Chris Angelico <ros...@gmail.com> wrote: > On Tue, Jan 10, 2012 at 9:35 AM, Dominic Binks <dbi...@codeaurora.org> wrote: > > The address database can have multiple names for the same IP address so > > gethostbyaddr returns a list - [0] being the first item in the list (and > > usually considered the canonical name) > > Point to note: The inverse operation, getting IP addresses from names, > quite frequently returns multiple results - and in that case, there's > no "canonical name". You'll still usually want to take the zeroth > (first) item in the list, but often the list will change in order. > Try, for instance: > > $ dig +short www.google.com > > which uses DNS rotation as a form of load-balancing. > > In any case, you normally just take the first entry. >
On both 2.6.6 and 2.7.1, the doc says: ,---- | >>> help(socket.gethostbyaddr) | Help on built-in function gethostbyaddr in module _socket: | | gethostbyaddr(...) | gethostbyaddr(host) -> (name, aliaslist, addresslist) | | Return the true host name, a list of aliases, and a list of IP addresses, | for a host. The host argument is a string giving a host name or IP number. `---- so socket.gethostbyaddr(os.environ['REMOTE_ADDR'])[0] is the 0th element of the tuple, i.e. the *name* that is returned. So (unless I'm confused which is always a distinct possibility) this has nothing to do with load-balancing or multiple addresses: it's the equivalent of dig +short -x <addr> Nick -- http://mail.python.org/mailman/listinfo/python-list
