Comment 1 for bug 251276

Per Raj's request, the fulltext api has to expose enough information for search highlighting in a scanned image viewer. I'm planning to insert a stored field with each indexed page, giving the byte offset into the relevant djvu_xml file where the xml for that page is. At least initially, it will then be up to the book viewer to send an http range request to get the relevant XML page out of the djvu, figure out where the requested words are and match the words up with pixel coordinates. I may be able to do something better later, but this is a start.