Struct RustTlsDuplexStream

Source

pub struct RustTlsDuplexStream<C, S>where
    C: DerefMut + Deref<Target = ConnectionCommon<S>> + Send,
{ /* private fields */ }

Implementations§

Source §

impl<C, S> RustTlsDuplexStream<C, S>
where C: DerefMut + Deref<Target = ConnectionCommon<S>> + Send, S: SideData,

Source

pub fn new_unpooled<R, W>(con: C, read: R, write: W) -> Result<Self>
where R: Read + Send + 'static, W: Write + Send + 'static,

Creates a new ‘unpooled’ Tls stream wrapper. This is a good choice for an application such as a client that does not create connections and doesn’t have a thread pool.

This fn will spawn 2 new threads using thread::Builder::new().spawn(...). The threads will terminate when the returned stream is dropped and the read/write errors out.

§Resource Leaks

Be aware that a Read which “blocks” forever will cause the thread that’s reading to stay alive forever. Either set a reasonable connection read timeout so that your Read will eventually return or call your connections shutdown fn like TcpStream::shutdown if such a method exists to ensure that all threads are stopped and no resources are leaked.

§Errors

if thread::Builder::new().spawn fails to spawn 2 threads.

Source

pub fn new<R, W, T>(con: C, read: R, write: W, spawner: T) -> Result<Self>
where R: Read + Send + 'static, W: Write + Send + 'static, T: FnMut(Box<dyn FnOnce() + Send>) -> Result<()>,

Creates a new Tls stream wrapper.

This fn will spawn 2 new threads using the provided thread spawner function. The thread spawner function is called exactly twice. If the first call yields an error then it’s not called again. The functions (“threads”) will end when the returned stream wrapper is dropped.

§Resource Leaks

Be aware that a Read which “blocks” forever will cause the thread that’s reading to stay alive forever. Either set a reasonable connection read timeout so that your Read will eventually return or call your connections shutdown fn like TcpStream::shutdown if such a method exists to ensure that all threads are stopped and no resources are leaked.

§Errors

propagated from the spawner fn.

Source

pub fn write(&self, buffer: &[u8]) -> Result<usize>

see Write::write

§Errors

propagated from Write::write once subsequent writes/flushes turn into BrokenPipe

Source

pub fn flush(&self) -> Result<()>

see Write::flush

§Errors

propagated from Write::flush once subsequent writes/flushes turn into BrokenPipe

Source

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

see Read::read

§Errors

propagated from Read::read once subsequent reads turn into BrokenPipe

Source

pub fn set_read_timeout(&self, timeout: Option<Duration>) -> Result<()>

sets the timeout for the writing operation. This has no effect on the underlying connection and purely deals with internal writing semantics. Calls to fns that writs data will return TimedOut if no plain text data could be written. Cause of this is likely to be that the underlying connection does not read data fast enough. This is never caused by writing too much data.

§Errors

In case of poisoned mutex

Source

pub fn read_timeout(&self) -> Result<Option<Duration>>

Returns the current read timeout if any

§Errors

In case of poisoned mutex

Source

pub fn set_read_non_block(&self, on: bool) -> Result<()>

sets non-blocking mode for read. This has no effect on the underlying connection and purely deals with internal reading semantics. Calls to fns that read data will return WouldBlock immediately if no plain text data is available to be read.

§Errors

In case of poisoned mutex

Source

pub fn set_write_timeout(&self, timeout: Option<Duration>) -> Result<()>

sets the timeout for the writing operation. This has no effect on the underlying connection and purely deals with internal writing semantics. Calls to fns that writs data will return TimedOut if no plain text data could be written. Cause of this is likely to be that the underlying connection does not send data fast enough. This is never caused by reading too much data.