>  The only major failing of the FreeBSD
> build scheme that I'm aware of is that it doesn't work as a cross-build
> system and as such does require that you be running a "current" kernel.

Those are in fact the hard parts.

FreeBSD's current system (according to your description) has two clear
advantages that I can see:

  - it's easy to roll back the system to a last stable set of sources,
	in the event of a ghastly failure late in the build that you decide
	is bad enough to merit re-SUP-ing and starting over.

  - it's in the tree and lots of people use it, therefore it gets worked on.

In practical terms, the second advantage has more leverage. While tracking
current I have almost never had to completely abort a build -- usually I
can get around the problems and continue after a reasonable amount of bug
hunting.

Generally the couple times I have had a system completely chew itself to
pieces, it was a serious toolchain problem (arm32 PIC code, anyone remember
that?) or instability caused by intrinsically risky development (alpha SMP,
late november, took out one of my multia's by slow corruption of boot disk,
requiring a re-install of 1.4.1 which brought it back to life). In both cases
the first 'make build' completed without failure, but a second never finished.

Lately I've been hammering on 'make snapshot' so it is usable on sparc and
sun3, results to be uploaded soon (sparc's happy, sun3 is on the final run).

Even a 20 mhz sun3 can produce a snapshot all by itself every two weeks,
if most of the process is automated.

Todd Whitesel
toddpw @ best.com