home *** CD-ROM | disk | FTP | other *** search
Text File | 1990-06-29 | 1.3 KB | 33 lines | [TEXT/GEOL] |
- Item 7712105 27-June-90 13:24PDT
-
- From: MM.XOBJ MacroMind, Haim Zamir,PRT
-
- To: D2605 Xanadu Operating,Marc Stiegler,PRT
-
- cc: CPLUS.DEV$ C++ Interest List--Developers
-
- Sub: re[1] walking the inheritance
-
-
- Marc,
-
- Thank you for your suggestions. The macro concept you refer to has occurred to
- me. Could you send me a simple example of how it works?
-
- In your link you neither confirmed nor denied the existance of inheritance
- chain information at runtime. It obviously isn't necessary for vtables. But
- is it used for an implementation of "new?" My understanding of constructors is
- that an object as it is initialized physically evolves from its root class to
- its destination, changing vtable references as it goes so that virtual
- functions are called correctly from within each constructor. This could either
- be done with a piece of core code or could be implemented inline for each
- "new." Do you know which it is?
-
- As for adding methods dynamically, the approach you suggest is interesting.
- Are you suggesting that I subclass this class with the "value" protocol for
- EACH method I wish to attach dynamically? Also, how would this value method
- operate on a class's data, as a "friend?"
-
- --Haim Zamir
-
-