This is the mail archive of the gcc@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: PR 6212


On Sun, May 05, 2002 at 10:43:15PM -0700, Mark Mitchell wrote:
> If the pointer value is merely being read, the assumption does not hold,
> 
> For example, we must not optimize:
> 
>   (int)(T*)x % alignof(T)
> 
> to 0.

Correct, and there is in fact a test for this in the suite.

The point at which we need to be presuming alignment is the
point at which we actually create a MEM.


r~


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