I've implemented a cache-to-disk mechanism for PDFStream that seems
promising. Unfortunately, I still have to hork up the entire byte
array in order to apply PDFFilters to it. I assume this is related to
how PDF files are compressed. Is there a way to:

a) apply a pdf filter incrementally, rather than all in one go
(e.g. on 1024-byte blocks), OR

b) store the filtered data in more than one chunk

Perhaps I'm wrong about the basic problem. But I am running out of
heap on large plots still; my cache file (what was formerly the
contents of PDFStream._data) runs up to 350K or more before it breaks,
and who knows how many copies of those bytes get passed around...


Paul Reavis                                      [EMAIL PROTECTED]
Design Lead
Partner Software, Inc.                        http://www.partnersoft.com

To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to