------- Additional Comments From tobi at gcc dot gnu dot org 2004-12-08 23:58 ------- For the second call we generate: atmp.12.dtype = 281; atmp.12.data = 0B; atmp.12.offset = 0; atmp.12.dim[0].stride = 0; ... (no changes to atmp.12 funname (&atmp.12, (int4 *) n, &atmp.16); which is clearly not correct, as atmp.12 should point to y (as it correctly does in the first call).
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18197