Am 01.11.2016 um 15:58 schrieb [email protected]:
On 2016-10-31 12:44 (-0400), Andreas Lehmkuehler <[email protected]> wrote:
Am 31.10.2016 um 16:10 schrieb [email protected]:
OBJ_NUMBER , OFFSET
49,12769
50,25217
51,25502
52,26034
53,116
54,-36
55,-36
56,-36
A negative offset indicates that the given object is part of a compressed object
stream. In your case that stream is object 36 0.
Thanks in advance for your time.
BR
Andreas
Thanks Andreas. Where can I find more documentation on compressed objects with
respect to negative offsets?
PDF spec, chapter "7.5.7 Object Streams"
BR
Andreas
Thanks. I read that section. I also read about cross-reference streams. Nowhere
in the entire PDF Spec does it mention negative offsets. I'm relatively new to
pdf parsing and where I'm stumped is I don't know exactly how to handle these
negative offsets.
OBJ_NUMBER , OFFSET
49,12769
50,25217
51,25502
52,26034
53,116
54,-36
55,-36
56,-36
Let's take Object Number 54 for example. In your answer you stated that this
would be Object [36,0] and that it is compressed. How did you know it's object
36,0? Where did you find this information?How - using the PDFBox API, would I
retrieve this object and decode it? I wish I could find documentation
(somewhere) on how to handle this scenario - maybe I'm missing something or not
searching for the right keywords.
I appreciate your patience in helping me out. I'm more than willing to
read/research anything necessary, but I just got through reading 3 sections of
the PDF Reference and while it discussed at length the structure of objects,
object streams, cross reference streams , it did not help me in dealing with
how to handle/ locate / parse objects with a negative offset.
OK, I understand your confusion. There is no negative offset within the spec. We
are using negative values to distinguish "direct" xref entries from those which
point to a compressed object stream.
Have a look at COSParser#parseObjectStream to see how PDFBox handles those
object streams.
BR
Andreas
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
