Bug#290438: produces wrong code on i386 compiling with -fno-automatic -O1

[Matthias Klose <doko@cs.tu-berlin.de>]

# submitted Debian report #290438 to gcc-gnats as PR 19469
# http://gcc.gnu.org/PR19469

forwarded 290438 http://gcc.gnu.org/PR19469
retitle 290438 [PR 19469] wrong code on i486 compiling with f77 -fno-automatic -O1
tags 290438 + upstream
thanks

Please could you subscribe to the upstream report?

Kevin B. McCarty writes:
> The above behavior is that of g77-3.3 and g77-2.95.  g77-3.4 does even
> worse.  It produces the same result as g77-3.3 for the -fno-automatic -O1
> case.  Furthermore, when g77-3.4 is called with only -O1 (no
> -fno-automatic flag), the resulting program segfaults.  (Run the
> attached testcase with "make F77=g77-3.4" to see.)

using the current g77-3.4 from unstable, I cannot reproduce the
segfault (you should link with g77-3.4 as well).

gfortran-4.0 from experimental:

With -O0 -fno-automatic flags:
 calling dadmul
 I=   3.87578458503752      +-  7.112138523428194E-012 %         495 evals error                  0
With -O1 flag:
 calling dadmul
 I=   3.87578458503752      +-  7.112205303755621E-012 %         495 evals error                  0
With -O2 flag:
 calling dadmul
 I=   3.87578458503752      +-  7.112267978935682E-012 %         495 evals error                  0
With -O1 -fno-automatic flags:
 calling dadmul
 I=   3.87578458503752      +-  7.112205303755621E-012 %         495 evals error                  0