// Copyright 2013 The Gorilla WebSocket Authors. All rights reserved. // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. // Package websocket implements the WebSocket protocol defined in RFC 6455. // // Overview // // The Conn type represents a WebSocket connection. A server application calls // the Upgrader.Upgrade method from an HTTP request handler to get a *Conn: // // var upgrader = websocket.Upgrader{ // ReadBufferSize: 1024, // WriteBufferSize: 1024, // } // // func handler(w http.ResponseWriter, r *http.Request) { // conn, err := upgrader.Upgrade(w, r, nil) // if err != nil { // log.Println(err) // return // } // ... Use conn to send and receive messages. // } // // Call the connection's WriteMessage and ReadMessage methods to send and // receive messages as a slice of bytes. This snippet of code shows how to echo // messages using these methods: // // for { // messageType, p, err := conn.ReadMessage() // if err != nil { // log.Println(err) // return // } // if err := conn.WriteMessage(messageType, p); err != nil { // log.Println(err) // return // } // } // // In above snippet of code, p is a []byte and messageType is an int with value // websocket.BinaryMessage or websocket.TextMessage. // // An application can also send and receive messages using the io.WriteCloser // and io.Reader interfaces. To send a message, call the connection NextWriter // method to get an io.WriteCloser, write the message to the writer and close // the writer when done. To receive a message, call the connection NextReader // method to get an io.Reader and read until io.EOF is returned. This snippet // shows how to echo messages using the NextWriter and NextReader methods: // // for { // messageType, r, err := conn.NextReader() // if err != nil { // return // } // w, err := conn.NextWriter(messageType) // if err != nil { // return err // } // if _, err := io.Copy(w, r); err != nil { // return err // } // if err := w.Close(); err != nil { // return err // } // } // // Data Messages // // The WebSocket protocol distinguishes between text and binary data messages. // Text messages are interpreted as UTF-8 encoded text. The interpretation of // binary messages is left to the application. // // This package uses the TextMessage and BinaryMessage integer constants to // identify the two data message types. The ReadMessage and NextReader methods // return the type of the received message. The messageType argument to the // WriteMessage and NextWriter methods specifies the type of a sent message. // // It is the application's responsibility to ensure that text messages are // valid UTF-8 encoded text. // // Control Messages // // The WebSocket protocol defines three types of control messages: close, ping // and pong. Call the connection WriteControl, WriteMessage or NextWriter // methods to send a control message to the peer. // // Connections handle received close messages by calling the handler function // set with the SetCloseHandler method and by returning a *CloseError from the // NextReader, ReadMessage or the message Read method. The default close // handler sends a close message to the peer. // // Connections handle received ping messages by calling the handler function // set with the SetPingHandler method. The default ping handler sends a pong // message to the peer. // // Connections handle received pong messages by calling the handler function // set with the SetPongHandler method. The default pong handler does nothing. // If an application sends ping messages, then the application should set a // pong handler to receive the corresponding pong. // // The control message handler functions are called from the NextReader, // ReadMessage and message reader Read methods. The default close and ping // handlers can block these methods for a short time when the handler writes to // the connection. // // The application must read the connection to process close, ping and pong // messages sent from the peer. If the application is not otherwise interested // in messages from the peer, then the application should start a goroutine to // read and discard messages from the peer. A simple example is: // // func readLoop(c *websocket.Conn) { // for { // if _, _, err := c.NextReader(); err != nil { // c.Close() // break // } // } // } // // Concurrency // // Connections support one concurrent reader and one concurrent writer. // // Applications are responsible for ensuring that no more than one goroutine // calls the write methods (NextWriter, SetWriteDeadline, WriteMessage, // WriteJSON, EnableWriteCompression, SetCompressionLevel) concurrently and // that no more than one goroutine calls the read methods (NextReader, // SetReadDeadline, ReadMessage, ReadJSON, SetPongHandler, SetPingHandler) // concurrently. // // The Close and WriteControl methods can be called concurrently with all other // methods. // // Origin Considerations // // Web browsers allow Javascript applications to open a WebSocket connection to // any host. It's up to the server to enforce an origin policy using the Origin // request header sent by the browser. // // The Upgrader calls the function specified in the CheckOrigin field to check // the origin. If the CheckOrigin function returns false, then the Upgrade // method fails the WebSocket handshake with HTTP status 403. // // If the CheckOrigin field is nil, then the Upgrader uses a safe default: fail // the handshake if the Origin request header is present and the Origin host is // not equal to the Host request header. // // The deprecated package-level Upgrade function does not perform origin // checking. The application is responsible for checking the Origin header // before calling the Upgrade function. // // Buffers // // Connections buffer network input and output to reduce the number // of system calls when reading or writing messages. // // Write buffers are also used for constructing WebSocket frames. See RFC 6455, // Section 5 for a discussion of message framing. A WebSocket frame header is // written to the network each time a write buffer is flushed to the network. // Decreasing the size of the write buffer can increase the amount of framing // overhead on the connection. // // The buffer sizes in bytes are specified by the ReadBufferSize and // WriteBufferSize fields in the Dialer and Upgrader. The Dialer uses a default // size of 4096 when a buffer size field is set to zero. The Upgrader reuses // buffers created by the HTTP server when a buffer size field is set to zero. // The HTTP server buffers have a size of 4096 at the time of this writing. // // The buffer sizes do not limit the size of a message that can be read or // written by a connection. // // Buffers are held for the lifetime of the connection by default. If the // Dialer or Upgrader WriteBufferPool field is set, then a connection holds the // write buffer only when writing a message. // // Applications should tune the buffer sizes to balance memory use and // performance. Increasing the buffer size uses more memory, but can reduce the // number of system calls to read or write the network. In the case of writing, // increasing the buffer size can reduce the number of frame headers written to // the network. // // Some guidelines for setting buffer parameters are: // // Limit the buffer sizes to the maximum expected message size. Buffers larger // than the largest message do not provide any benefit. // // Depending on the distribution of message sizes, setting the buffer size to // a value less than the maximum expected message size can greatly reduce memory // use with a small impact on performance. Here's an example: If 99% of the // messages are smaller than 256 bytes and the maximum message size is 512 // bytes, then a buffer size of 256 bytes will result in 1.01 more system calls // than a buffer size of 512 bytes. The memory savings is 50%. // // A write buffer pool is useful when the application has a modest number // writes over a large number of connections. when buffers are pooled, a larger // buffer size has a reduced impact on total memory use and has the benefit of // reducing system calls and frame overhead. // // Compression EXPERIMENTAL // // Per message compression extensions (RFC 7692) are experimentally supported // by this package in a limited capacity. Setting the EnableCompression option // to true in Dialer or Upgrader will attempt to negotiate per message deflate // support. // // var upgrader = websocket.Upgrader{ // EnableCompression: true, // } // // If compression was successfully negotiated with the connection's peer, any // message received in compressed form will be automatically decompressed. // All Read methods will return uncompressed bytes. // // Per message compression of messages written to a connection can be enabled // or disabled by calling the corresponding Conn method: // // conn.EnableWriteCompression(false) // // Currently this package does not support compression with "context takeover". // This means that messages must be compressed and decompressed in isolation, // without retaining sliding window or dictionary state across messages. For // more details refer to RFC 7692. // // Use of compression is experimental and may result in decreased performance. package websocket