53f3031005
C99 introduced macros of the form `INTn_C(v)` which expand to a signed or unsigned integer constant with the specified value `v` and the type `int_leastN_t`. Clang's initial implementation of these macros used token pasting to form the integer constants, but this means that users cannot define a macro named `L`, `U`, `UL`, etc before including `<stdint.h>` (in freestanding mode, where Clang's header is being used) because that could form invalid token pasting results. The new definitions now use the predefined `__INTn_C` macros instead of using token pasting. This matches the behavior of GCC. Fixes #85995
60 lines
2.3 KiB
C
60 lines
2.3 KiB
C
/* RUN: %clang_cc1 -std=c99 -ffreestanding -triple x86_64-unknown-linux -fsyntax-only -verify -pedantic -Wno-c11-extensions %s
|
|
RUN: %clang_cc1 -std=c99 -ffreestanding -triple x86_64-unknown-win32 -fms-compatibility -fsyntax-only -verify -pedantic -Wno-c11-extensions %s
|
|
RUN: %clang_cc1 -std=c11 -ffreestanding -fsyntax-only -verify -pedantic %s
|
|
RUN: %clang_cc1 -std=c17 -ffreestanding -fsyntax-only -verify -pedantic %s
|
|
RUN: %clang_cc1 -std=c2x -ffreestanding -fsyntax-only -verify -pedantic %s
|
|
*/
|
|
// expected-no-diagnostics
|
|
|
|
/* WG14 DR209: partial
|
|
* Problem implementing INTN_C macros
|
|
*/
|
|
#include <stdint.h>
|
|
|
|
#if INT8_C(0) != 0
|
|
#error "uh oh"
|
|
#elif INT16_C(0) != 0
|
|
#error "uh oh"
|
|
#elif INT32_C(0) != 0
|
|
#error "uh oh"
|
|
#elif INT64_C(0) != 0LL
|
|
#error "uh oh"
|
|
#elif UINT8_C(0) != 0U
|
|
#error "uh oh"
|
|
#elif UINT16_C(0) != 0U
|
|
#error "uh oh"
|
|
#elif UINT32_C(0) != 0U
|
|
#error "uh oh"
|
|
#elif UINT64_C(0) != 0ULL
|
|
#error "uh oh"
|
|
#endif
|
|
|
|
void dr209(void) {
|
|
(void)_Generic(INT8_C(0), __typeof__(+(int_least8_t){0}) : 1);
|
|
(void)_Generic(INT16_C(0), __typeof__(+(int_least16_t){0}) : 1);
|
|
(void)_Generic(INT32_C(0), __typeof__(+(int_least32_t){0}) : 1);
|
|
(void)_Generic(INT64_C(0), __typeof__(+(int_least64_t){0}) : 1);
|
|
// The type of the expanded value in both of these cases should be 'int',
|
|
//
|
|
// C99 7.18.4p3: The type of the expression shall have the same type as would
|
|
// an expression of the corresponding type converted according to the integer
|
|
// promotions.
|
|
//
|
|
// C99 7.18.4.1p1: The macro UINTN_C(value) shall expand to an integer
|
|
// constant expression corresponding to the type uint_leastN_t.
|
|
//
|
|
// C99 7.18.1.2p2: The typedef name uint_leastN_t designates an unsigned
|
|
// integer type with a width of at least N, ...
|
|
//
|
|
// So the value's type is the same underlying type as uint_leastN_t, which is
|
|
// unsigned char for uint_least8_t, and unsigned short for uint_least16_t,
|
|
// but then the value undergoes integer promotions which would convert both
|
|
// of those types to int.
|
|
//
|
|
(void)_Generic(UINT8_C(0), __typeof__(+(uint_least8_t){0}) : 1);
|
|
(void)_Generic(UINT16_C(0), __typeof__(+(uint_least16_t){0}) : 1);
|
|
(void)_Generic(UINT32_C(0), __typeof__(+(uint_least32_t){0}) : 1);
|
|
(void)_Generic(UINT64_C(0), __typeof__(+(uint_least64_t){0}) : 1);
|
|
}
|
|
|