You are viewing a plain text version of this content. The canonical link for it is here.
Posted to c-dev@xerces.apache.org by we...@ilt.fhg.de on 2000/11/28 12:30:54 UTC
Access Violation error
Hi,
I want to change the values of some attributes in a DOM-Document. For that I have to find
specified DOM_Element - nodes and than call setAttribute(key, value) on that elements to
change the attribute. The following shows the listing of a small sample program (I'm using
VC++6):
#include <string>
#include <iostream>
#include <iltxerces.h> //some utilities and a simple wrapper on top of Xerces
using namespace std;
DOM_Element * getElement(DOM_Element docRoot)
{
DOM_NodeList list = docRoot.getElementsByTagName("child");
DOM_Node node = list.item(0);
DOM_Element * el = (DOM_Element*)&node;
el->setAttribute("attribute", "2");
return el;
}
void main()
{
//the xml-document
string xmlDoc = "<?xml version='1.0' encoding='us-ascii'?>";
xmlDoc = xmlDoc + "<root> <child attribute=\"1\"> </child> </root>";
//defined in iltxerces.h
ILTParser iltParser;
DOM_Element docRoot = iltParser.parse("", xmlDoc);
DOM_Element * el = getElement(docRoot);
DOM_NodeList list = docRoot.getElementsByTagName("child");
DOM_Node node = list.item(0);
DOM_Element * el2 = (DOM_Element*)&node;
string str = U::DOMString2string(el2->getAttribute("attribute"));
//str no has the value "2"
el->setAttribute("attribute", "3");
//at this point an "Access Violation" error is thrown
}
The function getElement(...) obviously changes the entry in my DOM-tree but DOM_Element * el
seems not to point to a valid element within the DOM-tree after returning from getElement(..).
Could anybody explain me what is the reason for that?
Tanks in advance
Rolf
-------------------------------------
Rolf Wester
wester@ilt.fhg.de
Re: Access Violation error
Posted by we...@ilt.fhg.de.
> > DOM_Element * getElement(DOM_Element docRoot)
> > {
> > DOM_NodeList list = docRoot.getElementsByTagName("child");
> > DOM_Node node = list.item(0);
> > DOM_Element * el = (DOM_Element*)&node;
> >
> > el->setAttribute("attribute", "2");
> > return el;
> > }
> >
> > [ ... ]
> >
> > The function getElement(...) obviously changes the entry in my DOM-tree but DOM_Element * el
> > seems not to point to a valid element within the DOM-tree after returning from getElement(..).
>
> You are passing a pointer to an object on the stack. That object is
> destroyed when the getElement() function returns, thus the pointer is left
> pointing at a no-longer-initialized chunk of memory.
How is the attribute of the element in the xml-document changed? Doesn't el
point to this element?
el->setAttribute("attribute", "2");
>
> Since DOM_Node and children are light weight, the better choice is to pass
> the entire object rather than a pointer:
>
Could you please tell me what "light weight" means?
> DOM_Element getElement(DOM_Element docRoot)
> {
> // ....
> DOM_Element el = (DOM_Element &)node;
> el.setAttribute("attribute", "2");
> return el;
> }
>
>
> --Bill
>
Thanks for your answer.
--Rolf
-------------------------------------
Rolf Wester
wester@ilt.fhg.de
Re: Access Violation error
Posted by Bill Schindler <de...@bitranch.com>.
> DOM_Element * getElement(DOM_Element docRoot)
> {
> DOM_NodeList list = docRoot.getElementsByTagName("child");
> DOM_Node node = list.item(0);
> DOM_Element * el = (DOM_Element*)&node;
>
> el->setAttribute("attribute", "2");
> return el;
> }
>
> [ ... ]
>
> The function getElement(...) obviously changes the entry in my DOM-tree but DOM_Element * el
> seems not to point to a valid element within the DOM-tree after returning from getElement(..).
You are passing a pointer to an object on the stack. That object is
destroyed when the getElement() function returns, thus the pointer is left
pointing at a no-longer-initialized chunk of memory.
Since DOM_Node and children are light weight, the better choice is to pass
the entire object rather than a pointer:
DOM_Element getElement(DOM_Element docRoot)
{
// ....
DOM_Element el = (DOM_Element &)node;
el.setAttribute("attribute", "2");
return el;
}
--Bill