A quick follow-up. Guard malloc did, in fact, catch my test write to a memory location about 1,000,000 bytes beyond the allocated upper boundary. This is reassuring, as the normal code appears to have passed the Guard malloc test – at least for now.

Thank you "Polyphonic" for your clear and very helpful reply. Your suggestion to use the second row, "Environment Variables" of the arguments tab worked. I never knew what those were – now I am much wiser. At first, setting "MALLOC_PERMIT_INSANE_REQUESTS = 1", did not work. I got that from another blog where it was to be entered as a command line "set env MALLOC_PERMIT_INSANE_REQUESTS = 1". That blog implied that any value would work. In XCODE I removed the "= 1", and it then worked. Again, thank you. You have made my day!