On Thu, Nov 10, 2005 at 10:00:48AM -0800, Danny Howard wrote: > On Thu, Nov 10, 2005 at 03:14:25AM -0500, Kris Kennaway wrote: > > > > As I understand it, 6.0 is primarily concentrating on improving some > > > of the major stuff introduced in 5.x, and shouldn't take nearly as > > > long to become a "stable" platform. Even so, conventional wisdom > > > generally warns against using any X.0 release for critical > > > applications, but that depends on your definition of "critical" and > > > your level of tolerance for excitement. > > > > You really shouldn't think of 6.0 as "like a usual .0 release, so > > handle with care", but more like "5.4 plus extra optimization and > > stability fixes". We spent nearly 6 months during the release cycle > > on stress-testing and fixing stability bugs, and that hard work > > resulted in a lot of fixes to long-standing bugs that have existed > > since FreeBSD 5.x. In addition to the improved stability, performance > > is much better than 5.4 in several areas. > > > > Naturally there may be some regressions, but in the average case 6.0 > > seems to be an outstanding release of FreeBSD no matter what version > > number you give it. > > So ... I am genuinely curious ... if 6.0 is basically 5.4 plus > improvements, why isn't it called 5.5?
Because under the hood there are a few large changes to support the performance optimizations (e.g. VFS locking), and some that break compatibility. FreeBSD tries to keep compatibility of interfaces within a -STABLE branch, so if we called it 5.5 we'd have broken that rule. Kris
pgp7DHI6pNl9h.pgp
Description: PGP signature