This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Repair TYPE_REF_CAN_ALIAS_ALL machinery (take 2)


On 5/24/06, Eric Botcazou <ebotcazou@adacore.com> wrote:
> It looks like this will not work for
>
>  int *i;
>  float *f (can alias all);
>  *i = 1;
>  *f = 2;
>  return *i;
>
> because aliasing will have
>
>  # VDEF<SMT.int>
>  *i = 1;
>  # VDEF<SMT.can_alias_all>
>  *f = 2;
>  # VUSE<SMT.int>
>  return *i;
>
> And we happily propagate the 1 to the return statement.  no?  I.e. the
> can_alias_all will only alias theirself, not everything.
>
> Or am I missing something?

The assignment between f and i somewhere?

f = (float *)i;

It was just illustration ;) Make it two incoming parameters to a function, the assignment being in the caller. At least it looks fragile to me.

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]