Clint As far as I know, you currently can't compile Drizzle 7.1 / Fremont on CentOS 5 and it's been like this since for a long time (months). "Dropping support" has already happened.
We should of course continue to support Drizzle 7 on the platforms it has been released, but in reality we mostly fix bugs in 7.1 and don't really backport a lot. I think given the state where Drizzle is in, it is acceptable to drop older distros rather aggressively. Basically it means we make an assumption most Drizzle users at this point will be early adopters anyway, so they can also use a newer distro. This helps us efficiently focus our own resources too. At the same time, if we were a more mature project, with a large userbase (and then we would hopefully have more developers too), I would obviously agree with you. In the end, I don't think the real issue is whether there is a Centos 5 VM connected to Jenkins or not. First someone would have to make drizzle build on Centos 5 and also make a spec file for rpms. If that was done, I'm sure we could add it to the build matrix quite easily. henrik On Sun, Mar 18, 2012 at 6:50 PM, Clint Byrum <[email protected]> wrote: > Excerpts from Olaf van der Spek's message of Sun Mar 18 07:02:05 -0700 2012: >> On Fri, Mar 16, 2012 at 4:21 PM, Eric Bergen <[email protected]> wrote: >> > Just because it's inconvenient to support an OS doesn't mean support >> > should be dropped. CentOS 3 is *old*. CentOS 6 hasn't even been out >> > for a year. CentOS 5 is still actively released and widely used. >> >> Unfortunately, yes. >> The real question is: at what cost? >> > > Welcome to "DevOps". We've all seen the picture of the two guys with > the wall between them, "I want change || I want stability". > > The movement to have comprehensive, useful test suites and measurements > so that migrations can actually happen is helping to loosen things up > a bit between the change vs. stability camps quite a bit. But its not > universal and it still takes time, in some cases longer than a year. > > Abandoning the "old stable" releases of target platforms is a bit short > sighted IMO. It forces people who are not migrated yet to couple adoption > of Drizzle with their migrations. > > Dropping Ubuntu 10.10 is a no-brainer. Ubuntu will no longer support users > on 10.10, why should Drizzle? But CentOS 5 users will be supported for > quite some time, and I suspect users will continue using it for *years*. > Its not like we do a ton of bug fixes on the GA releases. How many were > actually made to 'elliot' ? 10-15 maybe? > > I think Drizzle can delay any backward incompatible breaks until after > the GA release, surely. The cost of doing this is developers being > frustrated at the inability to use a few cool new features of C++. Any > others? That cost is only there for a very short while, so I think its > worth it. > > _______________________________________________ > Mailing list: https://launchpad.net/~drizzle-discuss > Post to : [email protected] > Unsubscribe : https://launchpad.net/~drizzle-discuss > More help : https://help.launchpad.net/ListHelp -- [email protected] +358-40-8211286 skype: henrik.ingo irc: hingo www.openlife.cc My LinkedIn profile: http://www.linkedin.com/profile/view?id=9522559 _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
