The trust arugment may also be a single IP address string or an array of trusted addresses, as plain IP addresses, CIDR-formatted strings, or IP/netmask strings.
This module also supports IPv6. Your IPv6 addresses will be normalized automatically (i.e. fe80::00ed:1 equals fe80:0:0:0:0:0:ed:1).
This module will automatically work with IPv4-mapped IPv6 addresses as well to support node.js in IPv6-only mode. This means that you do not have to specify both ::ffff:a00:1 and 10.0.0.1.
As a convenience, this module also takes certain pre-defined names in addition to IP addresses, which expand into IP addresses:
loopback: IPv4 and IPv6 loopback addresses (like ::1 and 127.0.0.1).
linklocal: IPv4 and IPv6 link-local addresses (like fe80::1:1:1:1 and 169.254.0.1).
uniquelocal: IPv4 private addresses and IPv6 unique-local addresses (like fc00:ac:1ab5:fff::1 and 192.168.0.1).
When trust is specified as a function, it will be called for each address to determine if it is a trusted address. The function is given two arguments: addr and i, where addr is a string of the address to check and i is a number that represents the distance from the socket address.
proxyaddr.all(req, [trust])
Return all the addresses of the request, optionally stopping at the first untrusted. This array is ordered from closest to furthest (i.e. arr[0] === req.connection.remoteAddress).
The optional trust argument takes the same arguments as trust does in proxyaddr(req, trust).
proxyaddr.compile(val)
Compiles argument val into a trust function. This function takes the same arguments as trust does in proxyaddr(req, trust) and returns a function suitable for proxyaddr(req, trust).
This function is meant to be optimized for use against every request. It is recommend to compile a trust function up-front for the trusted configuration and pass that to proxyaddr(req, trust) for each request.