Non-standard macros and special forms
Making extra libraries and extensions available
require-library
[syntax] (require-library ID ...)
This form does all the necessary steps to make the libraries or extensions given
in ID ... available. It loads syntactic extensions, if needed and generates
code for loading/linking with core library modules or separately installed
extensions.
During interpretation/evaluation require-library performs one of the
following:
- If ID names a built-in feature chicken srfi-0 srfi-2 srfi-6 srfi-8 srfi-9 srfi-10 srfi-17 srfi-23 srfi-30 srfi-39 srfi-55, then nothing is done.
- If ID names one of the syntactic extensions chicken-syntax chicken-ffi-syntax, then this extension will be loaded.
- If ID names one of the core library units shipped with CHICKEN, then a (load-library 'ID) will be performed.
- If ID names an installed extension with the syntax or require-at-runtime attribute, then the extensions is loaded at compile-time, probably doing a run-time (require ...) for any run-time requirements.
- Otherwise, (require-library ID) is equivalent to (require 'ID).
During compilation, one of the following happens instead:
- If ID names a built-in feature chicken srfi-0 srfi-2 srfi-6 srfi-8 srfi-9 srfi-10 srfi-17 srfi-23 srfi-30 srfi-39 srfi-55, then nothing is done.
- If ID names one of the syntactic extensions chicken-syntax chicken-ffi-syntax, then this extension will be loaded at compile-time, making the syntactic extensions available in compiled code.
- If ID names one of the core library units shipped with CHICKEN, or if the option -uses ID has been passed to the compiler, then a (declare (uses ID)) is generated.
- If ID names an installed extension with the syntax or require-at-runtime attribute, then the extension is loaded at compile-time, and code is emitted to (require ...) any needed run-time requirements.
- Otherwise (require-library ID) is equivalent to (require 'ID).
To make long matters short - just use require-library and it will normally figure everything out for dynamically
loadable extensions and core library units.
ID should be a pure extension name and should not contain any path prefixes (for example dir/lib...) is illegal).
ID may also be a list that designates an extension-specifier. Currently the following extension specifiers are
defined:
- (srfi NUMBER ...) is required for SRFI-55 compatibility and is fully implemented
- (version ID NUMBER) is equivalent to ID, but checks at compile-time whether the extension named ID is installed and whether its version is equal or higher than NUMBER. NUMBER may be a string or a number, the comparison is done lexicographically (using string>=?).
See also: set-extension-specifier!
require-extension
[syntax] (require-extension ID ...)
This is equivalent to (require-library ID ...) but performs an implicit
import, if necessary.
This implementation of require-extension is compliant with SRFI-55
(see the SRFI-55 document for more information).
use
[syntax] (use ID ...)
use is just a shorter alias for require-extension.
Binding forms for optional arguments
optional
[syntax] (optional ARGS DEFAULT)
Use this form for procedures that take a single optional argument. If
ARGS is the empty list DEFAULT is evaluated and
returned, otherwise the first element of the list ARGS. It is
an error if ARGS contains more than one value.
(define (incr x . i) (+ x (optional i 1)))
(incr 10) ==> 11
(incr 12 5) ==> 17
case-lambda
[syntax] (case-lambda (LAMBDA-LIST1 EXP1 ...) ...)
Expands into a lambda that invokes the body following the first
matching lambda-list.
(define plus
(case-lambda
(() 0)
((x) x)
((x y) (+ x y))
((x y z) (+ (+ x y) z))
(args (apply + args))))
(plus) ==> 0
(plus 1) ==> 1
(plus 1 2 3) ==> 6
For more information see the documentation for
SRFI-16
let-optionals
[syntax] (let-optionals ARGS ((VAR1 DEFAULT1) ...) BODY ...)
Binding constructs for optional procedure arguments. ARGS should
be a rest-parameter taken from a lambda-list. let-optionals
binds VAR1 ... to available arguments in parallel, or
to DEFAULT1 ... if not enough arguments were provided.
let-optionals* binds VAR1 ... sequentially, so every
variable sees the previous ones. it is an error if any excess
arguments are provided.
(let-optionals '(one two) ((a 1) (b 2) (c 3))
(list a b c) ) ==> (one two 3)
let-optionals*
[syntax] (let-optionals* ARGS ((VAR1 DEFAULT1) ... [RESTVAR]) BODY ...)
Binding constructs for optional procedure arguments. ARGS should
be a rest-parameter taken from a lambda-list. let-optionals
binds VAR1 ... to available arguments in parallel, or
to DEFAULT1 ... if not enough arguments were provided.
let-optionals* binds VAR1 ... sequentially, so every
variable sees the previous ones. If a single variable RESTVAR
is given, then it is bound to any remaining arguments, otherwise it is
an error if any excess arguments are provided.
(let-optionals* '(one two) ((a 1) (b 2) (c a))
(list a b c) ) ==> (one two one)
Other binding forms
and-let*
[syntax] (and-let* (BINDING ...) EXP1 EXP2 ...)
SRFI-2. Bind sequentially and execute body. BINDING can
be a list of a variable and an expression, a list with a single
expression, or a single variable. If the value of an expression
bound to a variable is #f, the and-let* form
evaluates to #f (and the subsequent bindings and the body
are not executed). Otherwise the next binding is performed. If
all bindings/expressions evaluate to a true result, the body is
executed normally and the result of the last expression is the
result of the and-let* form. See also the documentation for
SRFI-2.
rec
[syntax] (rec NAME EXPRESSION)
[syntax] (rec (NAME VARIABLE ...) BODY ...)
Allows simple definition of recursive definitions. (rec NAME EXPRESSION) is
equivalent to (letrec ((NAME EXPRESSION)) NAME) and (rec (NAME VARIABLE ...) BODY ...)
is the same as (letrec ((NAME (lambda (VARIABLE ...) BODY ...))) NAME).
cut
[syntax] (cut SLOT ...)
[syntax] (cute SLOT ...)
Syntactic sugar for specializing parameters.
define-values
[syntax] (define-values (NAME ...) EXP)
Defines several variables at once, with the result values of expression
EXP.
fluid-let
[syntax] (fluid-let ((VAR1 X1) ...) BODY ...)
Binds the variables VAR1 ... dynamically to the values X1 ...
during execution of BODY ....
let-values
[syntax] (let-values (((NAME ...) EXP) ...) BODY ...)
Binds multiple variables to the result values of EXP ....
All variables are bound simultaneously.
let*-values
[syntax] (let*-values (((NAME ...) EXP) ...) BODY ...)
Binds multiple variables to the result values of EXP ....
The variables are bound sequentially.
(let*-values (((a b) (values 2 3))
((p) (+ a b)) )
p) ==> 5
letrec-values
[syntax] (letrec-values (((NAME ...) EXP) ...) BODY ...)
Binds the result values of EXP ... to multiple variables at once.
All variables are mutually recursive.
(letrec-values (((odd even)
(values
(lambda (n) (if (zero? n) #f (even (sub1 n))))
(lambda (n) (if (zero? n) #t (odd (sub1 n)))) ) ) )
(odd 17) ) ==> #t
parameterize
[syntax] (parameterize ((PARAMETER1 X1) ...) BODY ...)
Binds the parameters PARAMETER1 ... dynamically to the values
X1 ... during execution of BODY .... (see also:
make-parameter in Parameters). Note that PARAMETER may be any
expression that evaluates to a parameter procedure.
receive
[syntax] (receive (NAME1 ... [. NAMEn]) VALUEEXP BODY ...)
[syntax] (receive VALUEEXP)
SRFI-8. Syntactic sugar for call-with-values. Binds variables
to the result values of VALUEEXP and evaluates BODY ....
The syntax
(receive VALUEEXP)
is equivalent to
(receive _ VALUEEXP _)
set!-values
[syntax] (set!-values (NAME ...) EXP)
Assigns the result values of expression EXP to multiple
variables.
Substitution forms and macros
define-constant
[syntax] (define-constant NAME CONST)
Define a variable with a constant value, evaluated at compile-time.
Any reference to such a
constant should appear textually after its definition. This
construct is equivalent to define when evaluated or interpreted.
Constant definitions should only appear at toplevel. Note that constants
are local to the current compilation unit and are not available outside
of the source file in which they are defined. Names of constants still
exist in the Scheme namespace and can be lexically shadowed. If the
value is mutable, then the compiler is careful to preserve its identity.
CONST may be any constant expression, and may also refer to
constants defined via define-constant previously.
This for should only be used at top-level.
define-inline
[syntax] (define-inline (NAME VAR ... [. VAR]) BODY ...)
[syntax] (define-inline NAME EXP)
Defines an inline procedure. Any occurrence of NAME will be
replaced by EXP or (lambda (VAR ... [. VAR]) BODY ...).
This is similar to a macro, but variable-names and -scope will
be correctly handled. Inline substitutions take place after
macro-expansion. EXP should be a lambda-expression. Any
reference to NAME should appear textually after
its definition. Note that inline procedures are local to the current
compilation unit and are not available outside of the source file in
which they are defined. Names of inline procedures still exist in the
Scheme namespace and can be lexically shadowed. This construct is
equivalent to define when evaluated or interpreted. Inline
definitions should only appear at toplevel.
define-for-syntax
[syntax] (define-for-syntax (NAME VAR ... [. VAR]) EXP1 ...)
[syntax] (define-for-syntax NAME [VALUE])
Defines the toplevel variable NAME at macro-expansion time. This can
be helpful when you want to define support procedures for use in macro-transformers,
for example.
Conditional forms
select
[syntax] (select EXP ((KEY ...) EXP1 ...) ... [(else EXPn ...)])
This is similar to case, but the keys are evaluated.
unless
[syntax] (unless TEST EXP1 EXP2 ...)
Equivalent to:
(if (not TEST) (begin EXP1 EXP2 ...))
when
[syntax] (when TEST EXP1 EXP2 ...)
Equivalent to:
(if TEST (begin EXP1 EXP2 ...))
Record structures
define-record
[syntax] (define-record NAME SLOTNAME ...)
Defines a record type. Call make-NAME to create an instance
of the structure (with one initialization-argument for each slot).
(NAME? STRUCT) tests any object for being an instance of this
structure. Slots are accessed via (NAME-SLOTNAME STRUCT)
and updated using (NAME-SLOTNAME-set! STRUCT VALUE).
(define-record point x y)
(define p1 (make-point 123 456))
(point? p1) ==> #t
(point-x p1) ==> 123
(point-y-set! p1 99)
(point-y p1) ==> 99
define-record-type
[syntax] (define-record-type NAME
(CONSTRUCTOR TAG ...)
PREDICATE
(FIELD ACCESSOR [MODIFIER]) ...)
SRFI-9 record types. For more information see the documentation for
SRFI-9.
define-record-printer
[syntax] (define-record-printer (NAME RECORDVAR PORTVAR) BODY ...)
[syntax] (define-record-printer NAME PROCEDURE)
Defines a printing method for record of the type NAME by
associating a procedure with the record type. When a record of this
type is written using display, write or print, then
the procedure is called with two arguments: the record to be printed
and an output-port.
(define-record-type foo (make-foo x y z) foo?
(x foo-x)
(y foo-y)
(z foo-z))
(define f (make-foo 1 2 3))
(define-record-printer (foo x out)
(fprintf out "#,(foo ~S ~S ~S)"
(foo-x x) (foo-y x) (foo-z x)) )
(define-reader-ctor 'foo make-foo)
(define s (with-output-to-string
(lambda () (write f))))
s ==> "#,(foo 1 2 3)"
(equal? f (with-input-from-string
s read))) ==> #t
Other forms
assert
[syntax] (assert EXP [STRING ARG ...])
Signals an error if EXP evaluates to false. An optional message
STRING and arguments ARG ... may be supplied to give a
more informative error-message. If compiled in unsafe mode (either
by specifying the -unsafe compiler option or by declaring
(unsafe)), then this expression expands to an unspecified value.
The result is the value of EXP.
begin-for-syntax
[syntax] (begin-for-syntax EXP ...)
Equivalent to (begin EXP ...), but performs the evaluation of
the expression during macro-expansion time.
cond-expand
[syntax] (cond-expand FEATURE-CLAUSE ...)
Expands by selecting feature clauses. This form is allowed to appear in non-toplevel expressions.
Predefined feature-identifiers are "situation" specific:
- compile
- eval, library, match, compiling, srfi-11, srfi-15, srfi-31, srfi-26, srfi-16, utils, regex, srfi-4, match, srfi-1, srfi-69, srfi-28, extras, srfi-8, srfi-6, srfi-2, srfi-0, srfi-10, srfi-9, srfi-55, srfi-61 chicken, srfi-23, srfi-30, srfi-39, srfi-62, srfi-17, srfi-12.
- load
- srfi-69, srfi-28, extras, srfi-8, srfi-6, srfi-2, srfi-0, srfi-10, srfi-9, srfi-55, srfi-61, chicken, srfi-23, srfi-30, srfi-39, srfi-62, srfi-17, srfi-12. library is implicit.
- eval
- match, csi, srfi-11, srfi-15, srfi-31, srfi-26, srfi-16, srfi-69, srfi-28, extras, srfi-8, srfi-6, srfi-2, srfi-0, srfi-10, srfi-9, srfi-55, srfi-61, chicken, srfi-23, srfi-30, srfi-39, srfi-62, srfi-17, srfi-12. library is implicit.
The following feature-identifiers are available in all situations: (machine-byte-order), (machine-type), (software-type), (software-version), where the actual feature-identifier is platform dependent.
In addition the following feature-identifiers may exist: applyhook, extraslot, ptables, dload.
For further information, see the documentation for SRFI-0.
ensure
[syntax] (ensure PREDICATE EXP [ARGUMENTS ...])
Evaluates the expression EXP and applies the one-argument
procedure PREDICATE to the result. If the predicate returns
#f an error is signaled, otherwise the result of EXP
is returned. If compiled in unsafe mode (either by specifying
the -unsafe compiler option or by declaring (unsafe)),
then this expression expands to an unspecified value. If specified,
the optional ARGUMENTS are used as arguments to the invocation
of the error-signalling code, as in (error ARGUMENTS ...). If
no ARGUMENTS are given, a generic error message is displayed
with the offending value and PREDICATE expression.
eval-when
[syntax] (eval-when (SITUATION ...) EXP ...)
Controls evaluation/compilation of subforms. SITUATION should
be one of the symbols eval, compile or load.
When encountered in the evaluator, and the situation specifier
eval is not given, then this form is not evaluated and an
unspecified value is returned. When encountered while compiling code,
and the situation specifier compile is given, then this form is
evaluated at compile-time. When encountered while compiling code, and the
situation specifier load is not given, then this form is ignored
and an expression resulting into an unspecified value is compiled instead.
The following table should make this clearer:
| In compiled code | In interpreted code |
eval | ignore | evaluate |
compile | evaluate at compile time | ignore |
load | compile as normal | ignore |
include
[syntax] (include STRING)
Include toplevel-expressions from the given source file in the currently
compiled/interpreted program. If the included file has the extension
.scm, then it may be omitted. The file is searched in the
current directory and, if not found, in all directories specified in the
-include-path option.
nth-value
[syntax] (nth-value N EXP)
Returns the Nth value (counting from zero) of the values returned
by expression EXP.
time
[syntax] (time EXP1 ...)
Evaluates EXP1 ... and prints elapsed time and some
values about GC use, like time spent in major GCs, number of minor
and major GCs.
Previous: Non-standard read syntax
Next: Modules and macros