I noticed the following in the binary PDF format reportedly produced by OpenText Exstream Version 16.6.30:
A hex dump of the same file shows the first 16 characters of the file, where 25 50 44 46 are ascii for "%PDF", so it has an incorrect three-byte binary prefix of hex values ef bb bf:
0000000 ef bb bf 25 50 44 46 2d 31 2e 34 0a 25 e2 e3 cf
As such, the PDF incorrectly has data BEFORE the PDF header. This makes it an invalid PDF even though some PDF software may be able to ignore the header and locate the %PDF header later on and open the file.
Now I notice that the Opentext Brava Enterprise administration guide actually describes what this EF BB BF is, and I quote:
Having this Unicode character as a prefix in binary PDFs is quite wrong. Any chance this is a current or prior defect in Exstream in relation to PDF creation, inserting a Unicode byte order mark at the beginning of the PDF?
