------- Comment #4 from burnus at gcc dot gnu dot org  2009-06-22 20:15 -------
Thomas, I think your patch does not help.

(In reply to comment #2)
> This patchlet avoids the ICE:
>    gfc_array_size (source, &size);
> -  if (mpz_get_si (size)*ncopies > gfc_option.flag_max_array_constructor)
> +  if (ncopies && mpz_get_si (size)*ncopies
> +      > gfc_option.flag_max_array_constructor)

You only added a check for  ncopies == 0, but for
  print *, spread(1,dim=1,ncopies=4)
ncopies > 0 but there is still the problem:

SIZE is never initialized if SOURCE= is a scalar. I think the patch in PR 40472
is the correct one - and is has also been approved by Paul and tested by
Dominique. (Here, it fixes the valgrind error and regtests on x86-64.)


-- 

burnus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tkoenig at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40520

Reply via email to