FreeBSD 8.0 released!
in [FreeBSD]
|
From: Christoph Weber-Fahr on 7 Dec 2009 17:53 Chronos wrote: >> Everybody knows this. Do you really believe one instant that the >> developers who are now busy working on the 9.* branch will do anything >> non trivial on the 6.* or 7.* branch? > > http://svn.freebsd.org/viewvc/base/stable/ > > RELENG_6: 3 days since last commit. > RELENG_7: 75 minutes since last commit at time of posting. > > Both of these are MFCs. You were saying? Give you an example: driver for boradcomm's new 10-Gig chip set, which are appearing all over the place in quality servers. No FreeBD support at all. Some guy at Broadcomm working on it, no release date, apparently no 7.x or 6.x support planned ever. (Guy has stopped answering my bimontly inquiries what became of that issue, so I have no up-to-date knowledge). I'm sure there are more. Try reporting a ports bug that only applies to 6.x, for starters. Many ports maintainers will simply ignore you. I'm one of the folks who really would like to repeat the 4.x experience. Keep 7 alive until 7.11 ! Regards Christoph Weber-Fahr
From: Michel Talon on 7 Dec 2009 17:48 Chronos <me3(a)privacy.net> wrote: > Michel Talon wrote: > > > Everybody knows this. Do you really believe one instant that the > > developers who are now busy working on the 9.* branch will do anything > > non trivial on the 6.* or 7.* branch? > > http://svn.freebsd.org/viewvc/base/stable/ > > RELENG_6: 3 days since last commit. > RELENG_7: 75 minutes since last commit at time of posting. > > Both of these are MFCs. You were saying? I am saying that in my experience, dating to FreeBSD-2.2 when MFCs are done to systems that are no longer in the memory of developers, they do more harm than good. The present developement model, where the lifetime of series is short, and 3 series at least coexist is a recipe for having a system which never stabilizes. By the way i have checked this commit to releng_6 is a bug in the libc whose correction has been propagated down and the previous commit was a security commit by Colin Percival. I would be very surprised if substantial non critical things are MFCed to releng_6. On the other hand releng_7 receives obviously more extensive recent commits. Anyways the more we go the less we see improvements to the performance and stability of FreeBSD. It is quite troublesome to discover that a well-known "distribution" of FreeBSD envisions to relocate itself to Linux because there are "too many bugs" or that performance tests on multicore machines show NetBSD (which has a quite recent SMP implementation) in better position than FreeBSD, which claims to have improved its SMP during many many years and many revisions (4.*, 5.*, 6.*, 7.*, 8.*) each one being presented as major breakthrough in the fine grained locking or the efficient scheduling department. -- Michel TALON
From: Patrick Scheible on 7 Dec 2009 19:36 talon(a)lpthe.jussieu.fr (Michel Talon) writes: > Chronos <me3(a)privacy.net> wrote: > > Michel Talon wrote: > > > > > Everybody knows this. Do you really believe one instant that the > > > developers who are now busy working on the 9.* branch will do anything > > > non trivial on the 6.* or 7.* branch? > > > > http://svn.freebsd.org/viewvc/base/stable/ > > > > RELENG_6: 3 days since last commit. > > RELENG_7: 75 minutes since last commit at time of posting. > > > > Both of these are MFCs. You were saying? > > I am saying that in my experience, dating to FreeBSD-2.2 when > MFCs are done to systems that are no longer in the memory of developers, > they do more harm than good. The present developement model, where > the lifetime of series is short, and 3 series at least coexist is a > recipe for having a system which never stabilizes. > > By the way i have checked this commit to releng_6 is a bug in the libc > whose correction has been propagated down and the previous commit was a > security commit by Colin Percival. I would be very surprised if > substantial non critical things are MFCed to releng_6. On the other hand > releng_7 receives obviously more extensive recent commits. > > Anyways the more we go the less we see improvements to the performance > and stability of FreeBSD. It is quite troublesome to discover that > a well-known "distribution" of FreeBSD envisions to relocate itself to > Linux because there are "too many bugs" Which distribution is that? Thanks, -- Patrick
From: Andrew Reilly on 7 Dec 2009 20:08 On Mon, 07 Dec 2009 23:53:38 +0100, Christoph Weber-Fahr wrote: > Try reporting a ports bug > that only applies to 6.x, for starters. Many ports maintainers will > simply ignore you. If there is a bug that was fixed, isn't the right answer to use the version where the bug was fixed? Maybe the fix was too intrusive to easily retro-fit. Cheers, -- Andrew
From: Warren Block on 7 Dec 2009 23:33
Christoph Weber-Fahr <wefa2(a)gmx.de> wrote: > Chronos wrote: >>> Everybody knows this. Do you really believe one instant that the >>> developers who are now busy working on the 9.* branch will do anything >>> non trivial on the 6.* or 7.* branch? >> >> http://svn.freebsd.org/viewvc/base/stable/ >> >> RELENG_6: 3 days since last commit. >> RELENG_7: 75 minutes since last commit at time of posting. >> >> Both of these are MFCs. You were saying? > > Give you an example: > > driver for boradcomm's new 10-Gig chip set, which are appearing > all over the place in quality servers. > > No FreeBD support at all. Some guy at Broadcomm working on it, > no release date, apparently no 7.x or 6.x support planned > ever. (Guy has stopped answering my bimontly inquiries what > became of that issue, so I have no up-to-date knowledge). Are you saying that Broadcom's nonexistent support for their 10G cards on 8.x should be backported to 7.x and 6.x? If so, then good news! -- Warren Block * Rapid City, South Dakota * USA |