A Lisp program is composed mainly of Lisp functions. This chapter explains what functions are, how they accept arguments, and how to define them.
In a general sense, a function is a rule for carrying on a computation given several values called arguments. The result of the computation is called the value of the function. The computation can also have side effects: lasting changes in the values of variables or the contents of data structures.
Here are important terms for functions in Emacs Lisp and for other function-like objects.
car
or append
. These
functions are also called built-in functions or
subrs. (Special forms are also considered primitives.)
Usually the reason we implement a function as a primitive is either
because it is fundamental, because it provides a low-level
interface to operating system services, or because it needs to run
fast. Primitives can be modified or added only by changing the C
sources and recompiling the editor. See section Writing Emacs Primitives.
command-execute
can invoke; it is a
possible definition for a key sequence. Some functions are
commands; a function written in Lisp is a command if it contains an
interactive declaration (see section Defining Commands). Such a function
can be called from Lisp expressions like other functions; in this
case, the fact that the function is a command makes no difference.
Keyboard macros (strings and vectors) are commands also, even
though they are not functions. A symbol is a command if its
function definition is a command; such symbols can be invoked with
M-x. The symbol is a function as well if the definition
is a function. See section Command Loop
Overview.
t
if
object is any kind of function, or a special form or
macro.
t
if object is a built-in function (i.e., a
Lisp primitive).
(subrp 'message) ; message
is a symbol,
=> nil ; not a subr object.
(subrp (symbol-function 'message))
=> t
t
if object is a byte-code function. For
example: (byte-code-function-p (symbol-function 'next-line)) => t
A function written in Lisp is a list that looks like this:
(lambda (arg-variables...) [documentation-string] [interactive-declaration] body-forms...)
Such a list is called a lambda expression. In Emacs Lisp, it actually is valid as an expression--it evaluates to itself. In some other Lisp dialects, a lambda expression is not a valid expression at all. In either case, its main use is not to be evaluated as an expression, but to be called as a function.
The first element of a lambda
expression is always the symbol lambda
. This indicates
that the list represents a function. The reason functions are
defined to start with lambda
is so that other lists,
intended for other uses, will not accidentally be valid as
functions.
The second element is a list of symbols--the argument variable names. This is called the lambda list. When a Lisp function is called, the argument values are matched up against the variables in the lambda list, which are given local bindings with the values provided. See section Local Variables.
The documentation string is a Lisp string object placed within the function definition to describe the function for the Emacs help facilities. See section Documentation Strings of Functions.
The interactive declaration is a list of the form
(interactive code-string)
. This declares
how to provide arguments if the function is used interactively.
Functions with this declaration are called commands; they
can be called using M-x or bound to a key. Functions not
intended to be called in this way should not have interactive
declarations. See section Defining
Commands, for how to write an interactive declaration.
The rest of the elements are the body of the function: the Lisp code to do the work of the function (or, as a Lisp programmer would say, "a list of Lisp forms to evaluate"). The value returned by the function is the value returned by the last element of the body.
Consider for example the following function:
(lambda (a b c) (+ a b c))
We can call this function by writing it as the CAR of an expression, like this:
((lambda (a b c) (+ a b c)) 1 2 3)
This call evaluates the body of the lambda expression with the
variable a
bound to 1, b
bound to 2, and
c
bound to 3. Evaluation of the body adds these three
numbers, producing the result 6; therefore, this call to the
function returns the value 6.
Note that the arguments can be the results of other function calls, as in this example:
((lambda (a b c) (+ a b c)) 1 (* 2 3) (- 5 4))
This evaluates the arguments 1
, (* 2
3)
, and (- 5 4)
from left to right. Then it
applies the lambda expression to the argument values 1, 6 and 1 to
produce the value 8.
It is not often useful to write a lambda expression as the CAR
of a form in this way. You can get the same result, of making local
variables and giving them values, using the special form
let
(see section Local
Variables). And let
is clearer and easier to use.
In practice, lambda expressions are either stored as the function
definitions of symbols, to produce named functions, or passed as
arguments to other functions (see section Anonymous Functions).
However, calls to explicit lambda expressions were very useful
in the old days of Lisp, before the special form let
was invented. At that time, they were the only way to bind and
initialize local variables.
Our simple sample function, (lambda (a b c) (+ a b
c))
, specifies three argument variables, so it must be
called with three arguments: if you try to call it with only two
arguments or four arguments, you get a
wrong-number-of-arguments
error.
It is often convenient to write a function that allows certain
arguments to be omitted. For example, the function
substring
accepts three arguments--a string, the start
index and the end index--but the third argument defaults to the
length of the string if you omit it. It is also
convenient for certain functions to accept an indefinite number of
arguments, as the functions list
and +
do.
To specify optional arguments that may be
omitted when a function is called, simply include the keyword
&optional
before the optional arguments. To
specify a list of zero or more extra arguments, include the keyword
&rest
before one final argument.
Thus, the complete syntax for an argument list is as follows:
(required-vars... [&optional optional-vars...] [&rest rest-var])
The square brackets indicate that the &optional
and &rest
clauses, and the variables that follow
them, are optional.
A call to the function requires one actual argument for each of
the required-vars. There may be actual arguments for
zero or more of the optional-vars, and there cannot be
any actual arguments beyond that unless the lambda list uses
&rest
. In that case, there may be any number of
extra actual arguments.
If actual arguments for the optional and rest variables are
omitted, then they always default to nil
. There is no
way for the function to distinguish between an explicit argument of
nil
and an omitted argument. However, the body of the
function is free to consider nil
an abbreviation for
some other meaningful value. This is what substring
does; nil
as the third argument to
substring
means to use the length of the string
supplied.
Common Lisp note: Common Lisp allows the function to specify what default value to use when an optional argument is omitted; Emacs Lisp always uses
nil
. Emacs Lisp does not support "supplied-p" variables that tell you whether an argument was explicitly passed.
For example, an argument list that looks like this:
(a b &optional c d &rest e)
binds a
and b
to the first two actual
arguments, which are required. If one or two more arguments are
provided, c
and d
are bound to them
respectively; any arguments after the first four are collected into
a list and e
is bound to that list. If there are only
two arguments, c
is nil
; if two or three
arguments, d
is nil
; if four arguments or
fewer, e
is nil
.
There is no way to have required arguments following optional
ones--it would not make sense. To see why this must be so, suppose
that c
in the example were optional and d
were required. Suppose three actual arguments are given; which
variable would the third argument be for? Similarly, it makes no
sense to have any more arguments (either required or optional)
after a &rest
argument.
Here are some examples of argument lists and proper calls:
((lambda (n) (1+ n)) ; One required: 1) ; requires exactly one argument. => 2 ((lambda (n &optional n1) ; One required and one optional: (if n1 (+ n n1) (1+ n))) ; 1 or 2 arguments. 1 2) => 3 ((lambda (n &rest ns) ; One required and one rest: (+ n (apply '+ ns))) ; 1 or more arguments. 1 2 3 4 5) => 15
A lambda expression may optionally have a documentation string just after the lambda list. This string does not affect execution of the function; it is a kind of comment, but a systematized comment which actually appears inside the Lisp world and can be used by the Emacs help facilities. See section Documentation, for how the documentation-string is accessed.
It is a good idea to provide documentation strings for all the functions in your program, even those that are called only from within your program. Documentation strings are like comments, except that they are easier to access.
The first line of the documentation string should stand on its
own, because apropos
displays just this first line. It
should consist of one or two complete sentences that summarize the
function's purpose.
The start of the documentation string is usually indented in the source file, but since these spaces come before the starting double-quote, they are not part of the string. Some people make a practice of indenting any additional lines of the string so that the text lines up in the program source. This is a mistake. The indentation of the following lines is inside the string; what looks nice in the source code will look ugly when displayed by the help commands.
You may wonder how the documentation string could be optional, since there are required components of the function that follow it (the body). Since evaluation of a string returns that string, without any side effects, it has no effect if it is not the last form in the body. Thus, in practice, there is no confusion between the first form of the body and the documentation string; if the only body form is a string then it serves both as the return value and as the documentation.
In most computer languages, every function has a name; the idea
of a function without a name is nonsensical. In Lisp, a function in
the strictest sense has no name. It is simply a list whose first
element is lambda
, a byte-code function object, or a
primitive subr-object.
However, a symbol can serve as the name of a function. This happens when you put the function in the symbol's function cell (see section Symbol Components). Then the symbol itself becomes a valid, callable function, equivalent to the list or subr-object that its function cell refers to. The contents of the function cell are also called the symbol's function definition. The procedure of using a symbol's function definition in place of the symbol is called symbol function indirection; see section Symbol Function Indirection.
In practice, nearly all functions are given names in this way
and referred to through their names. For example, the symbol
car
works as a function and does what it does because
the primitive subr-object #<subr car>
is stored
in its function cell.
We give functions names because it is convenient to refer to
them by their names in Lisp expressions. For primitive subr-objects
such as #<subr car>
, names are the only way you
can refer to them: there is no read syntax for such objects. For
functions written in Lisp, the name is more convenient to use in a
call than an explicit lambda expression. Also, a function with a
name can refer to itself--it can be recursive. Writing the
function's name in its own definition is much more convenient than
making the function definition point to itself (something that is
not impossible but that has various disadvantages in practice).
We often identify functions with the symbols used to name them.
For example, we often speak of "the function car
", not
distinguishing between the symbol car
and the
primitive subr-object that is its function definition. For most
purposes, there is no need to distinguish.
Even so, keep in mind that a function need not have a unique
name. While a given function object usually appears in the
function cell of only one symbol, this is just a matter of
convenience. It is easy to store it in several symbols using
fset
; then each of the symbols is equally well a name
for the same function.
A symbol used as a function name may also be used as a variable; these two uses of a symbol are independent and do not conflict. (Some Lisp dialects, such as Scheme, do not distinguish between a symbol's value and its function definition; a symbol's value as a variable is also its function definition.) If you have not given a symbol a function definition, you cannot use it as a function; whether the symbol has a value as a variable makes no difference to this.
We usually give a name to a function when it is first created.
This is called defining a function, and it is done with
the defun
special form.
defun
is the
usual way to define new Lisp functions. It defines the symbol
name as a function that looks like this: (lambda argument-list . body-forms)
defun
stores this lambda expression in the function
cell of name. It returns the value name, but
usually we ignore this value.
As described previously (see section Lambda Expressions),
argument-list is a list of argument names and may
include the keywords &optional
and
&rest
. Also, the first two of the
body-forms may be a documentation string and an
interactive declaration.
There is no conflict if the same symbol name is also used as a variable, since the symbol's value cell is independent of the function cell. See section Symbol Components.
Here are some examples:
(defun foo () 5) => foo (foo) => 5 (defun bar (a &optional b &rest c) (list a b c)) => bar (bar 1 2 3 4 5) => (1 2 (3 4 5)) (bar 1) => (1 nil nil) (bar) error--> Wrong number of arguments. (defun capitalize-backwards () "Upcase the last letter of a word." (interactive) (backward-word 1) (forward-word 1) (backward-char 1) (capitalize-word 1)) => capitalize-backwards
Be careful not to redefine existing functions unintentionally.
defun
redefines even primitive functions such as
car
without any hesitation or notification. Redefining
a function already defined is often done deliberately, and there is
no way to distinguish deliberate redefinition from unintentional
redefinition.
The proper place to use defalias
is where a
specific function name is being defined--especially where that name
appears explicitly in the source file being loaded. This is because
defalias
records which file defined the function, just
like defun
(see section Unloading).
By contrast, in programs that manipulate function definitions
for other purposes, it is better to use fset
, which
does not keep such records.
See also defsubst
, which defines a function like
defun
and tells the Lisp compiler to open-code it. See
section Inline Functions.
Defining functions is only half the battle. Functions don't do anything until you call them, i.e., tell them to run. Calling a function is also known as invocation.
The most common way of invoking a function is by evaluating a
list. For example, evaluating the list (concat "a"
"b")
calls the function concat
with arguments
"a"
and "b"
. See section Evaluation, for a description of
evaluation.
When you write a list as an expression in your program, the
function name it calls is written in your program. This means that
you choose which function to call, and how many arguments to give
it, when you write the program. Usually that's just what you want.
Occasionally you need to compute at run time which function to
call. To do that, use the function funcall
. When you
also need to determine at run time how many arguments to pass, use
apply
.
funcall
calls
function with arguments, and returns whatever
function returns. Since funcall
is a function, all of its arguments,
including function, are evaluated before
funcall
is called. This means that you can use any
expression to obtain the function to be called. It also means that
funcall
does not see the expressions you write for the
arguments, only their values. These values are
not evaluated a second time in the act of calling
function; funcall
enters the normal
procedure for calling a function at the place where the arguments
have already been evaluated.
The argument function must be either a Lisp function
or a primitive function. Special forms and macros are not allowed,
because they make sense only when given the "unevaluated" argument
expressions. funcall
cannot provide these because, as
we saw above, it never knows them in the first place.
(setq f 'list) => list (funcall f 'x 'y 'z) => (x y z) (funcall f 'x 'y '(z)) => (x y (z)) (funcall 'and t nil) error--> Invalid function: #<subr and>
Compare these example with the examples of
apply
.
apply
calls
function with arguments, just like
funcall
but with one difference: the last of
arguments is a list of objects, which are passed to
function as separate arguments, rather than a single
list. We say that apply
spreads this list so
that each individual element becomes an argument. apply
returns the result of calling
function. As with funcall
,
function must either be a Lisp function or a primitive
function; special forms and macros do not make sense in
apply
.
(setq f 'list) => list (apply f 'x 'y 'z) error--> Wrong type argument: listp, z (apply '+ 1 2 '(3 4)) => 10 (apply '+ '(1 2 3 4)) => 10 (apply 'append '((a b c) nil (x y z) nil)) => (a b c x y z)
For an interesting example of using apply
, see the
description of mapcar
, in section Mapping Functions.
It is common for Lisp
functions to accept functions as arguments or find them in data
structures (especially in hook variables and property lists) and
call them using funcall
or apply
.
Functions that accept function arguments are often called
functionals.
Sometimes, when you call a functional, it is useful to supply a no-op function as the argument. Here are two different kinds of no-op function:
nil
.
A mapping function applies a given function to each
element of a list or other collection. Emacs Lisp has several such
functions; mapcar
and mapconcat
, which
scan a list, are described here. See section Creating and Interning Symbols, for
the function mapatoms
which maps over the symbols in
an obarray.
These mapping functions do not allow char-tables because a
char-table is a sparse array whose nominal range of indices is very
large. To map over a char-table in a way that deals properly with
its sparse nature, use the function map-char-table
(see section Char-Tables).
mapcar
applies
function to each element of sequence in turn,
and returns a list of the results. The argument sequence can be any kind of sequence except a char-table; that is, a list, a vector, a bool-vector, or a string. The result is always a list. The length of the result is the same as the length of sequence.
For example:
(mapcar 'car '((a b) (c d) (e f)))
=> (a c e)
(mapcar '1+ [1 2 3])
=> (2 3 4)
(mapcar 'char-to-string "abc")
=> ("a" "b" "c")
;; Call each function in my-hooks
.
(mapcar 'funcall my-hooks)
(defun mapcar* (function &rest args)
"Apply FUNCTION to successive cars of all ARGS.
Return the list of results."
;; If no list is exhausted,
(if (not (memq 'nil args))
;; apply function to CARs.
(cons (apply function (mapcar 'car args))
(apply 'mapcar* function
;; Recurse for rest of elements.
(mapcar 'cdr args)))))
(mapcar* 'cons '(a b c) '(1 2 3 4))
=> ((a . 1) (b . 2) (c . 3))
mapconcat
applies function to each element of sequence:
the results, which must be strings, are concatenated. Between each
pair of result strings, mapconcat
inserts the string
separator. Usually separator contains a space
or comma or other suitable punctuation. The argument function must be a function that can take one argument and return a string. The argument sequence can be any kind of sequence except a char-table; that is, a list, a vector, a bool-vector, or a string.
(mapconcat 'symbol-name '(The cat in the hat) " ") => "The cat in the hat" (mapconcat (function (lambda (x) (format "%c" (1+ x)))) "HAL-8000" "") => "IBM.9111"
In Lisp, a function is a list that starts with
lambda
, a byte-code function compiled from such a
list, or alternatively a primitive subr-object; names are "extra".
Although usually functions are defined with defun
and
given names at the same time, it is occasionally more concise to
use an explicit lambda expression--an anonymous function. Such a
list is valid wherever a function name is.
Any method of creating such a list makes a valid function. Even this:
(setq silly (append '(lambda (x)) (list (list '+ (* 3 4) 'x)))) => (lambda (x) (+ 12 x))
This computes a list that looks like (lambda (x) (+ 12
x))
and makes it the value (not the function
definition!) of silly
.
Here is how we might call this function:
(funcall silly 1) => 13
(It does not work to write (silly 1)
,
because this function is not the function definition of
silly
. We have not given silly
any
function definition, just a value as a variable.)
Most of the time, anonymous functions are constants that appear
in your program. For example, you might want to pass one as an
argument to the function mapcar
, which applies any
given function to each element of a list.
Here we define a function change-property
which
uses a function as its third argument:
(defun change-property (symbol prop function) (let ((value (get symbol prop))) (put symbol prop (funcall function value))))
Here we define a function that uses
change-property
, passing it a function to double a
number:
(defun double-property (symbol prop) (change-property symbol prop '(lambda (x) (* 2 x))))
In such cases, we usually use the special form
function
instead of simple quotation to quote the
anonymous function, like this:
(defun double-property (symbol prop) (change-property symbol prop (function (lambda (x) (* 2 x)))))
Using function
instead of quote
makes
a difference if you compile the function
double-property
. For example, if you compile the
second definition of double-property
, the anonymous
function is compiled as well. By contrast, if you compile the first
definition which uses ordinary quote
, the argument
passed to change-property
is the precise list
shown:
(lambda (x) (* x 2))
The Lisp compiler cannot assume this list is a function, even
though it looks like one, since it does not know what
change-property
will do with the list. Perhaps it will
check whether the CAR of the third element is the symbol
*
! Using function
tells the compiler it
is safe to go ahead and compile the constant function.
We sometimes write function
instead of
quote
when quoting the name of a function, but this
usage is just a sort of comment:
(function symbol) == (quote symbol) == 'symbol
The read syntax #'
is a short-hand for using
function
. For example,
#'(lambda (x) (* x x))
is equivalent to
(function (lambda (x) (* x x)))
quote
. However, it serves as a note to
the Emacs Lisp compiler that function-object is intended
to be used only as a function, and therefore can safely be
compiled. Contrast this with quote
, in section Quoting.
See documentation
in section Access to Documentation Strings, for a
realistic example using function
and an anonymous
function.
The function definition of a symbol is the object stored in the function cell of the symbol. The functions described here access, test, and set the function cell of symbols.
See also the function indirect-function
in section
Symbol Function Indirection.
void-function
error is signaled. This function does not check that the returned object is a legitimate function.
(defun bar (n) (+ n 2)) => bar (symbol-function 'bar) => (lambda (n) (+ n 2)) (fset 'baz 'bar) => bar (symbol-function 'baz) => bar
If you have never given a
symbol any function definition, we say that that symbol's function
cell is void. In other words, the function cell does not
have any Lisp object in it. If you try to call such a symbol as a
function, it signals a void-function
error.
Note that void is not the same as nil
or the symbol
void
. The symbols nil
and
void
are Lisp objects, and can be stored into a
function cell just as any other object can be (and they can be
valid functions if you define them in turn with
defun
). A void function cell contains no object
whatsoever.
You can test the voidness of a symbol's function definition with
fboundp
. After you have given a symbol a function
definition, you can make it void once more using
fmakunbound
.
t
if the symbol has an object in its function cell,
nil
otherwise. It does not check that the object is a
legitimate function.
void-function
error. (See also makunbound
, in section When a Variable is "Void".) (defun foo (x) x) => foo (foo 1) =>1 (fmakunbound 'foo) => foo (foo 1) error--> Symbol's function definition is void: foo
There are three normal uses of this function:
defalias
instead of fset
; see section Defining Functions.)
defun
. For example, you
can use fset
to give a symbol s1
a
function definition which is another symbol s2
; then
s1
serves as an alias for whatever definition
s2
presently has. (Once again use
defalias
instead of fset
if you think of
this as the definition of s1
.)
defun
were not a primitive, it could be written in
Lisp (as a macro) using fset
.
Here are examples of these uses:
;; Savefoo
's definition inold-foo
. (fset 'old-foo (symbol-function 'foo)) ;; Make the symbolcar
the function definition ofxfirst
. ;; (Most likely,defalias
would be better thanfset
here.) (fset 'xfirst 'car) => car (xfirst '(1 2 3)) => 1 (symbol-function 'xfirst) => car (symbol-function (symbol-function 'xfirst)) => #<subr car> ;; Define a named keyboard macro. (fset 'kill-two-lines "\^u2\^k") => "\^u2\^k" ;; Here is a function that alters other functions. (defun copy-function-definition (new old) "Define NEW with the same function definition as OLD." (fset new (symbol-function old)))
When writing a function that extends a previously defined function, the following idiom is sometimes used:
(fset 'old-foo (symbol-function 'foo)) (defun foo () "Just like old-foo, except more so." (old-foo) (more-so))
This does not work properly if foo
has been defined
to autoload. In such a case, when foo
calls
old-foo
, Lisp attempts to define old-foo
by loading a file. Since this presumably defines foo
rather than old-foo
, it does not produce the proper
results. The only way to avoid this problem is to make sure the
file is loaded before moving aside the old definition of
foo
.
But it is unmodular and unclean, in any case, for a Lisp file to redefine a function defined elsewhere. It is cleaner to use the advice facility (see section Advising Emacs Lisp Functions).
You can define an inline
function by using defsubst
instead of
defun
. An inline function works just like an ordinary
function except for one thing: when you compile a call to the
function, the function's definition is open-coded into the
caller.
Making a function inline makes explicit calls run faster. But it also has disadvantages. For one thing, it reduces flexibility; if you change the definition of the function, calls already inlined still use the old definition until you recompile them. Since the flexibility of redefining functions is an important feature of Emacs, you should not make a function inline unless its speed is really crucial.
Another disadvantage is that making a large function inline can increase the size of compiled code both in files and in memory. Since the speed advantage of inline functions is greatest for small functions, you generally should not make large functions inline.
It's possible to define a macro to expand into the same code
that an inline function would execute. (See section Macros.) But the macro would be
limited to direct use in expressions--a macro cannot be called with
apply
, mapcar
and so on. Also, it takes
some work to convert an ordinary function into a macro. To convert
it into an inline function is very easy; simply replace
defun
with defsubst
. Since each argument
of an inline function is evaluated exactly once, you needn't worry
about how many times the body uses the arguments, as you do for
macros. (See section Evaluating Macro
Arguments Repeatedly.)
Inline functions can be used and open-coded later on in the same file, following the definition, just like macros.
Here is a table of several functions that do things related to function calling and function definitions. They are documented elsewhere, but we provide cross references here.
apply
autoload
call-interactively
commandp
documentation
eval
funcall
function
ignore
indirect-function
interactive
interactive-p
mapatoms
mapcar
map-char-table
mapconcat
undefined