FreeBSD9 and the sheer number of problem reports

[Kevin Oberman]

On Sun, Feb 26, 2012 at 11:43 AM, Mark Linimon <linimon at lonesome.com> wrote:
> On Sun, Feb 26, 2012 at 06:32:17PM +0700, Erich Dollansky wrote:
>> > There's no such odd/even number policy with FreeBSD-- I think you're
>> > thinking of another OS ;)
>> >
>> maybe something got stuck in my head with the move from 4 to 5.
>
> Yes, 5 was the Great Leap where true SMP was introduced.  In the
> many-year-long development cycle, so many other things (IIRC geom
> and suspend/resume, among others) that the change from 4 to 5 was
> completely disruptive.  We resolved to release more often so as to
> never be in that situation again.  (Granted, probably no architectural
> change will ever be that sweeping again.)

Minor correction. Suspend/resume was around in late 3 and 4, though
the Nomad code in 3 was a bit unstable. It worked well in 4, but was
dependent on APM which was already being replaced by ACPI. By the time
5.2 was actually released, many systems were being shipped without
APM, so could not run FreeBSD. (APM was fairly optional, but ACPI
systems really need ACPI support.) ACPI was one of several things that
forced 5.0 to be released even though RE and everyone running current
knew it had big problems. It i also why 5.0 and 5.1 were clearly
marked as "development" releases not for production.

I really hope to never see a release as ugly as 5. 9.0 may have issues
as did 7.0 and 8.0, but for most, it works quite well. I m happily
running it on a couple of systems.
-- 
R. Kevin Oberman, Network Engineer
E-mail: kob6558 at gmail.com