[cairo] cairo_status_to_string and why I don't like CAIRO_OK

[Kristian Høgsberg]

Carl Worth wrote:
> On Thu, 09 Jun 2005 10:19:38 -0400, Owen Taylor wrote:
> 
>>I don't think we should write this. If we are frequently writing
>>
>> if (status) {
>> }
>>
>>Then we should simply write
>>
>> if (!status) {
>> }
>>
>>They are both pretty unreadable until you learn the idiom. 
> 
> 
> If both of these forms appear in the code, then there's no idiom to
> learn. The code is just plain hard to read.
> 
> My take on this is that "if (status)" is shorthand for "if (status !=
> CAIRO_STATUS_SUCCESS)", and the only way that idiom is learnable is
> that there's only shorthand for the test in one sense, (which is also
> overwhelmingly more common in the implementation).

I don't remember if this has been suggested, but what about

   #define CAIRO_ERROR(status) ((status) != CAIRO_STATUS_SUCCESS)

instead of CAIRO_OK()?  I don't like CAIRO_OK much myself, because it 
feels like just a different (shorter) name for CAIRO_STATUS_SUCCESS, and 
having two names for the same concept is a bad idea.  CAIRO_ERROR() 
doesn't have that problem since it covers a range of status codes.

This looks pretty intuitive:

	if (CAIRO_ERROR (status))
		...

and this is only slightly confusing:

	if (!CAIRO_ERROR (status))
		...

Kristian