- From: <allan.saywitz@pb.com>
- Date: Fri, 26 Dec 2003 07:26:22 -0600
- To: public-qt-comments@w3.org
- Message-ID: <OF5B27E1E2.E1E0345D-ON85256E08.004857A6-86256E08.004A01AF@pb.com>
I need to modify lots of existing xml documents (add new tags, delete other tags, rename some existing tags). Originally, I decided to use XSL because it would save me lots of programming time since all I have to do is create an an XSLT Document with all my changes and apply it to the list of xml documents. However, when I discovered that CDATA sections are not preserved, I had to give up on using XSL altogether and code everything by hand using DOM. I cannot predict where a user may decide to use a CDATA section and when they won't. It all depends on what they are trying to do. Couldn't you add the ability for an XSL Processor to preserve cdata sections? I tried disable-output-escaping, but that doesn't work for CDATA sections. I tried cdata-section-elements, but that assumes you know in advance which tags need to have CDATA sections. Here are some suggestions: 1. The XSL Processor knows when it encounters an CDATA section in the input XML. So at that point, why can't the processor check some new attribute on the xsl:output tag like preserve-cdata-section-elements. 2. Add a new attribute to the text() element called disable-cdata-escaping. 3. Add a new function cdata() and when disable-output-escaping is applied in it, it will disable escaping so the CDATA section is preserved. Without this functionality, I cannot see how anybody can ever choose to use XSL to modify an an existing XML document. thanks allan
Received on Friday, 26 December 2003 09:53:43 UTC