pdftopdf produces non-standard-conform pdf files.
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
cups (Ubuntu) |
Fix Released
|
Undecided
|
Martin Pitt | ||
pdftk (Ubuntu) |
Invalid
|
Undecided
|
Unassigned |
Bug Description
Binary package hint: pdftk
Since upgrading to 9.10 (Karmic), I've been getting an error processing some pdf files with pdftk.
I get the error trying to process pdf files which have been printed from OpenOffice Writer - but the same document 'exported' as a pdf file from OO Writer (via File -> Export as PDF...) - rather than being printed - works fine with pdftk.
I'm using the version of OpenOffice that comes with Karmic (Openoffice.org 3.1.1 OOO310m19 Build:4920, synaptic says openoffice.
My two test case pdf files were created by from a completely blank OO Writer document both exported then printed to pdf - one works with pdftk, the other doesn't. I've attached them to this bug.
I get this error when I try to process the printed one:
duncan@
Unhandled Java Exception:
java.lang.
at com.lowagie.
This command also produces a zero length file called printed_output.pdf. Doing the same thing with the exported pdf produces no error:
duncan@
and creates the export_output.pdf, as expected. All three pdf files appear to work fine in evince.
This bug is a problem because the 'Export as PDF' feature in open office doesn't produce faithful pdf's in all cases - printing to PDF seems to work better - but now pdftk doesn't like my printed pdf's, thus messing up my workflow. If there's any more information that I can provide that would help debug this issue, please just let me know.
CVE References
Changed in pdftk (Ubuntu): | |
status: | New → Confirmed |
summary: |
- java.lang.ClassCastException: com.lowagie.text.pdf.PdfDictionary cannot - be cast to com.lowagie.text.pdf.PRIndirectReference + pdftopdf produces non-standard-conform pdf files. |
tags: | added: patch |
pdftk uses the itext library and a mail on its mailing list gave me the right hint: thread. gmane.org/ gmane.comp. java.lib. itext.general/ 46855/focus= 46859
http://
The printed pdf file is not standard conform [1].
The information dictionary of the document is a direct object, not an indirect reference as in the pdf that works.
So this is mainly a bug in the used pdf generator, not in pdftk. However, pdftk should produce a more user-friendly error message.
Can you provide the version of the cups / cups-pdf package which you have installed, please?
[1]: See table 3.11 on page 61 in http:// partners. adobe.com/ public/ developer/ en/pdf/ PDFReference13. pdf :
"Info dictionary (Optional; must be an indirect reference)"