[Bug gdb/31532] gdb/gmp-utils.c:169: internal-error: export_bits: Assertion `word_countp == 1' failed.

["tromey at sourceware dot org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=31532

Tom Tromey <tromey at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tromey at sourceware dot org

--- Comment #2 from Tom Tromey <tromey at sourceware dot org> ---
This assert is a "shouldn't happen".
So, it might be a bug elsewhere in that function,
or it could be that your GMP is built in some
unusual way.

The function tries to export an MPZ value to a byte array.
It tries to be careful to ensure the value will actually
fit into the array.  This checking is expected to rule
out the case where more space is needed than what has
been provided.

The case where the result is 0 (and so word_countp==0)
is also handled explicitly by

  /* It's possible that one of the above results in zero, which has to
     be handled specially.  */
  if (exported_val->sgn () == 0)
    {
      memset (buf.data (), 0, buf.size ());
      return;
    }

Since I don't know how to reproduce this, best would be
if you could debug gdb and stop on the failure.  Then we
can see the actual value of word_countp, the other arguments,
and perhaps even the input data -- that would let us even
write a small reproducer invoking just mpz_export so we could
see if some assumption about GMP is wrong.

-- 
You are receiving this mail because:
You are on the CC list for the bug.