I don't believe the standard requires the following to be rejected, but as a QOI issue I believe it should be. The debatable point is whether you believe the composite of the first two is specified by a function definition; if you do then you must reject. GCC rejects it if the order of the first two is switched.
void r(x) int (*x)[2]; {} void r(); void r(int (*x)[3]); /* Ideally rejected. */ Flags are e.g. -Wall -std=c99 -pedantic -- Summary: GCC does not reject an incompatible type declaration Product: gcc Version: 4.1.0 Status: UNCONFIRMED Severity: minor Priority: P3 Component: c AssignedTo: unassigned at gcc dot gnu dot org ReportedBy: neil at gcc dot gnu dot org CC: gcc-bugs at gcc dot gnu dot org http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22249