JSONRPC2 v2.0.0 JSONRPC2.Clients.TCP View Source

A client for JSON-RPC 2.0 using a line-based TCP transport.

Link to this section Summary

Functions

Call the given method with params using the client pool named name with options.

Asynchronously call the given method with params using the client pool named name with options.

Send a notification with the given method and params using the client pool named name.

Receive the response for a previous cast/3 which returned a request_id.

Start a client pool named name, connected to host at port.

Stop the client pool with name name.

Link to this section Types

Specs

call_option() :: {:string_id, boolean()} | {:timeout, pos_integer()}

Specs

call_options() :: [call_option()]

Specs

cast_options() :: [{:string_id, boolean()}]

Specs

Specs

request_id() :: any()

Link to this section Functions

Link to this function

call(name, method, params, options \\ [])

View Source

Specs

call(atom(), JSONRPC2.method(), JSONRPC2.params(), boolean() | call_options()) ::
  {:ok, any()} | {:error, any()}

Call the given method with params using the client pool named name with options.

You can provide the option string_id: true for compatibility with pathological implementations, to force the request ID to be a string.

You can also provide the option timeout: 5_000 to set the timeout to 5000ms, for instance.

For backwards compatibility reasons, you may also provide a boolean for the options parameter, which will set string_id to the given boolean.

Link to this function

cast(name, method, params, options \\ [])

View Source

Specs

cast(atom(), JSONRPC2.method(), JSONRPC2.params(), boolean() | cast_options()) ::
  {:ok, request_id()} | {:error, :backlog_full}

Asynchronously call the given method with params using the client pool named name with options.

Use receive_response/1 with the request_id to get the response.

You can provide the option string_id: true for compatibility with pathological implementations, to force the request ID to be a string.

You can also provide the option timeout: 5_000 to set the timeout to 5000ms, for instance.

Additionally, you may provide the option pid: self() in order to specify which process should be sent the message which is returned by receive_response/1.

For backwards compatibility reasons, you may also provide a boolean for the options parameter, which will set string_id to the given boolean.

Link to this function

notify(name, method, params)

View Source

Specs

notify(atom(), JSONRPC2.method(), JSONRPC2.params()) ::
  {:ok, request_id()} | {:error, :backlog_full}

Send a notification with the given method and params using the client pool named name.

This function returns a request_id, but it should not be used with receive_response/1.

Link to this function

receive_response(request_id)

View Source

Specs

receive_response(request_id()) :: {:ok, any()} | {:error, any()}

Receive the response for a previous cast/3 which returned a request_id.

Link to this function

start(host, port, name, client_opts \\ [], pool_opts \\ [])

View Source

Specs

start(host(), :inet.port_number(), atom(), Keyword.t(), Keyword.t()) :: :ok

Start a client pool named name, connected to host at port.

You can optionally pass client_opts, detailed here, as well as pool_opts, detailed here.

In addition to the client_opts above, you can also pass:

  • line_packet - by default, packets consist of a 4 byte header containing an unsigned integer in big-endian byte order specifying the number of bytes in the packet, followed by that number of bytes (equivalent to the erlang inet packet type 4). If set to true, packets will instead be terminated by line-endings, for compatibility with older implementations.

Specs

stop(atom()) :: :ok | {:error, :shackle_not_started | :pool_not_started}

Stop the client pool with name name.