- From: Arved Sandstrom <asandstrom2@eastlink.ca>
- Date: Thu, 21 Mar 2013 07:44:21 -0300
- To: Jirka Kosek <jirka@kosek.cz>
- Cc: public-ppl@w3.org
On 03/21/2013 05:56 AM, Jirka Kosek wrote: > On 21.3.2013 3:02, Arved Sandstrom wrote: > >> After about 3 hours of reading Saxon Java API docs and hacking, the >> closest I ever got was locating the DOMEnvelope class, but evidently >> that's not the right thing either. I messed around with PJConverters but >> they don't appear to understand DOM Nodes. >> >> It appears that Saxon passes the FO node tree variable in as a >> TinyDocumentImpl, which implements no W3C DOM stuff.This is the nub of >> the problem. >> >> Anyone way more familiar with the Saxon 9 HE Java API who might have >> some ideas? Thanks. > I think that for such Saxon specific question it will be better to ask > at saxon-help mailing list > (https://lists.sourceforge.net/lists/listinfo/saxon-help) > > IIRC Saxon can be configured to operate on DOM trees instead of its > TinyTree object model -- then you will not be forced to convert back and > forth DOM in order to interface with FOP. > > Jirka > Just needed to sleep on it. :-) I was in the right ballpark, just too tired last night to decipher all the API details. TinyDocumentImpl item = (TinyDocumentImpl) sis[0].next(); Document foTree = (Document) DocumentOverNodeInfo.wrap(item); ... Source src = new DOMSource(foTree); (where 'sis' is a SequenceIterator[]) works nicely. Also, at the cost of somewhat more Java complexity the Saxon-based stylesheet is less XSLT-engine-specific. I need to do a bit of refactoring but I ought to have a Saxon9HE integrated extension function mailed out NLT tomorrow. Arved
Received on Thursday, 21 March 2013 10:44:48 UTC