> case POINTER_PLUS_EXPR: > /* Even though the sizetype mode and the pointer's mode can be different > expand is able to handle this correctly and get the correct result > out > of the PLUS_EXPR code. */
Me thinks that comment now needs some rethinking ;-)