Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!uwm.edu!cs.utexas.edu!milano!cadillac!vaughan@mcc.com
From: vaughan@mcc.com (Paul Vaughan)
Newsgroups: comp.lang.c++
Subject: Re: A solution to the multiple inclusion problem
Message-ID: <3531@cadillac.CAD.MCC.COM>
Date: 27 Oct 89 14:55:50 GMT
References: <14240@well.UUCP> <1989Oct23.191634.6345@cs.rochester.edu> <950@dutrun.UUCP> <1030@ncratl2.Atlanta.NCR.COM>
Sender: news@cadillac.CAD.MCC.COM
Reply-To: vaughan@mcc.com (Paul Vaughan)
Distribution: comp
Organization: MCC VLSI CAD Program
Lines: 43
In-reply-to: dspoon@ncratl2.Atlanta.NCR.COM (Dave Witherspoon)


   From: dspoon@ncratl2.Atlanta.NCR.COM (Dave Witherspoon)

   Say I have 2 classes, A and B.  A contains a B*, and B contains an A*.
   Thus, we have a circular dependency!  One proposal (that I've tried):

This may not solve your exact problem, but the way to do this is
simply

in "A.h"

class B;
class A {
  B* b;
};

and in "B.h"

class A;
class B {
  A* a;
};

	However, this does bring up a real issue.  This sort of thing
works, but not when you try to define inline functions inside the .h
files.  Putting inlines in .h files requires the inclusion of other .h
files, that would not otherwise be necessary.  Since all base class
and component types must be defined before a class definition, a
single inline function can force a whole tree of .h files to be
included, (easily 10 more files) where they would not have been
necessary otherwise.  Of course, there is no choice when working with
ATT's cfront, because inline functions must be defined within the
class definition.  Also, with GNU g++, inline functions must be
defined before they are used, or else they simply can't be inlined.
However, at least with the g++ approach, one might structure the files
differently, perhaps, including a .inlines.cc file in those files that
really need to call inline methods and have them inlined.

Does anybody have a reasonable solution for decoupling the .h file
dependencies introduced by inlines for either g++ or c++?

 Paul Vaughan, MCC CAD Program | ARPA: vaughan@mcc.com | Phone: [512] 338-3639
 Box 200195, Austin, TX 78720  | UUCP: ...!cs.utexas.edu!milano!cadillac!vaughan