Alpha is seriously broken

[Garance A Drosihn]

At 5:46 PM +0300 8/20/04, Ruslan Ermilov wrote:
>Garance,
>
>Did you read this thread before replying?  :-)

I did read all the emails that I had received before I wrote what I
wrote.  You may have written other emails by then, but I had not
received them so I did not read them.  From the emails I had read,
I thought you were hoping someone could suggest some alternative
ideas.  I was just trying to be helpful.  I'm not sure why you
had to respond with some smart-ass sarcasm.

>  > If that does not work, and would a cross-compile work well-enough to
>>  generate a new library (if the problem is a library), or a new kernel
>>  (if the error is in the kernel).  Or maybe just generate dynamically-
>>  linked versions of make & cc.  Even if cross-builds are not good
>>  enough to do a complete buildworld, might they work well enough to
>  > recover from this with less work than a complete reinstall?
>
>All static binaries on these two systems crash.  Log in to beast
>and give it a try.

I thought most of the system was dynamically linked these days.

In any case, I still think my suggestion *might* work out, but if you
are sure it is a waste of time to even try then I guess a full install
is the only way out.  If I can login, then I assume slogin/scp must be
working, and I assume beast still has some filesystems nfs-mounted from
other machines.  If it were *me* with my lone sparc64 machine, I would
try to create a "dry spot to stand on" by scp-ing some fixed-binaries
into the box, and then nfs-mount a successful buildworld done on some
other machine, and see how far that got me.

I admit that some times my attempts to avoid a re-install have probably
taken me more time than the re-install would have, but other times it
has been well worth the attempt.

-- 
Garance Alistair Drosehn            =   gad at gilead.netel.rpi.edu
Senior Systems Programmer           or  gad at freebsd.org
Rensselaer Polytechnic Institute    or  drosih at rpi.edu