I've recompiled a GENERIC 1.4 kernel with the addition of debug options.
I can independently confirm that this does clear up the zombie process
problem.  So that's interesting. 

Here's an example of the "longjmp botch" error.  I was logged into futplex
from vaxzilla as root, running a make.  I interrupted the make with Ctrl-C
at which point I got the "longjmp botch" message and was kicked out of
make and tcsh on futplex back to vaxzilla. I suppose I should probably
eliminate tcsh as the culprit on my side.  Has anyone else seen this?

  futplex# make
  longjmp botch.
                10:49am vaxzilla [~]%
  10:49am vaxzilla [~]%

-brian.
---
Brian "JARAI" Chase | http://world.std.com/~bdc/ | VAXZilla LIVES!!!