anymore, then by all means don't > support it* I have already said that if CMake 2.4.5 is possible without causing problem then I'm happy to see us support CMake 2.4.5. The key point is that if you want this support you (as in all users users using this out of date CMake version) have to be pro-active in making sure that there aren't any build regressions. > As I said, we don't have any money to pay you. If you consider it too > expensive to maintain, then drop it. It's only expensive because CMake 2.4.5 users aren't being pro-active in testing. It's repeated stable/release candidates that are really really expensive. When you don't test till after a stable release has gone out then it's very much in the really expensive bracket. The fact that stuff that could have easily been caught by end users testing in release candidates of 2.8.1 is what pisses me off, all my repeated calls for testing and the time I gave the community it seems wasn't enough. There are quite a few users who do make the effort, these are the hero's of the community, others who waited till it after the release are the ones who effectively casting all their efforts in vain, as we're reset back to square one with a new series of 2.8.2-rc's and final 2.8.2 now required. > If support for CMake 2.4.5 is contingent upon me alone testing every release > that you put out, I cannot make any promises to do that. I'll do what I > can, but I cannot guarantee anything. I like my job, and I intend to keep > it. We'll if you and others who want CMake 2.4.5 support can't afford to make sure the OSG releases build it then there is nothing I can do. If we don't know about problems we can't fix them. > I've already said that I'll try to set up a CDash target for the trunk here > in the lab. Hopefully that's enough to satisfy your needs. This will be a major step forward, having a nightly build of svn/trunk and the latest stable branch (OpenSceneGraph-2.8) is one of the most effectively ways of catching problems early. It's not a total replacement for actual human testing though, whether something compiles or not is only part of the story, we also need to know if the OSG actual runs well against peoples applications prior to releases. > As far as gratitude, look, I'm not looking for a trophy or anything. I just > think that your contributors should be given a little respect. I do appreciate your contributions, but that won't mean that I will shy away from calling out problems when I see them. The build problem was minor problem easily dealt with, but the lack of community testing prior to a release is a big problem here - as fixing problems after releases is far more expensive, both for me and all those who contribute to getting releases out the door. Robert. _______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
