On Feb 25, 2012, at 7:45 AM, Scott Ribe <scott_r...@elevated-dev.com> wrote:
> On Feb 25, 2012, at 8:38 AM, Kyle Sluder wrote: > >> Does Harbison & Steele say that pointers converted to integer have the value >> zero, or does it say that if(ptr) is identical to if(ptr != 0)? > > It explicitly states that null pointers converted to int are 0, and that > other conversions are implementation-defined but round-trip conversions > through a large-enough int type produce the original pointer. Yeah, that's all kinds of wrong. As the standard also explicitly states, converting a pointer to integer is not guaranteed to result in a representation that fits in any size of integer, and it might be a trap representation. The conversion from integer to pointer is similarly implementation-defined. The only round-trip guaranteed by the standard is pointer-type to pointer-type, and then only if the pointer value is correctly aligned for the second pointee type, or the second pointer type is pointer-to-void. > I have a suspicion that this is a normalization in behavior introduced in the > standardization process--note that I still have K&R 2nd edition here, and as > far as I can tell it simply does not specify what happens when a null pointer > is converted to int. K&R 2nd Ed. was updated for C89, I believe. But regardless, K&R is notoriously light on specifics like this. Which is great when you're teaching a language, but not nearly as great when you're implementing a compiler that intends to support most of the code in the modern world. --Kyle Sluder _______________________________________________ Cocoa-dev mailing list (Cocoa-dev@lists.apple.com) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to arch...@mail-archive.com
