Madhu wrote:

My confusion appears to be because I was calling functions (via ffi) which in turn used getenv(3) to query the environment to decide on behaviour.

This, however, is enough reason to have setenv/putenv to keep *environment-list* and the C environment synchronized. If the user manipulated *environment-list* directly, the changes wouldn't be seen by ffi. This doesn't solve the issue if some function modifies the environment, though. Ray