https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116823
--- Comment #5 from Andrew Pinski ---
Here is a reduced testcase for fold-const.cc case:
```
struct s2
{
int i;
s2(const int &a) : i (a) {}
};
int h(s2 b);
int g(int l, int r)
{
return h(l > r ? l : r);
}
```
Which shows up since r6-68
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116823
--- Comment #4 from Andrew Pinski ---
And yes this shows up in GCC itself.
>From fold-const.cc:
```
if (lr_bitpos.2401_412 < rr_bitpos.2402_413)
goto ; [INV]
else
goto ; [INV]
:
:
# iftmp.2400_515 = PHI <&lr_bitpos(196), &r
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116823
--- Comment #3 from Andrew Pinski ---
Here is a testcase for a well defined case:
```
void f(int *);
int g(int i, struct f *ff)
{
const int t = 10;
const int *a;
{
int b;
f(&b);
if (t < i)
a = &t;
else
a = &i;
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116823
--- Comment #2 from Andrew Pinski ---
(In reply to Andrew Pinski from comment #1)
> I have a simple patch which allows for this.
> Basically we need to ignore clobbers in 2 places.
Even though clobbers have a vdef, they have no effect on aliasi
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116823
Andrew Pinski changed:
What|Removed |Added
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed|
