[Bug tree-optimization/117307] [15 regression] ICE when building gavl-1.4.0 in vect_slp_analyze_operations

[sjames at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117307

--- Comment #1 from Sam James <sjames at gcc dot gnu.org> ---
Reduced:
```
int blend_graya_float_j;
float *blend_graya_float_ovl_ptr, *blend_graya_float_dst_ptr;
float blend_graya_float_dst_ptr_3;
void blend_graya_float() {
  float a_b;
  for (; blend_graya_float_j; blend_graya_float_j++) {
    if (blend_graya_float_dst_ptr_3) {
      blend_graya_float_dst_ptr[0] = blend_graya_float_ovl_ptr[0];
      blend_graya_float_dst_ptr[1] = blend_graya_float_ovl_ptr[1];
    } else if (blend_graya_float_ovl_ptr[1]) {
      a_b = blend_graya_float_dst_ptr[1];
      blend_graya_float_dst_ptr[0] =
          blend_graya_float_ovl_ptr[0] * blend_graya_float_dst_ptr[0] * a_b;
    }
    blend_graya_float_ovl_ptr += 2;
    blend_graya_float_dst_ptr += 2;
  }
}
```