Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!snorkelwacker.mit.edu!bloom-beacon!eru!hagbard!sunic!lth.se!newsuser
From: dag@control.lth.se (Dag Bruck)
Newsgroups: comp.lang.c++
Subject: Re: Common data structures
Keywords: list, data structures, c++
Message-ID: <1991Jun11.061402.21934@lth.se>
Date: 11 Jun 91 06:14:02 GMT
References: <1991Jun10.200429.26713@lynx.CS.ORST.EDU>
Sender: newsuser@lth.se (LTH network news server)
Organization: Department of Automatic Control, Lund, Sweden
Lines: 56

I agree: writing a list package is a non-trivial task, and we should
encourage anyone who's willing to share his/her interpretation of what
it should look like.  Here are three additional suggestions:

1.  Member of multiple lists

In most list packages an element can be the member of only one list
at a time.  For me and my colleagues this is quite unacceptable.  We
often keep a large group objects in a "master" list and then create
temporary sub-lists for some particular processing task.

The straigh-forward use of a base class "Link" which all other classes
are derived from won't work; some sort of indirection is needed.

2.  Constant objects

It should be possible to declare a list of `const' objects.  This may
not be easy depending on how you use the macro processor.  My own list
package does not handle this case gracefully.  A `typedef' may help:

	typedef const T cT;
	List(cT) clist;

(Note that I assume that a real template implementation handles this
case without any additional typedef.)

3.  Const compatibility

If I have a list of Foo objects, I would like to pass it to a function
than takes a list of `const Foo' objects:

	typedef const Foo cFoo;

	void f(const List(Foo) x) { .... }

	void g(List(cFoo) x) { .... }

	void h() {
		List(Foo) fl;
		f(fl);		// no problem
		g(fl);		// hmmm
	}

In function h(), `fl' is a mutable list of mutable Foo objects.  There
is no problem calling f(), which takes a constant list of mutable
objects.  The difficulty is to explain to the compiler that a list of
non-mutable objects is compatible with a list of mutable objects.

Comments would be appreciated.

Dag Bruck
--
Department of Automatic Control		E-mail: dag@control.lth.se
Lund Institute of Technology
P. O. Box 118				Phone:	+46 46-104287
S-221 00 Lund, SWEDEN			Fax:    +46 46-138118