Struct NativeVclResponseShim 

pub struct NativeVclResponseShim;

Placeholder response implementation for native Varnish backends.

This type exists only to satisfy the trait bounds for Backend<S, T> when wrapping native backends. None of its methods should ever be called.

Trait Implementations

impl Debug for NativeVclResponseShim

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

Formats the value using the given formatter.

impl VclBackend<NativeVclResponseShim> for NativeVclBackendShim

fn get_response( &self, _ctx: &mut Ctx<'_>, ) -> Result<Option<NativeVclResponseShim>, VclError>

If the VCL pick this backend (or a director ended up choosing it), this method gets called so that the VclBackend implementer can:

fn finish(&self, _ctx: &mut Ctx<'_>)

Once a backend transaction is finished, the Backend has a chance to clean up, collect data and others in the finish methods.

fn probe(&self, _ctx: &mut Ctx<'_>) -> (bool, SystemTime)

Is your backend healthy, and when did its health change for the last time.

fn pipe(&self, ctx: &mut Ctx<'_>, _tcp_stream: TcpStream) -> StreamClose

If your backend is used inside vcl_pipe, this method is in charge of sending the request headers that Varnish already read, and then the body. The second argument, a TcpStream is the raw client stream that Varnish was using (converted from a raw fd).

fn event(&self, _event: VclEvent)

The method will get called when the VCL changes temperature or is discarded. It’s notably a chance to start/stop probes to consume fewer resources.

fn panic(&self, _vsb: &mut Buffer<'_>)

fn report(&self, ctx: &mut Ctx<'_>, vsb: &mut Buffer<'_>)

Generate simple report output for varnishadm backend.list (no flags)

fn report_details(&self, _ctx: &mut Ctx<'_>, _vsb: &mut Buffer<'_>)

Generate detailed report output for varnishadm backend.list -p

fn report_json(&self, ctx: &mut Ctx<'_>, vsb: &mut Buffer<'_>)

Generate simple JSON report output for varnishadm backend.list -j

fn report_details_json(&self, _ctx: &mut Ctx<'_>, vsb: &mut Buffer<'_>)

Generate detailed JSON report output for varnishadm backend.list -j -p

impl VclResponse for NativeVclResponseShim

fn read(&mut self, buf: &mut [u8]) -> Result<usize, VclError>

The only mandatory method, it will be called repeated so that the VclResponse object can fill buf. The transfer will stop if any of its calls returns an error, and it will complete successfully when Ok(0) is returned.

fn len(&self) -> Option<usize>

If returning Some(_), we know the size of the body generated, and it’ll be used to fill the content-length header of the response. Otherwise, chunked encoding will be used, which is what’s assumed by default.

fn get_ip(&self) -> Result<Option<SocketAddr>, VclError>

Potentially return the IP:port pair that the backend is using to transfer the body. It might not make sense for your implementation.