[Bug fortran/51218] [4.7 Regression] Potential optimization bug due to implicit_pure?

[sgk at troutmask dot apl.washington.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51218

--- Comment #18 from Steve Kargl <sgk at troutmask dot apl.washington.edu> 2011-11-21 20:21:01 UTC ---
On Mon, Nov 21, 2011 at 08:02:20PM +0000, burnus at gcc dot gnu.org wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51218
> 
> --- Comment #17 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-11-21 20:02:20 UTC ---
> (In reply to comment #16)
> > One thing it relies on is that the compiler recognizes
> > that the bad function are not pure, as they have a
> > side effect (e.g. accessing module variable call_level).
> > If a side effect is able to disable critical optimizations,
> > then I'm optimistic that the code will work on most platforms.
> > 
> > Now as promised, here's the reduced example:
> 
> Thanks for the example!
> 
> Untested patch:
> 
> --- a/gcc/fortran/resolve.c
> +++ b/gcc/fortran/resolve.c
> @@ -3257,6 +3255,7 @@ pure_subroutine (gfc_code *c, gfc_symbol *sym)
>    else if (gfc_pure (NULL))
>      gfc_error ("Subroutine call to '%s' at %L is not PURE", sym->name,
>                &c->loc);
> +  gfc_current_ns->proc_name->attr.implicit_pure = 0;
>  }

Harald's example has function calls not subroutines.  I would
expect that you need to set this in pure_function as well.

Also, does this type of change inhibit reason why 
implicit_pure was added in the first place?