Is there still work being done on the ASCII85Decode filter or is this considered finished and should be debugged?
I have found a number of problems in the implementation: 1. An incomplete quintet at the end is not processed 2. The EOD sequence is "~>", NOT '~'; if "~>" is not found at the end, this is an error 3. The quintet is not tested for the overflow condition (> 2^32 -1) before conversion 4. The final quintet, if incomplete, is not tested for the condition (count > 1). - Cirilo
