Calling (machine-instance) returns the result of the first call to that function on subsequent calls. Given that it returns data that could change this seems wrong to me. Also the HyperSpec does not mention any "side effects" (caching is a side effect?).
I'm filing this bug so that this can be checked / discussed.
indeed, the ip address can change (and the name might too).
we could change it, I guess...
We could add a function
(machine-id &optional what)
which would do
if
what
isnil
, return the cacheif
what
ist
, the cache is recomputedif
what
is a string, the cache is set to that string.(maching-instance)
is the same as(machine-id)
.Bruno, what do you think we should do?
BTW, it is not obvious that IP address is TRT.
http://clhs.lisp.se/Body/f_mach_i.htm:
IOW, hardware MAC address seems like a better candidate.
It is much less useful, obviously.