llvm-project/clang/test/Sema/rounding-math.c
Richard Smith 552c6c2328 PR44406: Follow behavior of array bound constant folding in more recent versions of GCC.
Old GCC used to aggressively fold VLAs to constant-bound arrays at block
scope in GNU mode. That's non-conforming, and more modern versions of
GCC only do this at file scope. Update Clang to do the same.

Also promote the warning for this from off-by-default to on-by-default
in all cases; more recent versions of GCC likewise warn on this by
default.

This is still slightly more permissive than GCC, as pointed out in
PR44406, as we still fold VLAs to constant arrays in structs, but that
seems justifiable given that we don't support VLA-in-struct (and don't
intend to ever support it), but GCC does.

Differential Revision: https://reviews.llvm.org/D89523
2020-10-16 14:34:35 -07:00

36 lines
1.1 KiB
C

// RUN: %clang_cc1 -triple x86_64-linux -std=c17 -verify=expected,norounding %s
// RUN: %clang_cc1 -triple x86_64-linux -std=gnu17 -verify=expected,norounding %s
// RUN: %clang_cc1 -triple x86_64-linux -std=c17 -verify=expected,rounding %s -frounding-math
// RUN: %clang_cc1 -triple x86_64-linux -std=gnu17 -verify=expected,rounding %s -frounding-math
#define fold(x) (__builtin_constant_p(x) ? (x) : (x))
double a = 1.0 / 3.0;
#define f(n) ((n) * (1.0 / 3.0))
_Static_assert(fold((int)f(3)) == 1, "");
typedef int T[fold((int)f(3))];
typedef int T[1];
enum Enum { enum_a = (int)f(3) };
struct Bitfield {
unsigned int n : 1;
unsigned int m : fold((int)f(3));
};
void bitfield(struct Bitfield *b) {
b->n = (int)(6 * (1.0 / 3.0)); // norounding-warning {{changes value from 2 to 0}}
}
void vlas() {
// This is always a VLA due to its syntactic form.
typedef int vla1[(int)(-3 * (1.0 / 3.0))];
struct X1 { vla1 v; }; // expected-error {{fields must have a constant size}}
// This is always folded to a constant.
typedef int vla2[fold((int)(-3 * (1.0 / 3.0)))]; // expected-error {{negative size}}
struct X2 { vla2 v; };
}