🚀 go-pugleaf

X-Original-Date: 31 Dec 1995 07:54:48 GMT

In article <4c1n8o$2jn@netlab.cs.rpi.edu>, Bowden Wise <wiseb@cs.rpi.edu> wrote:
 
> BOOL DDETopic::DelAllItems ()
> {
>    for (_itemIter  =   _itemList.begin();
>         _itemIter  !=  _itemList.end();
>         _itemIter++)
>    {
>       _itemList.erase (_itemIter);
>    } 
>    return TRUE;
> }



After any one 'erase' your list is changed and iterator doesn't point to
the objrct that you want . _itemList.end() also will point on something new.
And not to the old value of  _itemList.end() which you have in 
'for' .
If  you want to erase all elements in list/array then use 
erase(iterator forst, iterator last) mem-function

    [ comp.lang.c++.moderated is a moderated newsgroup.  Submit articles ]
    [  to <c++-submit@netlab.cs.rpi.edu>.  The moderation policy can be  ]
    [   retrieved from <http://netlab.cs.rpi.edu/~cppmods/guide.html>.   ]
    [    Moderators can be reached at: c++-request@netlab.cs.rpi.edu.    ]

X-Original-Date: 1 Jan 1996 05:46:03 GMT

"erase" invalidates all active iterators on the container.  The results are 
undefined.

Do this instead:

_itemList.erase( _itemList.begin(), _itemList.end() );


- Terris

In article <4c1n8o$2jn@netlab.cs.rpi.edu>, wiseb@cs.rpi.edu says...
>
>X-Original-Date: Fri, 29 Dec 1995 16:21:37 -0800
>
>Hi,
>
>I am using the STL to hold a list of items for a DDEML topic.  When
>the topic is destroyed, all of the associated items with that
>topic are erased from the topic's item list.  Erasing the item
>also causes the item's destructor to be called:
>
>
>BOOL DDETopic::DelAllItems ()
>{
>#ifdef DEBUGON
>Tracer t("DDETopic::DelAllItems ()");
>#endif
>
>   for (_itemIter  =   _itemList.begin();
>        _itemIter  !=  _itemList.end();
>        _itemIter++)
>   {
>#ifdef DEBUGON
>if (__theLog) __theLog->OutputLn ("calling erase for item");
>#endif
>      _itemList.erase (_itemIter);
>   }
>
>   return TRUE;
>}
>
>In my particular application the list only contains a single 
>element.
>
>When DEBUGON is enabled, the single item is correctly deleted the
>first time through the loop.  However, I can also verify that
>the loop is entered a second time, thereby resulting in a GPF.
>
>Any ideas why the _itemIter != _itemList.end() test of the for
>loop is failing?  Anyone else encountered this?
>
>Bowden
>Dept of Computer Science, RPI
>wiseb@cs.rpi.edu
>http://www.cs.rpi.edu/~wiseb/
>
>    [ comp.lang.c++.moderated is a moderated newsgroup.  Submit articles ]
>    [  to <c++-submit@netlab.cs.rpi.edu>.  The moderation policy can be  ]
>    [   retrieved from <http://netlab.cs.rpi.edu/~cppmods/guide.html>.   ]
>    [    Moderators can be reached at: c++-request@netlab.cs.rpi.edu.    ]


    [ comp.lang.c++.moderated is a moderated newsgroup.  Submit articles ]
    [  to <c++-submit@netlab.cs.rpi.edu>.  The moderation policy can be  ]
    [   retrieved from <http://netlab.cs.rpi.edu/~cppmods/guide.html>.   ]
    [    Moderators can be reached at: c++-request@netlab.cs.rpi.edu.    ]

X-Original-Date: 02 Jan 1996 18:54:44 GMT

In article <4c1n8o$2jn@netlab.cs.rpi.edu> Bowden Wise <wiseb@cs.rpi.edu> writes:
>...
>    for (_itemIter  =   _itemList.begin();
>         _itemIter  !=  _itemList.end();
>         _itemIter++)
>    {
>...
>       _itemList.erase (_itemIter);
>    }
> ...
> Any ideas why the _itemIter != _itemList.end() test of the for
> loop is failing?  Anyone else encountered this?

The erase() invalidates _itemIter.  Continuing the loop is therefore
invalid.  Use _itemList.erase(_itemList.begin(), _itemList.end())
instead.

Doug
----------
Doug Morgan         TEL: (415) 960-7444 Booz-Allen & Hamilton, Inc. 
doug@ads.com (UNIX) FAX: (415) 960-7500 1500 Plymouth St.            
morgan_doug@bah.com (MS-Mail)           Mountain View, CA 94043-1230
----------

    [ comp.lang.c++.moderated is a moderated newsgroup.  Submit articles ]
    [  to <c++-submit@netlab.cs.rpi.edu>.  The moderation policy can be  ]
    [   retrieved from <http://netlab.cs.rpi.edu/~cppmods/guide.html>.   ]
    [    Moderators can be reached at: c++-request@netlab.cs.rpi.edu.    ]

X-Original-Date: 2 Jan 1996 19:18:24 GMT

terris@rahul.net (Terris Linenbach) wrote:
>X-Original-Date: 1 Jan 1996 05:46:03 GMT
>
>"erase" invalidates all active iterators on the container.  

I think it is not quite true. Erase invalidates some active iterators on the
container. In case of list, erase invalidates only those iterators that point
to the erased elements.

In this particular case:

    for (_itemIter  =   _itemList.begin();
         _itemIter  !=  _itemList.end();
         _itemIter++)  // <- itemIter is invalid now
    {
       _itemList.erase (_itemIter);
    }

But there is a simple fix:

   for (_itemIter  =   _itemList.begin();
         _itemIter  !=  _itemList.end(); )
    {
       _itemList.erase (_itemIter++);
    }

and, of course,

	_itemList.erase(_itemList.begin(),  _itemList.end());

will do the same thing.



Alex Stepanov
stepanov@mti.sgi.com

Silicon Graphics
2011 N. Shorline Boulevard
Mountain View, CA 94043-1389

tel.: (415)933-6121


    [ comp.lang.c++.moderated is a moderated newsgroup.  Submit articles ]
    [  to <c++-submit@netlab.cs.rpi.edu>.  The moderation policy can be  ]
    [   retrieved from <http://netlab.cs.rpi.edu/~cppmods/guide.html>.   ]
    [    Moderators can be reached at: c++-request@netlab.cs.rpi.edu.    ]