Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!apple!usc!elroy.jpl.nasa.gov!decwrl!pa.dec.com!shlump.nac.dec.com!bolt.enet.dec.com!minow
From: minow@bolt.enet.dec.com (Martin Minow)
Newsgroups: comp.sys.mac.programmer
Subject: Re: RE: removing objects from THINK CList class
Message-ID: <19302@shlump.nac.dec.com>
Date: 23 Jan 91 00:44:25 GMT
Sender: newsdaemon@shlump.nac.dec.com
Organization: Digital Equipment Corporation
Lines: 52

During the discussion of removing objects from a TCL list, I wrote that
I had used a sequence like:
> while ((myObject = myList->FirstItem()) != NULL) {
>   myList->Remove(myObject);
>   myObject->Dispose();
>}

And a respondant quite naturally asked why I didn't use myList->DisposeItems();

Actually, the real situation was a set of multiply-connected lists that
handle open PPC Toolbox sessions.  The PPC class contains a list of
open PPCPorts, the PPCPort class has a list of open sessions, and the
PPCSession class has a list of active (asychronous) I/O buffers.  Messages
have an instance variable linking them to their Session and sessions
have an instance variable pointing to their port.

Looking only at the session-port interaction; the application terminates
a session by mySession->Dispose().  This does the toolbox stuff, then
executes itsPort->itsSessionList->Remove(), then does an inherited::Dispose().

If the application closes the port, it must empty the session list
without using DoForEach or any method that uses that technique as the
actual removal is done by the list element's Dispose() method.

Here's the PPCPort disposal:

void
PPCPort::Dispose()
{
	PPCSession		*aSession;
			
	while ((aSession = (PPCSession *) itsSessions->FirstItem()) != NULL)
	    aSession->Dispose();
	itsSessions->Dispose();
	/* Toolbox stuff deleted */
	inherited::Dispose();
}

Here's the PPCSession disposal:

void
PPCSession::Dispose()
{	
	/* Toolbox stuff deleted */			
	itsPort->itsSessions->Remove(this);
	inherited::Dispose();
}

Hope this is clearer.

Martin Minow
minow@bolt.enet.dec.com