I'm using OS/390 and IBM JDK1.3.0 When i use FOP in order to render PDF, the PDF headers are created in EBCDIC, and the PDF "streams" are written in binary, so the genereated pdf file is unreadable in a ASCII or EBCDIC machine. I suppose you're using methods like writeBytes or writeChars or something like that. In EBCDIC machines the results are different when writen bytes or chars to a file.
I've been working on source files. I've found a possible soultion. The error is caused by the use of getBytes() in String translations. I've susbstituted getBytes() by getBytes("ISO-8859-1"), so the render of PDF files runs in my EBCDIC (S390) machine and in my ASCII (windows) machine. I think the PDFDocument class should include a "setEncoding(String enc)" and a "getEncoding()" methods.
One of the problems is that PDF is in part a binary format because of embedded images (in particular compressed images). *** This bug has been marked as a duplicate of 2532 ***
batch transition to closed remaining pre-FOP1.0 resolved bugs