process_socket_server: The function that processes a command coming from the socket

In SciViews/svSocket: 'SciViews' - Socket Server

The function that processes a command coming from the socket

Description

This is the default R function called each time data is send by a client through a socket. It is possible to customize this function and to use customized versions for particular R socket servers.

Usage

process_socket_server(msg, socket, server_port, ...)

processSocket(msg, socket, server_port, ...)

Arguments

msg	the message send by the client, to be processed.
socket	the client socket identifier, as in get_socket_clients(). This is passed by the calling function and can be used internally.
server_port	the port on which the server is running, this is passed by the calling function and can be used internally.
...	anything you want to pass to process_socket_server(), but it needs to rework start_socket_server() to use it).

Details

There are special code that one can send to R to easily turn the server (possibly temporarily) into a given configuration. First, if you want to persistently store parameters for your client in the R server and make sure you retrieve the same parameters the next time you reconnect, you should specify your own identifier. This is done by sending ⁠<<<id=myID>>>⁠ at the very beginning of each of your commands. Always remember that, if you do not specify an identifier, the name of your socket will be used. Since socket names can be reused, you should always reinitialize the configuration of your server the first time you connect to it.

Then, sending ⁠<<<esc>>>⁠ breaks current multiline code submission and flushes the multiline buffer.

The sequence ⁠<<<q>>>⁠ at the beginning of a command indicates that the server wants to disconnect once the command is fully treated by R. Similarly, the sequence ⁠<<<Q>>>⁠ tells the server to disconnect the client before processing the command (no error message is returned to the client!).

It is easy to turn the server to evaluate R code (including multiline code) and return the result and disconnect by using the ⁠<<<e>>>⁠ sequence at the beginning of a command. Using ⁠<<<h>>>⁠ or ⁠<<<H>>>⁠ configures that server to process a (single-line code only) command silently and disconnect before (uppercase H) or after (lowercase h) processing that command. It is the less intrusive mode that is very useful for all commands that should be executed behind the scene between R and a R editor or IDE, like contextual help, calltips, completion lists, etc.). Note that using these modes in a server that is, otherwise, configured as a multi-line server does not break current multi-line buffer.

The other sequences that can be used are: ⁠<<<s>>>⁠ for a placeholder to configure the current server (with configuration parameters after it), and ⁠<<<n>>>⁠ to indicate a newline in your code (submitting two lines of code as a single one; also works with servers configured as single-line evaluators).

To debug the R socket server and inspect how commands send by a client are interpreted by this function, use options(debug.Socket = TRUE). This function uses svMisc::parse_text() and svMisc::capture_all() in order to evaluate R code in character string almost exactly the same way as if it was typed at the command line of a R console.

Value

The results of processing msg in a character string vector.

See Also

start_socket_server(), send_socket_clients(), par_socket_server(), svMisc::parse_text(), svMisc::capture_all()

Examples

## Not run: 
# A simple REPL (R eval/process loop) using basic features of processSocket()
repl <- function() {
  pars <- par_socket_server("repl", "", bare = FALSE)  # Parameterize the loop
  cat("Enter R code, hit <CTRL-C> or <ESC> to exit\n> ")   # First prompt
  repeat {
    entry <- readLines(n = 1) 				 # Read a line of entry
    if (entry == "") entry <- "<<<esc>>>"    # Exit from multiline mode
    cat(process_socket_server(entry, "repl", ""))    # Process the entry
  }
}
repl()

## End(Not run)

SciViews/svSocket documentation built on March 2, 2024, 5:50 p.m.