There is a mismatch in behaviour of a library function (returns cluster on default port) and what the program expects (null value for port if no cluster was selected) In some circumstances, (-p specified and older version listening on default port ) this leads to the incorrect binary being run, despite comments in the (program) code saying it should be the most recent binary that is selected. The patch avoids that bug by implementing a TODO: note in the code. My reference to sid was because I saw that some variable names had been changed. I do not currently have a test machine to see which of the 2 versions I have seen is the one in testing. David Sent from my phone. Please excuse any errors On Tue, 9 Oct 2018, 15:01 Robie Basak,