Study Guide for Final

Windows NT Systems Programming: Spring 1999

Study Guide for Final

Document/View

Message routing (Command and non-command messages)
Responsibilities of the four main objects
CLine implementation and usage/relationship to CObjArray
- how is serialize called?
- how is a line redrawn (name all functions involved)
- Why need both delete and RemoveAll in DeleteContents
displaying state on menus
drawing "rubberband" temporary lines

DCOM

Why a binary standard needed? (motivation lesson)
- explain how an interface class solves the problem of updating a DLL without recompiling the application code
- explain role of v-table in compiler independence
- why is "new" not allowed on DLL objects in the application program?
- why should interfaces be immutable?
Chapter 2: pp.31-67
- in queryinterface for ISUM in listing 2-2 p. 46, how could the remove of the reference operator on the last parameter crach the system?
- why need both m_cRef and g_cComponents in the implementation class of ISum (p. 50)?
- what is the purpose of a classFactory?
- who calls LockServer? and why?
- what does CreateInstance create - interface object or implementation object?
- explain fig 2.5 (how is DLL loaded? when? role of registry)
- what is the purpose of GetProcAddress when using a DLL?
Chapter 5: advantage or use of connectable object?
Chapter 8: pp.289-294, pp298-302
How to convert an "InProc" COM object into a remote one.
Chapter 9 pp. 325-335
- What is parameter marshalling?
- What is the role of the proxy and stub?
- What is role of registry?
Chapter 10 pp. 373-387
- difference C++ prototype and IDL spec (in.out) necessary or optimzation
- what two things makes the C++ implementation (see notes containing code) so much shorter than the C implementation (generated by MIDL)
Appendix on RPC to p 518
- advantage of manual binding over automatic

Examples: