dpans/concept-conditions.tex

% -*- Mode: TeX -*-

%% Errors
%% CONDITION-RESTARTS hasn't been included

%!!! What to do with this use of "construct"? -kmp 3-Sep-91
Common Lisp constructs are described not only in terms of their
behavior in situations during which they are intended to be used (see
the ``Description'' part of each \term{operator} specification),
but in all other situations (see the ``Exceptional Situations''
part of each \term{operator} specification).
 
A situation is the evaluation of an expression in a specific context.
%at a particular point in time? -kmp 5-Sep-91
A \term{condition} is an \term{object} that
represents a specific situation that has been detected.  
\term{Conditions} are \instancesofclasses{condition}.
A hierarchy of \term{condition} classes is defined in \clisp.  
A \term{condition} has \term{slots} that contain data 
relevant to the situation that the \term{condition} represents.

An error is a situation in which normal program execution cannot
continue correctly without some form of intervention (either
interactively by the user or under program control).  Not all errors
are detected.  When an error goes undetected, the effects can be
\term{implementation-dependent}, \term{implementation-defined}, unspecified, or
undefined. \Seesection\Definitions.  All detected errors can
be represented by \term{conditions}, but not all 
\term{conditions} represent errors.
 
Signaling is the process by which a \term{condition} can alter
the flow of control in a program by raising the 
\term{condition} which can then be \term{handled}.  The functions
\funref{error}, \funref{cerror}, \funref{signal}, and
\funref{warn} are used to signal \term{conditions}.
 
The process of signaling involves the selection and invocation of a
\term{handler} from a set of \term{active} \term{handlers}.  
A \term{handler} is a \term{function} of one argument (the 
\term{condition}) that is invoked to handle a \term{condition}.
Each \term{handler} is associated with a \term{condition} \term{type},
and a \term{handler} will be invoked only on a \term{condition} of the
\term{handler}'s associated \term{type}.
 
\term{Active} \term{handlers} are \term{established} dynamically
(see \macref{handler-bind} or \macref{handler-case}).
\term{Handlers} are invoked in a \term{dynamic environment} 
equivalent to that of the signaler,
except that the set of \term{active} \term{handlers} 
is bound in such a way as to include only those that were \term{active} 
at the time the \term{handler} being invoked was \term{established}.
Signaling a \term{condition} has no side-effect on the \term{condition}, 
and there is no dynamic state contained in a \term{condition}.
 
If a \term{handler} is invoked, it can address the \term{situation} 
in one of three ways:
 
\beginlist
\itemitem{\b{Decline}}

It can decline to \term{handle} the \term{condition}.  It does this by
simply returning rather than transferring control.
When this happens, any values returned by the handler are
ignored and the next most recently established handler is invoked.
If there is no such handler and the signaling function is \funref{error}
or \funref{cerror}, the debugger is entered in the
\term{dynamic environment} of the signaler. If there is no such
handler and the signaling function is either \funref{signal} or
\funref{warn}, the signaling function simply returns~\nil.

\itemitem{\b{Handle}}
 
It can \term{handle} the \term{condition} by performing a non-local
transfer of control.  This can be done either primitively by using
\specref{go}, \macref{return}, \specref{throw} or more
abstractly by using a function such as \funref{abort} or
\funref{invoke-restart}.

\itemitem{\b{Defer}}
 
It can put off a decision about whether to \term{handle} or \term{decline},
by any of a number of actions, but most commonly by 
    signaling another condition,
    resignaling the same condition,
 or forcing entry into the debugger.

\endlist
%%KMP
%The latter two actions 
%are really just ways of putting off the decision to either \term{handle}
%or decline. Ultimately,
%all a handler can do is to \term{handle} or decline to \term{handle}.
 
\beginSubsection{Condition Types}
 
\Thenextfigure\ lists the \term{standardized} \term{condition} \term{types}.
Additional \term{condition} \term{types} can be defined by using \macref{define-condition}.

\issue{UNDEFINED-FUNCTIONS-AND-VARIABLES:COMPROMISE}
\issue{DATA-IO:ADD-SUPPORT}
\issue{PARSE-ERROR-STREAM:SPLIT-TYPES}
\issue{ACCESS-ERROR-NAME}
\issue{READER-ERROR:NEW-TYPE}
\issue{FLOATING-POINT-CONDITION-NAMES:X3J13-NOV-89}
\issue{COMPILER-DIAGNOSTICS:USE-HANDLER}
\DefineFigure{StandardizedConditionTypes}
\displaythree{Standardized Condition Types}{
arithmetic-error&floating-point-overflow&simple-type-error\cr
cell-error&floating-point-underflow&simple-warning\cr
condition&package-error&storage-condition\cr
control-error&parse-error&stream-error\cr
division-by-zero&print-not-readable&style-warning\cr
end-of-file&program-error&type-error\cr
error&reader-error&unbound-slot\cr
file-error&serious-condition&unbound-variable\cr
floating-point-inexact&simple-condition&undefined-function\cr
floating-point-invalid-operation&simple-error&warning\cr
}
\endissue{COMPILER-DIAGNOSTICS:USE-HANDLER}
\endissue{FLOATING-POINT-CONDITION-NAMES:X3J13-NOV-89}
\endissue{READER-ERROR:NEW-TYPE}
\endissue{ACCESS-ERROR-NAME}
\endissue{PARSE-ERROR-STREAM:SPLIT-TYPES}
\endissue{DATA-IO:ADD-SUPPORT}
\endissue{UNDEFINED-FUNCTIONS-AND-VARIABLES:COMPROMISE}

All \term{condition} types are \subtypesof{condition}.  That is,
 
\code
 (typep \param{c} 'condition) \EV \term{true}
\endcode
if and only if \param{c} is a \term{condition}.
 
\term{Implementations} must define all specified \term{subtype} relationships.
%!!! Barrett: I don't understand this sentence.
Except where noted, all \term{subtype} relationships indicated in 
this document are not mutually exclusive.
A \term{condition} inherits the structure of its \term{supertypes}.
%!!! Barrett: It would be easier to say that conditions are classes, 
%      but maybe there is some reason to avoid doing so?
 
The metaclass of \theclass{condition} is not specified.
\term{Names} of \term{condition} \term{types} may be used to specify
\term{supertype} relationships in \macref{define-condition}, 
but the consequences are not specified if an attempt is made to use
a \term{condition} \term{type} as a \term{superclass} in a \macref{defclass} \term{form}.

% \Thenextfigure\ lists the \term{condition} \term{types} defined in this specification.
%  
% \issue{UNDEFINED-FUNCTIONS-AND-VARIABLES:COMPROMISE}
% \issue{DATA-IO:ADD-SUPPORT}
% \issue{PARSE-ERROR-STREAM:SPLIT-TYPES}
% \issue{ACCESS-ERROR-NAME}
% \issue{READER-ERROR:NEW-TYPE}
% \issue{FLOATING-POINT-CONDITION-NAMES:X3J13-NOV-89}
% \issue{COMPILER-DIAGNOSTICS:USE-HANDLER}
% \displaythree{Condition Types}{
% arithmetic-error&package-error&storage-condition\cr
% cell-error&parse-error&stream-error\cr
% condition&print-not-readable&style-warning\cr
% control-error&program-error&type-error\cr
% division-by-zero&reader-error&unbound-slot\cr
% end-of-file&serious-condition&unbound-slot-instance\cr
% error&simple-condition&unbound-variable\cr
% file-error&simple-error&undefined-function\cr
% floating-point-overflow&simple-type-error&warning\cr
% floating-point-underflow&simple-warning&\cr
% }
% \endissue{COMPILER-DIAGNOSTICS:USE-HANDLER}
% \endissue{FLOATING-POINT-CONDITION-NAMES:X3J13-NOV-89}
% \endissue{READER-ERROR:NEW-TYPE}
% \endissue{ACCESS-ERROR-NAME}
% \endissue{PARSE-ERROR-STREAM:SPLIT-TYPES}
% \endissue{DATA-IO:ADD-SUPPORT}
% \endissue{UNDEFINED-FUNCTIONS-AND-VARIABLES:COMPROMISE}

\Thenextfigure\ shows \term{operators} that
define \term{condition} \term{types} and creating \term{conditions}.

\displaythree{Operators that define and create conditions.}{
define-condition&make-condition&\cr
}

\Thenextfigure\ shows \term{operators} that \term{read} 
the \term{value} of \term{condition} \term{slots}.

\issue{DATA-IO:ADD-SUPPORT}
\issue{FORMAT-STRING-ARGUMENTS:SPECIFY}
\displaytwo{Operators that read condition slots.}{
arithmetic-error-operands&simple-condition-format-arguments\cr
arithmetic-error-operation&simple-condition-format-control\cr
cell-error-name&stream-error-stream\cr
file-error-pathname&type-error-datum\cr
package-error-package&type-error-expected-type\cr
print-not-readable-object&unbound-slot-instance\cr
}
\endissue{FORMAT-STRING-ARGUMENTS:SPECIFY}
\endissue{DATA-IO:ADD-SUPPORT}

\beginsubsubsection{Serious Conditions}

A \term{serious condition} is a \term{condition} serious
enough to require interactive intervention if not handled.  
\term{Serious conditions} are typically signaled with \funref{error} or \funref{cerror};
non-serious \term{conditions} are typically signaled with \funref{signal} or \funref{warn}.
%%Barrett: Definitional.
%All \term{serious conditions} should be \subtypesof{serious-condition}.

\endsubsubsection%{Serious Conditions}

\endSubsection%{Condition Types}

\beginsubsection{Creating Conditions}

The function \funref{make-condition} can be used to construct
a \term{condition} \term{object} explicitly.  Functions such as \funref{error},
\funref{cerror}, \funref{signal}, and \funref{warn} operate on
\term{conditions} and might create \term{condition} \term{objects}
implicitly.  Macros such as \macref{ccase}, \macref{ctypecase},
\macref{ecase}, \macref{etypecase}, \macref{check-type}, and
\macref{assert} might also implicitly create (and \term{signal})
\term{conditions}.
 
\beginsubsubsection{Condition Designators}
\DefineSection{ConditionDesignators}

A number of the functions in the condition system take arguments which
are identified as \newtermidx{condition designators}{condition designator}.
By convention, those arguments are notated as

\ \param{datum} {\rest} \param{arguments}

Taken together, the \param{datum} and the \param{arguments} are 
``\term{designators} for a \term{condition} of default type \param{default-type}.''
How the denoted \term{condition} is computed depends on the type of the \param{datum}:

\beginlist

\item{{\bull} If the \param{datum} is a \term{symbol} 
              naming a \term{condition} \term{type} $\ldots$}

The denoted \term{condition} is the result of

\code
 (apply #'make-condition \param{datum} \param{arguments})
\endcode

\issue{FORMAT-STRING-ARGUMENTS:SPECIFY}
\item{{\bull} If the \param{datum} is a \term{format control} $\ldots$}
\endissue{FORMAT-STRING-ARGUMENTS:SPECIFY}

The denoted \term{condition} is the result of 

\issue{FORMAT-STRING-ARGUMENTS:SPECIFY}
\code
 (make-condition \param{defaulted-type} 
                 :format-control \param{datum}
                 :format-arguments \param{arguments})
\endcode
\endissue{FORMAT-STRING-ARGUMENTS:SPECIFY}

%Barrett wanted this added so that implementations could do something more
%elaborate than just the given type without interfering with user programs.
%There's a lot of mail about simple-condition-disjointness-bug which might or
%might not actually turn into a formal issue someday. -kmp 2-Feb-92
where the \param{defaulted-type} is a \term{subtype} of \param{default-type}.

\item{{\bull} If the \param{datum} is a \term{condition} $\ldots$}

The denoted \term{condition} is the \param{datum} itself.
In this case, unless otherwise specified by the description of the
\term{operator} in question, the \term{arguments} must be \term{null};
that is, the consequences are undefined if any \param{arguments} were supplied. 

\endlist

Note that the \param{default-type} gets used only in the case where
the \param{datum} \term{string} is supplied.  In the other situations,
the resulting condition is not necessarily of \term{type} \param{default-type}.

Here are some illustrations of how different \term{condition designators}
can denote equivalent \term{condition} \term{objects}:

\issue{FORMAT-STRING-ARGUMENTS:SPECIFY}
\code
(let ((c (make-condition 'arithmetic-error :operator '/ :operands '(7 0))))
  (error c))
\EQ (error 'arithmetic-error :operator '/ :operands '(7 0))

(error "Bad luck.")
\EQ (error 'simple-error :format-control "Bad luck." :format-arguments '())
\endcode
\endissue{FORMAT-STRING-ARGUMENTS:SPECIFY}

\endsubsubsection%{Condition Designators}

\endsubsection%{Creating Conditions}
 
\beginsubsection{Printing Conditions}
\DefineSection{PrintingConditions}
 
If the \kwd{report} argument to \macref{define-condition} is used,
a print function is defined that is called whenever 
the defined \term{condition} is printed while \thevalueof{*print-escape*} is \term{false}. 
This function is called the \newterm{condition reporter};
the text which it outputs is called a \newterm{report message}.

When a \term{condition} is printed and \varref{*print-escape*}
is \term{false}, the \term{condition reporter} for the \term{condition} is invoked.
\term{Conditions} are printed automatically by functions such as
\funref{invoke-debugger}, \funref{break}, and \funref{warn}.

When \varref{*print-escape*} is \term{true}, the \term{object} should print in an
abbreviated fashion according to the style of the implementation
(\eg by \funref{print-unreadable-object}).  It is not required that a
\term{condition} can be recreated by reading its printed representation.
 
No \term{function} is provided for directly \term{accessing} 
or invoking \term{condition reporters}.
 
\beginsubsubsection{Recommended Style in Condition Reporting}

In order to ensure a properly aesthetic result when presenting
\term{report messages} to the user, certain stylistic conventions are
recommended.

There are stylistic recommendations for the content of the messages
output by \term{condition reporters}, but there are no formal requirements 
on those \term{programs}.
If a \term{program} violates the recommendations for some message, the
display of that message might be less aesthetic than if the guideline
had been observed, but the \term{program} is still considered a
\term{conforming program}.

The requirements on a \term{program} or \term{implementation} which
invokes a \term{condition reporter} are somewhat stronger.  A \term{conforming
program} must be permitted to assume that if these style guidelines are
followed, proper aesthetics will be maintained.  Where appropriate, any
specific requirements on such routines are explicitly mentioned below.

\beginsubsubsubsection{Capitalization and Punctuation in Condition Reports}

It is recommended that a \term{report message} be a complete sentences, in the
proper case and correctly punctuated.  In English, for example, this
means the first letter should be uppercase, and there should be a
trailing period.

\code
 (error "This is a message")  ; Not recommended
 (error "this is a message.") ; Not recommended

 (error "This is a message.") ; Recommended instead
\endcode

\endsubsubsubsection%{Capitalization and Punctuation in Condition Reports}

\beginsubsubsubsection{Leading and Trailing Newlines in Condition Reports}

It is recommended that a \term{report message} not begin with any 
introductory text, such as ``\f{Error: }'' or ``\f{Warning: }''
or even just \term{freshline} or \term{newline}. 
Such text is added, if appropriate to the context,
by the routine invoking the \term{condition reporter}.

It is recommended that a \term{report message} not be followed 
by a trailing \term{freshline} or \term{newline}.
Such text is added, if appropriate to the context, 
by the routine invoking the \term{condition reporter}.

\code
 (error "This is a message.~%")   ; Not recommended
 (error "~&This is a message.")   ; Not recommended
 (error "~&This is a message.~%") ; Not recommended

 (error "This is a message.")     ; Recommended instead
\endcode

\endsubsubsubsection%{Leading and Trailing Newlines in Condition Reports}

\beginsubsubsubsection{Embedded Newlines in Condition Reports}

Especially if it is long, it is permissible and appropriate for 
a \term{report message} to contain one or more embedded \term{newlines}.

If the calling routine conventionally inserts some additional prefix
(such as ``\f{Error: }'' or ``\f{;; Error: }'') on the first line of
the message, it must also assure that an appropriate prefix will be
added to each subsequent line of the output, so that the left edge of
the message output by the \term{condition reporter} will still be properly
aligned.

\code
 (defun test ()
   (error "This is an error message.~\%It has two lines."))

 ;; Implementation A
 (test)
 This is an error message.
 It has two lines.

 ;; Implementation B
 (test)
 ;; Error: This is an error message.
 ;;        It has two lines.

 ;; Implementation C
 (test)
 >> Error: This is an error message. 
           It has two lines.
\endcode

\endsubsubsubsection%{Embedded Newlines in Condition Reports}

\beginsubsubsubsection{Note about Tabs in Condition Reports}

Because the indentation of a \term{report message} might be shifted to the right or
left by an arbitrary amount, special care should be taken with the
semi-standard \term{character} \TabChar\ 
(in those \term{implementations} that support such a \term{character}).  
Unless the \term{implementation} specifically defines its behavior 
in this context, its use should be avoided.

\endsubsubsubsection%{Note about Tabs in Condition Reports}

\beginsubsubsubsection{Mentioning Containing Function in Condition Reports}

The name of the containing function should generally not be mentioned in
\term{report messages}.  It is assumed that the \term{debugger} will make this
information accessible in situations where it is necessary and appropriate.

\endsubsubsubsection%{Mentioning Containing Function in Condition Reports}

\endsubsubsection%{Recommended Style in Condition Reporting}

\endsubsection%{Printing Conditions}
\goodbreak
\beginSubsection{Signaling and Handling Conditions}
\DefineSection{CondSignalHandle}
 
The operation of the condition system depends on the ordering of
active \term{applicable handlers} from most recent to least recent.

Each \term{handler} is associated with a \term{type specifier}
that must designate a \subtypeof{condition}.  A \term{handler}
is said to be \term{applicable} to a \term{condition} if that
\term{condition} is of the \term{type} designated by the associated
\term{type specifier}.

\term{Active} \term{handlers} are \term{established} by using 
\macref{handler-bind} (or an abstraction based on \macref{handler-bind}, 
such as \macref{handler-case} or \macref{ignore-errors}).
 
\term{Active} \term{handlers} can be \term{established} within the
dynamic scope of other \term{active} \term{handlers}.
At any point during program execution, there is a set of \term{active} \term{handlers}.
When a \term{condition} is signaled, the \term{most recent} active \term{applicable handler}
for that \term{condition} is selected from this set.
Given a \term{condition}, the order of recentness of 
active \term{applicable handlers} is defined by the following two rules:
 
\beginlist
 
 \itemitem{1.} Each handler in a set of active handlers $H\sub 1$ is
more recent than every handler in a set $H\sub 2$ if the
handlers in $H\sub 2$ were active when the handlers in $H\sub 1$ were
established.
 
 \itemitem{2.} Let $h\sub 1$ and $h\sub 2$ be two applicable active
handlers established by the same \term{form}. Then $h\sub 1$ is
more recent than $h\sub 2$ if $h\sub 1$ was defined to the left of
$h\sub 2$ in the \term{form} that established them.
 
\endlist
 
%!!! Barrett:  This doesn't match my reading of CSv18, p12.
%     I believe contradicts p21,p22 of that document.  It also differs from
%     previous paragraph (w/ item 2) and first paragraph under "signaling".
Once a handler in a handler binding \term{form} (such as 
\macref{handler-bind} or \macref{handler-case}) has been selected, all
handlers in that \term{form} become inactive for 
the remainder of the signaling process.
%--------------------------------------
While the selected \term{handler} runs, no other \term{handler} established
by that \term{form} is active. That is, if the \term{handler} declines, 
no other handler established by that \term{form} will be considered for possible invocation.
%-----------------------------------------------------------------------------------------

\Thenextfigure\ shows \term{operators} relating to 
the \term{handling} of \term{conditions}.

\displaythree{Operators relating to handling conditions.}{
handler-bind&handler-case&ignore-errors\cr
}
 
\beginsubsubsection{Signaling}
\DefineSection{Signaling}
 
When a \term{condition} is signaled, the most recent
applicable \term{active} \term{handler} is invoked.  
Sometimes a handler will decline by simply returning
without a transfer of control.
In such cases, the next most recent applicable active handler is
invoked. 
 
If there are no applicable handlers for a \term{condition} that
has been signaled, or if all applicable handlers decline, the
\term{condition} is unhandled.
 
The functions \funref{cerror} and \funref{error} invoke the
interactive \term{condition} handler (the debugger) rather than
return if the \term{condition} being signaled, regardless of
its \term{type}, is unhandled.  In contrast, \funref{signal}
returns \nil\ if the \term{condition} being signaled,
regardless of its \term{type}, is unhandled.

\Thevariable{*break-on-signals*} can be used to cause the
debugger to be entered before the signaling process begins.
 
\Thenextfigure\ shows \term{defined names} relating to
the \term{signaling} of \term{conditions}.

\displaythree{Defined names relating to signaling conditions.}{
*break-on-signals*&error&warn\cr
cerror&signal&\cr
}

\beginsubsubsubsection{Resignaling a Condition}

\issue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}
During the \term{dynamic extent} of the \term{signaling} process for
a particular \term{condition} \term{object}, 
\funref{signaling} the same \term{condition} \term{object} again
is permitted if and only if the \term{situation} represented in both
cases are the same.

For example, a \term{handler} might legitimately \term{signal} 
the \term{condition} \term{object} that is its \term{argument}
in order to allow outer \term{handlers} first opportunity to \term{handle} 
the condition.  (Such a \term{handlers} is sometimes called a ``default handler.'')
This action is permitted because the \term{situation} which the second
\term{signaling} process is addressing is really the same \term{situation}.

On the other hand, in an \term{implementation} that implemented asynchronous 
keyboard events by interrupting the user process with a call to \funref{signal},
it would not be permissible for two distinct asynchronous keyboard events
to \term{signal} \term{identical} \term{condition} \term{objects}
at the same time for different 
%% Per X3J13 (at request of Gadbois). -kmp 5-Oct-93
%the
situations.
\endissue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}
 
\endsubsubsubsection%{Resignaling a Condition}

\endsubsubsection%{Signaling}
\beginsubsubsection{Restarts}
\DefineSection{Restarts}
 
%!!! Barrett: Not true. Debugger may permit arbitrary returning, depending on implementation.
The interactive condition handler returns only through
non-local transfer of control to specially defined \term{restarts}
that can be set up either by the system or by user code.  Transferring
control to a restart is called ``invoking'' the restart.  Like
handlers, active \term{restarts} are \term{established}
dynamically, and 
only active \term{restarts}
can be invoked.  An active 
\term{restart} can be invoked by the user from
the debugger or by a program by using \funref{invoke-restart}.
 
%!!! Barrett: :TEST predicate also affects applicability.
A \term{restart} contains a 
\term{function} to be \term{called} when the \term{restart} is
invoked, an optional name that can be used to find or invoke the 
\term{restart}, and
an optional set of interaction information for the debugger to use to
enable the user to manually invoke a \term{restart}. 
%    some optional information that allows the debugger 
%    to manage the interactive selection of the \term{restart} in
%    situations where program handlers cannot select one.

The name of a \term{restart} is
used by \funref{invoke-restart}. \term{Restarts} that can be invoked
only within the debugger do not need names.
%Useless information - terms unnamed and anonymous are never used.
%\term{Restart} names provide a means 
%to access program interfaces
%such as 
%\funref{find-restart} and \funref{invoke-restart}.  
%\term{Restarts} named
%\nil\ are called ``unnamed'' or ``anonymous'' 
%\term{restarts}. Named \term{restarts} 
%can be used in both
%interactive and non-interactive situations, but unnamed
%\term{restarts} are typically useful only in interactive situations.
 
\term{Restarts} can be established by using \macref{restart-bind}, 
\macref{restart-case}, and \macref{with-simple-restart}.
A \term{restart} function can itself invoke any other \term{restart}
that was active at the time of establishment of the \term{restart} 
of which the \term{function} is part.

\issue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}
\issue{JUN90-TRIVIAL-ISSUES:14}
The \term{restarts} \term{established} by 
    a \macref{restart-bind} \term{form}, 
    a \macref{restart-case} \term{form},
 or a \macref{with-simple-restart} \term{form}
have \term{dynamic extent}
which extends for the duration of that \term{form}'s execution.
\endissue{JUN90-TRIVIAL-ISSUES:14}
\endissue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}

\term{Restarts} of the same name can be ordered from least recent to
most recent according to the following two rules:
 
\beginlist
 
 \itemitem{1.} Each \term{restart} in a set of active restarts
$R\sub 1$ is more recent than every \term{restart} in a
set $R\sub 2$ if the \term{restarts} 
in $R\sub 2$ were active when the  \term{restarts} in $R\sub 1$ were
established.
 
 \itemitem{2.} Let $r\sub 1$ and $r\sub 2$ be two active \term{restarts} with
the same name established by the same \term{form}. Then $r\sub 1$ is
more recent than $r\sub 2$ if $r\sub 1$ was defined to the
left of $r\sub 2$ in the \term{form} that established them.
 
\endlist

If a \term{restart} is invoked but does not transfer control,
the values resulting from the \term{restart} function are
returned by the function that invoked the restart, either
\funref{invoke-restart} or \funref{invoke-restart-interactively}.
%!!! Barrett: There are other functions that invoke restarts that we built
%      on these two; i.e., muffle-warning, etc.

\beginsubsubsubsection{Interactive Use of Restarts}

%%KMP replaced 
%Interaction information comprises two \term{functions}, called the
%\kwd{report} function and the \kwd{interactive}
%function. The \kwd{report} function is used to print a
%description of the \term{restart} and takes a 
%\term{stream} as an argument.  The
%\kwd{interactive} function is used to produce a list of
%arguments for the \term{restart} function if that 
%\term{restart} is invoked; the
%\kwd{interactive} function takes no arguments.
% 
% 
%When the debugger is entered, the set of active \term{restarts} 
%is presented to the user. If an active \term{restart} has interaction
%information, the \kwd{report} function is used to print a
%description of the \term{restart}. If the 
%\term{restart} is invoked (in an
%\term{implementation-dependent} manner), the \kwd{interactive} function
%is invoked to produce a list of arguments to which the \term{restart}
%function is applied as if with \funref{apply}. If an active
%\term{restart} does not have a \kwd{report} 
%function and a description
%of the \term{restart} is printed, the description is
%\term{implementation-dependent}.  If an active \term{restart} does have not an
%\kwd{interactive} function and the \term{restart} is invoked
%interactively, the \term{restart} 
%function is invoked with no arguments.
%\macref{restart-bind} and \macref{restart-case} supply
%defaults for interaction information.  
%If a \term{restart} 
%is invoked from
%within the debugger and the \term{restart} 
%function simply returns,
%the consequences are unspecified.
 
For interactive handling, two pieces of information are needed
from a \term{restart}: a report function and an interactive function.
 
The report function
%which can be specified using the 
%\kwd{report-function}
%   keyword in 
%\macref{restart-bind} or the 
%\kwd{report} keyword in \macref{restart-case},
is used by a program such as the debugger to
present a description of the action the \term{restart} will take.  
The report function is specified and established by the 
\kwd{report-function} keyword to
\macref{restart-bind} or the 
\kwd{report} keyword to \macref{restart-case}. 
%The
%   report function is a function of one argument, a stream on which
%   the output is to be done.  If no report function is specified by
%   the user, the restart will be reported in an \term{implementation-dependent}
%   way. 
 
The interactive function, which can be specified using the 
\kwd{interactive-function} keyword to 
\macref{restart-bind} or \kwd{interactive} keyword
to \macref{restart-case}, is used when the \term{restart}
is invoked
interactively, such as from the debugger, to produce a suitable
list of arguments. 
%The function takes no arguments, and may
%   prompt interactively on *QUERY-IO* if necessary.  The result 
%   should be a list of arguments suitable for use in the expression
%     (APPLY #'INVOKE-RESTART <restart> <arguments>).
%   If no interactive function is specified by the user, the argument
%   list NIL will be assumed.
 
\funref{invoke-restart} invokes the most recently \term{established}
%!!! Barrett: "active" -- :TEST could inhibit activation.
\term{restart} whose
name is the same as the first argument to \funref{invoke-restart}.
If a \term{restart} is invoked interactively by the debugger and  does
not transfer control but rather returns values, the precise
action of the debugger on those values is \term{implementation-defined}.

%!!! Barrett: This doesn't talk about supplying restart objects to INVOKE-RESTART.
 
\endsubsubsubsection%{Interactive Use of Restarts}
\beginsubsubsubsection{Interfaces to Restarts}
\DefineSection{InterfacesToRestarts}
 
%%KMP replaced
%It is possible to define a functional interface that hides the use of
%\funref{invoke-restart}.  The functions \funref{abort},
%\funref{continue}, \funref{muffle-warning}, \funref{store-value},
%and \funref{use-value} are such interfaces.
Some \term{restarts} have functional interfaces, 
%either for syntactic
%   convenience or to de-emphasize the use of restarts in their
%   implementation.  
such as \funref{abort}, \funref{continue}, 
\funref{muffle-warning}, \funref{store-value}, and 
\funref{use-value}.
They are ordinary functions that use 
 \funref{find-restart} and \funref{invoke-restart} internally,
that have the same name as the \term{restarts} they manipulate,
and that are provided simply for notational convenience.
 
\Thenextfigure\ shows \term{defined names} relating to
\term{restarts}.

\displaythree{Defined names relating to restarts.}{
abort&invoke-restart-interactively&store-value\cr
compute-restarts&muffle-warning&use-value\cr
continue&restart-bind&with-simple-restart\cr
find-restart&restart-case&\cr
invoke-restart&restart-name&\cr
}
 
\endsubsubsubsection%{Interfaces to Restarts}

\beginsubsubsubsection{Restart Tests}

\issue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}

Each \term{restart} has an associated test, which is a function of one
argument (a \term{condition} or \nil) which returns \term{true} if the \term{restart}
should be visible in the current \term{situation}.  This test is created by 
the \kwd{test-function} option to \macref{restart-bind} or 
the \kwd{test} option to \macref{restart-case}.
\endissue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}

\endsubsubsubsection%{Restart Tests}

\beginsubsubsubsection{Associating a Restart with a Condition}
\DefineSection{AssocRestartWithCond}

\issue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}
A \term{restart} can be ``associated with'' a \term{condition} explicitly
by \macref{with-condition-restarts}, or implicitly by \macref{restart-case}.
Such an assocation has \term{dynamic extent}.

A single \term{restart} may be associated with several \term{conditions} 
at the same time.
A single \term{condition} may have several associated \term{restarts}
at the same time.

Active restarts associated with a particular \term{condition} can be detected
by \term{calling} a \term{function} such as \funref{find-restart}, supplying
that \term{condition} as the \param{condition} \term{argument}.
Active restarts can also be detected without regard to any associated
\term{condition} by calling such a function without a \param{condition} \term{argument},
or by supplying a value of \nil\ for such an \term{argument}.
\endissue{CONDITION-RESTARTS:PERMIT-ASSOCIATION}

\endsubsubsubsection%{Associating a Restart with a Condition}

\endsubsubsection%{Restarts}
\endSubsection%{Signaling and Handling Conditions}
\goodbreak
\beginSubsection{Assertions}
 
Conditional signaling of \term{conditions}
based on such things as key match, form evaluation,
and \term{type} are handled by assertion \term{operators}.
\Thenextfigure\ shows \term{operators} relating to assertions.

\displaythree{Operators relating to assertions.}{
assert&check-type&ecase\cr
ccase&ctypecase&etypecase\cr
}
 
\endSubsection%{Assertions}

% Date: Tue, 4 Jun 91 10:39:28 EDT
% From: kab@chestnut.com (Kim Barrett)
% To: mueller@sumex-aim.stanford.edu
% Cc: common-lisp@mcc.com
% Message-Id: <9106041439.AA08564@chestnut.com>
% In-Reply-To: <2884982915-12705048@KSL-EXP-30>
% Subject: CLEH restarts...
% 
% [...]
% The rules for determining whether a (visible) restart is active are:
% 
% 1. If the test function for the restart returns false when applied to the
% condition argument (which may be NIL), then the restart is inactive.
% 
% 2. If testing for applicability to a specific condition (the condition argument
% is actually a condition object), then either
%   a. The restart is associated with the specified condition.
%   b. The restart is not associated with some other condition object.
%      [This case was accidentally left out of the Version 2 proposal.]
% or the restart is inactive.
% 
% 3. If neither of (1) and (2) indicate the restart to be inactive, then it is
% active.
% [...]


\beginsubsection{Notes about the Condition System's Background}

For a background reference to the abstract concepts detailed in this
section, see \CondSysPaper.  The details of that paper are not binding on
this document, but may be helpful in establishing a conceptual basis for
understanding this material.

\endsubsection%{Notes about the Condition System's Background}