On this page:
18.1.1 Interactive Mode
18.1.2 Module Mode
18.1.3 Load Mode
Version: 4.2.1

18.1 Running mzscheme and mred

Depending on command-line arguments, mzscheme or mred runs in interactive mode, module mode, or load mode.

18.1.1 Interactive Mode

When mzscheme is run with no command-line arguments (other than confguration options, like -j), then it starts a REPL with a >  prompt:

  Welcome to MzScheme

For information on GNU Readline support, see readline.

To initialize the REPL’s environment, mzscheme first requires the scheme/init module, which provides all of scheme, and also installs pretty-print for display results. Finally, mzscheme loads the file reported by (find-system-path 'init-file), if it exists, before starting the REPL.

If any command-line arguments are provided (other than configuration options), add -i or --repl to re-enable the REPL. For example,

  mzscheme -e '(display "hi\n")' -i

displays “hi” on start-up, but still presents a REPL.

If module-requiring flags appear before -i/--repl, they cancel the automatic requiring of scheme/init. This behavior can be used to initialize the REPL’s environment with a different language. For example,

  mzscheme -l scheme/base -i

starts a REPL using a much smaller initial language (that loads much faster). Beware that most modules do not provide the basic syntax of Scheme, including function-call syntax and require. For example,

  mzscheme -l scheme/date -i

produces a REPL that fails for every expression, because scheme/date provides only a few functions, and not the #%top-interaction and #%app bindings that are needed to evaluate top-level function calls in the REPL.

If a module-requiring flag appears after -i/--repl instead of before it, then the module is required after scheme/init to augment the initial environment. For example,

  mzscheme -i -l scheme/date

starts a useful REPL with scheme/date available in addition to the exports of scheme.

18.1.2 Module Mode

If a file argument is supplied to mzscheme before any command-line switch (other than configuration options), then the file is required as a module, and (unless -i/--repl is specified), no REPL is started. For example,

  mzscheme hello.ss

requires the "hello.ss" module and then exits. Any argument after the file name, flag or otherwise, is preserved as a command-line argument for use by the required module via current-command-line-arguments.

If command-line flags are used, then the -u or --require-script flag can be used to explicitly require a file as a module. The -t or --require flag is similar, except that additional command-line flags are processed by mzscheme, instead of preserved for the required module. For example,

  mzscheme -t hello.ss -t goodbye.ss

requires the "hello.ss" module, then requires the "goodbye.ss" module, and then exits.

The -l or --lib flag is similar to -t/--require, but it requires a module using a lib module path instead of a file path. For example,

  mzscheme -l compiler

is the same as running the mzc executable with no arguments, since the compiler module is the main mzc module.

Note that if you wanted to pass command-line flags to compiler above, you would need to protect the flags with a --, so that mzscheme doesn’t try to parse them itself:

  mzscheme -l compiler -- --make prog.ss

18.1.3 Load Mode

The -f or --load flag supports loading top-level expressions in a file directly, as opposed to expressions within a module file. This evaluation is like starting a REPL and typing the expressions directly, except that the results are not printed. For example,

  mzscheme -f hi.ss

loads "hi.ss" and exits. Note that load mode is generally a bad idea, for the reasons explained in A Note to Readers with Scheme/Lisp Experience; using module mode is typically better.

The -e or --eval flag accepts an expression to evaluate directly. Unlike file loading, the result of the expression is printed, as in a REPL. For example,

  mzscheme -e '(current-seconds)'

prints the number of seconds since January 1, 1970.

For file loading and expression evaluation, the top-level environment is created in the same way for interactive mode: scheme/init is required unless another module is specified first. For example,

  mzscheme -l scheme/base -e '(current-seconds)'

likely runs faster, because it initializes the environment for evaluation using the smaller scheme/base language, instead of scheme/init.