Following this great post by Dennis I believe that a few other things are worth adding.
First the OpenSolaris support at Blastwave : The situation at Blastwave is that we maintain four separate stacks of software. A stack for x86 and sparc both divided between a stable and unstable tree. Those four trees are maintained for compatibility across 4 major releases of the Solaris operating system. Solaris 8, 9, 10 and the various forms of OpenSolaris (SXCE etc.). Supporting a production quality state for such a broad range of targets is no easy feat, but we are still strongly attached to that goal reinforced by the impressive usage statistics. The truth however is that the OpenSolaris packaging format was engineered without much consultation with the community. Much less consultation with long term players. Now don't get me wrong, I'm fully behind the OpenSolaris project. In fact I have been using OSOL for a very long time now and it became my exclusive desktop OS since the first LiveCD release. But changing the package format created a break in our support chain. Much worst, it broke the smooth upgrade path that Blastwave.org aim to offer. This may not look important for a casual user, but one of the reason we support many releases of Solaris is that the Blastwave stack of software can be used as a solution to ease the upgrade from previous releases of Solaris. All the way down from Solaris 8 up to OpenSolaris in both production and desktop productivity environments. Being incompatible with the way we engineer packages the IPS required us to create two completely separate stack of softwares. Users then had the choice of using the supported method of installation with pkgutil or going the IPS way. Needless to say this is too much of a burden to support but we tried it nonetheless, for a while. Our decision to support it further or not would have been based on the long term support status of OpenSolaris. Up until now, long term support has never been available and as such prevented even the possibility of us supporting a separate IPS software stack or even OpenSolaris as a whole. As an example of the importance of long term support let me use an example. Today I diagnosed an issue of a Blastwave package not working in OpenSolaris 2009.11. At runtime the dynamic linker was complaining about an unsatisfied version dependency in a system library. Generally, since we build our softwares on previous versions of Solaris, in this case Solaris 10, backward compatibility would mostly assure us that it works on any higher versions. But in this case a bug in the versioning of a system library prevented a software from working. The fix in those cases is to report a bug and wait for an OS patch. But this is impossible at the moment in OpenSolaris, the only option for the user here is to leave 2009.11 for the development build which corrected this issue. Add this with the uncertainty on the exact future of Solaris and OpenSolaris it becomes even harder, and the uncertainty do not show signs of fading yet. Still, as a maintainer for Blastwave I continue to include OpenSolaris as part of my QA work and I as well as everyone else continue to ensure the greatest amount of support as possible for OpenSolaris. The future at Blastwave.org... Some have expressed concerns about the future of Blastwave. The reality is that we never intended to give up and are certainly far from doing so at this critical moment in the history of Solaris. Some of you may have noticed that we are in the process of upgrading our website along with our support infrastructure. This behind the scene work is progressing smoothly. Also, as stated by Dennis, work is under way to modernize our package distribution methods. This is critical to address the greater needs for security and reliability. Part of this is to ensure end-to-end integrity and security on par with modern requirements. A first step to this end, and by no mean the last is the inclusion of checksumming by default in our latest pkgutil package. Some other enhancements includes a more user friendly installations of packages, especially in the case of errors during package installations. Moreover, we are working hard on providing better documentations, release notes and upgrade guides for Blastwave packages. This should really bridge the gap between Solaris and OpenSource softwares both for commercial and casual users. And obviously we continuously work on increasing the quality and completeness of our software releases. All in all, Solaris and OpenSolaris both have a bright future to look at and we certainly hope that the full potential of Solaris will finally be achieved. As such, we are fully committed as we always were to Solaris and OpenSolaris. -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
