On Mon, Sep 9, 2013 at 4:19 AM, Marc Glisse <marc.gli...@inria.fr> wrote: > On Mon, 9 Sep 2013, Richard Biener wrote: > >> On Sat, Sep 7, 2013 at 11:00 PM, Marc Glisse <marc.gli...@inria.fr> wrote: >>> >>> On Sat, 7 Sep 2013, Mike Stump wrote: >>> >>>> On Sep 7, 2013, at 12:27 PM, Marc Glisse <marc.gli...@inria.fr> wrote: >>>>> >>>>> >>>>> Now flag_check_new should probably disable this optimization… >>>> >>>> >>>> >>>> Yes, this why my point. >>> >>> >>> >>> Ok, here it is (again, no proper testing until bootstrap is fixed) >> >> >> I wonder what happens on targets where 0 is a valid address of an object >> (stated by !flag_delete_null_pointer_checks)? > > > I am not at all familiar with those targets (I thought you had to write > asm to access 0 so the compiler doesn't mess with your code), but it > makes sense to me to test (flag_delete_null_pointer_checks && > !flag_check_new) instead of just !flag_check_new. Consider the patch > changed this way. (we have so many options, I wouldn't be surprised if > there is yet another one to check…)
If we have such a target (do we?) where 0 is a valid address of an object, I would not be surprised if it breaks in so many other ways, since the language explicitly states that can never happen (programmers write code depending on that). > > -- > Marc Glisse