Hi Christian, On Thu, 10 Mar 2005 17:13:21 +0100, Christian Zulehner <[EMAIL PROTECTED]> wrote:
> I'd say, that this is normal behaviour! If you don't "normaliz" these > text-nodes, > how would you be able to differ between them (if e.g. reading form a file) ? If you serialize the changed document to file you cannot differ between the nodes anyway because then the document becomes XML again and not DOM. I want to differ between the nodes using XPath queries on DOM documents. Then the textnodes before and after the <firstname> element would be /student/name/node()[1] and /student/name/node()[3]. Now if I delete the <firstname> element, /student/name/node()[1] and /student/name/node()[2] (corrected for the deleted node) do *not* point to the same textnodes, in fact /student/name/node()[2] now points to the <lastname> element. I looked around in Xerces's code and found that in ParentNode's internalRemoveChild() a call is made to checkNormalizationAfterRemove and that sets isNormalized(false) causing indeed the two text nodes to be normalized. regards, Jeroen --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
