NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/lang/cplus/18216

Internet Message Format | 1992-12-21 | 1.9 KB

Path: sparky!uunet!cs.utexas.edu!ut-emx!jamshid
From: [e-mail] (Jamshid Afshar)
Newsgroups: comp.lang.c++
Subject: Re: Overriding the order of constructors?
Summary: too complicated; RTFAQ
Keywords: constructor ordering
Message-ID: <[e-mail]>
Date: 17 Dec 92 18:32:34 GMT
References: <[e-mail]> <robison1.724307343@husc10>
Reply-To: [e-mail]
Organization: The University of Texas at Austin; Austin, Texas
Lines: 38

In article <robison1.724307343@husc10> [e-mail] (Keith Robison) writes:
|[e-mail] (Prof. Ron Dembo) writes:
|>What is the rational for not allowing a programmer to override the
|>default calling order of constructors and initializers?

If you're asking why you can't construct a derived class's data member
before constructing a base class, I think because that would just add
complexity and confusion.  At least one problem is how to handle the
order of destructor calls.  The class and data member destructors
should be called in the reverse order their constructors were called.
If different member initialization lists could specify different
construction orders, there would be no practical way to know which
order to destroy the members.

|    I believe you can effectively accomplish this dangerous
|goal by calling virtual functions in your constructors:
|
|class A {
|  A() { reroute(); do_A_ctor_stuff(); }
|  virtual void reroute(); { }
|}
|
|class B : public A {
|  B() {}
|  void reroute ()  { do_B_ctor_stuff(); }
|}
|
|This will cause B's reroute() to be executed before any
|A constructing has occurred.

No, it will not.  Any virtual functions called during a constructor
call that class's definition of the virtual functon, not a derived
class's.  See FAQ #65 "Why does base ctor get *base*'s virtual fn
instead of the derived version?".  The FAQL is available by anonymous
ftp at sun.soe.clarkson.edu in pub/C++/FAQ.

Jamshid Afshar
[e-mail]