I've begun to look at the creation of a PDFRender subclass that will do impositioning of the resultant pages.
There are three potential ways I can do this: 1. Use reference xobjects from PDF 1.4, I generate as normal except that I trap the original file into a embedded file stream. I then create the output pages based on the number of pages received. It seems to have the least potential for bugs, but it also relies on PDF 1.4 so I'm not sure what RIP support will be available. 2. Convert (actually they would never be pages as I would override the renderPage function) the pages to xobject forms and position them. There are two problems I see with this, one is that if the rip caches pdf form objects then it may cache the pages for no good reason causing an extra memory burden. A number of commercial products use this technique. Also some RIPS generate an inordinate number of gsave/grestore operations when nested xforms are in a document. 3. Do they same as above but convert the page information to content streams. It does mean that all pages need to share the same resource dictionary though which could cause some problems (though FOP currently does this) and is a more complex solution that the other two (pages and xobjects have a simliar structure). I'm sure there would also be other problems with this solution. I'm leaning toward 2 as it balances complexity with compatibility. Anyones thoughts? -k PS: does anyone but me want this :)? -- If you don't test then your code is only a collection of bugs which apparently behave like a working program. Website: http://radio.weblogs.com/0111457/ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]