>>>>> "mk" == Matthew Kolb <muk@msu.edu> writes:

    mk> there is no short cut here for you in this situation, afaik.

well sure.  All of the branches should build, though, for most of the
time.  so this is kind of a serious bug.

    mk> by cvs co'ing -rnetbsd-1-6 you are getting the "1.6" branch
    mk> which is going to equate to you getting the latest greatest
    mk> 1.6 (1.6.2, effectively).

That's my understanding as well.  I think you are getting the branch
that will eventually become 1.6.3.

    mk> additionally, if you're going to checkout src and build a new
    mk> kernel, you should build a new world. w/o fail.

eh.

I have several machines that run sorta okay with a much newer kernel
than userland.  Programs in general run ok, but netstat and ps can
have problems.  However the neat tradeoff for this is (besides being
spared a tiresome littany of procfs local root exploits), if the
kernel somehow miraculously manages to crash somewhere outside UVM or
UBC, you get a kernel coredump into the swap area.  And if you are
lucky enough to actually get a coredump, you can run netstat and ps on
these coredumps to see what was happening right before you crashed.

-- 
Le fascisme est la dictature ouverte de la bourgeoisie.
		-- Georg Dimitrov