Struct miltr_server::Server

pub struct Server<'m, M: Milter> { /* private fields */ }

The entry point to host a milter server

Implementations

impl<'m, M: Milter> Server<'m, M>

pub fn new( milter: &'m mut M, quit_on_abort: bool, max_buffer_size: usize ) -> Self

Create a new Server to handle connections

pub fn default_postfix(milter: &'m mut M) -> Self

Create a server with defaults working with postfix.

The main difference is treating the call to abort like a call to quit. See this comment as a source in the postfix docs

AFAIK, originally there where three use cases individual methods:

1. Abort
   The current smtp client that is connected to the milter client has finished. Next mail arrives.
2. Quit
   The current smtp client that was connected to the milter client has quit it’s connection and the milter client will now quit this connection.
3. Quit NC
   The current smtp client that was connected to the milter client has quit it’s connection but the milter client would like to re-use this connection for someone else.

Different implementation mix them up, making e.g. postfix just always opening up a new connection for every milter conversation.

pub async fn handle_connection<RW: AsyncRead + AsyncWrite + Unpin + Send>( &mut self, socket: RW ) -> Result<(), Error<M::Error>>

Handle a single milter connection.

Arguments

• milter: the object implementing [crate::server::Milter]. It’s methods will be called at the appropriate times.

Errors

This basically errors for three cases: Io Problems, Codec Problems and problems returned by the milter implementation.

Have a look at [enum@crate::server::Error] for more information.

Trait Implementations

impl<'m, M: Debug + Milter> Debug for Server<'m, M>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.