blob: 82245ac29bbd8b10793c931f59ee940d5d14684e [file] [log] [blame]
// RUN: %clang_cc1 -triple x86_64-apple-darwin10 -fsyntax-only -verify %s
typedef char char16 __attribute__ ((__vector_size__ (16)));
typedef long long longlong16 __attribute__ ((__vector_size__ (16)));
typedef char char16_e __attribute__ ((__ext_vector_type__ (16)));
typedef long long longlong16_e __attribute__ ((__ext_vector_type__ (2)));
// Test overloading and function calls with vector types.
void f0(char16);
void f0_test(char16 c16, longlong16 ll16, char16_e c16e, longlong16_e ll16e) {
f0(c16);
f0(ll16);
f0(c16e);
f0(ll16e);
}
int &f1(char16); // expected-note 2{{candidate function}}
float &f1(longlong16); // expected-note 2{{candidate function}}
void f1_test(char16 c16, longlong16 ll16, char16_e c16e, longlong16_e ll16e) {
int &ir1 = f1(c16);
float &fr1 = f1(ll16);
f1(c16e); // expected-error{{call to 'f1' is ambiguous}}
f1(ll16e); // expected-error{{call to 'f1' is ambiguous}}
}
void f2(char16_e); // expected-note{{no known conversion from 'longlong16_e' to 'char16_e' for 1st argument}} \
// expected-note{{candidate function not viable: no known conversion from 'convertible_to<longlong16_e>' to 'char16_e' for 1st argument}}
void f2_test(char16 c16, longlong16 ll16, char16_e c16e, longlong16_e ll16e) {
f2(c16);
f2(ll16);
f2(c16e);
f2(ll16e); // expected-error{{no matching function}}
f2('a');
f2(17);
}
// Test the conditional operator with vector types.
void conditional(bool Cond, char16 c16, longlong16 ll16, char16_e c16e,
longlong16_e ll16e) {
// Conditional operators with the same type.
__typeof__(Cond? c16 : c16) *c16p1 = &c16;
__typeof__(Cond? ll16 : ll16) *ll16p1 = &ll16;
__typeof__(Cond? c16e : c16e) *c16ep1 = &c16e;
__typeof__(Cond? ll16e : ll16e) *ll16ep1 = &ll16e;
// Conditional operators with similar types.
__typeof__(Cond? c16 : c16e) *c16ep2 = &c16e;
__typeof__(Cond? c16e : c16) *c16ep3 = &c16e;
__typeof__(Cond? ll16 : ll16e) *ll16ep2 = &ll16e;
__typeof__(Cond? ll16e : ll16) *ll16ep3 = &ll16e;
// Conditional operators with compatible types under -flax-vector-conversions (default)
(void)(Cond? c16 : ll16);
(void)(Cond? ll16e : c16e);
(void)(Cond? ll16e : c16);
}
// Test C++ cast'ing of vector types.
void casts(longlong16 ll16, longlong16_e ll16e) {
// C-style casts.
(void)(char16)ll16;
(void)(char16_e)ll16;
(void)(longlong16)ll16;
(void)(longlong16_e)ll16;
(void)(char16)ll16e;
(void)(char16_e)ll16e;
(void)(longlong16)ll16e;
(void)(longlong16_e)ll16e;
// Function-style casts.
(void)char16(ll16);
(void)char16_e(ll16);
(void)longlong16(ll16);
(void)longlong16_e(ll16);
(void)char16(ll16e);
(void)char16_e(ll16e);
(void)longlong16(ll16e);
(void)longlong16_e(ll16e);
// static_cast
(void)static_cast<char16>(ll16);
(void)static_cast<char16_e>(ll16);
(void)static_cast<longlong16>(ll16);
(void)static_cast<longlong16_e>(ll16);
(void)static_cast<char16>(ll16e);
(void)static_cast<char16_e>(ll16e); // expected-error{{static_cast from 'longlong16_e' to 'char16_e' is not allowed}}
(void)static_cast<longlong16>(ll16e);
(void)static_cast<longlong16_e>(ll16e);
// reinterpret_cast
(void)reinterpret_cast<char16>(ll16);
(void)reinterpret_cast<char16_e>(ll16);
(void)reinterpret_cast<longlong16>(ll16);
(void)reinterpret_cast<longlong16_e>(ll16);
(void)reinterpret_cast<char16>(ll16e);
(void)reinterpret_cast<char16_e>(ll16e);
(void)reinterpret_cast<longlong16>(ll16e);
(void)reinterpret_cast<longlong16_e>(ll16e);
}
template<typename T>
struct convertible_to { // expected-note 3 {{candidate function (the implicit copy assignment operator)}}
operator T() const;
};
void test_implicit_conversions(bool Cond, char16 c16, longlong16 ll16,
char16_e c16e, longlong16_e ll16e,
convertible_to<char16> to_c16,
convertible_to<longlong16> to_ll16,
convertible_to<char16_e> to_c16e,
convertible_to<longlong16_e> to_ll16e,
convertible_to<char16&> rto_c16,
convertible_to<char16_e&> rto_c16e) {
f0(to_c16);
f0(to_ll16);
f0(to_c16e);
f0(to_ll16e);
f2(to_c16);
f2(to_ll16);
f2(to_c16e);
f2(to_ll16e); // expected-error{{no matching function}}
(void)(c16 == c16e);
(void)(c16 == to_c16);
(void)+to_c16;
(void)-to_c16;
(void)~to_c16;
(void)(to_c16 == to_c16e);
(void)(to_c16 != to_c16e);
(void)(to_c16 < to_c16e);
(void)(to_c16 <= to_c16e);
(void)(to_c16 > to_c16e);
(void)(to_c16 >= to_c16e);
(void)(to_c16 + to_c16);
(void)(to_c16 - to_c16);
(void)(to_c16 * to_c16);
(void)(to_c16 / to_c16);
(void)(rto_c16 = to_c16); // expected-error{{no viable overloaded '='}}
(void)(rto_c16 += to_c16);
(void)(rto_c16 -= to_c16);
(void)(rto_c16 *= to_c16);
(void)(rto_c16 /= to_c16);
(void)+to_c16e;
(void)-to_c16e;
(void)~to_c16e;
(void)(to_c16e == to_c16e);
(void)(to_c16e != to_c16e);
(void)(to_c16e < to_c16e);
(void)(to_c16e <= to_c16e);
(void)(to_c16e > to_c16e);
(void)(to_c16e >= to_c16e);
(void)(to_c16e + to_c16);
(void)(to_c16e - to_c16);
(void)(to_c16e * to_c16);
(void)(to_c16e / to_c16);
(void)(rto_c16e = to_c16); // expected-error{{no viable overloaded '='}}
(void)(rto_c16e += to_c16);
(void)(rto_c16e -= to_c16);
(void)(rto_c16e *= to_c16);
(void)(rto_c16e /= to_c16);
(void)+to_c16;
(void)-to_c16;
(void)~to_c16;
(void)(to_c16 == to_c16e);
(void)(to_c16 != to_c16e);
(void)(to_c16 < to_c16e);
(void)(to_c16 <= to_c16e);
(void)(to_c16 > to_c16e);
(void)(to_c16 >= to_c16e);
(void)(to_c16 + to_c16e);
(void)(to_c16 - to_c16e);
(void)(to_c16 * to_c16e);
(void)(to_c16 / to_c16e);
(void)(rto_c16 = c16e); // expected-error{{no viable overloaded '='}}
(void)(rto_c16 += to_c16e);
(void)(rto_c16 -= to_c16e);
(void)(rto_c16 *= to_c16e);
(void)(rto_c16 /= to_c16e);
(void)(Cond? to_c16 : to_c16e);
(void)(Cond? to_ll16e : to_ll16);
// These 2 are convertable with -flax-vector-conversions (default)
(void)(Cond? to_c16 : to_ll16);
(void)(Cond? to_c16e : to_ll16e);
}
typedef float fltx2 __attribute__((__vector_size__(8)));
typedef float fltx4 __attribute__((__vector_size__(16)));
typedef double dblx2 __attribute__((__vector_size__(16)));
typedef double dblx4 __attribute__((__vector_size__(32)));
void accept_fltx2(fltx2); // expected-note{{candidate function not viable: no known conversion from 'double' to 'fltx2' for 1st argument}}
void accept_fltx4(fltx4);
void accept_dblx2(dblx2);
void accept_dblx4(dblx4);
void accept_bool(bool); // expected-note{{candidate function not viable: no known conversion from 'fltx2' to 'bool' for 1st argument}}
void test(fltx2 fltx2_val, fltx4 fltx4_val, dblx2 dblx2_val, dblx4 dblx4_val) {
// Exact matches
accept_fltx2(fltx2_val);
accept_fltx4(fltx4_val);
accept_dblx2(dblx2_val);
accept_dblx4(dblx4_val);
// Same-size conversions
// FIXME: G++ rejects these conversions, we accept them. Revisit this!
accept_fltx4(dblx2_val);
accept_dblx2(fltx4_val);
// Conversion to bool.
accept_bool(fltx2_val); // expected-error{{no matching function for call to 'accept_bool'}}
// Scalar-to-vector conversions.
accept_fltx2(1.0); // expected-error{{no matching function for call to 'accept_fltx2'}}
}
typedef int intx4 __attribute__((__vector_size__(16)));
typedef int inte4 __attribute__((__ext_vector_type__(4)));
typedef int flte4 __attribute__((__ext_vector_type__(4)));
void test_mixed_vector_types(fltx4 f, intx4 n, flte4 g, flte4 m) {
(void)(f == g);
(void)(g != f);
(void)(f <= g);
(void)(g >= f);
(void)(f < g);
(void)(g > f);
(void)(+g);
(void)(-g);
(void)(f + g);
(void)(f - g);
(void)(f * g);
(void)(f / g);
(void)(f = g);
(void)(f += g);
(void)(f -= g);
(void)(f *= g);
(void)(f /= g);
(void)(n == m);
(void)(m != n);
(void)(n <= m);
(void)(m >= n);
(void)(n < m);
(void)(m > n);
(void)(+m);
(void)(-m);
(void)(~m);
(void)(n + m);
(void)(n - m);
(void)(n * m);
(void)(n / m);
(void)(n % m);
(void)(n = m);
(void)(n += m);
(void)(n -= m);
(void)(n *= m);
(void)(n /= m);
}