The Resolver is responsible for performing recursive queries to lookup domain names.

This is a 100% in process DNS resolver. It does not use the Host OS' resolver. If what is desired is to use the Host OS' resolver, generally in the system's libc, then the std::net::ToSocketAddrs variant over &str is what should be used. As of the initial release, trust-dns-resolver it does not currently support search paths or ndot recursive lookups. It only supports FQDN, where the name must be specified with the final dot, e.g. www.example.com.. This limitation will be removed in future releases.

Unlike the trust-dns client, this tries to provide a simpler interface to perform DNS queries. For update options, i.e. Dynamic DNS, the trust-dns crate must be used directly.

There are two types for performing DNS queries, Resolver and ResolverFuture. Resolver is the easiest to work with, it is a wrapper around ResolverFuture. ResolverFuture is a Tokio based async resolver, and can be used inside any Tokio based system.

Notes on current limitations: DNSSec is not yet supported in the Resolver, use the trust-dns Client if DNSSec validation is needed. Many standard system options and configurations are not yet supported. The host system /etc/resolv.conf is not yet being used. Recursion is not yet built, i.e. CNAME chains will not be fully resolved if an address is not returned, though many upstream resolvers will perform this making it a none blocking issue.

This as best as possible attempts to abide by the the DNS RFCs, please file issues at https://github.com/bluejekyll/trust-dns .

Usage

Declare dependency

[dependency]
trust-dns-resolver = "^0.1"

Extern the crate for usage in the library

extern crate trust_dns_resolver;

Using the Synchronous Resolver

This uses the default configuration. Currently this sets the google resolvers as the upstream resolvers.

use std::net::*;
use trust_dns_resolver::Resolver;
use trust_dns_resolver::config::*;

// Construct a new Resolver with default configuration options
let mut resolver = Resolver::new(ResolverConfig::default(), ResolverOpts::default()).unwrap();

// Lookup the IP addresses associated with a name.
// NOTE: do not forget the final dot, as the resolver does not yet support search paths.
let mut response = resolver.lookup_ip("www.example.com.").unwrap();

// There can be many addresses associated with the name,
//  this can return IPv4 and/or IPv6 addresses
let address = response.next().expect("no addresses returned!");
if address.is_ipv4() {
    assert_eq!(address, IpAddr::V4(Ipv4Addr::new(93, 184, 216, 34)));
} else {
    assert_eq!(address, IpAddr::V6(Ipv6Addr::new(0x2606, 0x2800, 0x220, 0x1, 0x248, 0x1893, 0x25c8, 0x1946)));
}

Using the host system config

On Unix systems, the /etc/resolv.conf can be used for configuration. Not all options specified in the host systems resolv.conf are applicable or compatible with this software. In addition there may be additional options supported which the host system does not. Example:

# use std::net::*;
# use trust_dns_resolver::Resolver;
// Use the host OS'es `/etc/resolv.conf`
let mut resolver = Resolver::from_system_conf().unwrap();
let mut response = resolver.lookup_ip("www.example.com.").unwrap();

Using the Tokio/Async Resolver

For more advanced asynchronous usage, the ResolverFuture is integrated with Tokio. In fact, the ResolverFuture is used by the synchronous Resolver for all lookups.

# extern crate futures;
# extern crate tokio_core;
# extern crate trust_dns_resolver;
# fn main() {
use std::net::*;
use tokio_core::reactor::Core;
use trust_dns_resolver::ResolverFuture;
use trust_dns_resolver::config::*;

// We need a Tokio reactor::Core to run the resolver
//  this is responsible for running all Future tasks and registering interest in IO channels
let mut io_loop = Core::new().unwrap(); 

// Construct a new Resolver with default configuration options
let mut resolver = ResolverFuture::new(ResolverConfig::default(), ResolverOpts::default(), &io_loop.handle());

// Lookup the IP addresses associated with a name.
// NOTE: do not forget the final dot, as the resolver does not yet support search paths.
// This returns a future that will lookup the IP addresses, it must be run in the Core to
//  to get the actual result.
let lookup_future = resolver.lookup_ip("www.example.com.");

// Run the lookup until it resolves or errors
let mut response = io_loop.run(lookup_future).unwrap();

// There can be many addresses associated with the name,
//  this can return IPv4 and/or IPv6 addresses
let address = response.next().expect("no addresses returned!");
if address.is_ipv4() {
    assert_eq!(address, IpAddr::V4(Ipv4Addr::new(93, 184, 216, 34)));
} else {
    assert_eq!(address, IpAddr::V6(Ipv6Addr::new(0x2606, 0x2800, 0x220, 0x1, 0x248, 0x1893, 0x25c8, 0x1946)));
}
# }

Generally after a lookup in an asynchornous context, there would probably be a connection made to a server, for example:

let result = io_loop.run(lookup_future.and_then(|ips| {
                                 let ip = ips.next().unwrap();
                                 TcpStream::connect()
                             }).and_then(|conn| /* do something with the connection... */)
                         ).unwrap();