I narrowed down an example that fails and I've attached it. You'll need to Arial Black font file which should be readily available if you're on Windows. The problem is that under whatever circumstances the result of the first modulo operation in convertWord() is incorrect, always returning -1 (and thus the illegal char always being 32).
Hope it helps, Jason <<FOPTest.java>> -------------------- > I've switched from the Sun 1.3.1_05 JVM to the IBM 1.3.1 build > cn131-20020403 JVM for various reasons. My PDF generation via FOP that work > perfectly with the Sun JVM is failing with the IBM JVM. When it's running, I > get many "illegal char value XX" error statements (from the ASCII85Filter > class) and the PDF is corrupt. It's a well known bug in the IBM JVM. There should be an upgrade, ask your customer support. > > Under FOP 0.20.4 I got many, many of these error statements with various > char values in place of the XX. Under FOP 0.20.5rc I only get a handful, and > they are all for value 32. Hm, I thought the code change avoided the bug completely. I'll look into this later. J.Pietschmann
FOPTest.java
Description: Binary data
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
