Question

In cpp

EBD_HEADER* pHeader;

pHeader = reinterpret_cast <EBD_HEADER*>(reinterpret_cast<void*>(const_cast <char*> (GetHeader())));

In Objective C

EBD_HEADER* pHeader;

pHeader = (EBD_HEADER*)((void*)((char*)([self GetHeader])));

Why is these so many casts needed. Is it wrong to convert it into (EBD_HEADER*)[self GetHeader];

Answer 1

You don’t need that many casts in C/Objective-C, you should just use one cast for simplicity:

pHeader = (EBD_HEADER*)[self GetHeader];

With C++-style casts, you need two casts because one casts away the const-ness and the other reinterprets the pointer to one type to a pointer to another type; the intermediate cast to void* is redundant and should be removed.

However, in this case, you should try to refactor the code so that you don’t need any casts at all. Why does [self GetHeader] return a const char* instead of an EBD_HEADER*? Is the object returned actually constant? If so, you should declare pHeader as const EBD_HEADER*, and you should not cast away the const-ness; if not, then [self GetHeader] should not return a const type.