Hi Mike,
This should never happen under normal working conditions. It may be caused by a bug in the swap I/O routines that I have rewritten for version 1.8.4, although this is rare, since these routines have been exhaustively analyzed and checked for correctness. If we have a bug here, it is a maximum priority issue.
This problem may be caused also by a sporadic hardware and/or driver failure. One question: is the image in question very big? Which are its dimensions and pixel sample format?
To recover from this problem, you can force a rewrite of the corresponding swap file as follows:
- With History Explorer, make sure that you make a backup of the processes that you have applied. Simply create a ProcessContainer icon with the current image history.
- With History Explorer, jump to the step immediately before the checksum error.
- Re-apply the rest of processes, which you have now in the ProcessContainer icon.
Let me know if this helps.