On Thu, Dec 16, 2021 at 06:11:27AM +0530, Siddhesh Poyarekar wrote:
> On 12/16/21 00:13, Jakub Jelinek wrote:
> > On Wed, Dec 15, 2021 at 10:42:29PM +0530, Siddhesh Poyarekar wrote:
> > > On 12/15/21 20:51, Jakub Jelinek wrote:
> > > > Shouldn't this also tree_int_cst_compare (old_wholeval, wholeval) ?
> > > >
> > >
> > > AFAICT, there is no situation where wholeval changes but val doesn't, so I
> > > believe the val check should be sufficient. Do you think otherwise?
> >
> > Dunno, just something that caught my eye.
>
> How about if I add an assert like so:
>
> if (tree_int_cst_compare (oldval, val))
> return true;
> else
> {
> gcc_checking_assert (tree_int_cst_compare (old_wholeval,
> wholeval) == 0);
> return false;
> }
Yes, but please fix up formatting, wholeval should go below old_wholeval.
Though, perhaps it would be better to:
if (tree_int_cst_compare (oldval, val))
return true;
gcc_checking_assert (tree_int_cst_compare (old_wholeval, wholeval) == 0);
return false;
Ok with that change.
Jakub
