3.1 Modules: module, module*, ...

The module Form in The Racket Guide introduces module.

For a module-like form that works in definition contexts other than the top level or a module body, there’s define-package, but using a separate module or submodule is usually better.

The module-path form must be as for require, and it supplies the initial bindings for the body forms. That is, it is treated like a (require module-path) prefix before the forms, except that the bindings introduced by module-path can be shadowed by definitions and requires in the module body forms.

If a single form is provided, then it is partially expanded in a module-begin context. If the expansion leads to #%plain-module-begin, then the body of the #%plain-module-begin is the body of the module. If partial expansion leads to any other primitive form, then the form is wrapped with #%module-begin using the lexical context of the module body; this identifier must be bound by the initial module-path import, and its expansion must produce a #%plain-module-begin to supply the module body. Finally, if multiple forms are provided, they are wrapped with #%module-begin, as in the case where a single form does not expand to #%plain-module-begin.

After such wrapping, if any, and before any expansion, an 'enclosing-module-name property is attached to the #%module-begin syntax object (see Syntax Object Properties); the property’s value is a symbol corresponding to id.

Each form is partially expanded (see Partial Expansion) in a module context. Further action depends on the shape of the form:

• If it is a begin form, the sub-forms are flattened out into the module’s body and immediately processed in place of the begin.

• If it is a define-syntaxes form, then the right-hand side is evaluated (in phase 1), and the binding is immediately installed for further partial expansion within the module. Evaluation of the right-hand side is parameterized to set current-namespace as in let-syntax.

• If it is a begin-for-syntax form, then the body is expanded (in phase 1) and evaluated. Expansion within a begin-for-syntax form proceeds with the same partial-expansion process as for a module body, but in a higher phase, and saving all #%provide forms for all phases until the end of the module’s expansion. Evaluation of the body is parameterized to set current-namespace as in let-syntax.

• If the form is a #%require form, bindings are introduced immediately, and the imported modules are instantiated or visited as appropriate.

• If the form is a #%provide form, then it is recorded for processing after the rest of the body.

• If the form is a define-values form, then the binding is installed immediately, but the right-hand expression is not expanded further.

• If the form is a module form, then it is immediately expanded and declared for the extent of the current top-level enclosing module’s expansion.

• If the form is a module* form, then it is not expanded further.

• Similarly, if the form is an expression, it is not expanded further.

After all forms have been partially expanded this way, then the remaining expression forms (including those on the right-hand side of a definition) are expanded in an expression context. After all expression forms, #%provide forms are processed in the order in which they appear (independent of phase) in the expanded module. Finally, all module* forms are expanded in order, so that each becomes available for use by subsequent module* forms; the enclosing module itself is also available for use by module* submodules.

The scope of all imported identifiers covers the entire module body, except for nested module and module* forms (assuming a non-#f module-path in the latter case). The scope of any identifier defined within the module body similarly covers the entire module body except for such nested module and module* forms. The ordering of syntax definitions does not affect the scope of the syntax names; a transformer for A can produce expressions containing B, while the transformer for B produces expressions containing A, regardless of the order of declarations for A and B. However, a syntactic form that produces syntax definitions must be defined before it is used.

No identifier can be imported or defined more than once at any phase level within a single module, except that a definition via define-values or define-syntaxes can shadow an import via #%require—as long as no preceding #%declare form includes #:require=defined. Every exported identifier must be imported or defined. No expression can refer to a top-level variable. A module* form in which the enclosing module’s bindings are visible (i.e., a nested module* with #f instead of a module-path) can define or import bindings that shadow the enclosing module’s bindings.

The evaluation of a module form does not evaluate the expressions in the body of the module (except sometimes for redeclarations; see Module Redeclarations). Evaluation merely declares a module, whose full name depends both on id or (current-module-declare-name).

A module body is executed only when the module is explicitly instantiated via require or dynamic-require. On invocation, imported modules are instantiated in the order in which they are required into the module (although earlier instantiations or transitive requires can trigger the instantiation of a module before its order within a given module). Then, expressions and definitions are evaluated in order as they appear within the module. Each evaluation of an expression or definition is wrapped with a continuation prompt (see call-with-continuation-prompt) for the default prompt tag and using a prompt handler that re-aborts and propagates its argument to the next enclosing prompt. Each evaluation of a definition is followed, outside of the prompt, by a check that each of the definition’s variables has a value; if the portion of the prompt-delimited continuation that installs values is skipped, then the exn:fail:contract:variable? exception is raised.

Portions of a module body at higher phase levels are delimited similarly to run-time portions. For example, portions of a module within begin-for-syntax are delimited by a continuation prompt both as the module is expanded and when it is visited. The evaluation of a define-syntaxes form is delimited, but unlike define-values, there is no check that the syntax definition completed.

Accessing a module-level variable before it is defined signals a run-time error, just like accessing an undefined global variable. If a module (in its fully expanded form) does not contain a set! for an identifier that defined within the module, then the identifier is a constant after it is defined; its value cannot be changed afterward, not even through reflective mechanisms. The compile-enforce-module-constants parameter, however, can be used to disable enforcement of constants.

When a syntax object representing a module form has a 'module-language syntax property attached, and when the property value is a vector of three elements where the first is a module path (in the sense of module-path?) and the second is a symbol, then the property value is preserved in the corresponding compiled and/or declared module. The third component of the vector should be printable and readable, so that it can be preserved in marshaled bytecode. The racket/base and racket languages attach '#(racket/language-info get-info #f) to a module form. See also module-compiled-language-info, module->language-info, and racket/language-info.

See also Modules and Module-Level Variables, Module Expansion, Phases, and Visits, and Information on Expanded Modules.

Changed in version 6.3 of package base: Changed define-syntaxes and define-values to shadow any preceding import, and dropped the use of 'submodule syntax property values on nested module or module* forms.

Like module, but only for declaring a submodule within a module, and for submodules that may require the enclosing module.

Instead of a module-path after id, #f indicates that all bindings from the enclosing module are visible in the submodule. In that case, begin-for-syntax forms that wrap the module* form shift the phase level of the enclosing module’s bindings relative to the submodule. The macro expander handles such nesting by shifting the phase level of the module* form so that its body starts at phase level 0, expanding, and then reverting the phase level shift; beware that this process can leave syntax objects as 'origin syntax property values out-of-sync with the expanded module.

When a module* form has a module-path, the submodule expansion starts by removing the scopes of the enclosing module, the same as the module form. No shifting compensates for any begin-for-syntax forms that may wrap the submodule.

Declares and/or adds to a submodule named id.

Each addition for id is combined in order to form the entire submodule using (module* id #f ....) at the end of the enclosing module. If there is only one module+ for a given id, then (module+ id form ...) is equivalent to (module* id #f form ...), but still moved to the end of the enclosing module.

A syntax property on the module* form with the key 'origin-form-srcloc records the srcloc for every contributing module+ form.

When a module contains multiple submodules declared with module+, then the relative order of the initial module+ declarations for each submodule determines the relative order of the module* declarations at the end of the enclosing module.

A submodule must not be defined using module+ and module or module*. That is, if a submodule is made of module+ pieces, then it must be made only of module+ pieces.

Changed in version 8.9.0.1 of package base: Added 'origin-form-srcloc syntax property.

The #%module-begin form of racket/base wraps every top-level expression to print non-#<void> results using current-print.

The #%module-begin form of racket/base also declares a configure-runtime submodule (before any other form), unless some form is either an immediate module or module* form with the name configure-runtime. If a configure-runtime submodule is added, the submodule calls the configure function of racket/runtime-config.

Like #%module-begin, but without adding a configure-runtime submodule.

• #:cross-phase-persistent — declares the module as cross-phase persistent, and reports a syntax error if the module does not meet the import or syntactic constraints of a cross-phase persistent module.

• #:empty-namespace — declares that module->namespace for this module should produce a namespace with no bindings; limiting namespace support in this way can reduce the lexical information that otherwise must be preserved for the module.

• #:require=define — declares that no subsequent definition immediately with the module body is allowed to shadow a #%require (or require) binding. This declaration does not affect shadowing of a module’s initial imports (i.e., the module’s language).

• #:unsafe — declares that the module can be compiled without checks that could trigger exn:fail:contract, and the resulting behavior is unspecified for an evaluation where exn:fail:contract should have been raised; see also Unsafe Operations. For example, a use of car can be compiled as a use of unsafe-car, and the behavior is unspecified is unsafe-car is applied to a non-pair. The #:unsafe declaration keyword is allowed only when the current code inspector is the initial one. Macros can generate conditionally unsafe code, depending on the expansion context, by expanding to a use of (variable-reference-from-unsafe? (#%variable-reference)).

• #:realm identifier — declares that the module and any procedures within the module are given a realm that is the symbol form of identifier, effectively overriding the value of current-compile-realm.

A #%declare form must appear in a module context or a module-begin context. Each declaration-keyword can be declared at most once within a module body.

Changed in version 6.3 of package base: Added #:empty-namespace.
Changed in version 7.9.0.5: Added #:unsafe.
Changed in version 8.4.0.2: Added #:realm.
Changed in version 8.6.0.9: Added #:require=define.