Lee Kelvin was talking about analyzing the contents of the task metadata to look at memory usage and noted that his numbers had to be scaled by 1024.
This horrifies me because it now seems that we are storing task metadata (and have been for years) where you need to know where you ran it in order to work out how much memory it was using.
In another part of utils we have:
so the fix to logInfo is trivial. What I don't know is how I can fix it without everyone's scripts breaking. We can't let this situation persist though and nothing in the code says 1024 byte units. MacOS returns bytes (and I don't know what freebsd returns).
Do I put a version number into the logInfo data structure? Do I put a unit string in there?