Mercurial > emacs

@c -*-texinfo-*-
@c This is part of the GNU Emacs Lisp Reference Manual.
@c Copyright (C) 1990, 1991, 1992, 1993, 1994, 1995, 1998, 1999, 2001, 2002,
@c   2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011  Free Software Foundation, Inc.
@c See the file elisp.texi for copying conditions.
@setfilename ../../info/control
@node Control Structures, Variables, Evaluation, Top
@chapter Control Structures
@cindex special forms for control structures
@cindex control structures

  A Lisp program consists of expressions or @dfn{forms} (@pxref{Forms}).
We control the order of execution of these forms by enclosing them in
@dfn{control structures}.  Control structures are special forms which
control when, whether, or how many times to execute the forms they
contain.

@cindex textual order
  The simplest order of execution is sequential execution: first form
@var{a}, then form @var{b}, and so on.  This is what happens when you
write several forms in succession in the body of a function, or at top
level in a file of Lisp code---the forms are executed in the order
written.  We call this @dfn{textual order}.  For example, if a function
body consists of two forms @var{a} and @var{b}, evaluation of the
function evaluates first @var{a} and then @var{b}.  The result of
evaluating @var{b} becomes the value of the function.

  Explicit control structures make possible an order of execution other
than sequential.

  Emacs Lisp provides several kinds of control structure, including
other varieties of sequencing, conditionals, iteration, and (controlled)
jumps---all discussed below.  The built-in control structures are
special forms since their subforms are not necessarily evaluated or not
evaluated sequentially.  You can use macros to define your own control
structure constructs (@pxref{Macros}).

@menu
* Sequencing::             Evaluation in textual order.
* Conditionals::           @code{if}, @code{cond}, @code{when}, @code{unless}.
* Combining Conditions::   @code{and}, @code{or}, @code{not}.
* Iteration::              @code{while} loops.
* Nonlocal Exits::         Jumping out of a sequence.
@end menu

@node Sequencing
@section Sequencing

  Evaluating forms in the order they appear is the most common way
control passes from one form to another.  In some contexts, such as in a
function body, this happens automatically.  Elsewhere you must use a
control structure construct to do this: @code{progn}, the simplest
control construct of Lisp.

  A @code{progn} special form looks like this:

@example
@group
(progn @var{a} @var{b} @var{c} @dots{})
@end group
@end example

@noindent
and it says to execute the forms @var{a}, @var{b}, @var{c}, and so on, in
that order.  These forms are called the @dfn{body} of the @code{progn} form.
The value of the last form in the body becomes the value of the entire
@code{progn}.  @code{(progn)} returns @code{nil}.

@cindex implicit @code{progn}
  In the early days of Lisp, @code{progn} was the only way to execute
two or more forms in succession and use the value of the last of them.
But programmers found they often needed to use a @code{progn} in the
body of a function, where (at that time) only one form was allowed.  So
the body of a function was made into an ``implicit @code{progn}'':
several forms are allowed just as in the body of an actual @code{progn}.
Many other control structures likewise contain an implicit @code{progn}.
As a result, @code{progn} is not used as much as it was many years ago.
It is needed now most often inside an @code{unwind-protect}, @code{and},
@code{or}, or in the @var{then}-part of an @code{if}.

@defspec progn forms@dots{}
This special form evaluates all of the @var{forms}, in textual
order, returning the result of the final form.

@example
@group
(progn (print "The first form")
       (print "The second form")
       (print "The third form"))
     @print{} "The first form"
     @print{} "The second form"
     @print{} "The third form"
@result{} "The third form"
@end group
@end example
@end defspec

  Two other control constructs likewise evaluate a series of forms but return
a different value:

@defspec prog1 form1 forms@dots{}
This special form evaluates @var{form1} and all of the @var{forms}, in
textual order, returning the result of @var{form1}.

@example
@group
(prog1 (print "The first form")
       (print "The second form")
       (print "The third form"))
     @print{} "The first form"
     @print{} "The second form"
     @print{} "The third form"
@result{} "The first form"
@end group
@end example

Here is a way to remove the first element from a list in the variable
@code{x}, then return the value of that former element:

@example
(prog1 (car x) (setq x (cdr x)))
@end example
@end defspec

@defspec prog2 form1 form2 forms@dots{}
This special form evaluates @var{form1}, @var{form2}, and all of the
following @var{forms}, in textual order, returning the result of
@var{form2}.

@example
@group
(prog2 (print "The first form")
       (print "The second form")
       (print "The third form"))
     @print{} "The first form"
     @print{} "The second form"
     @print{} "The third form"
@result{} "The second form"
@end group
@end example
@end defspec

@node Conditionals
@section Conditionals
@cindex conditional evaluation

  Conditional control structures choose among alternatives.  Emacs Lisp
has four conditional forms: @code{if}, which is much the same as in
other languages; @code{when} and @code{unless}, which are variants of
@code{if}; and @code{cond}, which is a generalized case statement.

@defspec if condition then-form else-forms@dots{}
@code{if} chooses between the @var{then-form} and the @var{else-forms}
based on the value of @var{condition}.  If the evaluated @var{condition} is
non-@code{nil}, @var{then-form} is evaluated and the result returned.
Otherwise, the @var{else-forms} are evaluated in textual order, and the
value of the last one is returned.  (The @var{else} part of @code{if} is
an example of an implicit @code{progn}.  @xref{Sequencing}.)

If @var{condition} has the value @code{nil}, and no @var{else-forms} are
given, @code{if} returns @code{nil}.

@code{if} is a special form because the branch that is not selected is
never evaluated---it is ignored.  Thus, in the example below,
@code{true} is not printed because @code{print} is never called.

@example
@group
(if nil
    (print 'true)
  'very-false)
@result{} very-false
@end group
@end example
@end defspec

@defmac when condition then-forms@dots{}
This is a variant of @code{if} where there are no @var{else-forms},
and possibly several @var{then-forms}.  In particular,

@example
(when @var{condition} @var{a} @var{b} @var{c})
@end example

@noindent
is entirely equivalent to

@example
(if @var{condition} (progn @var{a} @var{b} @var{c}) nil)
@end example
@end defmac

@defmac unless condition forms@dots{}
This is a variant of @code{if} where there is no @var{then-form}:

@example
(unless @var{condition} @var{a} @var{b} @var{c})
@end example

@noindent
is entirely equivalent to

@example
(if @var{condition} nil
   @var{a} @var{b} @var{c})
@end example
@end defmac

@defspec cond clause@dots{}
@code{cond} chooses among an arbitrary number of alternatives.  Each
@var{clause} in the @code{cond} must be a list.  The @sc{car} of this
list is the @var{condition}; the remaining elements, if any, the
@var{body-forms}.  Thus, a clause looks like this:

@example
(@var{condition} @var{body-forms}@dots{})
@end example

@code{cond} tries the clauses in textual order, by evaluating the
@var{condition} of each clause.  If the value of @var{condition} is
non-@code{nil}, the clause ``succeeds''; then @code{cond} evaluates its
@var{body-forms}, and the value of the last of @var{body-forms} becomes
the value of the @code{cond}.  The remaining clauses are ignored.

If the value of @var{condition} is @code{nil}, the clause ``fails,'' so
the @code{cond} moves on to the following clause, trying its
@var{condition}.

If every @var{condition} evaluates to @code{nil}, so that every clause
fails, @code{cond} returns @code{nil}.

A clause may also look like this:

@example
(@var{condition})
@end example

@noindent
Then, if @var{condition} is non-@code{nil} when tested, the value of
@var{condition} becomes the value of the @code{cond} form.

The following example has four clauses, which test for the cases where
the value of @code{x} is a number, string, buffer and symbol,
respectively:

@example
@group
(cond ((numberp x) x)
      ((stringp x) x)
      ((bufferp x)
       (setq temporary-hack x) ; @r{multiple body-forms}
       (buffer-name x))        ; @r{in one clause}
      ((symbolp x) (symbol-value x)))
@end group
@end example

Often we want to execute the last clause whenever none of the previous
clauses was successful.  To do this, we use @code{t} as the
@var{condition} of the last clause, like this: @code{(t
@var{body-forms})}.  The form @code{t} evaluates to @code{t}, which is
never @code{nil}, so this clause never fails, provided the @code{cond}
gets to it at all.

For example,

@example
@group
(setq a 5)
(cond ((eq a 'hack) 'foo)
      (t "default"))
@result{} "default"
@end group
@end example

@noindent
This @code{cond} expression returns @code{foo} if the value of @code{a}
is @code{hack}, and returns the string @code{"default"} otherwise.
@end defspec

Any conditional construct can be expressed with @code{cond} or with
@code{if}.  Therefore, the choice between them is a matter of style.
For example:

@example
@group
(if @var{a} @var{b} @var{c})
@equiv{}
(cond (@var{a} @var{b}) (t @var{c}))
@end group
@end example

@node Combining Conditions
@section Constructs for Combining Conditions

  This section describes three constructs that are often used together
with @code{if} and @code{cond} to express complicated conditions.  The
constructs @code{and} and @code{or} can also be used individually as
kinds of multiple conditional constructs.

@defun not condition
This function tests for the falsehood of @var{condition}.  It returns
@code{t} if @var{condition} is @code{nil}, and @code{nil} otherwise.
The function @code{not} is identical to @code{null}, and we recommend
using the name @code{null} if you are testing for an empty list.
@end defun

@defspec and conditions@dots{}
The @code{and} special form tests whether all the @var{conditions} are
true.  It works by evaluating the @var{conditions} one by one in the
order written.

If any of the @var{conditions} evaluates to @code{nil}, then the result
of the @code{and} must be @code{nil} regardless of the remaining
@var{conditions}; so @code{and} returns @code{nil} right away, ignoring
the remaining @var{conditions}.

If all the @var{conditions} turn out non-@code{nil}, then the value of
the last of them becomes the value of the @code{and} form.  Just
@code{(and)}, with no @var{conditions}, returns @code{t}, appropriate
because all the @var{conditions} turned out non-@code{nil}.  (Think
about it; which one did not?)

Here is an example.  The first condition returns the integer 1, which is
not @code{nil}.  Similarly, the second condition returns the integer 2,
which is not @code{nil}.  The third condition is @code{nil}, so the
remaining condition is never evaluated.

@example
@group
(and (print 1) (print 2) nil (print 3))
     @print{} 1
     @print{} 2
@result{} nil
@end group
@end example

Here is a more realistic example of using @code{and}:

@example
@group
(if (and (consp foo) (eq (car foo) 'x))
    (message "foo is a list starting with x"))
@end group
@end example

@noindent
Note that @code{(car foo)} is not executed if @code{(consp foo)} returns
@code{nil}, thus avoiding an error.

@code{and} expressions can also be written using either @code{if} or
@code{cond}.  Here's how:

@example
@group
(and @var{arg1} @var{arg2} @var{arg3})
@equiv{}
(if @var{arg1} (if @var{arg2} @var{arg3}))
@equiv{}
(cond (@var{arg1} (cond (@var{arg2} @var{arg3}))))
@end group
@end example
@end defspec

@defspec or conditions@dots{}
The @code{or} special form tests whether at least one of the
@var{conditions} is true.  It works by evaluating all the
@var{conditions} one by one in the order written.

If any of the @var{conditions} evaluates to a non-@code{nil} value, then
the result of the @code{or} must be non-@code{nil}; so @code{or} returns
right away, ignoring the remaining @var{conditions}.  The value it
returns is the non-@code{nil} value of the condition just evaluated.

If all the @var{conditions} turn out @code{nil}, then the @code{or}
expression returns @code{nil}.  Just @code{(or)}, with no
@var{conditions}, returns @code{nil}, appropriate because all the
@var{conditions} turned out @code{nil}.  (Think about it; which one
did not?)

For example, this expression tests whether @code{x} is either
@code{nil} or the integer zero:

@example
(or (eq x nil) (eq x 0))
@end example

Like the @code{and} construct, @code{or} can be written in terms of
@code{cond}.  For example:

@example
@group
(or @var{arg1} @var{arg2} @var{arg3})
@equiv{}
(cond (@var{arg1})
      (@var{arg2})
      (@var{arg3}))
@end group
@end example

You could almost write @code{or} in terms of @code{if}, but not quite:

@example
@group
(if @var{arg1} @var{arg1}
  (if @var{arg2} @var{arg2}
    @var{arg3}))
@end group
@end example

@noindent
This is not completely equivalent because it can evaluate @var{arg1} or
@var{arg2} twice.  By contrast, @code{(or @var{arg1} @var{arg2}
@var{arg3})} never evaluates any argument more than once.
@end defspec

@node Iteration
@section Iteration
@cindex iteration
@cindex recursion

  Iteration means executing part of a program repetitively.  For
example, you might want to repeat some computation once for each element
of a list, or once for each integer from 0 to @var{n}.  You can do this
in Emacs Lisp with the special form @code{while}:

@defspec while condition forms@dots{}
@code{while} first evaluates @var{condition}.  If the result is
non-@code{nil}, it evaluates @var{forms} in textual order.  Then it
reevaluates @var{condition}, and if the result is non-@code{nil}, it
evaluates @var{forms} again.  This process repeats until @var{condition}
evaluates to @code{nil}.

There is no limit on the number of iterations that may occur.  The loop
will continue until either @var{condition} evaluates to @code{nil} or
until an error or @code{throw} jumps out of it (@pxref{Nonlocal Exits}).

The value of a @code{while} form is always @code{nil}.

@example
@group
(setq num 0)
     @result{} 0
@end group
@group
(while (< num 4)
  (princ (format "Iteration %d." num))
  (setq num (1+ num)))
     @print{} Iteration 0.
     @print{} Iteration 1.
     @print{} Iteration 2.
     @print{} Iteration 3.
     @result{} nil
@end group
@end example

To write a ``repeat...until'' loop, which will execute something on each
iteration and then do the end-test, put the body followed by the
end-test in a @code{progn} as the first argument of @code{while}, as
shown here:

@example
@group
(while (progn
         (forward-line 1)
         (not (looking-at "^$"))))
@end group
@end example

@noindent
This moves forward one line and continues moving by lines until it
reaches an empty line.  It is peculiar in that the @code{while} has no
body, just the end test (which also does the real work of moving point).
@end defspec

  The @code{dolist} and @code{dotimes} macros provide convenient ways to
write two common kinds of loops.

@defmac dolist (var list [result]) body@dots{}
This construct executes @var{body} once for each element of
@var{list}, binding the variable @var{var} locally to hold the current
element.  Then it returns the value of evaluating @var{result}, or
@code{nil} if @var{result} is omitted.  For example, here is how you
could use @code{dolist} to define the @code{reverse} function:

@example
(defun reverse (list)
  (let (value)
    (dolist (elt list value)
      (setq value (cons elt value)))))
@end example
@end defmac

@defmac dotimes (var count [result]) body@dots{}
This construct executes @var{body} once for each integer from 0
(inclusive) to @var{count} (exclusive), binding the variable @var{var}
to the integer for the current iteration.  Then it returns the value
of evaluating @var{result}, or @code{nil} if @var{result} is omitted.
Here is an example of using @code{dotimes} to do something 100 times:

@example
(dotimes (i 100)
  (insert "I will not obey absurd orders\n"))
@end example
@end defmac

@node Nonlocal Exits
@section Nonlocal Exits
@cindex nonlocal exits

  A @dfn{nonlocal exit} is a transfer of control from one point in a
program to another remote point.  Nonlocal exits can occur in Emacs Lisp
as a result of errors; you can also use them under explicit control.
Nonlocal exits unbind all variable bindings made by the constructs being
exited.

@menu
* Catch and Throw::     Nonlocal exits for the program's own purposes.
* Examples of Catch::   Showing how such nonlocal exits can be written.
* Errors::              How errors are signaled and handled.
* Cleanups::            Arranging to run a cleanup form if an error happens.
@end menu

@node Catch and Throw
@subsection Explicit Nonlocal Exits: @code{catch} and @code{throw}

  Most control constructs affect only the flow of control within the
construct itself.  The function @code{throw} is the exception to this
rule of normal program execution: it performs a nonlocal exit on
request.  (There are other exceptions, but they are for error handling
only.)  @code{throw} is used inside a @code{catch}, and jumps back to
that @code{catch}.  For example:

@example
@group
(defun foo-outer ()
  (catch 'foo
    (foo-inner)))

(defun foo-inner ()
  @dots{}
  (if x
      (throw 'foo t))
  @dots{})
@end group
@end example

@noindent
The @code{throw} form, if executed, transfers control straight back to
the corresponding @code{catch}, which returns immediately.  The code
following the @code{throw} is not executed.  The second argument of
@code{throw} is used as the return value of the @code{catch}.

  The function @code{throw} finds the matching @code{catch} based on the
first argument: it searches for a @code{catch} whose first argument is
@code{eq} to the one specified in the @code{throw}.  If there is more
than one applicable @code{catch}, the innermost one takes precedence.
Thus, in the above example, the @code{throw} specifies @code{foo}, and
the @code{catch} in @code{foo-outer} specifies the same symbol, so that
@code{catch} is the applicable one (assuming there is no other matching
@code{catch} in between).

  Executing @code{throw} exits all Lisp constructs up to the matching
@code{catch}, including function calls.  When binding constructs such as
@code{let} or function calls are exited in this way, the bindings are
unbound, just as they are when these constructs exit normally
(@pxref{Local Variables}).  Likewise, @code{throw} restores the buffer
and position saved by @code{save-excursion} (@pxref{Excursions}), and
the narrowing status saved by @code{save-restriction} and the window
selection saved by @code{save-window-excursion} (@pxref{Window
Configurations}).  It also runs any cleanups established with the
@code{unwind-protect} special form when it exits that form
(@pxref{Cleanups}).

  The @code{throw} need not appear lexically within the @code{catch}
that it jumps to.  It can equally well be called from another function
called within the @code{catch}.  As long as the @code{throw} takes place
chronologically after entry to the @code{catch}, and chronologically
before exit from it, it has access to that @code{catch}.  This is why
@code{throw} can be used in commands such as @code{exit-recursive-edit}
that throw back to the editor command loop (@pxref{Recursive Editing}).

@cindex CL note---only @code{throw} in Emacs
@quotation
@b{Common Lisp note:} Most other versions of Lisp, including Common Lisp,
have several ways of transferring control nonsequentially: @code{return},
@code{return-from}, and @code{go}, for example.  Emacs Lisp has only
@code{throw}.
@end quotation

@defspec catch tag body@dots{}
@cindex tag on run time stack
@code{catch} establishes a return point for the @code{throw} function.
The return point is distinguished from other such return points by
@var{tag}, which may be any Lisp object except @code{nil}.  The argument
@var{tag} is evaluated normally before the return point is established.

With the return point in effect, @code{catch} evaluates the forms of the
@var{body} in textual order.  If the forms execute normally (without
error or nonlocal exit) the value of the last body form is returned from
the @code{catch}.

If a @code{throw} is executed during the execution of @var{body},
specifying the same value @var{tag}, the @code{catch} form exits
immediately; the value it returns is whatever was specified as the
second argument of @code{throw}.
@end defspec

@defun throw tag value
The purpose of @code{throw} is to return from a return point previously
established with @code{catch}.  The argument @var{tag} is used to choose
among the various existing return points; it must be @code{eq} to the value
specified in the @code{catch}.  If multiple return points match @var{tag},
the innermost one is used.

The argument @var{value} is used as the value to return from that
@code{catch}.

@kindex no-catch
If no return point is in effect with tag @var{tag}, then a @code{no-catch}
error is signaled with data @code{(@var{tag} @var{value})}.
@end defun

@node Examples of Catch
@subsection Examples of @code{catch} and @code{throw}

  One way to use @code{catch} and @code{throw} is to exit from a doubly
nested loop.  (In most languages, this would be done with a ``go to.'')
Here we compute @code{(foo @var{i} @var{j})} for @var{i} and @var{j}
varying from 0 to 9:

@example
@group
(defun search-foo ()
  (catch 'loop
    (let ((i 0))
      (while (< i 10)
        (let ((j 0))
          (while (< j 10)
            (if (foo i j)
                (throw 'loop (list i j)))
            (setq j (1+ j))))
        (setq i (1+ i))))))
@end group
@end example

@noindent
If @code{foo} ever returns non-@code{nil}, we stop immediately and return a
list of @var{i} and @var{j}.  If @code{foo} always returns @code{nil}, the
@code{catch} returns normally, and the value is @code{nil}, since that
is the result of the @code{while}.

  Here are two tricky examples, slightly different, showing two
return points at once.  First, two return points with the same tag,
@code{hack}:

@example
@group
(defun catch2 (tag)
  (catch tag
    (throw 'hack 'yes)))
@result{} catch2
@end group

@group
(catch 'hack
  (print (catch2 'hack))
  'no)
@print{} yes
@result{} no
@end group
@end example

@noindent
Since both return points have tags that match the @code{throw}, it goes to
the inner one, the one established in @code{catch2}.  Therefore,
@code{catch2} returns normally with value @code{yes}, and this value is
printed.  Finally the second body form in the outer @code{catch}, which is
@code{'no}, is evaluated and returned from the outer @code{catch}.

  Now let's change the argument given to @code{catch2}:

@example
@group
(catch 'hack
  (print (catch2 'quux))
  'no)
@result{} yes
@end group
@end example

@noindent
We still have two return points, but this time only the outer one has
the tag @code{hack}; the inner one has the tag @code{quux} instead.
Therefore, @code{throw} makes the outer @code{catch} return the value
@code{yes}.  The function @code{print} is never called, and the
body-form @code{'no} is never evaluated.

@node Errors
@subsection Errors
@cindex errors

  When Emacs Lisp attempts to evaluate a form that, for some reason,
cannot be evaluated, it @dfn{signals} an @dfn{error}.

  When an error is signaled, Emacs's default reaction is to print an
error message and terminate execution of the current command.  This is
the right thing to do in most cases, such as if you type @kbd{C-f} at
the end of the buffer.

  In complicated programs, simple termination may not be what you want.
For example, the program may have made temporary changes in data
structures, or created temporary buffers that should be deleted before
the program is finished.  In such cases, you would use
@code{unwind-protect} to establish @dfn{cleanup expressions} to be
evaluated in case of error.  (@xref{Cleanups}.)  Occasionally, you may
wish the program to continue execution despite an error in a subroutine.
In these cases, you would use @code{condition-case} to establish
@dfn{error handlers} to recover control in case of error.

  Resist the temptation to use error handling to transfer control from
one part of the program to another; use @code{catch} and @code{throw}
instead.  @xref{Catch and Throw}.

@menu
* Signaling Errors::      How to report an error.
* Processing of Errors::  What Emacs does when you report an error.
* Handling Errors::       How you can trap errors and continue execution.
* Error Symbols::         How errors are classified for trapping them.
@end menu

@node Signaling Errors
@subsubsection How to Signal an Error
@cindex signaling errors

   @dfn{Signaling} an error means beginning error processing.  Error
processing normally aborts all or part of the running program and
returns to a point that is set up to handle the error
(@pxref{Processing of Errors}).  Here we describe how to signal an
error.

  Most errors are signaled ``automatically'' within Lisp primitives
which you call for other purposes, such as if you try to take the
@sc{car} of an integer or move forward a character at the end of the
buffer.  You can also signal errors explicitly with the functions
@code{error} and @code{signal}.

  Quitting, which happens when the user types @kbd{C-g}, is not
considered an error, but it is handled almost like an error.
@xref{Quitting}.

  Every error specifies an error message, one way or another.  The
message should state what is wrong (``File does not exist''), not how
things ought to be (``File must exist'').  The convention in Emacs
Lisp is that error messages should start with a capital letter, but
should not end with any sort of punctuation.

@defun error format-string &rest args
This function signals an error with an error message constructed by
applying @code{format} (@pxref{Formatting Strings}) to
@var{format-string} and @var{args}.

These examples show typical uses of @code{error}:

@example
@group
(error "That is an error -- try something else")
     @error{} That is an error -- try something else
@end group

@group
(error "You have committed %d errors" 10)
     @error{} You have committed 10 errors
@end group
@end example

@code{error} works by calling @code{signal} with two arguments: the
error symbol @code{error}, and a list containing the string returned by
@code{format}.

@strong{Warning:} If you want to use your own string as an error message
verbatim, don't just write @code{(error @var{string})}.  If @var{string}
contains @samp{%}, it will be interpreted as a format specifier, with
undesirable results.  Instead, use @code{(error "%s" @var{string})}.
@end defun

@defun signal error-symbol data
@anchor{Definition of signal}
This function signals an error named by @var{error-symbol}.  The
argument @var{data} is a list of additional Lisp objects relevant to
the circumstances of the error.

The argument @var{error-symbol} must be an @dfn{error symbol}---a symbol
bearing a property @code{error-conditions} whose value is a list of
condition names.  This is how Emacs Lisp classifies different sorts of
errors. @xref{Error Symbols}, for a description of error symbols,
error conditions and condition names.

If the error is not handled, the two arguments are used in printing
the error message.  Normally, this error message is provided by the
@code{error-message} property of @var{error-symbol}.  If @var{data} is
non-@code{nil}, this is followed by a colon and a comma separated list
of the unevaluated elements of @var{data}.  For @code{error}, the
error message is the @sc{car} of @var{data} (that must be a string).
Subcategories of @code{file-error} are handled specially.

The number and significance of the objects in @var{data} depends on
@var{error-symbol}.  For example, with a @code{wrong-type-argument} error,
there should be two objects in the list: a predicate that describes the type
that was expected, and the object that failed to fit that type.

Both @var{error-symbol} and @var{data} are available to any error
handlers that handle the error: @code{condition-case} binds a local
variable to a list of the form @code{(@var{error-symbol} .@:
@var{data})} (@pxref{Handling Errors}).

The function @code{signal} never returns (though in older Emacs versions
it could sometimes return).

@smallexample
@group
(signal 'wrong-number-of-arguments '(x y))
     @error{} Wrong number of arguments: x, y
@end group

@group
(signal 'no-such-error '("My unknown error condition"))
     @error{} peculiar error: "My unknown error condition"
@end group
@end smallexample
@end defun

@cindex CL note---no continuable errors
@quotation
@b{Common Lisp note:} Emacs Lisp has nothing like the Common Lisp
concept of continuable errors.
@end quotation

@node Processing of Errors
@subsubsection How Emacs Processes Errors

When an error is signaled, @code{signal} searches for an active
@dfn{handler} for the error.  A handler is a sequence of Lisp
expressions designated to be executed if an error happens in part of the
Lisp program.  If the error has an applicable handler, the handler is
executed, and control resumes following the handler.  The handler
executes in the environment of the @code{condition-case} that
established it; all functions called within that @code{condition-case}
have already been exited, and the handler cannot return to them.

If there is no applicable handler for the error, it terminates the
current command and returns control to the editor command loop.  (The
command loop has an implicit handler for all kinds of errors.)  The
command loop's handler uses the error symbol and associated data to
print an error message.  You can use the variable
@code{command-error-function} to control how this is done:

@defvar command-error-function
This variable, if non-@code{nil}, specifies a function to use to
handle errors that return control to the Emacs command loop.  The
function should take three arguments: @var{data}, a list of the same
form that @code{condition-case} would bind to its variable;
@var{context}, a string describing the situation in which the error
occurred, or (more often) @code{nil}; and @var{caller}, the Lisp
function which called the primitive that signaled the error.
@end defvar

@cindex @code{debug-on-error} use
An error that has no explicit handler may call the Lisp debugger.  The
debugger is enabled if the variable @code{debug-on-error} (@pxref{Error
Debugging}) is non-@code{nil}.  Unlike error handlers, the debugger runs
in the environment of the error, so that you can examine values of
variables precisely as they were at the time of the error.

@node Handling Errors
@subsubsection Writing Code to Handle Errors
@cindex error handler
@cindex handling errors

  The usual effect of signaling an error is to terminate the command
that is running and return immediately to the Emacs editor command loop.
You can arrange to trap errors occurring in a part of your program by
establishing an error handler, with the special form
@code{condition-case}.  A simple example looks like this:

@example
@group
(condition-case nil
    (delete-file filename)
  (error nil))
@end group
@end example

@noindent
This deletes the file named @var{filename}, catching any error and
returning @code{nil} if an error occurs@footnote{
Actually, you should use @code{ignore-errors} in such a simple case;
see below.}.

  The @code{condition-case} construct is often used to trap errors that
are predictable, such as failure to open a file in a call to
@code{insert-file-contents}.  It is also used to trap errors that are
totally unpredictable, such as when the program evaluates an expression
read from the user.

  The second argument of @code{condition-case} is called the
@dfn{protected form}.  (In the example above, the protected form is a
call to @code{delete-file}.)  The error handlers go into effect when
this form begins execution and are deactivated when this form returns.
They remain in effect for all the intervening time.  In particular, they
are in effect during the execution of functions called by this form, in
their subroutines, and so on.  This is a good thing, since, strictly
speaking, errors can be signaled only by Lisp primitives (including
@code{signal} and @code{error}) called by the protected form, not by the
protected form itself.

  The arguments after the protected form are handlers.  Each handler
lists one or more @dfn{condition names} (which are symbols) to specify
which errors it will handle.  The error symbol specified when an error
is signaled also defines a list of condition names.  A handler applies
to an error if they have any condition names in common.  In the example
above, there is one handler, and it specifies one condition name,
@code{error}, which covers all errors.

  The search for an applicable handler checks all the established handlers
starting with the most recently established one.  Thus, if two nested
@code{condition-case} forms offer to handle the same error, the inner of
the two gets to handle it.

  If an error is handled by some @code{condition-case} form, this
ordinarily prevents the debugger from being run, even if
@code{debug-on-error} says this error should invoke the debugger.

  If you want to be able to debug errors that are caught by a
@code{condition-case}, set the variable @code{debug-on-signal} to a
non-@code{nil} value.  You can also specify that a particular handler
should let the debugger run first, by writing @code{debug} among the
conditions, like this:

@example
@group
(condition-case nil
    (delete-file filename)
  ((debug error) nil))
@end group
@end example

@noindent
The effect of @code{debug} here is only to prevent
@code{condition-case} from suppressing the call to the debugger.  Any
given error will invoke the debugger only if @code{debug-on-error} and
the other usual filtering mechanisms say it should.  @xref{Error Debugging}.

  Once Emacs decides that a certain handler handles the error, it
returns control to that handler.  To do so, Emacs unbinds all variable
bindings made by binding constructs that are being exited, and
executes the cleanups of all @code{unwind-protect} forms that are
being exited.  Once control arrives at the handler, the body of the
handler executes normally.

  After execution of the handler body, execution returns from the
@code{condition-case} form.  Because the protected form is exited
completely before execution of the handler, the handler cannot resume
execution at the point of the error, nor can it examine variable
bindings that were made within the protected form.  All it can do is
clean up and proceed.

  Error signaling and handling have some resemblance to @code{throw} and
@code{catch} (@pxref{Catch and Throw}), but they are entirely separate
facilities.  An error cannot be caught by a @code{catch}, and a
@code{throw} cannot be handled by an error handler (though using
@code{throw} when there is no suitable @code{catch} signals an error
that can be handled).

@defspec condition-case var protected-form handlers@dots{}
This special form establishes the error handlers @var{handlers} around
the execution of @var{protected-form}.  If @var{protected-form} executes
without error, the value it returns becomes the value of the
@code{condition-case} form; in this case, the @code{condition-case} has
no effect.  The @code{condition-case} form makes a difference when an
error occurs during @var{protected-form}.

Each of the @var{handlers} is a list of the form @code{(@var{conditions}
@var{body}@dots{})}.  Here @var{conditions} is an error condition name
to be handled, or a list of condition names (which can include @code{debug}
to allow the debugger to run before the handler); @var{body} is one or more
Lisp expressions to be executed when this handler handles an error.
Here are examples of handlers:

@smallexample
@group
(error nil)

(arith-error (message "Division by zero"))

((arith-error file-error)
 (message
  "Either division by zero or failure to open a file"))
@end group
@end smallexample

Each error that occurs has an @dfn{error symbol} that describes what
kind of error it is.  The @code{error-conditions} property of this
symbol is a list of condition names (@pxref{Error Symbols}).  Emacs
searches all the active @code{condition-case} forms for a handler that
specifies one or more of these condition names; the innermost matching
@code{condition-case} handles the error.  Within this
@code{condition-case}, the first applicable handler handles the error.

After executing the body of the handler, the @code{condition-case}
returns normally, using the value of the last form in the handler body
as the overall value.

@cindex error description
The argument @var{var} is a variable.  @code{condition-case} does not
bind this variable when executing the @var{protected-form}, only when it
handles an error.  At that time, it binds @var{var} locally to an
@dfn{error description}, which is a list giving the particulars of the
error.  The error description has the form @code{(@var{error-symbol}
. @var{data})}.  The handler can refer to this list to decide what to
do.  For example, if the error is for failure opening a file, the file
name is the second element of @var{data}---the third element of the
error description.

If @var{var} is @code{nil}, that means no variable is bound.  Then the
error symbol and associated data are not available to the handler.

@cindex rethrow a signal
Sometimes it is necessary to re-throw a signal caught by
@code{condition-case}, for some outer-level handler to catch.  Here's
how to do that:

@smallexample
  (signal (car err) (cdr err))
@end smallexample

@noindent
where @code{err} is the error description variable, the first argument
to @code{condition-case} whose error condition you want to re-throw.
@xref{Definition of signal}.
@end defspec

@defun error-message-string error-descriptor
This function returns the error message string for a given error
descriptor.  It is useful if you want to handle an error by printing the
usual error message for that error.  @xref{Definition of signal}.
@end defun

@cindex @code{arith-error} example
Here is an example of using @code{condition-case} to handle the error
that results from dividing by zero.  The handler displays the error
message (but without a beep), then returns a very large number.

@smallexample
@group
(defun safe-divide (dividend divisor)
  (condition-case err
      ;; @r{Protected form.}
      (/ dividend divisor)
@end group
@group
    ;; @r{The handler.}
    (arith-error                        ; @r{Condition.}
     ;; @r{Display the usual message for this error.}
     (message "%s" (error-message-string err))
     1000000)))
@result{} safe-divide
@end group

@group
(safe-divide 5 0)
     @print{} Arithmetic error: (arith-error)
@result{} 1000000
@end group
@end smallexample

@noindent
The handler specifies condition name @code{arith-error} so that it will handle only division-by-zero errors.  Other kinds of errors will not be handled, at least not by this @code{condition-case}.  Thus,

@smallexample
@group
(safe-divide nil 3)
     @error{} Wrong type argument: number-or-marker-p, nil
@end group
@end smallexample

  Here is a @code{condition-case} that catches all kinds of errors,
including those signaled with @code{error}:

@smallexample
@group
(setq baz 34)
     @result{} 34
@end group

@group
(condition-case err
    (if (eq baz 35)
        t
      ;; @r{This is a call to the function @code{error}.}
      (error "Rats!  The variable %s was %s, not 35" 'baz baz))
  ;; @r{This is the handler; it is not a form.}
  (error (princ (format "The error was: %s" err))
         2))
@print{} The error was: (error "Rats!  The variable baz was 34, not 35")
@result{} 2
@end group
@end smallexample

@defmac ignore-errors body@dots{}
This construct executes @var{body}, ignoring any errors that occur
during its execution.  If the execution is without error,
@code{ignore-errors} returns the value of the last form in @var{body};
otherwise, it returns @code{nil}.

Here's the example at the beginning of this subsection rewritten using
@code{ignore-errors}:

@smallexample
@group
  (ignore-errors
   (delete-file filename))
@end group
@end smallexample
@end defmac


@node Error Symbols
@subsubsection Error Symbols and Condition Names
@cindex error symbol
@cindex error name
@cindex condition name
@cindex user-defined error
@kindex error-conditions

  When you signal an error, you specify an @dfn{error symbol} to specify
the kind of error you have in mind.  Each error has one and only one
error symbol to categorize it.  This is the finest classification of
errors defined by the Emacs Lisp language.

  These narrow classifications are grouped into a hierarchy of wider
classes called @dfn{error conditions}, identified by @dfn{condition
names}.  The narrowest such classes belong to the error symbols
themselves: each error symbol is also a condition name.  There are also
condition names for more extensive classes, up to the condition name
@code{error} which takes in all kinds of errors (but not @code{quit}).
Thus, each error has one or more condition names: @code{error}, the
error symbol if that is distinct from @code{error}, and perhaps some
intermediate classifications.

  In order for a symbol to be an error symbol, it must have an
@code{error-conditions} property which gives a list of condition names.
This list defines the conditions that this kind of error belongs to.
(The error symbol itself, and the symbol @code{error}, should always be
members of this list.)  Thus, the hierarchy of condition names is
defined by the @code{error-conditions} properties of the error symbols.
Because quitting is not considered an error, the value of the
@code{error-conditions} property of @code{quit} is just @code{(quit)}.

@cindex peculiar error
  In addition to the @code{error-conditions} list, the error symbol
should have an @code{error-message} property whose value is a string to
be printed when that error is signaled but not handled.  If the
error symbol has no @code{error-message} property or if the
@code{error-message} property exists, but is not a string, the error
message @samp{peculiar error} is used.  @xref{Definition of signal}.

  Here is how we define a new error symbol, @code{new-error}:

@example
@group
(put 'new-error
     'error-conditions
     '(error my-own-errors new-error))
@result{} (error my-own-errors new-error)
@end group
@group
(put 'new-error 'error-message "A new error")
@result{} "A new error"
@end group
@end example

@noindent
This error has three condition names: @code{new-error}, the narrowest
classification; @code{my-own-errors}, which we imagine is a wider
classification; and @code{error}, which is the widest of all.

  The error string should start with a capital letter but it should
not end with a period.  This is for consistency with the rest of Emacs.

  Naturally, Emacs will never signal @code{new-error} on its own; only
an explicit call to @code{signal} (@pxref{Definition of signal}) in
your code can do this:

@example
@group
(signal 'new-error '(x y))
     @error{} A new error: x, y
@end group
@end example

  This error can be handled through any of the three condition names.
This example handles @code{new-error} and any other errors in the class
@code{my-own-errors}:

@example
@group
(condition-case foo
    (bar nil t)
  (my-own-errors nil))
@end group
@end example

  The significant way that errors are classified is by their condition
names---the names used to match errors with handlers.  An error symbol
serves only as a convenient way to specify the intended error message
and list of condition names.  It would be cumbersome to give
@code{signal} a list of condition names rather than one error symbol.

  By contrast, using only error symbols without condition names would
seriously decrease the power of @code{condition-case}.  Condition names
make it possible to categorize errors at various levels of generality
when you write an error handler.  Using error symbols alone would
eliminate all but the narrowest level of classification.

  @xref{Standard Errors}, for a list of all the standard error symbols
and their conditions.

@node Cleanups
@subsection Cleaning Up from Nonlocal Exits

  The @code{unwind-protect} construct is essential whenever you
temporarily put a data structure in an inconsistent state; it permits
you to make the data consistent again in the event of an error or
throw.  (Another more specific cleanup construct that is used only for
changes in buffer contents is the atomic change group; @ref{Atomic
Changes}.)

@defspec unwind-protect body-form cleanup-forms@dots{}
@cindex cleanup forms
@cindex protected forms
@cindex error cleanup
@cindex unwinding
@code{unwind-protect} executes @var{body-form} with a guarantee that
the @var{cleanup-forms} will be evaluated if control leaves
@var{body-form}, no matter how that happens.  @var{body-form} may
complete normally, or execute a @code{throw} out of the
@code{unwind-protect}, or cause an error; in all cases, the
@var{cleanup-forms} will be evaluated.

If @var{body-form} finishes normally, @code{unwind-protect} returns the
value of @var{body-form}, after it evaluates the @var{cleanup-forms}.
If @var{body-form} does not finish, @code{unwind-protect} does not
return any value in the normal sense.

Only @var{body-form} is protected by the @code{unwind-protect}.  If any
of the @var{cleanup-forms} themselves exits nonlocally (via a
@code{throw} or an error), @code{unwind-protect} is @emph{not}
guaranteed to evaluate the rest of them.  If the failure of one of the
@var{cleanup-forms} has the potential to cause trouble, then protect
it with another @code{unwind-protect} around that form.

The number of currently active @code{unwind-protect} forms counts,
together with the number of local variable bindings, against the limit
@code{max-specpdl-size} (@pxref{Definition of max-specpdl-size,, Local
Variables}).
@end defspec

  For example, here we make an invisible buffer for temporary use, and
make sure to kill it before finishing:

@smallexample
@group
(let ((buffer (get-buffer-create " *temp*")))
  (with-current-buffer buffer
    (unwind-protect
        @var{body-form}
      (kill-buffer buffer))))
@end group
@end smallexample

@noindent
You might think that we could just as well write @code{(kill-buffer
(current-buffer))} and dispense with the variable @code{buffer}.
However, the way shown above is safer, if @var{body-form} happens to
get an error after switching to a different buffer!  (Alternatively,
you could write a @code{save-current-buffer} around @var{body-form},
to ensure that the temporary buffer becomes current again in time to
kill it.)

  Emacs includes a standard macro called @code{with-temp-buffer} which
expands into more or less the code shown above (@pxref{Definition of
with-temp-buffer,, Current Buffer}).  Several of the macros defined in
this manual use @code{unwind-protect} in this way.

@findex ftp-login
  Here is an actual example derived from an FTP package.  It creates a
process (@pxref{Processes}) to try to establish a connection to a remote
machine.  As the function @code{ftp-login} is highly susceptible to
numerous problems that the writer of the function cannot anticipate, it
is protected with a form that guarantees deletion of the process in the
event of failure.  Otherwise, Emacs might fill up with useless
subprocesses.

@smallexample
@group
(let ((win nil))
  (unwind-protect
      (progn
        (setq process (ftp-setup-buffer host file))
        (if (setq win (ftp-login process host user password))
            (message "Logged in")
          (error "Ftp login failed")))
    (or win (and process (delete-process process)))))
@end group
@end smallexample

  This example has a small bug: if the user types @kbd{C-g} to
quit, and the quit happens immediately after the function
@code{ftp-setup-buffer} returns but before the variable @code{process} is
set, the process will not be killed.  There is no easy way to fix this bug,
but at least it is very unlikely.
author	Paul Eggert <eggert@cs.ucla.edu>
date	Fri, 21 Jan 2011 23:18:23 -0800
parents	ef719132ddfa
children