Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!munnari.oz.au!manuel!ccadfa!Christopher-Vance@adfa.oz.au
From: Christopher-Vance@adfa.oz.au (Christopher JS Vance)
Newsgroups: comp.lang.c++
Subject: pointer to member function
Message-ID: <2243@ccadfa.adfa.oz.au>
Date: 11 Feb 91 05:01:53 GMT
Sender: cjsv@ccadfa.adfa.oz.au
Organization: Computer Science, Univ. of NSW, ADFA Canberra, Australia
Lines: 46

I am using cfront 1.2.1 (dated 2/16/87), and am having trouble getting
pointers to member functions to work.

| #include <stream.h>
| 
| struct a;
| typedef void (a::*h)();
| 
| struct a {
| 	void c();
| 	void d();
| 	void (a::*e)();
| 	h f;
| 	a() { e = h(&a::c); f = e; }
| };
| 
| void a::c() { cerr << "this is a::c\n"; e = &a::d; }
| 
| void a::d() { cerr << "this is a::d\n"; }
| 
| int main() {
| 	a anA;
| 	(anA.e)();	// line A

line A: error:  object missing in call through pointer to member function

| 	(anA.*f)();	// line B

line B: error:  f is undefined
line B: error:  pointer to member function expected in .* expression:  any 

| }

And then I get a core dump from cfront, presumably before it's finished
that last error message. 

From reading Stroustrup's `The C++ Programming Language', it seems to me
that line A could reasonable be expected to have an error message,
although perhaps not the one I see.  It also seems that there should be
no error in line B. 

Have I done something stupid, or should I wait 18 months for AT&T to get
around to delivering the sources of cfront 2.x?  (It took *that long* to
get them to send us the version we have, so I'm not holding my breath.)

-- Christopher