rust-imap/src/extensions/idle.rs
2018-11-22 14:00:58 -05:00

171 lines
6.3 KiB
Rust

//! Adds support for the IMAP IDLE command specificed in [RFC
//! 2177](https://tools.ietf.org/html/rfc2177).
use crate::client::Session;
use crate::error::{Error, Result};
use native_tls::TlsStream;
use std::io::{self, Read, Write};
use std::net::TcpStream;
use std::time::Duration;
/// `Handle` allows a client to block waiting for changes to the remote mailbox.
///
/// The handle blocks using the [`IDLE` command](https://tools.ietf.org/html/rfc2177#section-3)
/// specificed in [RFC 2177](https://tools.ietf.org/html/rfc2177) until the underlying server state
/// changes in some way. While idling does inform the client what changes happened on the server,
/// this implementation will currently just block until _anything_ changes, and then notify the
///
/// Note that the server MAY consider a client inactive if it has an IDLE command running, and if
/// such a server has an inactivity timeout it MAY log the client off implicitly at the end of its
/// timeout period. Because of that, clients using IDLE are advised to terminate the IDLE and
/// re-issue it at least every 29 minutes to avoid being logged off. [`Handle::wait_keepalive`]
/// does this. This still allows a client to receive immediate mailbox updates even though it need
/// only "poll" at half hour intervals.
///
/// As long as a [`Handle`] is active, the mailbox cannot be otherwise accessed.
#[derive(Debug)]
pub struct Handle<'a, T: Read + Write + 'a> {
session: &'a mut Session<T>,
keepalive: Duration,
done: bool,
}
/// Must be implemented for a transport in order for a `Session` using that transport to support
/// operations with timeouts.
///
/// Examples of where this is useful is for `Handle::wait_keepalive` and
/// `Handle::wait_timeout`.
pub trait SetReadTimeout {
/// Set the timeout for subsequent reads to the given one.
///
/// If `timeout` is `None`, the read timeout should be removed.
///
/// See also `std::net::TcpStream::set_read_timeout`.
fn set_read_timeout(&mut self, timeout: Option<Duration>) -> Result<()>;
}
impl<'a, T: Read + Write + 'a> Handle<'a, T> {
pub(crate) fn new(session: &'a mut Session<T>) -> Result<Self> {
let mut h = Handle {
session,
keepalive: Duration::from_secs(29 * 60),
done: false,
};
h.init()?;
Ok(h)
}
fn init(&mut self) -> Result<()> {
// https://tools.ietf.org/html/rfc2177
//
// The IDLE command takes no arguments.
self.session.run_command("IDLE")?;
// A tagged response will be sent either
//
// a) if there's an error, or
// b) *after* we send DONE
let mut v = Vec::new();
self.session.readline(&mut v)?;
if v.starts_with(b"+") {
self.done = false;
return Ok(());
}
self.session.read_response_onto(&mut v)?;
// We should *only* get a continuation on an error (i.e., it gives BAD or NO).
unreachable!();
}
fn terminate(&mut self) -> Result<()> {
if !self.done {
self.done = true;
self.session.write_line(b"DONE")?;
self.session.read_response().map(|_| ())
} else {
Ok(())
}
}
/// Internal helper that doesn't consume self.
///
/// This is necessary so that we can keep using the inner `Session` in `wait_keepalive`.
fn wait_inner(&mut self) -> Result<()> {
let mut v = Vec::new();
match self.session.readline(&mut v).map(|_| ()) {
Err(Error::Io(ref e))
if e.kind() == io::ErrorKind::TimedOut || e.kind() == io::ErrorKind::WouldBlock =>
{
// we need to refresh the IDLE connection
self.terminate()?;
self.init()?;
self.wait_inner()
}
r => r,
}
}
/// Block until the selected mailbox changes.
pub fn wait(mut self) -> Result<()> {
self.wait_inner()
}
}
impl<'a, T: SetReadTimeout + Read + Write + 'a> Handle<'a, T> {
/// Set the keep-alive interval to use when `wait_keepalive` is called.
///
/// The interval defaults to 29 minutes as dictated by RFC 2177.
pub fn set_keepalive(&mut self, interval: Duration) {
self.keepalive = interval;
}
/// Block until the selected mailbox changes.
///
/// This method differs from [`Handle::wait`] in that it will periodically refresh the IDLE
/// connection, to prevent the server from timing out our connection. The keepalive interval is
/// set to 29 minutes by default, as dictated by RFC 2177, but can be changed using
/// [`Handle::set_keepalive`].
///
/// This is the recommended method to use for waiting.
pub fn wait_keepalive(self) -> Result<()> {
// The server MAY consider a client inactive if it has an IDLE command
// running, and if such a server has an inactivity timeout it MAY log
// the client off implicitly at the end of its timeout period. Because
// of that, clients using IDLE are advised to terminate the IDLE and
// re-issue it at least every 29 minutes to avoid being logged off.
// This still allows a client to receive immediate mailbox updates even
// though it need only "poll" at half hour intervals.
let keepalive = self.keepalive;
self.wait_timeout(keepalive)
}
/// Block until the selected mailbox changes, or until the given amount of time has expired.
pub fn wait_timeout(mut self, timeout: Duration) -> Result<()> {
self.session
.stream
.get_mut()
.set_read_timeout(Some(timeout))?;
let res = self.wait_inner();
self.session.stream.get_mut().set_read_timeout(None).is_ok();
res
}
}
impl<'a, T: Read + Write + 'a> Drop for Handle<'a, T> {
fn drop(&mut self) {
// we don't want to panic here if we can't terminate the Idle
self.terminate().is_ok();
}
}
impl<'a> SetReadTimeout for TcpStream {
fn set_read_timeout(&mut self, timeout: Option<Duration>) -> Result<()> {
TcpStream::set_read_timeout(self, timeout).map_err(Error::Io)
}
}
impl<'a> SetReadTimeout for TlsStream<TcpStream> {
fn set_read_timeout(&mut self, timeout: Option<Duration>) -> Result<()> {
self.get_ref().set_read_timeout(timeout).map_err(Error::Io)
}
}