New failures related to recent developments

[Torbjorn Granlund]

Three powerpc systems report failure this morning.  I suspect they would
have reported failure already yesterday, if compilation hadn't failed
due to a missing file...

The failures seem to be for small cases, for which the code
(mpz/lucnum2_ui.c) uses a dumpmp/mini-gmp generated table.

I therefore suspect a problem with the new bootstrap.c or mini-gmp.c.

I compared mpn/fib_table.c with a system that did not report any
failures (but this table to was generated with mini-gmp.c).  They have
the same contents.

But fib-table.h has

#define FIB_TABLE_LIMIT         47
#define FIB_TABLE_LUCNUM_LIMIT  47

on one of the failing systems and

#define FIB_TABLE_LIMIT         47
#define FIB_TABLE_LUCNUM_LIMIT  46

on a non-failing system.

(These are from 32-bit builds.  Corresponding differences can be
observed on failing 64-bit builds.)

-- 
Torbjörn