Re: Fastest minimum-width integer types and limits macros

[Ian Lance Taylor <iant at google dot com>]

Gunther Nikl <gnikl@users.sourceforge.net> writes:

> Ian Lance Taylor wrote:
>> Gunther Nikl <gnikl@users.sourceforge.net> writes:
>> 
>>> Here is an example for int_fast8_t:
>>>
>>>   typedef signed char int_fast8_t;
>>>   #define INT_FAST8_MIN INT8_MIN
>>>   #define INT_FAST8_MAX INT8_MAX
>>>
>>> Now if int_fast8_t uses "int" as underlying type like this:
>>>
>>>   typedef int int_fast8_t;
>>>
>>> which value is its MIN/MAX macros supposed to have? Is it
>>> allowed to use the same "char" limits as above or do I need
>>> to use
>>>
>>>   #define INT_FAST8_MIN INT32_MIN
>>>   #define INT_FAST8_MAX INT32_MAX
>>>
>>> in that case?
>>>
>>> So this is not about user code but an implementation question.
>> 
>> The value of INT_FAST8_MIN is implementation defined, so you can choose
>> whatever value you like (as long as it is <= -128).  That said, the
>> standard seems to suggest that the value should correspond to the value
>> of the underlying type.  That is, if int_fast8_t is defined as int, then
>> the value of INT_FAST8_MIN should be INT_MIN.  That at least makes
>> INT_FAST8_MIN useful, as otherwise it would always simply have the
>> value -128.
>
> Thanks for the information about the standard. FWIW, I assumed that if
> a fast 8bit type was requested than its only about a 8bit value range
> and thus the 8bit limits are sufficient.

Yes, it's hard to understand why a program would ever want to test
INT_FAST8_MIN.  I tried Google CodeSearch but it could only find
definitions, not uses.

Ian