Hi Michael,

How strange, AFAICS the attribute values Proced uses are straight from 'process-attributes' so perhaps the issue is there? - we could add '(process-attributes pid)' to the ert explainer to confirm it.

On the other hand, I don't see any harm in checking for degenerate values like -nan in 'proced-<' (the doc of 'process-attributes' doesn't appear to mention nan, not sure about the internals of that function but it doesn't look like a legal value to me)?

Thanks, Laurence