On Sun, 04 Mar 2012, Julio Merino wrote:
Portability is one of the reasons why we use __printflike instead of __attribute__((__format__(whatever))). Readability is another reason. You can define __printflike differently for different compilers just as easily as you can define ATF_DEFS_ATTRIBUTE_FORMAT_PRINTF differently for different compilers, so ATF_DEFS_ATTRIBUTE_FORMAT_PRINTF·does not seem to have any advantage.Except that the code lives in external/bsd/atf/ and is shipped in other systems that do not have __printflike.__printflike is a NetBSD-ism, right? If so, why do you think NetBSD is more important than any other system and thus __printflike has to be used?
I did not intend to imply that __printflike "has to be used" or that NetBSD is more important than other systems; I simply intended to imply that "portability" was not a valid reason for avoiding the use of __printflike.
You already have
#define ATF_DEFS_ATTRIBUTE_FORMAT_PRINTF(a, b) ...
in the atf sources. What I meant in my earlier message
was that you could change it to
#define __printflike(a, b) ...
with no loss of portability to non-NetBSD systems, and with an
increase in readability.
Given the the issue with identifiers starting with underscores, which you mentioned in another message, I may have been wrong about the portability issue. In my defence, you did not mention underscores until after I had written my message.
--apb (Alan Barrett)