This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Libjava failures status
On Tue, Mar 12, 2002 at 11:25:16AM +0000, Andrew Haley wrote:
> > > a = foo.a;
> > > b = foo.b;
> > >
> > > Then obviously the second load from foo can not trap.
>
> Um, how do you know what order the memory accesses happen in? There's
> nothing to stop sched from moving them around.
Yes there is -- the EH edge from foo.a to the handler. If there
is no handler in the current function, then we don't have any
edges and we don't care what order.
r~