Cascading build breaks (was: Re: x86 GENERIC kernels unlobotomized)

[Andreas Gustafsson <gson%gson.org@localhost>]

Jonathan A. Kollasch wrote:
> I've reverted the modularization of the x86 GENERIC kernels.

One thing that seems to have been lost among all the debate is that
whether or not this change is desirable, it broke the build, and needs
to be either fixed or reverted for that reason alone.  See for
example:

  http://www.gson.org/netbsd/bugs/build/build/2011.08.09.00.39.35/build.log.tail

Note that this build break may be hidden by a second one that jmcneill
subsequently introduced and that is now breaking the build in an earlier
place:

  http://www.gson.org/netbsd/bugs/build/commits-2011.08.html#2011.08.09.01.42.24

There is also a third build problem introduced by mrg a couple of days
ago that is affecting TNF's automated test server but not my own:

  
http://releng.netbsd.org/b5reports/i386/commits-2011.08.html#2011.08.07.08.15.02

Really, all three changes should be reverted if they can't be fixed
immediately, as they are keeping new commits from being tested.
-- 
Andreas Gustafsson, gson%gson.org@localhost