Jim Meyering wrote:
> Paul Eggert wrote:
>> On 06/15/11 08:17, Jim Meyering wrote:
>>
>>> + vasnprintf: avoid a warning from gcc -Wsign-compare
>>> + * lib/vasnprintf.c (VASNPRINTF): Cast a known-positive quantity to
>>> + size_t in order to avoid a warning from gcc -Wsign-compare.
>>
>> Given that the context is this:
>>
>> if (count >= 0)
>> {
>> /* Verify that snprintf() has NUL-terminated its
>> result. */
>> if (count < maxlen
>> && ((TCHAR_T *) (result + length)) [count] !=
>> '\0')
>> abort ();
>>
>> wouldn't it be better to send the GCC maintainers a bug report?
>> Since maxlen is unsigned, and since it is obvious to
>> the compiler that count is nonnegative, there is no reason for
>> a sign-compare warning here.
>
> Yes, but that takes longer...
> Thanks for the prod.
> I've done it:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49426
That has been marked as a duplicate of this report from 2008-12-10:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38470
The comments there suggest that this will not be fixed any time soon,
so I re-propose the cast-adding patch.
If you prefer, I'll be happy to add a FIXME comment suggesting to
reexamine this case (hopefully removing the cast) in a year or two.
