make check fails on powerpc-ibm-aix5.3.0.0[Scanned]

Eric Fernandez efernandez at physiomics-plc.com
Tue May 26 17:17:49 CEST 2009


> You mention your gcc version.  Does configure output confirm 
> that it chooses gcc?  Look for lines like these:
> 
> using ABI="32"
>       CC="gcc -maix32"
>       CFLAGS="-O2 -pedantic -mpowerpc"
>       CPPFLAGS=""
>       MPN_PATH=" powerpc32 generic"

There is a slight variation:
using ABI="32"
      CC="gcc"
      CFLAGS="-O2 -pedantic -mpowerpc"
      CPPFLAGS=""
      MPN_PATH=" powerpc32 generic"

"-maix32" is missing.

> 
> Please also make sure no races of old builds are present in 
> the source directory.  We generally recommend that you build 
> in a separate bject directory, not in the source directory, 
> since the latter is prone to errors.  (You need a "make 
> distclean" between builds.)

I have the same issue when building in a separate object directory.

> 
> If you configure for target "none", does it still fail?  If 
> it does, then a C file cannot access a C symbol from a 
> different file.  Pretty basic compiler tools problem in that case.

I used:
/hpc/work/fernandz/gnusrc/gmp-4.4.-1-20090526/configure
--prefix=$PREFIXGNU --build=none
(ABI="32" made the configure fail with: configure: error: ABI=32 is not
among the following valid choices: long longlong).

The configure output:
using ABI="long"
      CC="gcc"
      CFLAGS="-O2 -pedantic"
      CPPFLAGS="-DNO_ASM"
      MPN_PATH="  generic"

All tests passed. However, for some reason, I do require ABI="32".

I did further testing using 64bit mode:
/hpc/work/fernandz/gnusrc/gmp-4.4.-1-20090526/configure
--prefix=$PREFIXGNU

using ABI="aix64"
      CC="gcc"
      CFLAGS="-O2 -maix64 -mpowerpc64"
      CPPFLAGS=""
      MPN_PATH=" powerpc64/mode64  powerpc64 generic"

then make check also fails:
make[4]: Entering directory `/hpc/work/fernandz/gnusrc/gmpobj44/tests'
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-bswap because of the
following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-bswap
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-constants because of
the following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-constants
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-count_zeros because
of the following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-count_zeros
PASS: t-gmpmax
PASS: t-hightomask
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-modlinv because of
the following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-modlinv
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-popc because of the
following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-popc
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-parity because of
the following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-parity
exec(): 0509-036 Cannot load program
/hpc/work/fernandz/gnusrc/gmpobj44/tests/.libs/lt-t-sub because of the
following errors:
        0509-150   Dependent module
/hpc/work/fernandz/gnu/lib/libgmp.a(libgmp.so.3) could not be loaded.
        0509-103   The module has an invalid magic number.
FAIL: t-sub
====================================
7 of 9 tests failed
Please report to gmp-bugs at gmplib.org
====================================

Thanks,
Eric


More information about the gmp-bugs mailing list