Apropos

*debug-io*, *error-output*, *query-io*, *standard-input*, *standard-output*, *trace-output*Variable

    Value Type

    For *standard-input*: an input stream

    For *error-output*, *standard-output*, and *trace-output*: an output stream.

    For *debug-io*, *query-io*: a bidirectional stream.

    Initial Value

    implementation-dependent, but it must be an open stream that is not a generalized synonym stream to an I/O customization variables but that might be a generalized synonym stream to the value of some I/O customization variable. The initial value might also be a generalized synonym stream to either the symbol *terminal-io* or to the stream that is its value.

    Description

    These variables are collectively called the standardized I/O customization variables. They can be bound or assigned in order to change the default destinations for input and/or output used by various standardized operators and facilities.

    The value of *debug-io*, called debug I/O, is a stream to be used for interactive debugging purposes.

    The value of *error-output*, called error output, is a stream to which warnings and non-interactive error messages should be sent.

    The value of *query-io*, called query I/O, is a bidirectional stream to be used when asking questions of the user. The question should be output to this stream, and the answer read from it.

    The value of *standard-input*, called standard input, is a stream that is used by many operators as a default source of input when no specific input stream is explicitly supplied.

    The value of *standard-output*, called standard output, is a stream that is used by many operators as a default destination for output when no specific output stream is explicitly supplied.

    The value of *trace-output*, called trace output, is the stream on which traced functions (see trace) and the time macro print their output.

    Examples
     (with-output-to-string (*error-output*) 
       (warn "this string is sent to *error-output*")) 
      "Warning: this string is sent to *error-output* 
    " ;The exact format of this string is implementation-dependent. 
    
    
     (with-input-from-string (*standard-input* "1001") 
        (+ 990 (read)))  1991 
    
    
     (progn (setq out (with-output-to-string (*standard-output*) 
                         (print "print and format t send things to") 
                         (format t "*standard-output* now going to a string"))) 
            :done) 
     :DONE 
     out 
     " 
    \"print and format t send things to\" *standard-output* now going to a string" 
    
    
     (defun fact (n) (if (< n 2) 1 (* n (fact (- n 1))))) 
     FACT 
     (trace fact) 
     (FACT) 
    ;; Of course, the format of traced output is implementation-dependent. 
     (with-output-to-string (*trace-output*) 
       (fact 3)) 
     " 
    1 Enter FACT 3 
    | 2 Enter FACT 2 
    |   3 Enter FACT 1 
    |   3 Exit FACT 1 
    | 2 Exit FACT 2 
    1 Exit FACT 6"
    See Also

    *terminal-io*, synonym-stream, time, trace, Chapter 9 (Conditions), Chapter 23 (Reader), Chapter 22 (Printer)

    Notes

    The intent of the constraints on the initial value of the I/O customization variables is to ensure that it is always safe to bind or assign such a variable to the value of another I/O customization variable, without unduly restricting implementation flexibility.

    It is common for an implementation to make the initial values of *debug-io* and *query-io* be the same stream, and to make the initial values of *error-output* and *standard-output* be the same stream.

    The functions y-or-n-p and yes-or-no-p use query I/O for their input and output.

    In the normal Lisp read-eval-print loop, input is read from standard input. Many input functions, including read and read-char, take a stream argument that defaults to standard input.

    In the normal Lisp read-eval-print loop, output is sent to standard output. Many output functions, including print and write-char, take a stream argument that defaults to standard output.

    A program that wants, for example, to divert output to a file should do so by binding *standard-output*; that way error messages sent to *error-output* can still get to the user by going through *terminal-io* (if *error-output* is bound to *terminal-io*), which is usually what is desired.