Tom Palmer wrote: > Anyway, all I do is create a DOMParser, parse an XML file, > get a Text node, and call getData(). The '&', '<', and so on > are nowhere to be seen.
I just fixed this. The cvs report shows a lot more changes than are really there (my editor strips trailing whitespace -- I've turned that *off*). The comment for the commit explains the problem. Basically the parser pieces decide how they want character data delivered: either as a character array with offset and length; or as a data index in the string pool. The callback point for character entity references was not obeying this property and *always* sent the data as a character array. This worked on the SAX parser since it always wants its data this way but failed on the DOM parser since it would rather have them in clumps for construction of the DOM tree. -- Andy Clark * IBM, JTC - Silicon Valley * [EMAIL PROTECTED]
