I was testing out some of the data compression options and caught on how to get it to compress data then be able to uncompress it. However, on another test it (subject function) kept giving error 32 about the source buffer less than expected. In this case it was 17 (saved from when 8 bytes were asked to be encoded with flushfinish - same flushfinish used for the decode), just on a hunch, I added an if inputsize==17 then inputsize++. Then it was happy. Other odd (inputsize & 1) values that came through were much larger and it didn't complain. Using the latest IPP version 7 I downloaded last night. Looks like it's using the w7_ version of things interally.
Is this a known issue or bug? I didn't read anything about having to have the data aligned or even number of bytes, etc..?