emacs/lispref/positions.texi

@c -*-texinfo-*-
@c This is part of the GNU Emacs Lisp Reference Manual.
@c Copyright (C) 1990, 1991, 1992, 1993, 1994, 1995, 1998 Free Software Foundation, Inc. 
@c See the file elisp.texi for copying conditions.
@setfilename ../info/positions
@node Positions, Markers, Frames, Top
@chapter Positions
@cindex position (in buffer)

  A @dfn{position} is the index of a character in the text of a buffer.
More precisely, a position identifies the place between two characters
(or before the first character, or after the last character), so we can
speak of the character before or after a given position.  However, we
often speak of the character ``at'' a position, meaning the character
after that position.

  Positions are usually represented as integers starting from 1, but can
also be represented as @dfn{markers}---special objects that relocate
automatically when text is inserted or deleted so they stay with the
surrounding characters.  @xref{Markers}.

@menu
* Point::         The special position where editing takes place.
* Motion::        Changing point.
* Excursions::    Temporary motion and buffer changes.
* Narrowing::     Restricting editing to a portion of the buffer.
@end menu

@node Point
@section Point
@cindex point

  @dfn{Point} is a special buffer position used by many editing
commands, including the self-inserting typed characters and text
insertion functions.  Other commands move point through the text
to allow editing and insertion at different places.

  Like other positions, point designates a place between two characters
(or before the first character, or after the last character), rather
than a particular character.  Usually terminals display the cursor over
the character that immediately follows point; point is actually before
the character on which the cursor sits.

@cindex point with narrowing
  The value of point is a number between 1 and the buffer size plus 1.
If narrowing is in effect (@pxref{Narrowing}), then point is constrained
to fall within the accessible portion of the buffer (possibly at one end
of it).

  Each buffer has its own value of point, which is independent of the
value of point in other buffers.  Each window also has a value of point,
which is independent of the value of point in other windows on the same
buffer.  This is why point can have different values in various windows
that display the same buffer.  When a buffer appears in only one window,
the buffer's point and the window's point normally have the same value,
so the distinction is rarely important.  @xref{Window Point}, for more
details.

@defun point
@cindex current buffer position
This function returns the value of point in the current buffer,
as an integer.

@need 700
@example
@group
(point)
     @result{} 175
@end group
@end example
@end defun

@defun point-min
This function returns the minimum accessible value of point in the
current buffer.  This is normally 1, but if narrowing is in effect, it
is the position of the start of the region that you narrowed to.
(@xref{Narrowing}.)
@end defun

@defun point-max
This function returns the maximum accessible value of point in the
current buffer.  This is @code{(1+ (buffer-size))}, unless narrowing is
in effect, in which case it is the position of the end of the region
that you narrowed to.  (@xref{Narrowing}).
@end defun

@defun buffer-end flag
This function returns @code{(point-min)} if @var{flag} is less than 1,
@code{(point-max)} otherwise.  The argument @var{flag} must be a number.
@end defun

@defun buffer-size
This function returns the total number of characters in the current
buffer.  In the absence of any narrowing (@pxref{Narrowing}),
@code{point-max} returns a value one larger than this.

@example
@group
(buffer-size)
     @result{} 35
@end group
@group
(point-max)
     @result{} 36
@end group
@end example
@end defun

@node Motion
@section Motion

  Motion functions change the value of point, either relative to the
current value of point, relative to the beginning or end of the buffer,
or relative to the edges of the selected window.  @xref{Point}.

@menu
* Character Motion::       Moving in terms of characters.
* Word Motion::            Moving in terms of words.
* Buffer End Motion::      Moving to the beginning or end of the buffer.
* Text Lines::             Moving in terms of lines of text.
* Screen Lines::           Moving in terms of lines as displayed.
* List Motion::            Moving by parsing lists and sexps.
* Skipping Characters::    Skipping characters belonging to a certain set.
@end menu

@node Character Motion
@subsection Motion by Characters

  These functions move point based on a count of characters.
@code{goto-char} is the fundamental primitive; the other functions use
that.

@deffn Command goto-char position
This function sets point in the current buffer to the value
@var{position}.  If @var{position} is less than 1, it moves point to the
beginning of the buffer.  If @var{position} is greater than the length
of the buffer, it moves point to the end.

If narrowing is in effect, @var{position} still counts from the
beginning of the buffer, but point cannot go outside the accessible
portion.  If @var{position} is out of range, @code{goto-char} moves
point to the beginning or the end of the accessible portion.

When this function is called interactively, @var{position} is the
numeric prefix argument, if provided; otherwise it is read from the
minibuffer.

@code{goto-char} returns @var{position}.
@end deffn

@deffn Command forward-char &optional count
@c @kindex beginning-of-buffer
@c @kindex end-of-buffer
This function moves point @var{count} characters forward, towards the
end of the buffer (or backward, towards the beginning of the buffer, if
@var{count} is negative).  If the function attempts to move point past
the beginning or end of the buffer (or the limits of the accessible
portion, when narrowing is in effect), an error is signaled with error
code @code{beginning-of-buffer} or @code{end-of-buffer}.

In an interactive call, @var{count} is the numeric prefix argument.
@end deffn

@deffn Command backward-char &optional count
This function moves point @var{count} characters backward, towards the
beginning of the buffer (or forward, towards the end of the buffer, if
@var{count} is negative).  If the function attempts to move point past
the beginning or end of the buffer (or the limits of the accessible
portion, when narrowing is in effect), an error is signaled with error
code @code{beginning-of-buffer} or @code{end-of-buffer}.

In an interactive call, @var{count} is the numeric prefix argument.
@end deffn

@node Word Motion
@subsection Motion by Words

  These functions for parsing words use the syntax table to decide
whether a given character is part of a word.  @xref{Syntax Tables}.

@deffn Command forward-word count
This function moves point forward @var{count} words (or backward if
@var{count} is negative).  ``Moving one word'' means moving until point
crosses a word-constituent character and then encounters a
word-separator character (or the boundary of the accessible part of the
buffer).

If it is possible to move @var{count} words, without being stopped by
the buffer boundary (except perhaps after the last word), the value is
@code{t}.  Otherwise, the return value is @code{nil} and point stops
at the buffer boundary.

In an interactive call, @var{count} is set to the numeric prefix
argument.
@end deffn

@deffn Command backward-word count
This function is just like @code{forward-word}, except that it moves
backward until encountering the front of a word, rather than forward.

In an interactive call, @var{count} is set to the numeric prefix
argument.

This function is rarely used in programs, as it is more efficient to
call @code{forward-word} with a negative argument.
@end deffn

@defvar words-include-escapes
@c Emacs 19 feature
This variable affects the behavior of @code{forward-word} and everything
that uses it.  If it is non-@code{nil}, then characters in the
``escape'' and ``character quote'' syntax classes count as part of
words.  Otherwise, they do not.
@end defvar

@node Buffer End Motion
@subsection Motion to an End of the Buffer

  To move point to the beginning of the buffer, write:

@example
@group
(goto-char (point-min))
@end group
@end example

@noindent
Likewise, to move to the end of the buffer, use:

@example
@group
(goto-char (point-max))
@end group
@end example

  Here are two commands that users use to do these things.  They are
documented here to warn you not to use them in Lisp programs, because
they set the mark and display messages in the echo area.

@deffn Command beginning-of-buffer &optional n
This function moves point to the beginning of the buffer (or the limits
of the accessible portion, when narrowing is in effect), setting the
mark at the previous position.  If @var{n} is non-@code{nil}, then it
puts point @var{n} tenths of the way from the beginning of the buffer.

In an interactive call, @var{n} is the numeric prefix argument,
if provided; otherwise @var{n} defaults to @code{nil}.

@strong{Warning:} Don't use this function in Lisp programs!
@end deffn

@deffn Command end-of-buffer &optional n
This function moves point to the end of the buffer (or the limits of
the accessible portion, when narrowing is in effect), setting the mark
at the previous position.  If @var{n} is non-@code{nil}, then it puts
point @var{n} tenths of the way from the end of the buffer.

In an interactive call, @var{n} is the numeric prefix argument,
if provided; otherwise @var{n} defaults to @code{nil}.

@strong{Warning:} Don't use this function in Lisp programs!
@end deffn

@node Text Lines
@subsection Motion by Text Lines
@cindex lines

  Text lines are portions of the buffer delimited by newline characters,
which are regarded as part of the previous line.  The first text line
begins at the beginning of the buffer, and the last text line ends at
the end of the buffer whether or not the last character is a newline.
The division of the buffer into text lines is not affected by the width
of the window, by line continuation in display, or by how tabs and
control characters are displayed.

@deffn Command goto-line line
This function moves point to the front of the @var{line}th line,
counting from line 1 at beginning of the buffer.  If @var{line} is less
than 1, it moves point to the beginning of the buffer.  If @var{line} is
greater than the number of lines in the buffer, it moves point to the
end of the buffer---that is, the @emph{end of the last line} of the
buffer.  This is the only case in which @code{goto-line} does not
necessarily move to the beginning of a line.

If narrowing is in effect, then @var{line} still counts from the
beginning of the buffer, but point cannot go outside the accessible
portion.  So @code{goto-line} moves point to the beginning or end of the
accessible portion, if the line number specifies an inaccessible
position.

The return value of @code{goto-line} is the difference between
@var{line} and the line number of the line to which point actually was
able to move (in the full buffer, before taking account of narrowing).
Thus, the value is positive if the scan encounters the real end of the
buffer before finding the specified line.  The value is zero if scan
encounters the end of the accessible portion but not the real end of the
buffer.

In an interactive call, @var{line} is the numeric prefix argument if
one has been provided.  Otherwise @var{line} is read in the minibuffer.
@end deffn

@deffn Command beginning-of-line &optional count
This function moves point to the beginning of the current line.  With an
argument @var{count} not @code{nil} or 1, it moves forward
@var{count}@minus{}1 lines and then to the beginning of the line.

If this function reaches the end of the buffer (or of the accessible
portion, if narrowing is in effect), it positions point there.  No error
is signaled.
@end deffn

@deffn Command end-of-line &optional count
This function moves point to the end of the current line.  With an
argument @var{count} not @code{nil} or 1, it moves forward
@var{count}@minus{}1 lines and then to the end of the line.

If this function reaches the end of the buffer (or of the accessible
portion, if narrowing is in effect), it positions point there.  No error
is signaled.
@end deffn

@deffn Command forward-line &optional count
@cindex beginning of line
This function moves point forward @var{count} lines, to the beginning of
the line.  If @var{count} is negative, it moves point
@minus{}@var{count} lines backward, to the beginning of a line.  If
@var{count} is zero, it moves point to the beginning of the current
line.

If @code{forward-line} encounters the beginning or end of the buffer (or
of the accessible portion) before finding that many lines, it sets point
there.  No error is signaled.

@code{forward-line} returns the difference between @var{count} and the
number of lines actually moved.  If you attempt to move down five lines
from the beginning of a buffer that has only three lines, point stops at
the end of the last line, and the value will be 2.

In an interactive call, @var{count} is the numeric prefix argument.
@end deffn

@defun count-lines start end
@cindex lines in region
This function returns the number of lines between the positions
@var{start} and @var{end} in the current buffer.  If @var{start} and
@var{end} are equal, then it returns 0.  Otherwise it returns at least
1, even if @var{start} and @var{end} are on the same line.  This is
because the text between them, considered in isolation, must contain at
least one line unless it is empty.

Here is an example of using @code{count-lines}:

@example
@group
(defun current-line ()
  "Return the vertical position of point@dots{}"
  (+ (count-lines (window-start) (point))
     (if (= (current-column) 0) 1 0)
     -1))
@end group
@end example
@end defun

@ignore
@c ================
The @code{previous-line} and @code{next-line} commands are functions
that should not be used in programs.  They are for users and are
mentioned here only for completeness.

@deffn Command previous-line count
@cindex goal column
This function moves point up @var{count} lines (down if @var{count}
is negative).  In moving, it attempts to keep point in the ``goal column''
(normally the same column that it was at the beginning of the move).

If there is no character in the target line exactly under the current
column, point is positioned after the character in that line which
spans this column, or at the end of the line if it is not long enough.

If it attempts to move beyond the top or bottom of the buffer (or clipped
region), then point is positioned in the goal column in the top or
bottom line.  No error is signaled.

In an interactive call, @var{count} will be the numeric
prefix argument.

The command @code{set-goal-column} can be used to create a semipermanent
goal column to which this command always moves.  Then it does not try to
move vertically.

If you are thinking of using this in a Lisp program, consider using
@code{forward-line} with a negative argument instead.  It is usually easier
to use and more reliable (no dependence on goal column, etc.).
@end deffn

@deffn Command next-line count
This function moves point down @var{count} lines (up if @var{count}
is negative).  In moving, it attempts to keep point in the ``goal column''
(normally the same column that it was at the beginning of the move).

If there is no character in the target line exactly under the current
column, point is positioned after the character in that line which
spans this column, or at the end of the line if it is not long enough.

If it attempts to move beyond the top or bottom of the buffer (or clipped
region), then point is positioned in the goal column in the top or
bottom line.  No error is signaled.

In the case where the @var{count} is 1, and point is on the last
line of the buffer (or clipped region), a new empty line is inserted at the
end of the buffer (or clipped region) and point moved there.

In an interactive call, @var{count} will be the numeric
prefix argument.

The command @code{set-goal-column} can be used to create a semipermanent
goal column to which this command always moves.  Then it does not try to
move vertically.

If you are thinking of using this in a Lisp program, consider using
@code{forward-line} instead.  It is usually easier
to use and more reliable (no dependence on goal column, etc.).
@end deffn

@c ================
@end ignore

  Also see the functions @code{bolp} and @code{eolp} in @ref{Near Point}.
These functions do not move point, but test whether it is already at the
beginning or end of a line.

@node Screen Lines
@subsection Motion by Screen Lines

  The line functions in the previous section count text lines, delimited
only by newline characters.  By contrast, these functions count screen
lines, which are defined by the way the text appears on the screen.  A
text line is a single screen line if it is short enough to fit the width
of the selected window, but otherwise it may occupy several screen
lines.

  In some cases, text lines are truncated on the screen rather than
continued onto additional screen lines.  In these cases,
@code{vertical-motion} moves point much like @code{forward-line}.
@xref{Truncation}.

  Because the width of a given string depends on the flags that control
the appearance of certain characters, @code{vertical-motion} behaves
differently, for a given piece of text, depending on the buffer it is
in, and even on the selected window (because the width, the truncation
flag, and display table may vary between windows).  @xref{Usual
Display}.

  These functions scan text to determine where screen lines break, and
thus take time proportional to the distance scanned.  If you intend to
use them heavily, Emacs provides caches which may improve the
performance of your code.  @xref{Truncation, cache-long-line-scans}.


@defun vertical-motion count &optional window
This function moves point to the start of the screen line @var{count}
screen lines down from the screen line containing point.  If @var{count}
is negative, it moves up instead.

@code{vertical-motion} returns the number of screen lines over which it
moved point.  The value may be less in absolute value than @var{count}
if the beginning or end of the buffer was reached.

The window @var{window} is used for obtaining parameters such as the
width, the horizontal scrolling, and the display table.  But
@code{vertical-motion} always operates on the current buffer, even if
@var{window} currently displays some other buffer.
@end defun

@deffn Command move-to-window-line count
This function moves point with respect to the text currently displayed
in the selected window.  It moves point to the beginning of the screen
line @var{count} screen lines from the top of the window.  If
@var{count} is negative, that specifies a position
@w{@minus{}@var{count}} lines from the bottom (or the last line of the
buffer, if the buffer ends above the specified screen position).

If @var{count} is @code{nil}, then point moves to the beginning of the
line in the middle of the window.  If the absolute value of @var{count}
is greater than the size of the window, then point moves to the place
that would appear on that screen line if the window were tall enough.
This will probably cause the next redisplay to scroll to bring that
location onto the screen.

In an interactive call, @var{count} is the numeric prefix argument.

The value returned is the window line number point has moved to, with
the top line in the window numbered 0.
@end deffn

@defun compute-motion from frompos to topos width offsets window
This function scans the current buffer, calculating screen positions.
It scans the buffer forward from position @var{from}, assuming that is
at screen coordinates @var{frompos}, to position @var{to} or coordinates
@var{topos}, whichever comes first.  It returns the ending buffer
position and screen coordinates.

The coordinate arguments @var{frompos} and @var{topos} are cons cells of
the form @code{(@var{hpos} . @var{vpos})}.

The argument @var{width} is the number of columns available to display
text; this affects handling of continuation lines.  Use the value
returned by @code{window-width} for the window of your choice;
normally, use @code{(window-width @var{window})}.

The argument @var{offsets} is either @code{nil} or a cons cell of the
form @code{(@var{hscroll} . @var{tab-offset})}.  Here @var{hscroll} is
the number of columns not being displayed at the left margin; most
callers get this by calling @code{window-hscroll}.  Meanwhile,
@var{tab-offset} is the offset between column numbers on the screen and
column numbers in the buffer.  This can be nonzero in a continuation
line, when the previous screen lines' widths do not add up to a multiple
of @code{tab-width}.  It is always zero in a non-continuation line.

The window @var{window} serves only to specify which display table to
use.  @code{compute-motion} always operates on the current buffer,
regardless of what buffer is displayed in @var{window}.

The return value is a list of five elements:

@example
(@var{pos} @var{vpos} @var{hpos} @var{prevhpos} @var{contin})
@end example

@noindent
Here @var{pos} is the buffer position where the scan stopped, @var{vpos}
is the vertical screen position, and @var{hpos} is the horizontal screen
position.

The result @var{prevhpos} is the horizontal position one character back
from @var{pos}.  The result @var{contin} is @code{t} if the last line
was continued after (or within) the previous character.

For example, to find the buffer position of column @var{col} of screen line
@var{line} of a certain window, pass the window's display start location
as @var{from} and the window's upper-left coordinates as @var{frompos}.
Pass the buffer's @code{(point-max)} as @var{to}, to limit the scan to
the end of the accessible portion of the buffer, and pass @var{line} and
@var{col} as @var{topos}.  Here's a function that does this:

@example
(defun coordinates-of-position (col line)
  (car (compute-motion (window-start)
                       '(0 . 0)
                       (point-max)
                       (cons col line)
                       (window-width)
                       (cons (window-hscroll) 0)
                       (selected-window))))
@end example

When you use @code{compute-motion} for the minibuffer, you need to use
@code{minibuffer-prompt-width} to get the horizontal position of the
beginning of the first screen line.  @xref{Minibuffer Misc}.
@end defun

@node List Motion
@comment  node-name,  next,  previous,  up
@subsection Moving over Balanced Expressions 
@cindex sexp motion
@cindex Lisp expression motion
@cindex list motion

  Here are several functions concerned with balanced-parenthesis
expressions (also called @dfn{sexps} in connection with moving across
them in Emacs).  The syntax table controls how these functions interpret
various characters; see @ref{Syntax Tables}.  @xref{Parsing
Expressions}, for lower-level primitives for scanning sexps or parts of
sexps.  For user-level commands, see @ref{Lists Commands,,, emacs, GNU
Emacs Manual}.

@deffn Command forward-list arg
This function moves forward across @var{arg} balanced groups of
parentheses.  (Other syntactic entities such as words or paired string
quotes are ignored.)
@end deffn

@deffn Command backward-list arg
This function moves backward across @var{arg} balanced groups of
parentheses.  (Other syntactic entities such as words or paired string
quotes are ignored.)
@end deffn

@deffn Command up-list arg
This function moves forward out of @var{arg} levels of parentheses.
A negative argument means move backward but still to a less deep spot.
@end deffn

@deffn Command down-list arg
This function moves forward into @var{arg} levels of parentheses.  A
negative argument means move backward but still go
deeper in parentheses (@minus{}@var{arg} levels).
@end deffn

@deffn Command forward-sexp arg
This function moves forward across @var{arg} balanced expressions.
Balanced expressions include both those delimited by parentheses and
other kinds, such as words and string constants.  For example,

@example
@group
---------- Buffer: foo ----------
(concat@point{} "foo " (car x) y z)
---------- Buffer: foo ----------
@end group

@group
(forward-sexp 3)
     @result{} nil

---------- Buffer: foo ----------
(concat "foo " (car x) y@point{} z)
---------- Buffer: foo ----------
@end group
@end example
@end deffn

@deffn Command backward-sexp arg
This function moves backward across @var{arg} balanced expressions.
@end deffn

@deffn Command beginning-of-defun arg
This function moves back to the @var{arg}th beginning of a defun.  If
@var{arg} is negative, this actually moves forward, but it still moves
to the beginning of a defun, not to the end of one.
@end deffn

@deffn Command end-of-defun arg
This function moves forward to the @var{arg}th end of a defun.  If
@var{arg} is negative, this actually moves backward, but it still moves
to the end of a defun, not to the beginning of one.
@end deffn

@defopt defun-prompt-regexp
If non-@code{nil}, this variable holds a regular expression that
specifies what text can appear before the open-parenthesis that starts a
defun.  That is to say, a defun begins on a line that starts with a
match for this regular expression, followed by a character with
open-parenthesis syntax.
@end defopt

@node Skipping Characters
@comment  node-name,  next,  previous,  up
@subsection Skipping Characters
@cindex skipping characters

  The following two functions move point over a specified set of
characters.  For example, they are often used to skip whitespace.  For
related functions, see @ref{Motion and Syntax}.

@defun skip-chars-forward character-set &optional limit
This function moves point in the current buffer forward, skipping over a
given set of characters.  It examines the character following point,
then advances point if the character matches @var{character-set}.  This
continues until it reaches a character that does not match.  The
function returns the number of characters moved over.

The argument @var{character-set} is like the inside of a
@samp{[@dots{}]} in a regular expression except that @samp{]} is never
special and @samp{\} quotes @samp{^}, @samp{-} or @samp{\}.  Thus,
@code{"a-zA-Z"} skips over all letters, stopping before the first
nonletter, and @code{"^a-zA-Z"} skips nonletters stopping before the
first letter.  @xref{Regular Expressions}.

If @var{limit} is supplied (it must be a number or a marker), it
specifies the maximum position in the buffer that point can be skipped
to.  Point will stop at or before @var{limit}.

In the following example, point is initially located directly before the
@samp{T}.  After the form is evaluated, point is located at the end of
that line (between the @samp{t} of @samp{hat} and the newline).  The
function skips all letters and spaces, but not newlines.

@example
@group
---------- Buffer: foo ----------
I read "@point{}The cat in the hat
comes back" twice.
---------- Buffer: foo ----------
@end group

@group
(skip-chars-forward "a-zA-Z ")
     @result{} nil

---------- Buffer: foo ----------
I read "The cat in the hat@point{}
comes back" twice.
---------- Buffer: foo ----------
@end group
@end example
@end defun

@defun skip-chars-backward character-set &optional limit
This function moves point backward, skipping characters that match
@var{character-set}, until @var{limit}.  It is just like
@code{skip-chars-forward} except for the direction of motion.

The return value indicates the distance traveled.  It is an integer that
is zero or less.
@end defun

@node Excursions
@section Excursions
@cindex excursion

  It is often useful to move point ``temporarily'' within a localized
portion of the program, or to switch buffers temporarily.  This is
called an @dfn{excursion}, and it is done with the @code{save-excursion}
special form.  This construct saves the current buffer and its values of
point and the mark so they can be restored after the completion of the
excursion.

  The forms for saving and restoring the configuration of windows are
described elsewhere (see @ref{Window Configurations}, and @pxref{Frame
Configurations}).

@defspec save-excursion forms@dots{}
@cindex mark excursion
@cindex point excursion
@cindex current buffer excursion
The @code{save-excursion} special form saves the identity of the current
buffer and the values of point and the mark in it, evaluates
@var{forms}, and finally restores the buffer and its saved values of
point and the mark.  All three saved values are restored even in case of
an abnormal exit via @code{throw} or error (@pxref{Nonlocal Exits}).

The @code{save-excursion} special form is the standard way to switch
buffers or move point within one part of a program and avoid affecting
the rest of the program.  It is used more than 4000 times in the Lisp
sources of Emacs.

@code{save-excursion} does not save the values of point and the mark for
other buffers, so changes in other buffers remain in effect after
@code{save-excursion} exits.

@cindex window excursions
Likewise, @code{save-excursion} does not restore window-buffer
correspondences altered by functions such as @code{switch-to-buffer}.
One way to restore these correspondences, and the selected window, is to
use @code{save-window-excursion} inside @code{save-excursion}
(@pxref{Window Configurations}).

The value returned by @code{save-excursion} is the result of the last of
@var{forms}, or @code{nil} if no @var{forms} are given.

@example
@group
(save-excursion @var{forms})
@equiv{}
(let ((old-buf (current-buffer))
      (old-pnt (point-marker))
@end group
      (old-mark (copy-marker (mark-marker))))
  (unwind-protect
      (progn @var{forms})
    (set-buffer old-buf)
@group
    (goto-char old-pnt)
    (set-marker (mark-marker) old-mark)))
@end group
@end example
@end defspec

  @strong{Warning:} Ordinary insertion of text adjacent to the saved
point value relocates the saved value, just as it relocates all markers.
Therefore, when the saved point value is restored, it normally comes
before the inserted text.

  Although @code{save-excursion} saves the location of the mark, it does
not prevent functions which modify the buffer from setting
@code{deactivate-mark}, and thus causing the deactivation of the mark
after the command finishes.  @xref{The Mark}.

@node Narrowing
@section Narrowing
@cindex narrowing
@cindex restriction (in a buffer)
@cindex accessible portion (of a buffer)

  @dfn{Narrowing} means limiting the text addressable by Emacs editing
commands to a limited range of characters in a buffer.  The text that
remains addressable is called the @dfn{accessible portion} of the
buffer.

  Narrowing is specified with two buffer positions which become the
beginning and end of the accessible portion.  For most editing commands
and most Emacs primitives, these positions replace the values of the
beginning and end of the buffer.  While narrowing is in effect, no text
outside the accessible portion is displayed, and point cannot move
outside the accessible portion.

  Values such as positions or line numbers, which usually count from the
beginning of the buffer, do so despite narrowing, but the functions
which use them refuse to operate on text that is inaccessible.

  The commands for saving buffers are unaffected by narrowing; they save
the entire buffer regardless of any narrowing.

@deffn Command narrow-to-region start end
This function sets the accessible portion of the current buffer to start
at @var{start} and end at @var{end}.  Both arguments should be character
positions.

In an interactive call, @var{start} and @var{end} are set to the bounds
of the current region (point and the mark, with the smallest first).
@end deffn

@deffn Command narrow-to-page move-count
This function sets the accessible portion of the current buffer to
include just the current page.  An optional first argument
@var{move-count} non-@code{nil} means to move forward or backward by
@var{move-count} pages and then narrow to one page.  The variable
@code{page-delimiter} specifies where pages start and end
(@pxref{Standard Regexps}).

In an interactive call, @var{move-count} is set to the numeric prefix
argument.
@end deffn

@deffn Command widen
@cindex widening
This function cancels any narrowing in the current buffer, so that the
entire contents are accessible.  This is called @dfn{widening}.
It is equivalent to the following expression:

@example
(narrow-to-region 1 (1+ (buffer-size)))
@end example
@end deffn

@defspec save-restriction body@dots{}
This special form saves the current bounds of the accessible portion,
evaluates the @var{body} forms, and finally restores the saved bounds,
thus restoring the same state of narrowing (or absence thereof) formerly
in effect.  The state of narrowing is restored even in the event of an
abnormal exit via @code{throw} or error (@pxref{Nonlocal Exits}).
Therefore, this construct is a clean way to narrow a buffer temporarily.

The value returned by @code{save-restriction} is that returned by the
last form in @var{body}, or @code{nil} if no body forms were given.

@c Wordy to avoid overfull hbox.  --rjc 16mar92
@strong{Caution:} it is easy to make a mistake when using the
@code{save-restriction} construct.  Read the entire description here
before you try it.

If @var{body} changes the current buffer, @code{save-restriction} still
restores the restrictions on the original buffer (the buffer whose
restrictions it saved from), but it does not restore the identity of the
current buffer.

@code{save-restriction} does @emph{not} restore point and the mark; use
@code{save-excursion} for that.  If you use both @code{save-restriction}
and @code{save-excursion} together, @code{save-excursion} should come
first (on the outside).  Otherwise, the old point value would be
restored with temporary narrowing still in effect.  If the old point
value were outside the limits of the temporary narrowing, this would
fail to restore it accurately.

The @code{save-restriction} special form records the values of the
beginning and end of the accessible portion as distances from the
beginning and end of the buffer.  In other words, it records the amount
of inaccessible text before and after the accessible portion.

This method yields correct results if @var{body} does further narrowing.
However, @code{save-restriction} can become confused if the body widens
and then makes changes outside the range of the saved narrowing.  When
this is what you want to do, @code{save-restriction} is not the right
tool for the job.  Here is what you must use instead:

@example
@group
(let ((beg (point-min-marker))
      (end (point-max-marker)))
  (unwind-protect
      (progn @var{body})
    (save-excursion
      (set-buffer (marker-buffer beg))
      (narrow-to-region beg end))))
@end group
@end example

Here is a simple example of correct use of @code{save-restriction}:

@example
@group
---------- Buffer: foo ----------
This is the contents of foo
This is the contents of foo
This is the contents of foo@point{}
---------- Buffer: foo ----------
@end group

@group
(save-excursion
  (save-restriction
    (goto-char 1)
    (forward-line 2)
    (narrow-to-region 1 (point))
    (goto-char (point-min))
    (replace-string "foo" "bar")))

---------- Buffer: foo ----------
This is the contents of bar
This is the contents of bar
This is the contents of foo@point{}
---------- Buffer: foo ----------
@end group
@end example
@end defspec