henry@zoo.toronto.edu (Henry Spencer) (01/06/91)
Could some kind soul who is on the mailing list for X3J11's interpretation phase respond to a question that was asked, and not answered, early in the autumn: when is the size of a type needed? 3.5.2.3 says that an incomplete struct/union type can be used "only when the size of an object of the specified type is not needed". If I recall correctly, it was reported that an interpretation request for this was answered by saying roughly, "this means when it is required by the standard, not when it is needed by a particular implementation". I would very much like to see the exact wording of this interpretation, and any supporting references, because with the sole exception of sizeof, the standard makes no such requirement, dammit! There *are* certain circumstances when a type is required to be an object type, which excludes any form of incomplete type. But when incomplete types are admissible, there is no indication of whether "the size of an object of the specified type" is needed. 1.6 is most explicit that "the omission of any explicit definition of behavior" means that the behavior is undefined, i.e. "the standard imposes no requirements". -- If the Space Shuttle was the answer, | Henry Spencer at U of Toronto Zoology what was the question? | henry@zoo.toronto.edu utzoo!henry