Thanks, that makes sense. So there is information in the earlier parts of the resulting file that would need to be updated as new pages are added to the tail, which is kind of the architectural problem I was worried about. If I were clever enough, I could write them out separately, update the resulting files to take the total end document into account, then stream them one after another as one result.
Not the answer I was hoping for, but thanks a bunch for everyone’s input. :-) Regards, Mark > On Jun 5, 2015, at 10:33 AM, Brzrk One <[email protected]> wrote: > > Impossible? > Nothing (:*mostly* nothing:) is impossible, it's just a matter of how much > programming it will require! > :) > > I've used perl scripts to check/regenerate the xref table after I used a > text editor to twiddle objects... on one level, the PDF file is just text... > > In the simple case, one could write a script that iterated over the > files, renumbering all the object ids, replacing the > references, concatenating all the pieces, and building a combined xref > table at the end. I think the output could be streamed, and the files would > not need to be kept in memory. > > Of course, there are those *non*-simple cases which seem to comprise the > bulk of PDF examples... > > On Fri, Jun 5, 2015 at 11:14 AM, Morris, Mark <[email protected]> > wrote: > >> Since I’ve gotten no response, is it safe to assume this is something >> that’s architecturally impossible? >> >> Thanks, >> Mark >> >>> On Jun 3, 2015, at 11:12 AM, Morris, Mark <[email protected]> >> wrote: >>> >>> Hello! I’m generating multiple pdf documents using a commercial tool, >> then merging them into one document to deliver to the user. This all works, >> but I’m trying to reduce the memory footprint, and the current approach >> builds everything in memory. >>> >>> So I wondered if this is possible: Create a document that streams to a >> file, then keep concatenating more pdf documents to the end, without >> needing to have the whole thing in memory at any time. I’ve looked through >> the API and didn’t see a way, but it’s new to me so I’m sure I could have >> missed something. >>> >>> Thanks for any assistance! >>> >>> Regards, >>> Mark >> >> --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
