Readline: Terminal Interaction

The "readline" collection (not to be confused with Racket’s read-line function) provides glue for using the Editline library or GNU’s Readline library with the Racket read-eval-print-loop.

Due to licensing issues, the Readline collection is by default backed by the Editline library. To switch to GNU’s Readline library, either install the "readline-gpl" package, or set the PLT_READLINE_LIB environment variable to the library, which configures the readline collection to use Readline.

1 Normal Use of Readline

The readline library installs a Readline-based input port (whose name is 'readline-input) and hooks the prompt-and-read part of Racket’s read-eval-print-loop to interact with it.

You can start Racket with

  racket -il readline

or evaluate

(require readline)

in the Racket read-eval-print-loop to load Readline manually. You can also put (require readline) in your "~/.racketrc", so that Racket automatically loads Readline support in interactive mode.

If you want to enable Readline support only sometimes—such as only when you use an xterm, and not when you use an Emacs shell—then you can use dynamic-require, as in the following example:

(when (regexp-match? #rx"xterm"

(getenv "TERM"))

(dynamic-require 'readline #f))

The readline library automatically checks whether the current input port is a terminal, as determined by terminal-port?, and also checks that its name, as determined by object-name, is 'stdin, and it installs Readline only to replace the stdin terminal port. The readline/rep-start module installs Readline without a terminal check.

By default, Readline’s completion is set to use the visible bindings in the current namespace. This is far from ideal, but it’s better than Readline’s default filename completion which is rarely useful. In addition, the Readline history is stored across invocations in Racket’s preferences file, assuming that Racket exits normally.

The readline library adjusts read-eval-print-loop by setting the prompt read handler as determined by current-prompt-read. The call to the read interaction handler (as determined by current-read-interaction) is parameterized to set readline-prompt, so that a prompt will be printed when reading starts. To compensate for the prompt printed via readline-prompt, when the interaction input port’s name (as produced by function in the current-get-interaction-input-port parameter) is 'readline-input, the prompt read handler skips printing a prompt; otherwise, it displays a prompt as determined by current-prompt.

For more fine-grained control, such as conditionally loading Readline based on an environment variable, edit "~/.racketrc" manually.

If required through readline, pre-readline-input-port will be an input port only when the current-input-port is actually replaced. Otherwise, it is #f.

Using pre-readline-input-port is useful for sending the original stdin to subprocesses. Subprocesses generally require an input port backed by a file descriptor, and many interactive programs behave differently when they have a terminal file descriptor. Otherwise, pre-readline-input-port should not be used, as reading from it will interfere with the readline port.

Added in version 1.1 of package readline-lib.

2 Interacting with the Readline-Enabled Input Port

The reading facility that the new input port provides can be customized with the following parameters.

(parameterize ([readline-prompt some-byte-string])

...code-that-reads...)

This expression makes the first call to Readline use the prompt, and subsequent calls will use an all-spaces prompt of the same length (for example, when you’re reading an S-expression). The normal value of readline-prompt is #f for an empty prompt (and spaces after the prompt is used, which is why you should use parameterize to restore it to #f).

A proper solution would be to install a custom output port, too, which keeps track of text that is displayed without a trailing newline. As a cheaper solution, if line-counting is enabled for the terminal’s output-port, then a newline is printed before reading if the column is not 0. (The readline library enables line-counting for the output port.)

Warning: The Readline library uses the output port directly. You should not use it when current-input-port has been modified, or when it was not a terminal port when Racket was started (eg, when reading input from a pipe). Expect some problems if you ignore this warning (not too bad, mostly problems with detecting an EOF).

3 Direct Bindings for Readline Hackers

The rl_completion_append_character is used by the readline library whenever a completion function returns a single option, and it is therefore chosen. The choice will be filled in on the command line, and then the rl_completion_append_character is appended.

As an example, you could set the completion append character to a slash when completing the name of a directory, a space when you know the user will want to write another argument, or #\nul to avoid appending when a user might want to write something directly after the completion, such as punctuation or an extension of a word.

If you want to make a completion function to more easily write the names of your favorite characters (and share your excitement about them), a use of set-completion-append-character! may look like this:

(define (christmas-character-complete name-str)

(set-completion-append-character! #\!)

(filter (λ (x) (string-prefix? x name-str))

'("Rudolf" "Hermie" "Bumble" "Yukon" "Clarise" "Santa")))

Added in version 1.1 of package readline-lib.

The readline-redisplay function can be used together with readline-newline to prevent a background thread from cluttering up the user input by interleaving its output. For example, an unsafe wrapper function for the thread’s output might look like the following:

(define (with-thread-safe-output output-thunk)

(dynamic-wind

(lambda ()

(start-atomic)

(readline-newline))

output-thunk

(lambda ()

(readline-redisplay)

(end-atomic))))