On Wed, Mar 15, 2006 at 09:02:22PM -0800, Ben Pfaff wrote: > Ron <[EMAIL PROTECTED]> writes: > > > Do you know of some compiler that will choke on the more unabbreviated > > form I posted earlier? > > You mean "int main(void)"? Yes. K&R C compilers will choke on > it. Autoconf 2.59 still has some support for K&R C. If you go > around replacing () by (void) you will break that support.
No I mean using int main(int,char**), with named arguments, and enough trivial boilerplate using them to avoid the other typical warnings. Hmm. I guess a pure K&R would choke on that too. Ok, there's a clue that was eluding me... I suppose the only real answer then is another layer of indirection. Some sort of AC_MAIN variable that controls (read: lets me control) the boilerplate to output there. A fairly simple 'expert' switch could ensure the right default for gcc vs' ISO vs' K&R compilers, while the user can always have the final say. We're into more work now than I originally could see, but I'd hope not a lot more. In which case I might nag upstream when I've had time to look at making a patch along those lines first. There is nothing Debian specific to warrant it being a 'bug' here, but you've justified my decision filter it here before hitting upstream about it just the same. Thanks for the sanity check. I get to see a lot of things I don't always have time to fully appreciate, and if I don't report them while the problem is current, history says I'll smother over the symptoms and forget it was a problem at all -- until someone else reports it again... or it snowballs back onto the scene. The options open to addressing this one are at least a bit clearer now... cheers, Ron -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
