On Wed, 2008-04-30 at 22:51 +0200, Branko Čibej wrote: > He's proposing to make NULL a > well-defined parameter for this function, not an incorrect one.
Yeah, that's another way to see it. I was referring to the current behaviour, where NULL is considered an incorrect value (together with a value pointing nowhere), therefore causing a segfault. So, handling NULL specially is avoiding the segfault. -- Bojan
