gmp-4.2.1 build feedback

[Torbjorn Granlund]

"Nelson H. F. Beebe" <beebe at math.utah.edu> writes:

  Yesterday, I built gmp-4.2.1 in 32 build environments on about 20
  flavors of Unix.  A fair number were successful and passed all of
  their tests, but there were also some failures, notably a single test
  that fails on multiple platforms and multiple compilers.  This makes
  me think that the error may be in the test itself, rather than in
  libgmp.
  
Well, it makes me think that there is something in your
environment that trips the test.

I had problems with t-sub on a sparc system before, when the
sysadmins of that system had gotten carried away with
LD_LIBRARY_PATH.

I would assume there is something like that that is causing such
massive failures for you.

Please investigate this and get back to us with the outcome.

-- 
Torbjörn