Use `=daikanwa/ho' instead of `morohashi-daikanwa' for M-Hdddd.

[chise/xemacs-chise.git] / man / lispref / edebug-inc.texi

@comment -*-texinfo-*-

@node Edebug, , Compilation Errors, Top
@section Edebug
@cindex Edebug mode

@cindex Edebug
  Edebug is a source-level debugger for XEmacs Lisp programs that
provides the following features:

@itemize @bullet
@item
Step through evaluation, stopping before and after each expression.

@item
Set conditional or unconditional breakpoints, install embedded
breakpoints, or a global break event.

@item
Trace slow or fast stopping briefly at each stop point, or
each breakpoint.

@item
Display expression results and evaluate expressions as if outside of
Edebug.  Interface with the custom printing package
for printing circular structures.

@item
Automatically reevaluate a list of expressions and
display their results each time Edebug updates the display.

@item
Output trace info on function enter and exit.

@item
Errors stop before the source causing the error.

@item
Display backtrace without Edebug calls.

@item
Allow specification of argument evaluation for macros and defining forms.

@item
Provide rudimentary coverage testing and display of frequency counts.

@end itemize

The first three sections should tell you enough about Edebug to enable
you to use it.

@menu
* Using Edebug::                Introduction to use of Edebug.
* Instrumenting::               You must first instrument code.
* Edebug Execution Modes::      Execution modes, stopping more or less often.
* Jumping::                     Commands to jump to a specified place.
* Edebug Misc::                 Miscellaneous commands.
* Breakpoints::                 Setting breakpoints to make the program stop.
* Trapping Errors::             trapping errors with Edebug.
* Edebug Views::                Views inside and outside of Edebug.
* Edebug Eval::                 Evaluating expressions within Edebug.
* Eval List::                   Automatic expression evaluation.
* Reading in Edebug::           Customization of reading.
* Printing in Edebug::          Customization of printing.
* Tracing::                     How to produce tracing output.
* Coverage Testing::            How to test evaluation coverage.
* The Outside Context::         Data that Edebug saves and restores.
* Instrumenting Macro Calls::   Specifying how to handle macro calls.
* Edebug Options::              Option variables for customizing Edebug.
@end menu

@node Using Edebug
@subsection Using Edebug

  To debug an XEmacs Lisp program with Edebug, you must first
@dfn{instrument} the Lisp code that you want to debug.  If you want to
just try it now, load @file{edebug.el}, move point into a definition and
do @kbd{C-u C-M-x} (@code{eval-defun} with a prefix argument).
See @ref{Instrumenting} for alternative ways to instrument code.

  Once a function is instrumented, any call to the function activates
Edebug.  Activating Edebug may stop execution and let you step through
the function, or it may update the display and continue execution while
checking for debugging commands, depending on the selected Edebug
execution mode.  The initial execution mode is @code{step}, by default,
which does stop execution.  @xref{Edebug Execution Modes}.

  Within Edebug, you normally view an XEmacs buffer showing the source of
the Lisp function you are debugging.  This is referred to as the
@dfn{source code buffer}---but note that it is not always the same
buffer depending on which function is currently being executed.

  An arrow at the left margin indicates the line where the function is
executing.  Point initially shows where within the line the function is
executing, but you can move point yourself.

  If you instrument the definition of @code{fac} (shown below) and then
execute @code{(fac 3)}, here is what you normally see.  Point is at the
open-parenthesis before @code{if}.

@example
(defun fac (n)
=>@point{}(if (< 0 n)
      (* n (fac (1- n)))
    1))
@end example

@cindex stop points
The places within a function where Edebug can stop execution are called
@dfn{stop points}.  These occur both before and after each subexpression
that is a list, and also after each variable reference.
Here we show with periods the stop points found in the function
@code{fac}:

@example
(defun fac (n)
  .(if .(< 0 n.).
      .(* n. .(fac (1- n.).).).
    1).)
@end example

While the source code buffer is selected, the special commands of Edebug
are available in it, in addition to the commands of XEmacs Lisp mode.
(The buffer is temporarily made read-only, however.)  For example, you
can type the Edebug command @key{SPC} to execute until the next stop
point.  If you type @key{SPC} once after entry to @code{fac}, here is
the display you will see:

@example
(defun fac (n)
=>(if @point{}(< 0 n)
      (* n (fac (1- n)))
    1))
@end example

When Edebug stops execution after an expression, it displays the
expression's value in the echo area.

Other frequently used commands are @kbd{b} to set a breakpoint at a stop
point, @kbd{g} to execute until a breakpoint is reached, and @kbd{q} to
exit to the top-level command loop.  Type @kbd{?} to display a list of
all Edebug commands.


@node Instrumenting
@subsection Instrumenting for Edebug

  In order to use Edebug to debug Lisp code, you must first
@dfn{instrument} the code.  Instrumenting a form inserts additional code
into it which invokes Edebug at the proper places.  Furthermore, if
Edebug detects a syntax error while instrumenting, point is left at the
erroneous code and an @code{invalid-read-syntax} error is signaled.

@kindex C-M-x
@findex eval-defun (Edebug)
@findex edebug-all-defs
  Once you have loaded Edebug, the command @kbd{C-M-x}
(@code{eval-defun}) is redefined so that when invoked with a prefix
argument on a definition, it instruments the definition before
evaluating it.  (The source code itself is not modified.)  If the
variable @code{edebug-all-defs} is non-@code{nil}, that inverts the
meaning of the prefix argument: then @kbd{C-M-x} instruments the
definition @emph{unless} it has a prefix argument.  The default value of
@code{edebug-all-defs} is @code{nil}.  The command @kbd{M-x
edebug-all-defs} toggles the value of the variable
@code{edebug-all-defs}.

@findex edebug-all-forms
@findex eval-region (Edebug)
@findex eval-current-buffer (Edebug)
  If @code{edebug-all-defs} is non-@code{nil}, then the commands
@code{eval-region}, @code{eval-current-buffer}, and @code{eval-buffer}
also instrument any definitions they evaluate.  Similarly,
@code{edebug-all-forms} controls whether @code{eval-region} should
instrument @emph{any} form, even non-defining forms.  This doesn't apply
to loading or evaluations in the minibuffer.  The command @kbd{M-x
edebug-all-forms} toggles this option.

@findex edebug-eval-top-level-form
Another command, @kbd{M-x edebug-eval-top-level-form}, is available to
instrument any top-level form regardless of the value of
@code{edebug-all-defs} or @code{edebug-all-forms}.

Just before Edebug instruments any code, it calls any functions in the
variable @code{edebug-setup-hook} and resets its value to @code{nil}.
You could use this to load up Edebug specifications associated with a
package you are using but only when you also use Edebug.  For example,
@file{my-specs.el} may be loaded automatically when you use
@code{my-package} with Edebug by including the following code in
@file{my-package.el}.

@example
(add-hook 'edebug-setup-hook
  (function (lambda () (require 'my-specs))))
@end example

While Edebug is active, the command @kbd{I}
(@code{edebug-instrument-callee}) instruments the definition of the
function or macro called by the list form after point, if is not already
instrumented.  If the location of the definition is not known to Edebug,
this command cannot be used.  After loading Edebug, @code{eval-region}
records the position of every definition it evaluates, even if not
instrumenting it.  Also see the command @kbd{i} (@ref{Jumping}) which
steps into the callee.

@cindex special forms (Edebug)
@cindex interactive commands (Edebug)
@cindex anonymous lambda expressions (Edebug)
@cindex Common Lisp (Edebug)
@pindex cl.el (Edebug)
@pindex cl-specs.el
  Edebug knows how to instrument all the standard special forms, an
interactive form with an expression argument, anonymous lambda
expressions, and other defining forms.  (Specifications for macros
defined by @file{cl.el} (version 2.03) are provided in
@file{cl-specs.el}.)  Edebug cannot know what a user-defined macro will
do with the arguments of a macro call so you must tell it.  See
@ref{Instrumenting Macro Calls} for the details.

@findex eval-expression (Edebug)
  Note that a couple ways remain to evaluate expressions without
instrumenting them.  Loading a file via the @code{load} subroutine does
not instrument expressions for Edebug.  Evaluations in the minibuffer
via @code{eval-expression} (@kbd{M-ESC}) are not instrumented.

  To remove instrumentation from a definition, simply reevaluate it with
one of the non-instrumenting commands, or reload the file.

See @ref{Edebug Eval} for other evaluation functions available
inside of Edebug.


@node Edebug Execution Modes
@subsection Edebug Execution Modes

@cindex Edebug execution modes
Edebug supports several execution modes for running the program you are
debugging.  We call these alternatives @dfn{Edebug execution modes}; do
not confuse them with major or minor modes.  The current Edebug
execution mode determines how Edebug displays the progress of the
evaluation, whether it stops at each stop point, or continues to the
next breakpoint, for example.

Normally, you specify the Edebug execution mode by typing a command
to continue the program in a certain mode.  Here is a table of these
commands.  All except for @kbd{S} resume execution of the program, at
least for a certain distance.

@table @kbd
@item S
Stop: don't execute any more of the program for now, just wait for more
Edebug commands (@code{edebug-stop}).

@item @key{SPC}
Step: stop at the next stop point encountered (@code{edebug-step-mode}).

@item n
Next: stop at the next stop point encountered after an expression
(@code{edebug-next-mode}).  Also see @code{edebug-forward-sexp} in
@ref{Edebug Misc}.

@item t
Trace: pause one second at each Edebug stop point (@code{edebug-trace-mode}).

@item T
Rapid trace: update at each stop point, but don't actually
pause (@code{edebug-Trace-fast-mode}).

@item g
Go: run until the next breakpoint (@code{edebug-go-mode}).  @xref{Breakpoints}.

@item c
Continue: pause for one second at each breakpoint, but don't stop
(@code{edebug-continue-mode}).

@item C
Rapid continue: update at each breakpoint, but don't actually pause
(@code{edebug-Continue-fast-mode}).

@item G
Go non-stop: ignore breakpoints (@code{edebug-Go-nonstop-mode}).  You
can still stop the program by hitting any key.
@end table

In general, the execution modes earlier in the above list run the
program more slowly or stop sooner.

When you enter a new Edebug level, the initial execution mode comes from
the value of the variable @code{edebug-initial-mode}.  By default, this
specifies @code{step} mode.  Note that you may reenter the same Edebug
level several times if, for example, an instrumented function is called
several times from one command.

While executing or tracing, you can interrupt the execution by typing
any Edebug command.  Edebug stops the program at the next stop point and
then executes the command that you typed.  For example, typing @kbd{t}
during execution switches to trace mode at the next stop point.  You can
use @kbd{S} to stop execution without doing anything else.

If your function happens to read input, a character you hit intending to
interrupt execution may be read by the function instead.  You can avoid
such unintended results by paying attention to when your program wants
input.

@cindex keyboard macros (Edebug)
Keyboard macros containing Edebug commands do not work; when you exit
from Edebug, to resume the program, whether you are defining or
executing a keyboard macro is forgotten.  Also, defining or executing a
keyboard macro outside of Edebug does not affect the command loop inside
Edebug.  This is usually an advantage.  But see
@code{edebug-continue-kbd-macro}.


@node Jumping
@subsection Jumping

Commands described here let you jump to a specified location.
All, except @kbd{i}, use temporary breakpoints to establish the stop
point and then switch to @code{go} mode.  Any other breakpoint reached
before the intended stop point will also stop execution.  See
@ref{Breakpoints} for the details on breakpoints.

@table @kbd
@item f
Run the program forward over one expression
(@code{edebug-forward-sexp}).  More precisely, set a temporary
breakpoint at the position that @kbd{C-M-f} would reach, then execute in
@code{go} mode so that the program will stop at breakpoints.

With a prefix argument @var{n}, the temporary breakpoint is placed
@var{n} sexps beyond point.  If the containing list ends before @var{n}
more elements, then the place to stop is after the containing
expression.

Be careful that the position @kbd{C-M-f} finds is a place that the
program will really get to; this may not be true in a
@code{cond}, for example.

This command does @code{forward-sexp} starting at point rather than the
stop point.  If you want to execute one expression from the current stop
point, type @kbd{w} first, to move point there.

@item o
Continue ``out of'' an expression (@code{edebug-step-out}).  It places a
temporary breakpoint at the end of the sexp containing point.

If the containing sexp is a function definition itself, it continues
until just before the last sexp in the definition.  If that is where you
are now, it returns from the function and then stops.  In other words,
this command does not exit the currently executing function unless you
are positioned after the last sexp.

@item I
Step into the function or macro after point after first ensuring that it
is instrumented.  It does this by calling @code{edebug-on-entry} and
then switching to @code{go} mode.

Although the automatic instrumentation is convenient, it is not
later automatically uninstrumented.

@item h
Proceed to the stop point near where point is using a temporary
breakpoint (@code{edebug-goto-here}).

@end table

All the commands in this section may fail to work as expected in case
of nonlocal exit, because a nonlocal exit can bypass the temporary
breakpoint where you expected the program to stop.

@node Edebug Misc
@subsection Miscellaneous

Some miscellaneous commands are described here.

@table @kbd
@item ?
Display the help message for Edebug (@code{edebug-help}).

@item C-]
Abort one level back to the previous command level
(@code{abort-recursive-edit}).

@item q
Return to the top level editor command loop (@code{top-level}).  This
exits all recursive editing levels, including all levels of Edebug
activity.  However, instrumented code protected with
@code{unwind-protect} or @code{condition-case} forms may resume
debugging.

@item Q
Like @kbd{q} but don't stop even for protected code
(@code{top-level-nonstop}).

@item r
Redisplay the most recently known expression result in the echo area
(@code{edebug-previous-result}).

@item d
Display a backtrace, excluding Edebug's own functions for clarity
(@code{edebug-backtrace}).

You cannot use debugger commands in the backtrace buffer in Edebug as
you would in the standard debugger.

The backtrace buffer is killed automatically when you continue
execution.
@end table

From the Edebug recursive edit, you may invoke commands that activate
Edebug again recursively.  Any time Edebug is active, you can quit to
the top level with @kbd{q} or abort one recursive edit level with
@kbd{C-]}.  You can display a backtrace of all the
pending evaluations with @kbd{d}.


@node Breakpoints
@subsection Breakpoints

@cindex breakpoints
There are three more ways to stop execution once it has started:
breakpoints, the global break condition, and embedded breakpoints.

While using Edebug, you can specify @dfn{breakpoints} in the program you
are testing: points where execution should stop.  You can set a
breakpoint at any stop point, as defined in @ref{Using Edebug}.  For
setting and unsetting breakpoints, the stop point that is affected is
the first one at or after point in the source code buffer.  Here are the
Edebug commands for breakpoints:

@table @kbd
@item b
Set a breakpoint at the stop point at or after point
(@code{edebug-set-breakpoint}).  If you use a prefix argument, the
breakpoint is temporary (it turns off the first time it stops the
program).

@item u
Unset the breakpoint (if any) at the stop point at or after the current
point (@code{edebug-unset-breakpoint}).

@item x @var{condition} @key{RET}
Set a conditional breakpoint which stops the program only if
@var{condition} evaluates to a non-@code{nil} value
(@code{edebug-set-conditional-breakpoint}).  If you use a prefix
argument, the breakpoint is temporary (it turns off the first time it
stops the program).

@item B
Move point to the next breakpoint in the definition
(@code{edebug-next-breakpoint}).
@end table

While in Edebug, you can set a breakpoint with @kbd{b} and unset one
with @kbd{u}.  First you must move point to a position at or before the
desired Edebug stop point, then hit the key to change the breakpoint.
Unsetting a breakpoint that has not been set does nothing.

Reevaluating or reinstrumenting a definition clears all its breakpoints.

A @dfn{conditional breakpoint} tests a condition each time the program
gets there.  To set a conditional breakpoint, use @kbd{x}, and specify
the condition expression in the minibuffer.  Setting a conditional
breakpoint at a stop point that already has a conditional breakpoint
puts the current condition expression in the minibuffer so you can edit
it.

You can make both conditional and unconditional breakpoints
@dfn{temporary} by using a prefix arg to the command to set the
breakpoint.  After breaking at a temporary breakpoint, it is
automatically cleared.

Edebug always stops or pauses at a breakpoint except when the Edebug
mode is @code{Go-nonstop}.  In that mode, it ignores breakpoints entirely.

To find out where your breakpoints are, use @kbd{B}, which
moves point to the next breakpoint in the definition following point, or
to the first breakpoint if there are no following breakpoints.  This
command does not continue execution---it just moves point in the buffer.

@menu
* Global Break Condition::      Breaking on an event.
* Embedded Breakpoints::        Embedding breakpoints in code.
@end menu


@node Global Break Condition
@subsubsection Global Break Condition

@cindex stopping on events
@cindex global break condition
In contrast to breaking when execution reaches specified locations,
you can also cause a break when a certain event occurs.  The @dfn{global
break condition} is a condition that is repeatedly evaluated at every
stop point.  If it evaluates to a non-@code{nil} value, then execution
is stopped or paused depending on the execution mode, just like a
breakpoint.  Any errors that might occur as a result of evaluating the
condition are ignored, as if the result were @code{nil}.

@findex edebug-set-global-break-condition
@vindex edebug-global-break-condition
You can set or edit the condition expression, stored in
@code{edebug-global-break-condition}, using @kbd{X}
(@code{edebug-set-global-break-condition}).

Using the global break condition is perhaps the fastest way
to find where in your code some event occurs, but since it is rather
expensive you should reset the condition to @code{nil} when not in use.


@node Embedded Breakpoints
@subsubsection Embedded Breakpoints

@findex edebug
@cindex embedded breakpoints
Since all breakpoints in a definition are cleared each time you
reinstrument it, you might rather create an @dfn{embedded breakpoint}
which is simply a call to the function @code{edebug}.  You can, of
course, make such a call conditional.  For example, in the @code{fac}
function, insert the first line as shown below to stop when the argument
reaches zero:

@example
(defun fac (n)
  (if (= n 0) (edebug))
  (if (< 0 n)
      (* n (fac (1- n)))
    1))
@end example

When the @code{fac} definition is instrumented and the function is
called, Edebug will stop before the call to @code{edebug}.  Depending on
the execution mode, Edebug will stop or pause.

However, if no instrumented code is being executed, calling
@code{edebug} will instead invoke @code{debug}.  Calling @code{debug}
will always invoke the standard backtrace debugger.


@node Trapping Errors
@subsection Trapping Errors

@vindex edebug-on-error
@vindex edebug-on-quit
An error may be signaled by subroutines or XEmacs Lisp code.  If a signal
is not handled by a @code{condition-case}, this indicates an
unrecognized situation has occurred.  If Edebug is not active when an
unhandled error is signaled, @code{debug} is run normally (if
@code{debug-on-error} is non-@code{nil}).  But while Edebug is active,
@code{debug-on-error} and @code{debug-on-quit} are bound to
@code{edebug-on-error} and @code{edebug-on-quit}, which are both
@code{t} by default.  Actually, if @code{debug-on-error} already has
a non-@code{nil} value, that value is still used.

It is best to change the values of @code{edebug-on-error} or
@code{edebug-on-quit} when Edebug is not active since their values won't
be used until the next time Edebug is invoked at a deeper command level.
If you only change @code{debug-on-error} or @code{debug-on-quit} while
Edebug is active, these changes will be forgotten when Edebug becomes
inactive.  Furthermore, during Edebug's recursive edit, these variables
are bound to the values they had outside of Edebug.

Edebug shows you the last stop point that it knew about before the
error was signaled.  This may be the location of a call to a function
which was not instrumented, within which the error actually occurred.
For an unbound variable error, the last known stop point might be quite
distant from the offending variable.  If the cause of the error is not
obvious at first, note that you can also get a full backtrace inside of
Edebug (see @ref{Edebug Misc}).

Edebug can also trap signals even if they are handled.  If
@code{debug-on-error} is a list of signal names, Edebug will stop when
any of these errors are signaled.  Edebug shows you the last known stop
point just as for unhandled errors.  After you continue execution, the
error is signaled again (but without being caught by Edebug).  Edebug
can only trap errors that are handled if they are signaled in Lisp code
(not subroutines) since it does so by temporarily replacing the
@code{signal} function.


@node Edebug Views
@subsection Edebug Views

The following Edebug commands let you view aspects of the buffer and
window status that obtained before entry to Edebug.

@table @kbd
@item v
View the outside window configuration (@code{edebug-view-outside}).

@item p
Temporarily display the outside current buffer with point at its outside
position (@code{edebug-bounce-point}). If prefix arg is supplied, sit for
that many seconds instead.

@item w
Move point back to the current stop point (@code{edebug-where}) in the
source code buffer.  Also, if you use this command in another window
displaying the same buffer, this window will be used instead to
display the buffer in the future.

@item W
Toggle the @code{edebug-save-windows} variable which indicates whether
the outside window configuration is saved and restored
(@code{edebug-toggle-save-windows}).  Also, each time it is toggled on,
make the outside window configuration the same as the current window
configuration.

With a prefix argument, @code{edebug-toggle-save-windows} only toggles
saving and restoring of the selected window.  To specify a window that
is not displaying the source code buffer, you must use @kbd{C-xXW} from
the global keymap.


@end table

You can view the outside window configuration with @kbd{v} or just
bounce to the current point in the current buffer with @kbd{p}, even if
it is not normally displayed.  After moving point, you may wish to pop
back to the stop point with @kbd{w} from a source code buffer.

By using @kbd{W} twice, Edebug again saves and restores the
outside window configuration, but to the current configuration.  This is
a convenient way to, for example, add another buffer to be displayed
whenever Edebug is active.  However, the automatic redisplay of
@samp{*edebug*} and @samp{*edebug-trace*} may conflict with the buffers
you wish to see unless you have enough windows open.


@node Edebug Eval
@subsection Evaluation

While within Edebug, you can evaluate expressions ``as if'' Edebug were
not running.  Edebug tries to be invisible to the expression's
evaluation and printing.  Evaluation of expressions that cause side
effects will work as expected except for things that Edebug explicitly
saves and restores.  See @ref{The Outside Context} for details on this
process.  Also see @ref{Reading in Edebug} and @ref{Printing in Edebug}
for topics related to evaluation.

@table @kbd
@item e @var{exp} @key{RET}
Evaluate expression @var{exp} in the context outside of Edebug
(@code{edebug-eval-expression}).  In other words, Edebug tries to avoid
altering the effect of @var{exp}.

@item M-@key{ESC} @var{exp} @key{RET}
Evaluate expression @var{exp} in the context of Edebug itself.

@item C-x C-e
Evaluate the expression before point, in the context outside of Edebug
(@code{edebug-eval-last-sexp}).
@end table

@cindex lexical binding (Edebug)
Edebug supports evaluation of expressions containing references to
lexically bound symbols created by the following constructs in
@file{cl.el} (version 2.03 or later): @code{lexical-let},
@code{macrolet}, and @code{symbol-macrolet}.


@node Eval List
@subsection Evaluation List Buffer

You can use the @dfn{evaluation list buffer}, called @samp{*edebug*}, to
evaluate expressions interactively.  You can also set up the
@dfn{evaluation list} of expressions to be evaluated automatically each
time Edebug updates the display.

@table @kbd
@item E
Switch to the evaluation list buffer @samp{*edebug*}
(@code{edebug-visit-eval-list}).
@end table

In the @samp{*edebug*} buffer you can use the commands of Lisp
Interaction as well as these special commands:

@table @kbd
@item LFD
Evaluate the expression before point, in the outside context, and insert
the value in the buffer (@code{edebug-eval-print-last-sexp}).

@item C-x C-e
Evaluate the expression before point, in the context outside of Edebug
(@code{edebug-eval-last-sexp}).

@item C-c C-u
Build a new evaluation list from the first expression of each group,
reevaluate and redisplay (@code{edebug-update-eval-list}).  Groups are
separated by comment lines.

@item C-c C-d
Delete the evaluation list group that point is in
(@code{edebug-delete-eval-item}).

@item C-c C-w
Switch back to the source code buffer at the current stop point
(@code{edebug-where}).
@end table

You can evaluate expressions in the evaluation list window with
@kbd{LFD} or @kbd{C-x C-e}, just as you would in @samp{*scratch*};
but they are evaluated in the context outside of Edebug.

@cindex evaluation list (Edebug)
The expressions you enter interactively (and their results) are lost
when you continue execution unless you add them to the
evaluation list with @kbd{C-c C-u}.  This command builds a new list from
the first expression of each @dfn{evaluation list group}.  Groups are
separated by comment lines.  Be careful not to add expressions that
execute instrumented code otherwise an infinite loop will result.

When the evaluation list is redisplayed, each expression is displayed
followed by the result of evaluating it, and a comment line.  If an
error occurs during an evaluation, the error message is displayed in a
string as if it were the result.  Therefore expressions that, for
example, use variables not currently valid do not interrupt your
debugging.

Here is an example of what the evaluation list window looks like after
several expressions have been added to it:

@smallexample
(current-buffer)
#<buffer *scratch*>
;---------------------------------------------------------------
(selected-window)
#<window 16 on *scratch*>
;---------------------------------------------------------------
(point)
196
;---------------------------------------------------------------
bad-var
"Symbol's value as variable is void: bad-var"
;---------------------------------------------------------------
(recursion-depth)
0
;---------------------------------------------------------------
this-command
eval-last-sexp
;---------------------------------------------------------------
@end smallexample

To delete a group, move point into it and type @kbd{C-c C-d}, or simply
delete the text for the group and update the evaluation list with
@kbd{C-c C-u}.  When you add a new group, be sure it is separated from
its neighbors by a comment line.

After selecting @samp{*edebug*}, you can return to the source code
buffer with @kbd{C-c C-w}.  The @samp{*edebug*} buffer is killed when
you continue execution, and recreated next time it is needed.


@node Reading in Edebug
@subsection Reading in Edebug

@cindex reading (Edebug)
To instrument a form, Edebug first reads the whole form.  Edebug
replaces the standard Lisp Reader with its own reader that remembers the
positions of expressions.  This reader is used by the Edebug
replacements for @code{eval-region}, @code{eval-defun},
@code{eval-buffer}, and @code{eval-current-buffer}.

@pindex cl-read
Another package, @file{cl-read.el}, replaces the standard reader with
one that understands Common Lisp reader macros.  If you use that
package, Edebug will automatically load @file{edebug-cl-read.el} to
provide corresponding reader macros that remember positions of
expressions.  If you define new reader macros, you will have to define
similar reader macros for Edebug.


@node Printing in Edebug
@subsection Printing in Edebug

@cindex printing (Edebug)
@cindex printing circular structures
@pindex cust-print
If the result of an expression in your program contains a circular
reference, you may get an error when Edebug attempts to print it.  You
can set @code{print-length} to a non-zero value to limit the print
length of lists (the number of cdrs), and in Emacs 19, set
@code{print-level} to a non-zero value to limit the print depth of
lists.  But you can print such circular structures and structures that
share elements more informatively by using the @file{cust-print}
package.

To load @file{cust-print} and activate custom printing only for Edebug,
simply use the command @kbd{M-x edebug-install-custom-print}.  To
restore the standard print functions, use @kbd{M-x
edebug-uninstall-custom-print}.  You can also activate custom printing
for printing in any Lisp code; see the package for details.

Here is an example of code that creates a circular structure:

@example
(progn
  (edebug-install-custom-print)
  (setq a '(x y))
  (setcar a a))
@end example

Edebug will print the result of the @code{setcar} as @samp{Result:
#1=(#1# y)}.  The @samp{#1=} notation names the structure that follows
it, and the @samp{#1#} notation references the previously named
structure.  This notation is used for any shared elements of lists or
vectors.

@vindex edebug-print-length
@vindex edebug-print-level
@vindex edebug-print-circle
@vindex print-readably
Independent of whether @file{cust-print} is active, while printing
results Edebug binds @code{print-length}, @code{print-level}, and
@code{print-circle} to @code{edebug-print-length} (@code{50}),
@code{edebug-print-level} (@code{50}), and @code{edebug-print-circle}
(@code{t}) respectively, if these values are non-@code{nil}.  Also,
@code{print-readably} is bound to @code{nil} since some objects simply
cannot be printed readably.


@node Tracing
@subsection Tracing

@cindex tracing
In addition to automatic stepping through source code, which is also
called @emph{tracing} (see @ref{Edebug Execution Modes}), Edebug can
produce a traditional trace listing of execution in a separate buffer,
@samp{*edebug-trace*}.

@findex edebug-print-trace-before
@findex edebug-print-trace-after
If the variable @code{edebug-trace} is non-@code{nil}, each function entry and
exit adds lines to the trace buffer.  On function entry, Edebug prints
@samp{::::@{} followed by the function name and argument values.  On
function exit, Edebug prints @samp{::::@}} followed by the function name
and result of the function.  The number of @samp{:}s is computed from
the recursion depth.  The balanced braces in the trace buffer can be
used to find the matching beginning or end of function calls. These
displays may be customized by replacing the functions
@code{edebug-print-trace-before} and @code{edebug-print-trace-after},
which take an arbitrary message string to print.

@findex edebug-tracing
The macro @code{edebug-tracing} provides tracing similar to function
enter and exit tracing, but for arbitrary expressions.  This macro
should be explicitly inserted by you around expressions you wish to
trace the execution of.  The first argument is a message string
(evaluated), and the rest are expressions to evaluate.  The result of
the last expression is returned.

@findex edebug-trace
Finally, you can insert arbitrary strings into the trace buffer with
explicit calls to @code{edebug-trace}.  The arguments of this function
are the same as for @code{message}, but a newline is always inserted
after each string printed in this way.

@code{edebug-tracing} and @code{edebug-trace} insert lines in the trace
buffer even if Edebug is not active.  Every time the trace buffer is
added to, the window is scrolled to show the last lines inserted.
(There may be some display problems if you use tracing along with the
evaluation list.)


@node Coverage Testing
@subsection Coverage Testing

@cindex coverage testing
@cindex frequency counts
@cindex performance analysis
Edebug provides a rudimentary coverage tester and display of execution
frequency.  Frequency counts are always accumulated, both before and
after evaluation of each instrumented expression, even if the execution
mode is @code{Go-nonstop}.  Coverage testing is only done if the option
@code{edebug-test-coverage} is non-@code{nil} because this is relatively
expensive.  Both data sets are displayed by @kbd{M-x
edebug-display-freq-count}.

@deffn Command edebug-display-freq-count
Display the frequency count data for each line of the current
definition.  The frequency counts are inserted as comment lines after
each line, and you can undo all insertions with one @code{undo} command.
The counts are inserted starting under the @kbd{(} before an expression
or the @kbd{)} after an expression, or on the last char of a symbol.
The counts are only displayed when they differ from previous counts on
the same line.

If coverage is being tested, whenever all known results of an expression
are @code{eq}, the char @kbd{=} will be appended after the count
for that expression.  Note that this is always the case for an
expression only evaluated once.

To clear the frequency count and coverage data for a definition,
reinstrument it.

@end deffn

For example, after evaluating @code{(fac 5)} with an embedded
breakpoint, and setting @code{edebug-test-coverage} to @code{t}, when
the breakpoint is reached, the frequency data is looks like this:

@example
(defun fac (n)
  (if (= n 0) (edebug))
;#6           1      0 =5
  (if (< 0 n)
;#5         =
      (* n (fac (1- n)))
;#    5               0
    1))
;#   0
@end example

The comment lines show that @code{fac} has been called 6 times.  The
first @code{if} statement has returned 5 times with the same result each
time, and the same is true for the condition on the second @code{if}.
The recursive call of @code{fac} has not returned at all.


@node The Outside Context
@subsection The Outside Context

Edebug tries to be transparent to the program you are debugging.  In
addition, most evaluations you do within Edebug (see @ref{Edebug Eval})
occur in the same outside context which is temporarily restored for the
evaluation.  But Edebug is not completely successful and this section
explains precisely how it fails.  Edebug operation unavoidably alters
some data in XEmacs, and this can interfere with debugging certain
programs.  Also notice that Edebug's protection against change of
outside data means that any side effects @emph{intended} by the user in
the course of debugging will be defeated.

@menu
* Checking Whether to Stop::    When Edebug decides what to do.
* Edebug Display Update::       When Edebug updates the display.
* Edebug Recursive Edit::       When Edebug stops execution.
@end menu


@node Checking Whether to Stop
@subsubsection Checking Whether to Stop

Whenever Edebug is entered just to think about whether to take some
action, it needs to save and restore certain data.

@itemize @bullet
@item
@code{max-lisp-eval-depth} and @code{max-specpdl-size} are both
incremented one time to reduce Edebug's impact on the stack.
You could, however, still run out of stack space when using Edebug.

@item
The state of keyboard macro execution is saved and restored.  While
Edebug is active, @code{executing-macro} is bound to
@code{edebug-continue-kbd-macro}.

@end itemize


@node Edebug Display Update
@subsubsection Edebug Display Update

When Edebug needs to display something (e.g., in trace mode), it saves
the current window configuration from ``outside'' Edebug.  When you exit
Edebug (by continuing the program), it restores the previous window
configuration.

XEmacs redisplays only when it pauses.  Usually, when you continue
execution, the program comes back into Edebug at a breakpoint or after
stepping without pausing or reading input in between.  In such cases,
XEmacs never gets a chance to redisplay the ``outside'' configuration.
What you see is the same window configuration as the last time Edebug
was active, with no interruption.

Entry to Edebug for displaying something also saves and restores the
following data, but some of these are deliberately not restored if an
error or quit signal occurs.

@itemize @bullet
@item
@cindex current buffer point and mark (Edebug)
Which buffer is current, and where point and mark are in the current
buffer are saved and restored.

@item
@cindex window configuration (Edebug)
@findex save-excursion (Edebug)
@vindex edebug-save-windows
The Edebug Display Update, is saved and restored if
@code{edebug-save-windows} is non-@code{nil}.  It is not restored on
error or quit, but the outside selected window @emph{is} reselected even
on error or quit in case a @code{save-excursion} is active.
If the value of @code{edebug-save-windows} is a list, only the listed
windows are saved and restored.

The window start and horizontal scrolling of the source code buffer are
not restored, however, so that the display remains coherent.

@item
@vindex edebug-save-displayed-buffer-points
The value of point in each displayed buffer is saved and restored if
@code{edebug-save-displayed-buffer-points} is non-@code{nil}.

@item
The variables @code{overlay-arrow-position} and
@code{overlay-arrow-string} are saved and restored.  So you can safely
invoke Edebug from the recursive edit elsewhere in the same buffer.

@item
@code{cursor-in-echo-area} is locally bound to @code{nil} so that
the cursor shows up in the window.

@end itemize


@node Edebug Recursive Edit
@subsubsection Edebug Recursive Edit

When Edebug is entered and actually reads commands from the user, it
saves (and later restores) these additional data:

@itemize @bullet
@item
The current match data, for whichever buffer was current.

@item
@code{last-command}, @code{this-command}, @code{last-command-char},
@code{last-input-char}, @code{last-input-event},
@code{last-command-event},
@code{last-event-frame}, @code{last-nonmenu-event}, and
@code{track-mouse} .  Commands used within Edebug do not affect these
variables outside of Edebug.

The key sequence returned by @code{this-command-keys} is changed by
executing commands within Edebug and there is no way to reset
the key sequence from Lisp.

For Emacs 18, Edebug cannot save and restore the value of
@code{unread-command-char}.  Entering Edebug while this variable has
a nontrivial value can interfere with execution of the program you are
debugging.

@item
Complex commands executed while in Edebug are added to the variable
@code{command-history}.  In rare cases this can alter execution.

@item
Within Edebug, the recursion depth appears one deeper than the recursion
depth outside Edebug.  This is not true of the automatically updated
evaluation list window.

@item
@code{standard-output} and @code{standard-input} are bound to @code{nil}
by the @code{recursive-edit}, but Edebug temporarily restores them during
evaluations.

@item
The state of keyboard macro definition is saved and restored.  While
Edebug is active, @code{defining-kbd-macro} is bound to
@code{edebug-continue-kbd-macro}.

@end itemize


@node Instrumenting Macro Calls
@subsection Instrumenting Macro Calls

When Edebug instruments an expression that calls a Lisp macro, it needs
additional advice to do the job properly.  This is because there is no
way to tell which subexpressions of the macro call may be evaluated.
(Evaluation may occur explicitly in the macro body, or when the
resulting expansion is evaluated, or any time later.)  You must explain
the format of macro call arguments by using @code{def-edebug-spec} to
define an @dfn{Edebug specification} for each macro.

@defmac def-edebug-spec macro specification
Specify which expressions of a call to macro @var{macro} are forms to be
evaluated.  For simple macros, the @var{specification} often looks very
similar to the formal argument list of the macro definition, but
specifications are much more general than macro arguments.

The @var{macro} argument may actually be any symbol, not just a macro
name.

Unless you are using Emacs 19 or XEmacs, this macro is only defined
in Edebug, so you may want to use the following which is equivalent:
@code{(put '@var{macro} 'edebug-form-spec '@var{specification})}
@end defmac

Here is a simple example that defines the specification for the
@code{for} macro described in the XEmacs Lisp Reference Manual, followed
by an alternative, equivalent specification.

@example
(def-edebug-spec for
  (symbolp "from" form "to" form "do" &rest form))

(def-edebug-spec for
  (symbolp ['from form] ['to form] ['do body]))
@end example

Here is a table of the possibilities for @var{specification} and how each
directs processing of arguments.

@table @bullet

@item @code{t}
All arguments are instrumented for evaluation.

@item @code{0}
None of the arguments is instrumented.

@item a symbol
The symbol must have an Edebug specification which is used instead.
This indirection is repeated until another kind of specification is
found.  This allows you to inherit the specification for another macro.

@item a list
The elements of the list describe the types of the arguments of a
calling form.  The possible elements of a specification list are
described in the following sections.
@end table

@menu
* Specification List::          How to specify complex patterns of evaluation.
* Backtracking::                What Edebug does when matching fails.
* Debugging Backquote:: Debugging Backquote
* Specification Examples::      To help understand specifications.
@end menu


@node Specification List
@subsubsection Specification List

@cindex Edebug specification list
A @dfn{specification list} is required for an Edebug specification if
some arguments of a macro call are evaluated while others are not.  Some
elements in a specification list match one or more arguments, but others
modify the processing of all following elements.  The latter, called
@dfn{keyword specifications}, are symbols beginning with @samp{@code{&}}
(e.g.  @code{&optional}).

A specification list may contain sublists which match arguments that are
themselves lists, or it may contain vectors used for grouping.  Sublists
and groups thus subdivide the specification list into a hierarchy of
levels.  Keyword specifications only apply to the remainder of the
sublist or group they are contained in and there is an implicit grouping
around a keyword specification and all following elements in the
sublist or group.

If a specification list fails
at some level, then backtracking may be invoked to find some alternative
at a higher level, or if no alternatives remain, an error will be
signaled.  See @ref{Backtracking} for more details.

Edebug specifications provide at least the power of regular expression
matching.  Some context-free constructs are also supported: the matching
of sublists with balanced parentheses, recursive processing of forms,
and recursion via indirect specifications.

Each element of a specification list may be one of the following, with
the corresponding type of argument:

@table @code

@item sexp
A single unevaluated expression.

@item form
A single evaluated expression, which is instrumented.

@item place
@findex edebug-unwrap
A place as in the Common Lisp @code{setf} place argument.  It will be
instrumented just like a form, but the macro is expected to strip the
instrumentation.  Two functions, @code{edebug-unwrap} and
@code{edebug-unwrap*}, are provided to strip the instrumentation one
level or recursively at all levels.

@item body
Short for @code{&rest form}.  See @code{&rest} below.

@item function-form
A function form: either a quoted function symbol, a quoted lambda expression,
or a form (that should evaluate to a function symbol or lambda
expression).  This is useful when function arguments might be quoted
with @code{quote} rather than @code{function} since the body of a lambda
expression will be instrumented either way.

@item lambda-expr
An unquoted anonymous lambda expression.

@item &optional
@cindex &optional (Edebug)
All following elements in the specification list are optional; as soon
as one does not match, Edebug stops matching at this level.

To make just a few elements optional followed by non-optional elements,
use @code{[&optional @var{specs}@dots{}]}.  To specify that several
elements should all succeed together, use @code{&optional
[@var{specs}@dots{}]}.  See the @code{defun} example below.

@item &rest
@cindex &rest (Edebug)
All following elements in the specification list are repeated zero or
more times.  All the elements need not match in the last repetition,
however.

To repeat only a few elements, use @code{[&rest @var{specs}@dots{}]}.
To specify all elements must match on every repetition, use @code{&rest
[@var{specs}@dots{}]}.

@item &or
@cindex &or (Edebug)
Each of the following elements in the specification list is an
alternative, processed left to right until one matches.  One of the
alternatives must match otherwise the @code{&or} specification fails.

Each list element following @code{&or} is a single alternative even if
it is a keyword specification. (This breaks the implicit grouping rule.)
To group two or more list elements as a single alternative, enclose them
in @code{[@dots{}]}.

@item &not
@cindex &not (Edebug)
Each of the following elements is matched as alternatives as if by using
@code{&or}, but if any of them match, the specification fails.  If none
of them match, nothing is matched, but the @code{&not} specification
succeeds.

@item &define
@cindex &define (Edebug)
Indicates that the specification is for a defining form.  The defining
form itself is not instrumented (i.e. Edebug does not stop before and
after the defining form), but forms inside it typically will be
instrumented.  The @code{&define} keyword should be the first element in
a list specification.

Additional specifications that may only appear after @code{&define} are
described here.  See the @code{defun} example below.

@table @code

@item name
The argument, a symbol, is the name of the defining form.
But a defining form need not be named at all, in which
case a unique name will be created for it.

The @code{name} specification may be used more than once in the
specification and each subsequent use will append the corresponding
symbol argument to the previous name with @samp{@code{@@}} between them.
This is useful for generating unique but meaningful names for
definitions such as @code{defadvice} and @code{defmethod}.

@item :name
The element following @code{:name} should be a symbol; it is used as an
additional name component for the definition.  This is useful to add a
unique, static component to the name of the definition.  It may be used
more than once.  No argument is matched.

@item arg
The argument, a symbol, is the name of an argument of the defining form.
However, lambda list keywords (symbols starting with @samp{@code{&}})
are not allowed.  See @code{lambda-list} and the example below.

@item lambda-list
@cindex lambda-list (Edebug)
This matches the whole argument list of an XEmacs Lisp lambda
expression, which is a list of symbols and the keywords
@code{&optional} and @code{&rest}

@item def-body
The argument is the body of code in a definition.  This is like
@code{body}, described above, but a definition body must be instrumented
with a different Edebug call that looks up information associated with
the definition.  Use @code{def-body} for the highest level list of forms
within the definition.

@item def-form
The argument is a single, highest-level form in a definition.  This is
like @code{def-body}, except use this to match a single form rather than
a list of forms.  As a special case, @code{def-form} also means that
tracing information is not output when the form is executed.  See the
@code{interactive} example below.

@end table

@item nil
This is successful when there are no more arguments to match at the
current argument list level; otherwise it fails.  See sublist
specifications and the backquote example below.

@item gate
@cindex preventing backtracking
No argument is matched but backtracking through the gate is disabled
while matching the remainder of the specifications at this level.  This
is primarily used to generate more specific syntax error messages.  See
@ref{Backtracking} for more details.  Also see the @code{let} example
below.

@item @var{other-symbol}
@cindex indirect specifications
Any other symbol in a specification list may be a predicate or an
indirect specification.

If the symbol has an Edebug specification, this @dfn{indirect
specification} should be either a list specification that is used in
place of the symbol, or a function that is called to process the
arguments.  The specification may be defined with @code{def-edebug-spec}
just as for macros. See the @code{defun} example below.

Otherwise, the symbol should be a predicate.  The predicate is called
with the argument and the specification fails if the predicate fails.
The argument is not instrumented.

@findex keywordp
@findex lambda-list-keywordp
Predicates that may be used include: @code{symbolp}, @code{integerp},
@code{stringp}, @code{vectorp}, @code{atom} (which matches a number,
string, symbol, or vector), @code{keywordp}, and
@code{lambda-list-keywordp}.  The last two, defined in @file{edebug.el},
test whether the argument is a symbol starting with @samp{@code{:}} and
@samp{@code{&}} respectively.

@item [@var{elements}@dots{}]
@cindex [@dots{}] (Edebug)
Rather than matching a vector argument, a vector treats
the @var{elements} as a single @dfn{group specification}.

@item "@var{string}"
The argument should be a symbol named @var{string}.  This specification
is equivalent to the quoted symbol, @code{'@var{symbol}}, where the name
of @var{symbol} is the @var{string}, but the string form is preferred.

@item  '@var{symbol} @r{or} (quote @var{symbol})
The argument should be the symbol @var{symbol}.  But use a string
specification instead.

@item (vector @var{elements}@dots{})
The argument should be a vector whose elements must match the
@var{elements} in the specification.  See the backquote example below.

@item (@var{elements}@dots{})
Any other list is a @dfn{sublist specification} and the argument must be
a list whose elements match the specification @var{elements}.

@cindex dotted lists (Edebug)
A sublist specification may be a dotted list and the corresponding list
argument may then be a dotted list.  Alternatively, the last cdr of a
dotted list specification may be another sublist specification (via a
grouping or an indirect specification, e.g. @code{(spec .  [(more
specs@dots{})])}) whose elements match the non-dotted list arguments.
This is useful in recursive specifications such as in the backquote
example below.  Also see the description of a @code{nil} specification
above for terminating such recursion.

Note that a sublist specification of the form @code{(specs .  nil)}
means the same as @code{(specs)}, and @code{(specs .
(sublist-elements@dots{}))} means the same as @code{(specs
sublist-elements@dots{})}.

@end table

@c Need to document extensions with &symbol and :symbol

@node Backtracking
@subsubsection Backtracking

@cindex backtracking
@cindex syntax error (Edebug)
If a specification fails to match at some point, this does not
necessarily mean a syntax error will be signaled; instead,
@dfn{backtracking} will take place until all alternatives have been
exhausted.  Eventually every element of the argument list must be
matched by some element in the specification, and every required element
in the specification must match some argument.

Backtracking is disabled for the remainder of a sublist or group when
certain conditions occur, described below.  Backtracking is reenabled
when a new alternative is established by @code{&optional}, @code{&rest},
or @code{&or}.  It is also reenabled initially when processing a
sublist or group specification or an indirect specification.

You might want to disable backtracking to commit to some alternative so
that Edebug can provide a more specific syntax error message.  Normally,
if no alternative matches, Edebug reports that none matched, but if one
alternative is committed to, Edebug can report how it failed to match.

First, backtracking is disabled while matching any of the form
specifications (i.e. @code{form}, @code{body}, @code{def-form}, and
@code{def-body}).  These specifications will match any form so any error
must be in the form itself rather than at a higher level.

Second, backtracking is disabled after successfully matching a quoted
symbol or string specification, since this usually indicates a
recognized construct.  If you have a set of alternative constructs that
all begin with the same symbol, you can usually work around this
constraint by factoring the symbol out of the alternatives, e.g.,
@code{["foo" &or [first case] [second case] ...]}.

Third, backtracking may be explicitly disabled by using the
@code{gate} specification.  This is useful when you know that
no higher alternatives may apply.


@node Debugging Backquote
@subsubsection Debugging Backquote

@findex ` (Edebug)
@cindex backquote (Edebug)
Backquote (@kbd{`}) is a macro that results in an expression that may or
may not be evaluated.  It is often used to simplify the definition of a
macro to return an expression that is evaluated, but Edebug does not know
when this is the case.  However, the forms inside unquotes (@code{,} and
@code{,@@}) are evaluated and Edebug instruments them.

Nested backquotes are supported by Edebug, but there is a limit on the
support of quotes inside of backquotes.  Quoted forms (with @code{'})
are not normally evaluated, but if the quoted form appears immediately
within @code{,} and @code{,@@} forms, Edebug treats this as a backquoted
form at the next higher level (even if there is not a next higher level
- this is difficult to fix).

@findex edebug-`
If the backquoted forms happen to be code intended to be evaluated, you
can have Edebug instrument them by using @code{edebug-`} instead of the
regular @code{`}.  Unquoted forms can always appear inside
@code{edebug-`} anywhere a form is normally allowed.  But @code{(,
@var{form})} may be used in two other places specially recognized by
Edebug: wherever a predicate specification would match, and at the head
of a list form in place of a function name or lambda expression.  The
@var{form} inside a spliced unquote, @code{(,@@ @var{form})}, will be
wrapped, but the unquote form itself will not be wrapped since this
would interfere with the splicing.

There is one other complication with using @code{edebug-`}.  If the
@code{edebug-`} call is in a macro and the macro may be called from code
that is also instrumented, and if unquoted forms contain any macro
arguments bound to instrumented forms, then you should modify the
specification for the macro as follows: the specifications for those
arguments must use @code{def-form} instead of @code{form}.  (This is to
reestablish the Edebugging context for those external forms.)

For example, the @code{for} macro
@c (@pxref{Problems with Macros}) @c in XEmacs Lisp Reference Manual
(@pxref{Problems with Macros,,,, XEmacs Lisp Reference Manual}) @c Edebug Doc
is shown here but with @code{edebug-`}
substituted for regular @code{`}.

@example
(defmacro inc (var)
  (list 'setq var (list '1+ var)))

(defmacro for (var from init to final do &rest body)
  (let ((tempvar (make-symbol "max")))
    (edebug-` (let (((, var) (, init))
                    ((, tempvar) (, final)))
                (while (<= (, var) (, tempvar))
                  (,@ body)
                  (inc (, var)))))))
@end example

Here is the corresponding modified Edebug specification and some code
that calls the macro:

@example
(def-edebug-spec for
  (symbolp "from" def-form "to" def-form "do" &rest def-form))

(let ((n 5))
  (for i from n to (* n (+ n 1)) do
    (message "%s" i)))
@end example

After instrumenting the @code{for} macro and the macro call, Edebug
first steps to the beginning of the macro call, then into the macro
body, then through each of the unquoted expressions in the backquote
showing the expressions that will be embedded in the backquote form.
Then when the macro expansion is evaluated, Edebug will step through the
@code{let} form and each time it gets to an unquoted form, it will jump
back to an argument of the macro call to step through that expression.
Finally stepping will continue after the macro call.  Even more
convoluted execution paths may result when using anonymous functions.

@vindex edebug-unwrap-results
When the result of an expression is an instrumented expression, it is
difficult to see the expression inside the instrumentation.  So
you may want to set the option @code{edebug-unwrap-results} to a
non-@code{nil} value while debugging such expressions, but it would slow
Edebug down to always do this.


@node Specification Examples
@subsubsection Specification Examples

Here we provide several examples of Edebug specifications to show
many of its capabilities.

A @code{let} special form has a sequence of bindings and a body.  Each
of the bindings is either a symbol or a sublist with a symbol and
optional value.  In the specification below, notice the @code{gate}
inside of the sublist to prevent backtracking.

@example
(def-edebug-spec let
  ((&rest
    &or symbolp (gate symbolp &optional form))
   body))
@end example

Edebug uses the following specifications for @code{defun} and
@code{defmacro} and the associated argument list and @code{interactive}
specifications.  It is necessary to handle the expression argument of an
interactive form specially since it is actually evaluated outside of the
function body.

@example
(def-edebug-spec defmacro defun)      ; @r{Indirect ref to @code{defun} spec}
(def-edebug-spec defun
  (&define name lambda-list
           [&optional stringp]        ; @r{Match the doc string, if present.}
           [&optional ("interactive" interactive)]
           def-body))

(def-edebug-spec lambda-list
  (([&rest arg]
    [&optional ["&optional" arg &rest arg]]
    &optional ["&rest" arg]
    )))

(def-edebug-spec interactive
  (&optional &or stringp def-form))    ; @r{Notice: @code{def-form}}
@end example

The specification for backquote below illustrates how to match
dotted lists and use @code{nil} to terminate recursion.  It also
illustrates how components of a vector may be matched.  (The actual
specification provided by Edebug does not support dotted lists because
doing so causes very deep recursion that could fail.)

@example
(def-edebug-spec ` (backquote-form))  ;; alias just for clarity

(def-edebug-spec backquote-form
  (&or ([&or "," ",@@"] &or ("quote" backquote-form) form)
       (backquote-form . [&or nil backquote-form])
       (vector &rest backquote-form)
       sexp))
@end example


@node Edebug Options
@subsection Edebug Options

These options affect the behavior of Edebug:

@defopt edebug-setup-hook
Functions to call before Edebug is used.  Each time it is set to a new
value, Edebug will call those functions once and then
@code{edebug-setup-hook} is reset to @code{nil}.  You could use this to
load up Edebug specifications associated with a package you are using
but only when you also use Edebug.
See @ref{Instrumenting}.
@end defopt

@defopt edebug-all-defs
If non-@code{nil}, normal evaluation of any defining forms (e.g.
@code{defun} and @code{defmacro}) will instrument them for Edebug.  This
applies to @code{eval-defun}, @code{eval-region}, and
@code{eval-current-buffer}.

Use the command @kbd{M-x edebug-all-defs} to toggle the value of
this variable. You may want to make this variable local to each
buffer by calling @code{(make-local-variable 'edebug-all-defs)} in your
@code{emacs-lisp-mode-hook}.
See @ref{Instrumenting}.
@end defopt

@defopt edebug-all-forms
If non-@code{nil}, normal evaluation of any forms by @code{eval-defun},
@code{eval-region}, and @code{eval-current-buffer} will instrument them
for Edebug.

Use the command @kbd{M-x edebug-all-forms} to toggle the value of this
option.
See @ref{Instrumenting}.
@end defopt

@defopt edebug-save-windows
If non-@code{nil}, save and restore window configuration on Edebug
calls.  It takes some time to do this, so if your program does not care
what happens to data about windows, you may want to set this variable to
@code{nil}.

If the value is a list, only the listed windows are saved and
restored.

@kbd{M-x edebug-toggle-save-windows} may be used to change this variable.
This command is bound to @kbd{W} in source code buffers.
See @ref{Edebug Display Update}.
@end defopt

@defopt edebug-save-displayed-buffer-points
If non-@code{nil}, save and restore point in all displayed buffers.
This is necessary if you are debugging code that changes the point of a
buffer which is displayed in a non-selected window.  If Edebug or the
user then selects the window, the buffer's point will be changed to the
window's point.

This is an expensive operation since it visits each window and therefore
each displayed buffer twice for each Edebug activation, so it is best to
avoid it if you can.
See @ref{Edebug Display Update}.
@end defopt


@defopt edebug-initial-mode
If this variable is non-@code{nil}, it specifies the initial execution
mode for Edebug when it is first activated.  Possible values are
@code{step}, @code{next}, @code{go}, @code{Go-nonstop}, @code{trace},
@code{Trace-fast}, @code{continue}, and @code{Continue-fast}.

The default value is @code{step}.
See @ref{Edebug Execution Modes}.
@end defopt

@defopt edebug-trace
@findex edebug-print-trace-before
@findex edebug-print-trace-after
Non-@code{nil} means display a trace of function entry and exit.
Tracing output is displayed in a buffer named @samp{*edebug-trace*}, one
function entry or exit per line, indented by the recursion level.

The default value is @code{nil}.

Also see @code{edebug-tracing}.
See @ref{Tracing}.
@end defopt

@defopt edebug-test-coverage
If non-@code{nil}, Edebug tests coverage of all expressions debugged.
This is done by comparing the result of each expression
with the previous result. Coverage is considered OK if two different
results are found.  So to sufficiently test the coverage of your code,
try to execute it under conditions that evaluate all expressions more
than once, and produce different results for each expression.

Use @kbd{M-x edebug-display-freq-count} to display the frequency count
and coverage information for a definition.
See @ref{Coverage Testing}.
@end defopt

@defopt edebug-continue-kbd-macro
If non-@code{nil}, continue defining or executing any keyboard macro
that is executing outside of Edebug.   Use this with caution since it is not
debugged.
See @ref{Edebug Execution Modes}.
@end defopt

@defopt edebug-print-length
  If non-@code{nil}, bind @code{print-length} to this while printing
results in Edebug.  The default value is @code{50}.
See @ref{Printing in Edebug}.
@end defopt

@defopt edebug-print-level
  If non-@code{nil}, bind @code{print-level} to this while printing
results in Edebug.  The default value is @code{50}.
@end defopt

@defopt edebug-print-circle
  If non-@code{nil}, bind @code{print-circle} to this while printing
results in Edebug.  The default value is @code{nil}.
@end defopt

@defopt edebug-on-error
  @code{debug-on-error} is bound to this while Edebug is active.
See @ref{Trapping Errors}.
@end defopt

@defopt edebug-on-quit
  @code{debug-on-quit} is bound to this while Edebug is active.
See @ref{Trapping Errors}.
@end defopt

@defopt edebug-unwrap-results
  Non-@code{nil} if Edebug should unwrap results of expressions.
This is useful when debugging macros where the results of expressions
are instrumented expressions.  But don't do this when results might be
circular or an infinite loop will result.
See @ref{Debugging Backquote}.
@end defopt

@defopt edebug-global-break-condition
  If non-@code{nil}, an expression to test for at every stop point.
If the result is non-@code{nil}, then break.  Errors are ignored.
See @ref{Global Break Condition}.
@end defopt