It turns out this patch isn't actually needed once the fix for bug #960283 is in; the number of operands, once computed correctly, is only half of slp_node->vec_stmts_size, so the off-by-one doesn't actually matter.
Marking this bug as a duplicate.
It turns out this patch isn't actually needed once the fix for bug #960283 is in; the number of operands, once computed correctly, is only half of slp_node- >vec_stmts_ size, so the off-by-one doesn't actually matter.
Marking this bug as a duplicate.