[cfe-commits] r63977 - in /cfe/trunk/lib/Headers: limits.h stdint.h

[Eli Friedman]

On Fri, Feb 6, 2009 at 2:59 PM, Chris Lattner <sabre at nondot.org> wrote:
> +#define INT8_MAX    ((int8_t)127)
> +#define INT8_MIN   ((int8_t)-128)
> +#define UINT8_MAX  ((uint8_t)255)

Gah!  I can't believe I didn't spot this.  Casts aren't allowed here.
In this case, they're unnecessary anyway.

> +#define INT16_MAX    ((int16_t)32767)
> +#define INT16_MIN   ((int16_t)-32768)
> +#define UINT16_MAX ((uint16_t)(65535))

Same issue.  As long as the 65535 is marked unsigned, the casts are
unnecessary due to integer promotions.

> +#define INT32_MAX         ((int32_t)2147483647LL)
> +#define INT32_MIN    ((int32_t)(-2147483647LL-1))
> +#define UINT32_MAX      ((uint32_t)4294967295ULL)

Same issue; you'll need some ifdefs to get this right for platforms
where int32_t isn't int.  Note that the rule is that INT32_MAX is
required to have type int32_t.

> +#define INT64_MAX      ((int64_t)9223372036854775807LL)
> +#define INT64_MIN ((int64_t)(-9223372036854775807LL-1))
> +#define UINT64_MAX  ((uint64_t)18446744073709551615ULL)

Same issue.  Also, this is wrong if int64_t doesn't map to long long;
do we make any guarantees here?

> +/* C99 7.18.2.5 Limits of greatest-width integer types. */
> +#define INTMAX_MIN  (-__INTMAX_MAX__-1)
> +#define INTMAX_MAX   __INTMAX_MAX__
> +#define UINTMAX_MAX (__INTMAX_MAX__*2+1)

Not that there's an issue here, but it's probably worth
double-checking that __INTMAX_MAX__ has the right type.

> +#define INT64_C(v)  (v##LL)
> +#define UINT64_C(v) (v##ULL)

Do we guarantee int64_t == long long?

> +/* 7.18.4.2 Macros for greatest-width integer constants. */
> +#define INTMAX_C(v)  (v ## LL)
> +#define UINTMAX_C(v) (v ## ULL)

This is wrong for targets where intmax_t != long long.

-Eli