I'd suggest storing the objects themselves elsewhere, and keeping a void * pointer to the object inside the node. Functions which operate on nodes would cast the void * to the appropriate type based on the value of iType.Originally Posted by pronecracker
This is basic polymorphism, which is much easier to handle in C++ than in C.
I know, please don't take my signature too literallyI sometimes have thoughts like that, because C seems so simple and I sometimes get confused because of all the functionality of C++.
But, why is using a void* safer then what I'm doing now? And what about
Code:union { Type1 *pType1; Type2 *pType2; };
Don't call me stupid.
I'd say it's safer because it's more likely to work on some screwy compiler. The standard indicates that what you are doing SHOULD work, but I wouldn't count on it. The union thing is fine, too. I do something very similar when writing parsers that build parse trees -- each node of the tree is of type Node, with some code which identifies what kind of node it is, and the data specific to that node type is kept in a field of a union.
If you're going to use a union, no need to use pointers. Just embed the type-specific data directly in there.
Fine details for varying systems.
http://www.open-std.org/jtc1/sc22/wg...24.pdf#page=59
Implementations have existed in which pointers to different types have different representations and/or sizes.
http://c-faq.com/null/machexamp.html
7. It is easier to write an incorrect program than understand a correct one.
40. There are two ways to write error-free programs; only the third one works.*
