Jonathan Lennox wrote: >> Jonathan Lennox wrote: > Andrew Haley writes: > >>> I was under the impression it was valid to cast between a pointer >>> to a structure element and a pointer to the containing structure, >>> so long as the underlying memory really is the containing >>> structure -- this is how you can implement something equivalent to >>> casting between C++ base and derived classes, e.g. Was I wrong? >> That's not what you're doing here. Of course, you can cast a >> structure pointer to the type of its first element, and back. But >> here you are casting a pointer to a type that the object to which >> it points does not have: > I see. So even though the ultimate type (through which I'm actually > accessing the pointer) really is the type of the structure's first > element, the fact that I'm casting through an "incorrect" > intermediate type causes strict-aliasing violations? Yes. But rather than follow my explanation, I urge you actually to *read the standard*, in this case, Section 6.3.2.3 Para 7. While people will always try to do their best to explain, there is no substitute for going to the source. > I find that if I annotate the derived_union type with > __attribute__((__may_alias__)), gcc 4.4 does not warn, and compiles my > code correctly. Is this a correct use of may_alias? I hesitate to describe any use of __may_alias__ as "correct", but I think so. :-) Andrew.
