On Sat, 2003-11-15 at 21:12, Thomas DeWeese wrote: > > The CDataSection of batik SVGDOM is splitted into multiple CDataSection > > on each new line? is it a bug? > > I guess you could argue this either way we create a CDATA node > every time the SAX parser calls us with a chunk of characters (see > batik.dom.util.SAXDocumentFactory.characters). It would be possible > to store these in a StringBuffer and only create the CDATA/TextNode > when something other than a character is parsed. This is tricky > as there are lots of things that might end it (this is especially > problematic for text not such much CDATA which is explicitly delimited).
Thanks for the response. I agree with you, however normaly people are keeping their embedded binary data encoded in base64 such as image or font, etc. Therefore the structure of CDATA is quite important :) For the TextNode it maybe no problem, if needed can be done with node.normalize(), I look at batik DOM implementation the normalize method is empty :) > > The problem occur when the SVGDom generated by batik is to be > > serialized/writing back to text file. Because the result of serialized > > is not the same with the loaded one due to different in CDataSection. > > Batik itself also unable to open/load the result of serialization. > > This is (IMHO) a much bigger deal I'll look into it (could you > file this in Bugzilla). here is the bugzilla id #24731 Regards Tonny Kohar --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
