cvs commit: src/sys/cam/scsi scsi_da.c

[Kenneth D. Merry]

On Wed, Apr 30, 2003 at 17:40:06 +0200, Poul-Henning Kamp wrote:
> In message <20030430092040.A2864 at panzer.kdm.org>, "Kenneth D. Merry" writes:
> 
> >> >Oops.  I did a 'make universe', but wrongly assumed that it would stop on
> >> >any errors.
> >> 
> >> No, universe chugs, based on the philosophy that it probably ran for
> >> N hours already, and you might be after somethinn else entirely...
> >
> >True, I suppose it would be a major pain to blow up a make universe.  With
> >the failures, it took 8.5 hours on my box, and probably would have taken
> >a bit longer if the 64 bit kernel builds had succeeded.
> 
> It's the kind of thing I usually kick off when I go to bed :-)

Yep.

> >> Well, if you just look at the stdout/stderr from the make universe, it
> >> should say that something failed, you can then traverse the relevant
> >> log file for what exactly failed.
> >
> >It didn't mention any failures.  (output attached)
> 
> Actually, that's interesting.  I seem to have caught it in my private
> extension of universe:  I build a native and a sparc64 cross release.
> 
> I'm sure there is a good explanation why it breaks the kernel in the
> cross-release but not in the cross-world...

It does break the kernel builds, it's just that you have to dig through the
logs for each kernel build to see the breakage.  It doesn't show up in the
top level make universe output.

It seems like a release would have to break, since you're actually trying
to do something with the kernels (stick them in the release directory, I
suppose).

Ken
-- 
Kenneth Merry
ken at kdm.org