Re: [Podofo-users] PoDoFo cannot load some PDFs with XMP metadata

[Uli Z. <ul...@ri...>]

Am 18.04.2014 um 03:56 schrieb Leonard Rosenthol <lro...@ad...>:

> I haven’t looked at the code in Load() in a while, but I can’t think of a
> reason that it is loading the XMP.  Are you SURE that this where the
> exception is being thrown - XMP parsing??

Yep.

> I would think it’s somewhere else in the PDF.  Can you post a sample?

You can try the 3 test files attached in a zip archive with podofoxmp.

1. XMP-Test.pdf is a simple PDF test file without XMP metadata.

2. XMP-Test-'?'.pdf is the same file as 1), but with XMP metadata added by the aforementioned buggy Java application, which writes begin='?' in the packet header.

3. XMP-Test-''.pdf is the same file as 2), but with the faulty "?" removed by a hex editor.


podofoxmp handles 1) and 2) fine, but throws an exception on 3).

Adobe's XMP Toolkit can handles 1) and 3) correctly, but does not recognize the XMP metadata in 2) (because begin='?' is out of spec).


Bye

Uli