https://gcc.gnu.org/bugzilla/show_bug.cgi?id=55591
--- Comment #8 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Testing the updated patch
--- ../_clean/gcc/fortran/options.c 2019-03-11 15:11:11.000000000 +0100
+++ gcc/fortran/options.c 2019-04-04 18:55:50.000000000 +0200
@@ -166,6 +166,8 @@ gfc_init_options (unsigned int decoded_o
set_default_std_flags ();
+ flag_strict_aliasing = -1;
+
/* Initialize cpp-related options. */
gfc_cpp_init_options (decoded_options_count, decoded_options);
gfc_diagnostics_init ();
@@ -266,6 +268,10 @@ gfc_post_options (const char **pfilename
sorry ("%<-fexcess-precision=standard%> for Fortran");
flag_excess_precision_cmdline = EXCESS_PRECISION_FAST;
+ /* By default use strict-aliasing semantics. */
+ if (flag_strict_aliasing == -1)
+ flag_strict_aliasing = 1;
+
/* Fortran allows associative math - but we cannot reassociate if
we want traps or signed zeros. Cf. also flag_protect_parens. */
if (flag_associative_math == -1)
