Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!lavaca.uh.edu!uhnix1!moray!urchin!f506.n106.z1.fidonet.org!Roy.Browning
From: Roy.Browning@f506.n106.z1.fidonet.org (Roy Browning)
Newsgroups: comp.lang.c++
Subject: Re: lint++ -> really "AutoCodeReview" -- Ideas requested.
Message-ID: <5866.2587A9E2@urchin.fidonet.org>
Date: 14 Dec 89 04:54:28 GMT
Sender: ufgate@urchin.fidonet.org (newsout1.26)
Organization: FidoNet node 1:106/506 - Fulcrum's Edge, Spring TX
Lines: 47


 > > > ComIdent::~ComIdent()
 > > > {
 > > >   if ( string )
 > > >     delete[strlen(string)+1] string;
 > > > }

 > Actually, delete[X] p; is used when p points at an
 > array of X objects with destructors.  It calls the
 > destructor on each of the X objects but only passes
 > "p" to free() (or the equivalent function.)
 > 
 > Since 'char' doesn't have a destructor, the above
 > code is equivalent to just 'delete string;'

Joe:

        Ouch, guess I deserved that.  Couched my answer in very loose 'C'  
jargon in the C++ newsgroup and got my head handed to me <grin>.  I habitually  
use imprecise terminology relying on the overall meaning rather than the exact  
words.  Will have to revert to a policy of only posting tested code.  

        Spent a couple of hours proving that the code below would loose random  
pointers to free().  Even modified my library to to demonstrate exactly what  
was occurring.

main()
{
        char* str = "Test string";

        String* a = (String*) new String;
        *a = str;
        cout << a;

        delete[5] a;

        exit(0);
}

Then I looked at your reply and the original message and observed that I had  
only proven that I had misread the original missive.  You are quite correct,  
"delete[x] string" calls free only once deallocating the string.

                             I stand corrected,

                               Roy Browning