Re: [STYLE] use true/false in C++, not TRUE/FALSE

[MARK CALLAGHAN]

On Tue, Aug 31, 2010 at 10:42 AM, Roy Lyseng <roy.lyseng@stripped> wrote:
> On 31.08.10 19.26, MARK CALLAGHAN wrote:
>>
>> On Tue, Aug 31, 2010 at 10:23 AM, Davi Arnaut<davi.arnaut@stripped>
>>  wrote:
>>>
>>> I haven't keep up with the thread, but I think the my_bool for C code is
>>> a
>>> bit of silliness. We should just stick to int where applicable and be
>>> done
>>> with it.
>>
>> I hope the code moves in that direction. I understand using 0 for
>> success and != 0 for error. I get confused when TRUE means error and
>> FALSE means success.
>>
> The call pattern is almost the same, so no need to be confused:
>
>  if (func(a, b, c))       // Bool return
>    DBUG_RETURN(true);
>
>  if (func(a, b, c))       // Int return
>    DBUG_RETURN(1);
>
>  if (func(a, b, c) != 0)  // Int return - more specific testing
>    DBUG_RETURN(1);
>
> The good thing with bool return is that you cannot overload it with
> additional information...
>

That isn't the problem. The problem is when I read frequently
undocumented header files.
my_bool do_this();

The standard used by libc and others is to return an int with 0 as
success and != 0 as an error. MySQL is the only system on which I have
worked that uses bool and my_bool for this.

-- 
Mark Callaghan
mdcallag@stripped