This chapter describes the functions that deal with the text in a buffer. Most examine, insert, or delete text in the current buffer, often in the vicinity of point. Many are interactive. All the functions that change the text provide for undoing the changes (see section Undo).
Many text-related functions operate on a region of text defined
by two buffer positions passed in arguments named start
and end. These arguments should be either markers (see
section Markers) or numeric
character positions (see section Positions). The order of these
arguments does not matter; it is all right for start to
be the end of the region and end the beginning. For
example, (delete-region 1 10)
and (delete-region
10 1)
are equivalent. An args-out-of-range
error is signaled if either start or end is
outside the accessible portion of the buffer. In an interactive
call, point and the mark are used for these arguments.
Throughout this chapter, "text" refers to the characters in the buffer, together with their properties (when relevant).
Many functions are provided to look at the characters around
point. Several simple functions are described here. See also
looking-at
in section Regular Expression Searching.
nil
. The
default for position is point. In the following example, assume that the first character in the buffer is `@':
(char-to-string (char-after 1)) => "@"
nil
. The default for
position is point.
(char-after (point))
. However, if point is at the end
of the buffer, then following-char
returns 0. Remember that point is always between characters, and the
terminal cursor normally appears over the character following
point. Therefore, the character returned by
following-char
is the character the cursor is
over.
In this example, point is between the `a' and the `c'.
---------- Buffer: foo ---------- Gentlemen may cry ``Pea-!-ce! Peace!,'' but there is no peace. ---------- Buffer: foo ---------- (char-to-string (preceding-char)) => "a" (char-to-string (following-char)) => "c"
following-char
, for an example. If point is at the
beginning of the buffer, preceding-char
returns
0.
t
if point is at the beginning of the buffer. If
narrowing is in effect, this means the beginning of the accessible
portion of the text. See also point-min
in section Point.
t
if point is at the end of the buffer. If narrowing
is in effect, this means the end of accessible portion of the text.
See also point-max
in See section Point.
t
if point is at the beginning of a line. See section
Motion by Text Lines. The beginning
of the buffer (or of its accessible portion) always counts as the
beginning of a line.
t
if point is at the end of a line. The end of the
buffer (or of its accessible portion) is always considered the end
of a line.
This section describes two functions that allow a Lisp program to convert any portion of the text in the buffer into a string.
buffer-substring
signals an
args-out-of-range
error. It is not necessary for start to be less than end; the arguments can be given in either order. But most often the smaller argument is written first.
If the text being copied has any text properties, these are copied into the string along with the characters they belong to. See section Text Properties. However, overlays (see section Overlays) in the buffer and their properties are ignored, not copied.
---------- Buffer: foo ---------- This is the contents of buffer foo ---------- Buffer: foo ---------- (buffer-substring 1 10) => "This is t" (buffer-substring (point-max) 10) => "he contents of buffer foo "
buffer-substring
, except that it does not copy text
properties, just the characters themselves. See section Text Properties.
(buffer-substring (point-min) (point-max))
---------- Buffer: foo ---------- This is the contents of buffer foo ---------- Buffer: foo ---------- (buffer-string) => "This is the contents of buffer foo "
The argument thing is a symbol which specifies a kind
of syntactic entity. Possibilities include symbol
,
list
, sexp
, defun
,
filename
, url
, word
,
sentence
, whitespace
, line
,
page
, and others.
---------- Buffer: foo ---------- Gentlemen may cry ``Pea-!-ce! Peace!,'' but there is no peace. ---------- Buffer: foo ---------- (thing-at-point 'word) => "Peace" (thing-at-point 'line) => "Gentlemen may cry ``Peace! Peace!,''\n" (thing-at-point 'whitespace) => nil
This function lets you compare portions of the text in a buffer, without copying them into strings first.
nil
for buffer1, buffer2, or
both to stand for the current buffer. The value is negative if the first substring is less, positive if the first is greater, and zero if they are equal. The absolute value of the result is one plus the index of the first differing characters within the substrings.
This function ignores case when comparing characters if
case-fold-search
is non-nil
. It always
ignores text properties.
Suppose the current buffer contains the text `foobarbar haha!rara!'; then in this example the two substrings are `rbar ' and `rara!'. The value is 2 because the first substring is greater at the second character.
(compare-buffer-substring nil 6 11 nil 16 21) => 2
Insertion means adding new text to a buffer. The inserted text goes at point--between the character before point and the character after point. Some insertion functions leave point before the inserted text, while other functions leave it after. We call the former insertion after point and the latter insertion before point.
Insertion relocates markers that point at positions after the
insertion point, so that they stay with the surrounding text (see
section Markers). When a marker
points at the place of insertion, insertion may or may not relocate
the marker, depending on the marker's insertion type (see section
Marker Insertion Types). Certain
special functions such as insert-before-markers
relocate all such markers to point after the inserted text,
regardless of the markers' insertion type.
Insertion functions signal an error if the current buffer is read-only.
These functions copy text characters from strings and buffers along with their properties. The inserted characters have exactly the same properties as the characters they were copied from. By contrast, characters specified as separate arguments, not part of a string or buffer, inherit their text properties from the neighboring text.
The insertion functions convert text from unibyte to multibyte in order to insert in a multibyte buffer, and vice versa--if the text comes from a string or from a buffer. However, they do not convert unibyte character codes 128 through 255 to multibyte characters, not even if the current buffer is a multibyte buffer. See section Converting Text Representations.
nil
.
nil
. This function is unlike the other insertion functions in that it relocates markers initially pointing at the insertion point, to point after the inserted text. If an overlay begins the insertion point, the inserted text falls outside the overlay; if a nonempty overlay ends at the insertion point, the inserted text falls inside that overlay.
nil
means 1), and character must be
a character. The value is nil
. This function does not convert unibyte character codes 128 through 255 to multibyte characters, not even if the current buffer is a multibyte buffer. See section Converting Text Representations.
If inherit is non-nil
, then the inserted
characters inherit sticky text properties from the two characters
before and after the insertion point. See section Stickiness of Text Properties.
nil
.
In this example, the form is executed with buffer `bar' as the current buffer. We assume that buffer `bar' is initially empty.
---------- Buffer: foo ---------- We hold these truths to be self-evident, that all ---------- Buffer: foo ---------- (insert-buffer-substring "foo" 1 20) => nil ---------- Buffer: bar ---------- We hold these truth-!- ---------- Buffer: bar ----------
See section Stickiness of Text Properties, for other insertion functions that inherit text properties from the nearby text in addition to inserting it. Whitespace inserted by indentation functions also inherits text properties.
This section describes higher-level commands for inserting text, commands intended primarily for the user but useful also in Lisp programs.
nil
.
nil
.
Most printing characters are bound to this command. In routine use,
self-insert-command
is the most frequently called
function in Emacs, but programs rarely use it except to install it
on a keymap. In an interactive call, count is the numeric prefix argument.
This command calls auto-fill-function
whenever that
is non-nil
and the character inserted is a space or a
newline (see section Auto
Filling).
This command performs abbrev expansion if Abbrev mode is enabled and the inserted character does not have word-constituent syntax. (See section Abbrevs And Abbrev Expansion, and section Table of Syntax Classes.)
This is also responsible for calling
blink-paren-function
when the inserted character has
close parenthesis syntax (see section Blinking Parentheses).
This function calls
auto-fill-function
if the current column number is
greater than the value of fill-column
and
number-of-newlines is nil
. Typically what
auto-fill-function
does is insert a newline; thus, the
overall result in this case is to insert two newlines at different
places: one at point, and another earlier in the line.
newline
does not auto-fill if
number-of-newlines is non-nil
.
This command indents to the left margin if that is not zero. See section Margins for Filling.
The value returned is nil
. In an interactive call,
count is the numeric prefix argument.
indent-to
function.
split-line
returns the position of point. Programs hardly ever use this function.
overwrite-mode-textual
,
overwrite-mode-binary
, or nil
.
overwrite-mode-textual
specifies textual overwrite
mode (treats newlines and tabs specially), and
overwrite-mode-binary
specifies binary overwrite mode
(treats newlines and tabs like any other characters).
Deletion means removing part of the text in a buffer, without saving it in the kill ring (see section The Kill Ring). Deleted text can't be yanked, but can be reinserted using the undo mechanism (see section Undo). Some deletion functions do save text in the kill ring in some special cases.
All of the deletion functions operate on the current buffer, and
all return a value of nil
.
buffer-read-only
error.
Otherwise, it deletes the text without asking for any confirmation.
It returns nil
. Normally, deleting a large amount of text from a buffer inhibits
further auto-saving of that buffer "because it has shrunk".
However, erase-buffer
does not do this, the idea being
that the future text is not really related to the former text, and
its size should not be compared with that of the former text.
nil
.
If point was inside the deleted region, its value afterward is
start. Otherwise, point relocates with the surrounding
text, as markers do.
nil
, then it saves the deleted characters in the
kill ring. In an interactive call, count is the numeric prefix argument, and killp is the unprocessed prefix argument. Therefore, if a prefix argument is supplied, the text is saved in the kill ring. If no prefix argument is supplied, then one character is deleted, but not saved in the kill ring.
The value returned is always nil
.
nil
, then it saves the deleted characters in the
kill ring. In an interactive call, count is the numeric prefix argument, and killp is the unprocessed prefix argument. Therefore, if a prefix argument is supplied, the text is saved in the kill ring. If no prefix argument is supplied, then one character is deleted, but not saved in the kill ring.
The value returned is always nil
.
nil
, then the command saves the deleted characters
in the kill ring. Conversion of tabs to spaces happens only if count is positive. If it is negative, exactly -count characters after point are deleted.
In an interactive call, count is the numeric prefix argument, and killp is the unprocessed prefix argument. Therefore, if a prefix argument is supplied, the text is saved in the kill ring. If no prefix argument is supplied, then one character is deleted, but not saved in the kill ring.
The value returned is always nil
.
backward-delete-char-untabify
should deal with
whitespace. Possible values include untabify
, the
default, meaning convert a tab to many spaces and delete one;
hungry
, meaning delete all the whitespace characters
before point with one command, and nil
, meaning do
nothing special for whitespace characters.
This section describes higher-level commands for deleting text, commands intended primarily for the user but useful also in Lisp programs.
nil
. In the following examples, we call
delete-horizontal-space
four times, once on each line,
with point between the second and third characters on the line each
time.
---------- Buffer: foo ---------- I -!-thought I -!- thought We-!- thought Yo-!-u thought ---------- Buffer: foo ---------- (delete-horizontal-space) ; Four times. => nil ---------- Buffer: foo ---------- Ithought Ithought Wethought You thought ---------- Buffer: foo ----------
nil
,
delete-indentation
joins this line to the following
line instead. The function returns nil
. If there is a fill prefix, and the second of the lines being
joined starts with the prefix, then delete-indentation
deletes the fill prefix before joining the lines. See section Margins for Filling.
In the example below, point is located on the line starting `events', and it makes no difference if there are trailing spaces in the preceding line.
---------- Buffer: foo ---------- When in the course of human -!- events, it becomes necessary ---------- Buffer: foo ---------- (delete-indentation) => nil ---------- Buffer: foo ---------- When in the course of human-!- events, it becomes necessary ---------- Buffer: foo ----------
After the lines are joined, the function
fixup-whitespace
is responsible for deciding whether
to leave a space at the junction.
nil
. At the beginning or end of a line, the appropriate amount of space is none. Before a character with close parenthesis syntax, or after a character with open parenthesis or expression-prefix syntax, no space is also appropriate. Otherwise, one space is appropriate. See section Table of Syntax Classes.
In the example below, fixup-whitespace
is called
the first time with point before the word `spaces' in
the first line. For the second invocation, point is directly after
the `('.
---------- Buffer: foo ---------- This has too many -!-spaces This has too many spaces at the start of (-!- this list) ---------- Buffer: foo ---------- (fixup-whitespace) => nil (fixup-whitespace) => nil ---------- Buffer: foo ---------- This has too many spaces This has too many spaces at the start of (this list) ---------- Buffer: foo ----------
nil
.
A blank line is defined as a line containing only tabs and spaces.
delete-blank-lines
returns nil
.
Kill functions delete text like the deletion functions, but save it so that the user can reinsert it by yanking. Most of these functions have `kill-' in their name. By contrast, the functions whose names start with `delete-' normally do not save text for yanking (though they can still be undone); these are "deletion" functions.
Most of the kill commands are primarily for interactive use, and are not described here. What we do describe are the functions provided for use in writing such commands. You can use these functions to write commands for killing text. When you need to delete text for internal purposes within a Lisp function, you should normally use deletion functions, so as not to disturb the kill ring contents. See section Deleting Text.
Killed text is saved for later yanking in the kill
ring. This is a list that holds a number of recent kills, not
just the last text kill. We call this a "ring" because yanking
treats it as having elements in a cyclic order. The list is kept in
the variable kill-ring
, and can be operated on with
the usual functions for lists; there are also specialized
functions, described in this section, that treat it as a ring.
Some people think this use of the word "kill" is unfortunate, since it refers to operations that specifically do not destroy the entities "killed". This is in sharp contrast to ordinary life, in which death is permanent and "killed" entities do not come back to life. Therefore, other metaphors have been proposed. For example, the term "cut ring" makes sense to people who, in pre-computer days, used scissors and paste to cut up and rearrange manuscripts. However, it would be difficult to change the terminology now.
The kill ring records killed text as strings in a list, most recent first. A short kill ring, for example, might look like this:
("some text" "a different piece of text" "even older text")
When the list reaches kill-ring-max
entries in
length, adding a new entry automatically deletes the last
entry.
When kill commands are interwoven with other commands, each kill command makes a new entry in the kill ring. Multiple kill commands in succession build up a single kill-ring entry, which would be yanked as a unit; the second and subsequent consecutive kill commands add text to the entry made by the first one.
For yanking, one entry in the kill ring is designated the "front" of the ring. Some yank commands "rotate" the ring by designating a different element as the "front." But this virtual rotation doesn't change the list itself--the most recent entry always comes first in the list.
kill-region
is the usual subroutine for killing
text. Any command that calls this function is a "kill command" (and
should probably have `kill' in its name).
kill-region
puts the newly killed text in a new
element at the beginning of the kill ring or adds it to the most
recent element. It determines automatically (using
last-command
) whether the previous command was a kill
command, and if so appends the killed text to the most recent
entry.
nil
. In an interactive call, start and end are point and the mark.
If the buffer is read-only, kill-region
modifies
the kill ring just the same, then signals an error without
modifying the buffer. This is convenient because it lets the user
use all the kill commands to copy text into the kill ring from a
read-only buffer.
nil
, kill-region
does not get an
error if the buffer is read-only. Instead, it simply returns,
updating the kill ring but not changing the buffer.
nil
. It also indicates the
extent of the text copied by moving the cursor momentarily, or by
displaying a message in the echo area. The command does not set this-command
to
kill-region
, so a subsequent kill command does not
append to the same kill ring entry.
Don't call copy-region-as-kill
in Lisp programs
unless you aim to support Emacs 18. For newer Emacs versions, it is
better to use kill-new
or kill-append
instead. See section Low-Level Kill
Ring.
Yanking means reinserting an entry of previously killed text from the kill ring. The text properties are copied too.
If arg is a list (which occurs interactively when the
user types C-u with no digits), then yank
inserts the text as described above, but puts point before the
yanked text and puts the mark after it.
If arg is a number, then yank
inserts
the argth most recently killed text--the
argth element of the kill ring list.
yank
does not alter the contents of the kill ring
or rotate it. It returns nil
.
This is allowed only immediately after a yank
or
another yank-pop
. At such a time, the region contains
text that was just inserted by yanking. yank-pop
deletes that text and inserts in its place a different piece of
killed text. It does not add the deleted text to the kill ring,
since it is already in the kill ring somewhere.
If arg is nil
, then the replacement text
is the previous element of the kill ring. If arg is
numeric, the replacement is the argth previous kill. If
arg is negative, a more recent kill is the
replacement.
The sequence of kills in the kill ring wraps around, so that after the oldest one comes the newest one, and before the newest one goes the oldest.
The return value is always nil
.
These functions and variables provide access to the kill ring at a lower level, but still convenient for use in Lisp programs, because they take care of interaction with window system selections (see section Window System Selections).
current-kill
rotates the yanking pointer, which
designates the "front" of the kill ring, by n places
(from newer kills to older ones), and returns the text at that
place in the ring. If the optional second argument do-not-move is
non-nil
, then current-kill
doesn't alter
the yanking pointer; it just returns the nth kill,
counting from the current yanking pointer.
If n is zero, indicating a request for the latest
kill, current-kill
calls the value of
interprogram-paste-function
(documented below) before
consulting the kill ring.
interprogram-cut-function
(see
below).
nil
, it goes at the
beginning. This function also invokes the value of
interprogram-cut-function
(see below).
nil
or a
function of no arguments. If the value is a function, current-kill
calls it
to get the "most recent kill". If the function returns a
non-nil
value, then that value is used as the "most
recent kill". If it returns nil
, then the first
element of kill-ring
is used.
The normal use of this hook is to get the window system's primary selection as the most recent kill, even if the selection belongs to another application. See section Window System Selections.
nil
or a
function of one argument. If the value is a function, kill-new
and
kill-append
call it with the new first element of the
kill ring as an argument.
The normal use of this hook is to set the window system's primary selection from the newly killed text. See section Window System Selections.
The variable kill-ring
holds the kill ring
contents, in the form of a list of strings. The most recent kill is
always at the front of the list.
The kill-ring-yank-pointer
variable points to a
link in the kill ring list, whose CAR is the text to yank next. We
say it identifies the "front" of the ring. Moving
kill-ring-yank-pointer
to a different link is called
rotating the kill ring. We call the kill ring a "ring"
because the functions that move the yank pointer wrap around from
the end of the list to the beginning, or vice-versa. Rotation of
the kill ring is virtual; it does not change the value of
kill-ring
.
Both kill-ring
and
kill-ring-yank-pointer
are Lisp variables whose values
are normally lists. The word "pointer" in the name of the
kill-ring-yank-pointer
indicates that the variable's
purpose is to identify one element of the list for use by the next
yank command.
The value of kill-ring-yank-pointer
is always
eq
to one of the links in the kill ring list. The
element it identifies is the CAR of that link. Kill commands, which
change the kill ring, also set this variable to the value of
kill-ring
. The effect is to rotate the ring so that
the newly killed text is at the front.
Here is a diagram that shows the variable
kill-ring-yank-pointer
pointing to the second entry in
the kill ring ("some text" "a different piece of text" "yet
older text")
.
kill-ring ---- kill-ring-yank-pointer | | | v | --- --- --- --- --- --- --> | | |------> | | |--> | | |--> nil --- --- --- --- --- --- | | | | | | | | -->"yet older text" | | | --> "a different piece of text" | --> "some text"
This state of affairs might occur after C-y
(yank
) immediately followed by M-y
(yank-pop
).
kill-ring
, and its CAR is the kill string that
C-y should yank.
kill-ring-max
is 30.
Most buffers have an undo list, which records all
changes made to the buffer's text so that they can be undone. (The
buffers that don't have one are usually special-purpose buffers for
which Emacs assumes that undoing is not useful.) All the primitives
that modify the text in the buffer automatically add elements to
the front of the undo list, which is in the variable
buffer-undo-list
.
t
disables the recording of undo information.
Here are the kinds of elements an undo list can have:
position
(beg . end)
(text . position)
(abs
position)
.
(t high . low)
primitive-undo
uses those values to determine whether
to mark the buffer as unmodified once again; it does so only if the
file's modification time matches those numbers.
(nil property value beg
. end)
(put-text-property beg end property value)
(marker . adjustment)
nil
nil
. The editor command loop automatically creates an undo boundary before each key sequence is executed. Thus, each undo normally undoes the effects of one command. Self-inserting input characters are an exception. The command loop makes a boundary for the first such character; the next 19 consecutive self-inserting input characters do not make boundaries, and then the 20th does, and so on as long as self-inserting characters continue.
All buffer modifications add a boundary whenever the previous undoable change was made in some other buffer. This is to ensure that each command makes a boundary in each buffer where it makes changes.
Calling this function explicitly is useful for splitting the
effects of a command into more than one unit. For example,
query-replace
calls undo-boundary
after
each replacement, so that the user can undo individual replacements
one by one.
primitive-undo
adds elements to the buffer's undo
list when it changes the buffer. Undo commands avoid confusion by
saving the undo list value at the beginning of a sequence of undo
operations. Then the undo operations use and update the saved
value. The new elements added by undoing are not part of this saved
value, so they don't interfere with continuing to undo.
This section describes how to enable and disable undo information for a given buffer. It also explains how the undo list is truncated automatically so it doesn't get too big.
Recording of undo information in a newly created buffer is
normally enabled to start with; but if the buffer name starts with
a space, the undo recording is initially disabled. You can
explicitly enable or disable undo recording with the following two
functions, or by setting buffer-undo-list
yourself.
nil
. In an interactive call, buffer-or-name is the current buffer. You cannot specify any other buffer.
This function returns nil
.
The name buffer-flush-undo
is not considered
obsolete, but the preferred name is
buffer-disable-undo
.
As editing continues, undo lists get longer and longer. To
prevent them from using up all available memory space, garbage
collection trims them back to size limits you can set. (For this
purpose, the "size" of an undo list measures the cons cells that
make up the list, plus the strings of deleted text.) Two variables
control the range of acceptable sizes: undo-limit
and
undo-strong-limit
.
Filling means adjusting the lengths of lines (by moving
the line breaks) so that they are nearly (but no greater than) a
specified maximum width. Additionally, lines can be
justified, which means inserting spaces to make the left
and/or right margins line up precisely. The width is controlled by
the variable fill-column
. For ease of reading, lines
should be no longer than 70 or so columns.
You can use Auto Fill mode (see section Auto Filling) to fill text automatically as you insert it, but changes to existing text may leave it improperly filled. Then you must fill the text explicitly.
Most of the commands in this section return values that are not
meaningful. All the functions that do filling take note of the
current left margin, current right margin, and current
justification style (see section Margins for Filling). If the current
justification style is none
, the filling functions
don't actually do anything.
Several of the filling functions have an argument
justify. If it is non-nil
, that requests
some kind of justification. It can be left
,
right
, full
, or center
, to
request a specific style of justification. If it is t
,
that means to use the current justification style for this part of
the text (see current-justification
, below). Any other
value is treated as full
.
When you call the filling functions interactively, using a
prefix argument implies the value full
for
justify.
nil
, each line is
justified as well. It uses the ordinary paragraph motion commands
to find paragraph boundaries. See section `Paragraphs' in The
Emacs Manual.
nil
. If nosqueeze is non-nil
, that means to
leave whitespace other than line breaks untouched. If
to-eop is non-nil
, that means to keep
filling to the end of the paragraph--or the next hard newline, if
use-hard-newlines
is enabled (see below).
The variable paragraph-separate
controls how to
distinguish paragraphs. See section Standard Regular Expressions Used in
Editing.
The first two arguments, start and end,
are the beginning and end of the region to be filled. The third and
fourth arguments, justify and mail-flag, are
optional. If justify is non-nil
, the
paragraphs are justified as well as filled. If mail-flag
is non-nil
, it means the function is operating on a
mail message and therefore should not fill the header lines.
Ordinarily, fill-individual-paragraphs
regards each
change in indentation as starting a new paragraph. If
fill-individual-varying-indent
is
non-nil
, then only separator lines separate
paragraphs. That mode can handle indented paragraphs with
additional indentation on the first line.
fill-individual-paragraphs
as described
above.
nil
. In an interactive call, any prefix argument requests justification.
If nosqueeze is non-nil
, that means to
leave whitespace other than line breaks untouched. If
squeeze-after is non-nil
, it specifies a
position in the region, and means don't canonicalize spaces before
that position.
In Adaptive Fill mode, this command calls
fill-context-prefix
to choose a fill prefix by
default. See section Adaptive Fill
Mode.
fill-column
. It returns nil
. The argument how, if non-nil
specifies
explicitly the style of justification. It can be left
,
right
, full
, center
, or
none
. If it is t
, that means to do follow
specified justification style (see
current-justification
, below). nil
means
to do full justification.
If eop is non-nil
, that means do
left-justification if current-justification
specifies
full justification. This is used for the last line of a paragraph;
even if the paragraph as a whole is fully justified, the last line
should not be.
If nosqueeze is non-nil
, that means do
not change interior whitespace.
left
, right
, full
,
center
, or none
. The default value is
left
.
nil
, a period followed by just one space does not
count as the end of a sentence, and the filling functions avoid
breaking the line at such a place.
nil
, fill-paragraph
calls
this function to do the work. If the function returns a
non-nil
value, fill-paragraph
assumes the
job is done, and immediately returns that value. The usual use of this feature is to fill comments in programming language modes. If the function needs to fill a paragraph in the usual way, it can do so as follows:
(let ((fill-paragraph-function nil)) (fill-paragraph arg))
nil
, the filling functions do not delete newlines
that have the hard
text property. These "hard
newlines" act as paragraph separators.
The fill prefix follows the left margin whitespace, if any.
As a practical matter, if you are writing text for other people
to read, you should set fill-column
to no more than
70. Otherwise the line will be too long for people to read
comfortably, and this can make the text seem clumsy.
fill-column
in buffers that
do not override it. This is the same as (default-value
'fill-column)
. The default value for default-fill-column
is
70.
left-margin
property on the text from from
to to to the value margin. If Auto Fill mode
is enabled, this command also refills the region to fit the new
margin.
right-margin
property on the text from from
to to to the value margin. If Auto Fill mode
is enabled, this command also refills the region to fit the new
margin.
left-margin
property of
the character at the start of the current line (or zero if none),
and the value of the variable left-margin
.
fill-column
variable,
minus the value of the right-margin
property of the
character after point.
current-left-margin
. If the argument n is
non-nil
, move-to-left-margin
moves
forward n-1 lines first. If force is non-nil
, that says to fix
the line's indentation if that doesn't match the left margin
value.
current-left-margin
. In no case does this
function delete non-whitespace.
indent-line-function
, used in Fundamental mode, Text
mode, etc. Its effect is to adjust the indentation at the beginning
of the current line to the value specified by the variable
left-margin
. This may involve either inserting or
deleting whitespace.
nil
,
then the line won't be broken there.
Adaptive Fill mode chooses a fill prefix automatically from the text in each paragraph being filled.
nil
. It is
t
by default.
comment-start-skip
, then it is
used--otherwise, spaces amounting to the same width are used
instead. However, the fill prefix is never taken from a one-line paragraph if it would act as a paragraph starter on subsequent lines.
adaptive-fill-regexp
does not match, with point after
the left margin of a line, and it should return the appropriate
fill prefix based on that line. If it returns nil
,
that means it sees no fill prefix in that line.
Auto Fill mode is a minor mode that fills lines automatically as text is inserted. This section describes the hook used by Auto Fill mode. For a description of functions that you can call explicitly to fill and justify existing text, see section Filling.
Auto Fill mode also enables the functions that change the margins and justification style to refill portions of the text. See section Margins for Filling.
nil
, in
which case nothing special is done in that case. The value of auto-fill-function
is
do-auto-fill
when Auto-Fill mode is enabled. That is a
function whose sole purpose is to implement the usual strategy for
breaking a line.
In older Emacs versions, this variable was named
auto-fill-hook
, but since it is not called with the standard convention for hooks, it was renamed toauto-fill-function
in version 19.
auto-fill-function
, if and when
Auto Fill is turned on. Major modes can set buffer-local values for
this variable to alter how Auto Fill works.
The sorting functions described in this section all rearrange
text in a buffer. This is in contrast to the function
sort
, which rearranges the order of the elements of a
list (see section Functions that
Rearrange Lists). The values returned by these functions are
not meaningful.
To understand how sort-subr
works, consider the
whole accessible portion of the buffer as being divided into
disjoint pieces called sort records. The records may or
may not be contiguous, but they must not overlap. A portion of each
sort record (perhaps all of it) is designated as the sort key.
Sorting rearranges the records in order by their sort keys.
Usually, the records are rearranged in order of ascending sort
key. If the first argument to the sort-subr
function,
reverse, is non-nil
, the sort records are
rearranged in order of descending sort key.
The next four arguments to sort-subr
are functions
that are called to move point across a sort record. They are called
many times from within sort-subr
.
sort-subr
is called. Therefore, you should usually
move point to the beginning of the buffer before calling
sort-subr
. This function can indicate there are no
more sort records by leaving point at the end of the buffer.
nil
value to be used as the sort key, or return nil
to
indicate that the sort key is in the buffer starting at point. In
the latter case, endkeyfun is called to find the end of
the sort key.
nil
and this
argument is omitted (or nil
), then the sort key
extends to the end of the record. There is no need for
endkeyfun if startkeyfun returns a
non-nil
value.
As an example of sort-subr
, here is the complete
function definition for sort-lines
:
;; Note that the first two lines of doc string ;; are effectively one line when viewed by a user. (defun sort-lines (reverse beg end) "Sort lines in region alphabetically;\ argument means descending order. Called from a program, there are three arguments: REVERSE (non-nil means reverse order),\ BEG and END (region to sort). The variable `sort-fold-case' determines\ whether alphabetic case affects the sort order. (interactive "P\nr") (save-excursion (save-restriction (narrow-to-region beg end) (goto-char (point-min)) (sort-subr reverse 'forward-line 'end-of-line))))
Here forward-line
moves point to the start of the
next record, and end-of-line
moves point to the end of
record. We do not pass the arguments startkeyfun and
endkeyfun, because the entire record is used as the sort
key.
The sort-paragraphs
function is very much the same,
except that its sort-subr
call looks like this:
(sort-subr reverse (function (lambda () (while (and (not (eobp)) (looking-at paragraph-separate)) (forward-line 1)))) 'forward-paragraph)
Markers pointing into any sort records are left with no useful
position after sort-subr
returns.
nil
, sort-subr
and the other buffer
sorting functions ignore case when comparing strings.
Alphabetical sorting means that two sort keys are compared by comparing the first characters of each, the second characters of each, and so on. If a mismatch is found, it means that the sort keys are unequal; the sort key whose character is less at the point of first mismatch is the lesser sort key. The individual characters are compared according to their numerical character codes in the Emacs character set.
The value of the record-regexp argument specifies how to divide the buffer into sort records. At the end of each record, a search is done for this regular expression, and the text that matches it is taken as the next record. For example, the regular expression `^.+$', which matches lines with at least one character besides a newline, would make each such line into a sort record. See section Regular Expressions, for a description of the syntax and meaning of regular expressions.
The value of the key-regexp argument specifies what part of each record is the sort key. The key-regexp could match the whole record, or only a part. In the latter case, the rest of the record has no effect on the sorted order of records, but it is carried along when the record moves to its new position.
The key-regexp argument can refer to the text matched by a subexpression of record-regexp, or it can be a regular expression on its own.
If key-regexp is:
sort-regexp-fields
searches for a match for
the regular expression within the record. If such a match is found,
it is the sort key. If there is no match for key-regexp
within a record then that record is ignored, which means its
position in the buffer is not changed. (The other records may move
around it.)
For example, if you plan to sort all the lines in the region by the first word on each line starting with the letter `f', you should set record-regexp to `^.*$' and set key-regexp to `\<f\w*\>'. The resulting expression looks like this:
(sort-regexp-fields nil "^.*$" "\\<f\\w*\\>" (region-beginning) (region-end))
If you call sort-regexp-fields
interactively, it
prompts for record-regexp and key-regexp in
the minibuffer.
nil
, the
sort is in reverse order.
nil
, the
sort is in reverse order.
nil
, the
sort is in reverse order.
If reverse is non-nil
, the sort is in
reverse order.
One unusual thing about this command is that the entire line containing position beg, and the entire line containing position end, are included in the region sorted.
Note that sort-columns
uses the sort
utility program, and so cannot work properly on text containing tab
characters. Use M-x untabify to convert tabs to spaces
before sorting.
The column functions convert between a character position (counting characters from the beginning of the buffer) and a column position (counting screen characters from the beginning of a line).
These functions count each character according to the number of
columns it occupies on the screen. This means control characters
count as occupying 2 or 4 columns, depending upon the value of
ctl-arrow
, and tabs count as occupying a number of
columns that depends on the value of tab-width
and on
the column where the tab begins. See section Usual Display Conventions.
Column number computations ignore the width of the window and the amount of horizontal scrolling. Consequently, a column value can be arbitrarily high. The first (or leftmost) column is numbered 0.
For an example of using current-column
, see the
description of count-lines
in section Motion by Text Lines.
If column column is beyond the end of the line, point moves to the end of the line. If column is negative, point moves to the beginning of the line.
If it is impossible to move to column column because
that is in the middle of a multicolumn character such as a tab,
point moves to the end of that character. However, if
force is non-nil
, and column is
in the middle of a tab, then move-to-column
converts
the tab into spaces so that it can move precisely to column
column. Other multicolumn characters can cause anomalies
despite force, since there is no way to split them.
The argument force also has an effect if the line isn't long enough to reach column column; in that case, it says to add whitespace at the end of the line to reach that column.
If column is not an integer, an error is signaled.
The return value is the column number actually moved to.
The indentation functions are used to examine, move to, and change whitespace that is at the beginning of a line. Some of the functions can also change whitespace elsewhere on a line. Columns and indentation count from zero at the left margin.
This section describes the primitive functions used to count and insert indentation. The functions in the following sections use these primitives. See section Width, for related functions.
nil
, then at
least that many spaces are inserted even if this requires going
beyond column. Otherwise the function does nothing if
point is already beyond column. The value is the column
at which the inserted indentation ends. The inserted whitespace characters inherit text properties from the surrounding text (usually, from the preceding text only). See section Stickiness of Text Properties.
nil
, indentation functions can insert tabs as well
as spaces. Otherwise, they insert only spaces. Setting this
variable automatically makes it buffer-local in the current
buffer.
An important function of each major mode is to customize the TAB key to indent properly for the language being edited. This section describes the mechanism of the TAB key and how to control it. The functions in this section return unpredictable values.
indent-according-to-mode
does no more than call this
function. In Lisp mode, the value is the symbol
lisp-indent-line
; in C mode,
c-indent-line
; in Fortran mode,
fortran-indent-line
. In Fundamental mode, Text mode,
and many other modes with no standard for indentation, the value is
indent-to-left-margin
(which is the default
value).
indent-line-function
to indent the current
line in a way appropriate for the current major mode.
indent-line-function
to indent the current
line; however, if that function is
indent-to-left-margin
, insert-tab
is
called instead. (That is a trivial command that inserts a tab
character.)
It does indentation by calling the current
indent-line-function
. In programming language modes,
this is the same thing TAB does, but in some text modes,
where TAB inserts a tab, newline-and-indent
indents to the column specified by left-margin
.
This command does indentation on both lines according to the
current major mode, by calling the current value of
indent-line-function
. In programming language modes,
this is the same thing TAB does, but in some text modes,
where TAB inserts a tab,
reindent-then-newline-and-indent
indents to the column
specified by left-margin
.
This section describes commands that indent all the lines in the region. They return unpredictable values.
nil
, indent-region
indents each nonblank
line by calling the current mode's indentation function, the value
of indent-line-function
. If to-column is non-nil
, it should be an
integer specifying the number of columns of indentation; then this
function gives each line exactly that much indentation, by either
adding or deleting whitespace.
If there is a fill prefix, indent-region
indents
each line by making it start with the fill prefix.
indent-region
as a
short cut. It should take two arguments, the start and end of the
region. You should design the function so that it will produce the
same results as indenting the lines of the region one by one, but
presumably faster. If the value is nil
, there is no short cut, and
indent-region
actually works line by line.
A short-cut function is useful in modes such as C mode and Lisp
mode, where the indent-line-function
must scan from
the beginning of the function definition: applying it to each line
would be quadratic in time. The short cut can update the scan
information as it moves through the lines indenting them; this
takes linear time. In a mode where indenting a line individually is
fast, there is no need for a short cut.
indent-region
with a non-nil
argument
to-column has a different meaning and does not use this
variable.
For example, if count is 3, this command adds 3 columns of indentation to each of the lines beginning in the region specified.
In Mail mode, C-c C-y
(mail-yank-original
) uses indent-rigidly
to indent the text copied from the message being replied to.
indent-rigidly
, except that it doesn't alter lines
that start within strings or comments. In addition, it doesn't alter a line if
nochange-regexp matches at the beginning of the line (if
nochange-regexp is non-nil
).
This section describes two commands that indent the current line based on the contents of previous lines.
If the previous nonblank line has no next indent point (i.e.,
none at a great enough column position),
indent-relative
either does nothing (if
unindented-ok is non-nil
) or calls
tab-to-tab-stop
. Thus, if point is underneath and to
the right of the last column of a short line of text, this command
ordinarily moves point to the next tab stop by inserting
whitespace.
The return value of indent-relative
is
unpredictable.
In the following example, point is at the beginning of the second line:
This line is indented twelve spaces. -!-The quick brown fox jumped.
Evaluation of the expression (indent-relative nil)
produces the following:
This line is indented twelve spaces. -!-The quick brown fox jumped.
In this next example, point is between the `m' and `p' of `jumped':
This line is indented twelve spaces. The quick brown fox jum-!-ped.
Evaluation of the expression (indent-relative nil)
produces the following:
This line is indented twelve spaces. The quick brown fox jum -!-ped.
indent-relative
with t
as the
unindented-ok argument. The return value is
unpredictable. If the previous nonblank line has no indent points beyond the current column, this command does nothing.
This section explains the mechanism for user-specified "tab stops" and the mechanisms that use and set them. The name "tab stops" is used because the feature is similar to that of the tab stops on a typewriter. The feature works by inserting an appropriate number of spaces and tab characters to reach the next tab stop column; it does not affect the display of tab characters in the buffer (see section Usual Display Conventions). Note that the TAB character as input uses this tab stop feature only in a few major modes, such as Text mode.
tab-stop-list
. It searches the list for an element
greater than the current column number, and uses that element as
the column to indent to. It does nothing if no such element is
found.
tab-to-tab-stops
. The
elements should be integers in increasing order. The tab stop
columns need not be evenly spaced. Use M-x edit-tab-stops to edit the location of tab stops interactively.
These commands, primarily for interactive use, act based on the indentation in the text.
nil
.
nil
.
nil
.
The case change commands described here work on text in the current buffer. See section Case Conversion in Lisp, for case conversion functions that work on strings and characters. See section The Case Table, for how to customize which characters are upper or lower case and how to convert them.
nil
. If one end of the region is in the middle of a word, the part of the word within the region is treated as an entire word.
When capitalize-region
is called interactively,
start and end are point and the mark, with
the smallest first.
---------- Buffer: foo ---------- This is the contents of the 5th foo. ---------- Buffer: foo ---------- (capitalize-region 1 44) => nil ---------- Buffer: foo ---------- This Is The Contents Of The 5th Foo. ---------- Buffer: foo ----------
nil
. When downcase-region
is called interactively,
start and end are point and the mark, with
the smallest first.
nil
. When upcase-region
is called interactively,
start and end are point and the mark, with
the smallest first.
nil
. If point is in the middle of a word, the part of the word before point is ignored when moving forward. The rest is treated as an entire word.
When capitalize-word
is called interactively,
count is set to the numeric prefix argument.
nil
. When downcase-word
is called interactively,
count is set to the numeric prefix argument.
nil
. When upcase-word
is called interactively,
count is set to the numeric prefix argument.
Each character position in a buffer or a string can have a text property list, much like the property list of a symbol (see section Property Lists). The properties belong to a particular character at a particular place, such as, the letter `T' at the beginning of this sentence or the first `o' in `foo'---if the same character occurs in two different places, the two occurrences generally have different properties.
Each property has a name and a value. Both of these can be any Lisp object, but the name is normally a symbol. The usual way to access the property list is to specify a name and ask what value corresponds to it.
If a character has a category
property, we call it
the category of the character. It should be a symbol. The
properties of the symbol serve as defaults for the properties of
the character.
Copying text between strings and buffers preserves the
properties along with the characters; this includes such diverse
functions as substring
, insert
, and
buffer-substring
.
The simplest way to examine text properties is to ask for the
value of a particular property of a particular character. For that,
use get-text-property
. Use
text-properties-at
to get the entire property list of
a character. See section Text Property
Search Functions, for functions to examine the properties of a
number of characters at once.
These functions handle both strings and buffers. Keep in mind that positions in a string start from 0, whereas positions in a buffer start from 1.
If there is no prop property strictly speaking, but
the character has a category that is a symbol, then
get-text-property
returns the prop property
of that symbol.
get-text-property
, except that it checks overlays
first and then text properties. See section Overlays. The argument object may be a string, a buffer, or a window. If it is a window, then the buffer displayed in that window is used for text properties and overlays, but only the overlays active for that window are considered. If object is a buffer, then all overlays in that buffer are considered, as well as text properties. If object is a string, only text properties are considered, since strings never have overlays.
nil
, it defaults to the current buffer.
(setq default-text-properties '(foo 69)) ;; Make sure character 1 has no properties of its own. (set-text-properties 1 2 nil) ;; What we get, when we ask, is the default value. (get-text-property 1 'foo) => 69
The primitives for changing properties apply to a specified
range of text in a buffer or string. The function
set-text-properties
(see end of section) sets the
entire property list of the text in that range; more often, it is
useful to add, change, or delete just certain properties specified
by name.
Since text properties are considered part of the contents of the buffer (or string), and can affect how a buffer looks on the screen, any change in buffer text properties mark the buffer as modified. Buffer text property changes are undoable also (see section Undo).
nil
, it
defaults to the current buffer.
nil
, it defaults to the current
buffer. The argument props specifies which properties to add. It should have the form of a property list (see section Property Lists): a list whose elements include the property names followed alternately by the corresponding values.
The return value is t
if the function actually
changed some property's value; nil
otherwise (if
props is nil
or its values agree with those
in the text).
For example, here is how to set the comment
and
face
properties of a range of text:
(add-text-properties start end '(comment t face highlight))
nil
, it defaults to the current
buffer. The argument props specifies which properties to
delete. It should have the form of a property list (see section Property Lists): a list whose elements
are property names alternating with corresponding values. But only
the names matter--the values that accompany them are ignored. For
example, here's how to remove the face
property.
(remove-text-properties start end '(face nil))
The return value is t
if the function actually
changed some property's value; nil
otherwise (if
props is nil
or if no character in the
specified text had any of those properties).
To remove all text properties from certain text, use
set-text-properties
and specify nil
for
the new property list.
nil
, it
defaults to the current buffer. The argument props is the new property list. It should be a list whose elements are property names alternating with corresponding values.
After set-text-properties
returns, all the
characters in the specified range have identical properties.
If props is nil
, the effect is to get
rid of all properties from the specified range of text. Here's an
example:
(set-text-properties start end nil)
See also the function
buffer-substring-no-properties
(see section Examining Buffer Contents) which
copies text from the buffer but does not copy its properties.
In typical use of text properties, most of the time several or many consecutive characters have the same value for a property. Rather than writing your programs to examine characters one by one, it is much faster to process chunks of text that have the same property value.
Here are functions you can use to do this. They use
eq
for comparing property values. In all cases,
object defaults to the current buffer.
For high performance, it's very important to use the limit argument to these functions, especially the ones that search for a single property--otherwise, they may spend a long time scanning to the end of the buffer, if the property you are interested in does not change.
These functions do not move point; instead, they return a
position (or nil
). Remember that a position is always
between two characters; the position returned by these functions is
between two characters with different properties.
If limit is non-nil
, then the scan ends
at position limit. If there is no property change before
that point, next-property-change
returns
limit.
The value is nil
if the properties remain unchanged
all the way to the end of object and limit is
nil
. If the value is non-nil
, it is a
position greater than or equal to pos. The value equals
pos only when limit equals
pos.
Here is an example of how to scan the buffer by chunks of text within which all properties are constant:
(while (not (eobp)) (let ((plist (text-properties-at (point))) (next-change (or (next-property-change (point) (current-buffer)) (point-max)))) Process text from point to next-change... (goto-char next-change)))
If limit is non-nil
, then the scan ends
at position limit. If there is no property change before
that point, next-single-property-change
returns
limit.
The value is nil
if the property remains unchanged
all the way to the end of object and limit is
nil
. If the value is non-nil
, it is a
position greater than or equal to pos; it equals
pos only if limit equals pos.
next-property-change
, but scans back from
pos instead of forward. If the value is
non-nil
, it is a position less than or equal to
pos; it equals pos only if limit
equals pos.
next-single-property-change
, but scans back from
pos instead of forward. If the value is
non-nil
, it is a position less than or equal to
pos; it equals pos only if limit
equals pos.
next-property-change
except that it considers overlay properties as well as text
properties. There is no object operand because this
function operates only on the current buffer. It returns the next
address at which either kind of property changes.
next-char-property-change
, but scans back from
position instead of forward.
nil
if at least one character between
start and end has a property prop
whose value is value. More precisely, it returns the
position of the first such character. Otherwise, it returns
nil
. The optional fifth argument, object, specifies the string or buffer to scan. Positions are relative to object. The default for object is the current buffer.
nil
if at least one character between
start and end does not have a property
prop with value value. More precisely, it
returns the position of the first such character. Otherwise, it
returns nil
. The optional fifth argument, object, specifies the string or buffer to scan. Positions are relative to object. The default for object is the current buffer.
Here is a table of text property names that have special built-in meanings. The following sections list a few additional special property names that control filling and property inheritance. All other names have no standard meaning, and you can use them as you like.
category
category
property, we call it the category of the character. It
should be a symbol. The properties of the symbol serve as defaults
for the properties of the character.
face
face
to
control the font and color of text. Its value is a face name or a
list of face names. See section Faces, for more information. If the
property value is a list, elements may also have the form
(foreground-color . color-name)
or
(background-color . color-name)
. These
elements specify just the foreground color or just the background
color; therefore, there is no need to create a face for each color
that you want to use. See section Font
Lock Mode, for information on how to update face
properties automatically based on the contents of the text.
mouse-face
mouse-face
is used instead of face
when
the mouse is on or near the character. For this purpose, "near"
means that all text between the character and where the mouse is
have the same mouse-face
property value.
local-map
local-map
property. The property's value for the character after point, if
non-nil
, is used for key lookup instead of the
buffer's local map. If the property value is a symbol, the symbol's
function definition is used as the keymap. See section Active Keymaps.
syntax-table
syntax-table
property overrides what the
syntax table says about this particular character. See section Syntax Properties.
read-only
read-only
, then modifying that character is not
allowed. Any command that would do so gets an error. Insertion next
to a read-only character is an error if inserting ordinary text
there would inherit the read-only
property due to
stickiness. Thus, you can control permission to insert next to
read-only text by controlling the stickiness. See section Stickiness of Text Properties. Since
changing properties counts as modifying the buffer, it is not
possible to remove a read-only
property unless you
know the special trick: bind inhibit-read-only
to a
non-nil
value and then remove the property. See
section Read-Only Buffers.
invisible
nil
invisible
property can make a character invisible on
the screen. See section Invisible
Text, for details.
intangible
nil
intangible
properties, then you cannot place point
between them. If you try to move point forward into the group,
point actually moves to the end of the group. If you try to move
point backward into the group, point actually moves to the start of
the group. When the variable
inhibit-point-motion-hooks
is non-nil
,
the intangible
property is ignored.
modification-hooks
modification-hooks
, then
its value should be a list of functions; modifying that character
calls all of those functions. Each function receives two arguments:
the beginning and end of the part of the buffer being modified.
Note that if a particular modification hook function appears on
several characters being modified by a single primitive, you can't
predict how many times the function will be called.
insert-in-front-hooks
insert-behind-hooks
insert-in-front-hooks
property of the following
character and in the insert-behind-hooks
property of
the preceding character. These functions receive two arguments, the
beginning and end of the inserted text. The functions are called
after the actual insertion takes place. See also section
Change Hooks, for other hooks that
are called when you change text in a buffer.
point-entered
point-left
point-entered
and point-left
record hook functions that report motion of point. Each time point
moves, Emacs compares these two property values:
point-left
property of the character after the
old location, and
point-entered
property of the character after
the new location.
nil
) with two arguments: the old value of point, and
the new one. The same comparison is made for the characters before
the old and new locations. The result may be to execute two
point-left
functions (which may be the same function)
and/or two point-entered
functions (which may be the
same function). In any case, all the point-left
functions are called first, followed by all the
point-entered
functions. It is possible using
char-after
to examine characters at various positions
without moving point to those positions. Only an actual change in
the value of point runs these hook functions.
nil
, point-left
and
point-entered
hooks are not run, and the
intangible
property has no effect. Do not set this
variable globally; bind it with let
.
These text properties affect the behavior of the fill commands. They are used for representing formatted text. See section Filling, and section Margins for Filling.
hard
use-hard-newlines
is
non-nil
.
right-margin
left-margin
justification
Self-inserting characters normally take on the same properties as the preceding character. This is called inheritance of properties.
In a Lisp program, you can do insertion with inheritance or
without, depending on your choice of insertion primitive. The
ordinary text insertion functions such as insert
do
not inherit any properties. They insert text with precisely the
properties of the string being inserted, and no others. This is
correct for programs that copy text from one context to
another--for example, into or out of the kill ring. To insert with
inheritance, use the special primitives described in this section.
Self-inserting characters inherit properties because they work
using these primitives.
When you do insertion with inheritance, which
properties are inherited depends on two specific properties:
front-sticky
and rear-nonsticky
.
Insertion after a character inherits those of its properties that are rear-sticky. Insertion before a character inherits those of its properties that are front-sticky. By default, a text property is rear-sticky but not front-sticky. Thus, the default is to inherit all the properties of the preceding character, and nothing from the following character. You can request different behavior by specifying the stickiness of certain properties.
If a character's front-sticky
property is
t
, then all its properties are front-sticky. If the
front-sticky
property is a list, then the sticky
properties of the character are those whose names are in the list.
For example, if a character has a front-sticky
property whose value is (face read-only)
, then
insertion before the character can inherit its face
property and its read-only
property, but no
others.
The rear-nonsticky
works the opposite way. Every
property is rear-sticky by default, so the
rear-nonsticky
property says which properties are
not rear-sticky. If a character's
rear-nonsticky
property is t
, then none
of its properties are rear-sticky. If the
rear-nonsticky
property is a list, properties are
rear-sticky unless their names are in the list.
When you insert text with inheritance, it inherits all the rear-sticky properties of the preceding character, and all the front-sticky properties of the following character. The previous character's properties take precedence when both sides offer different sticky values for the same property.
Here are the functions that insert text with inheritance of properties:
insert
, but
inherit any sticky properties from the adjoining text.
insert-before-markers
, but inherit any sticky
properties from the adjoining text.
See section Inserting Text, for the ordinary insertion functions which do not inherit.
You can save text properties in files (along with the text itself), and restore the same text properties when visiting or inserting the files, using these two hooks:
write-region
to run to encode
text properties in some fashion as annotations to the text being
written in the file. See section Writing to Files. Each function in the list is called with two arguments: the start and end of the region to be written. These functions should not alter the contents of the buffer. Instead, they should return lists indicating annotations to write in the file in addition to the text in the buffer.
Each function should return a list of elements of the form
(position . string)
, where
position is an integer specifying the relative position
within the text to be written, and string is the
annotation to add there.
Each list returned by one of these functions must be already
sorted in increasing order by position. If there is more
than one function, write-region
merges the lists
destructively into one sorted list.
When write-region
actually writes the text from the
buffer to the file, it intermixes the specified annotations at the
corresponding positions. All this takes place without modifying the
buffer.
insert-file-contents
to call after
inserting a file's contents. These functions should scan the
inserted text for annotations, and convert them to the text
properties they stand for. Each function receives one argument, the length of the inserted text; point indicates the start of that text. The function should scan that text for annotations, delete them, and create the text properties that the annotations specify. The function should return the updated length of the inserted text, as it stands after those changes. The value returned by one function becomes the argument to the next function.
These functions should always return with point at the beginning of the inserted text.
The intended use of after-insert-file-functions
is
for converting some sort of textual annotations into actual text
properties. But other uses may be possible.
We invite users to write Lisp programs to store and retrieve text properties in files, using these hooks, and thus to experiment with various data formats and find good ones. Eventually we hope users will produce good, general extensions we can install in Emacs.
We suggest not trying to handle arbitrary Lisp objects as text property names or values--because a program that general is probably difficult to write, and slow. Instead, choose a set of possible data types that are reasonably flexible, and not too hard to encode.
See section File Format Conversion, for a related feature.
Instead of computing text properties for all the text in the buffer, you can arrange to compute the text properties for parts of the text when and if something depends on them.
The primitive that extracts text from the buffer along with its
properties is buffer-substring
. Before examining the
properties, this function runs the abnormal hook
buffer-access-fontify-functions
.
buffer-substring
copies the text and text properties
for a portion of the buffer, it calls all the functions in this
list. Each of the functions receives two arguments that specify the
range of the buffer being accessed. (The buffer itself is always
the current buffer.)
The function buffer-substring-no-properties
does
not call these functions, since it ignores text properties
anyway.
In order to prevent the hook functions from being called more
than once for the same part of the buffer, you can use the variable
buffer-access-fontified-property
.
nil
, it is a symbol which is used as a text
property name. A non-nil
value for that text property
means, "the other text properties for this character have already
been computed." If all the characters in the range specified for
buffer-substring
have a non-nil
value for
this property, buffer-substring
does not call the
buffer-access-fontify-functions
functions. It assumes
these characters already have the right text properties, and just
copies the properties they already have.
The normal way to use this feature is that the
buffer-access-fontify-functions
functions add this
property, as well as others, to the characters they operate on.
That way, they avoid being called over and over for the same
text.
There are two ways to set up clickable text in a buffer. There are typically two parts of this: to make the text highlight when the mouse is over it, and to make a mouse button do something when you click it on that part of the text.
Highlighting is done with the mouse-face
text
property. Here is an example of how Dired does it:
(condition-case nil (if (dired-move-to-filename) (put-text-property (point) (save-excursion (dired-move-to-end-of-filename) (point)) 'mouse-face 'highlight)) (error nil))
The first two arguments to put-text-property
specify the beginning and end of the text.
The usual way to make the mouse do something when you click it
on this text is to define mouse-2
in the major mode's
keymap. The job of checking whether the click was on clickable text
is done by the command definition. Here is how Dired does it:
(defun dired-mouse-find-file-other-window (event) "In dired, visit the file or directory name you click on." (interactive "e") (let (file) (save-excursion (set-buffer (window-buffer (posn-window (event-end event)))) (save-excursion (goto-char (posn-point (event-end event))) (setq file (dired-get-filename)))) (select-window (posn-window (event-end event))) (find-file-other-window (file-name-sans-versions file t))))
The reason for the outer save-excursion
construct
is to avoid changing the current buffer; the reason for the inner
one is to avoid permanently altering point in the buffer you click
on. In this case, Dired uses the function
dired-get-filename
to determine which file to visit,
based on the position found in the event.
Instead of defining a mouse command for the major mode, you can
define a key binding for the clickable text itself, using the
local-map
text property:
(let ((map (make-sparse-keymap))) (define-key-binding map [mouse-2] 'operate-this-button) (put-text-property (point) (save-excursion (dired-move-to-end-of-filename) (point)) 'local-map map))
This method makes it possible to define different commands for various clickable pieces of text. Also, the major mode definition (or the global definition) remains available for the rest of the text in the buffer.
Some editors that support adding attributes to text in the buffer do so by letting the user specify "intervals" within the text, and adding the properties to the intervals. Those editors permit the user or the programmer to determine where individual intervals start and end. We deliberately provided a different sort of interface in Emacs Lisp to avoid certain paradoxical behavior associated with text modification.
If the actual subdivision into intervals is meaningful, that means you can distinguish between a buffer that is just one interval with a certain property, and a buffer containing the same text subdivided into two intervals, both of which have that property.
Suppose you take the buffer with just one interval and kill part of the text. The text remaining in the buffer is one interval, and the copy in the kill ring (and the undo list) becomes a separate interval. Then if you yank back the killed text, you get two intervals with the same properties. Thus, editing does not preserve the distinction between one interval and two.
Suppose we "fix" this problem by coalescing the two intervals when the text is inserted. That works fine if the buffer originally was a single interval. But suppose instead that we have two adjacent intervals with the same properties, and we kill the text of one interval and yank it back. The same interval-coalescence feature that rescues the other case causes trouble in this one: after yanking, we have just one interval. One again, editing does not preserve the distinction between one interval and two.
Insertion of text at the border between intervals also raises questions that have no satisfactory answer.
However, it is easy to arrange for editing to behave consistently for questions of the form, "What are the properties of this character?" So we have decided these are the only questions that make sense; we have not implemented asking questions about where intervals start or end.
In practice, you can usually use the text property search functions in place of explicit interval boundaries. You can think of them as finding the boundaries of intervals, assuming that intervals are always coalesced whenever possible. See section Text Property Search Functions.
Emacs also provides explicit intervals as a presentation feature; see section Overlays.
The following functions replace characters within a specified region based on their character codes.
If noundo is non-nil
,
then subst-char-in-region
does not record the change
for undo and does not mark the buffer as modified. This feature is
used for controlling selective display (see section Selective Display).
subst-char-in-region
does not move point and
returns nil
.
---------- Buffer: foo ---------- This is the contents of the buffer before. ---------- Buffer: foo ---------- (subst-char-in-region 1 20 ?i ?X) => nil ---------- Buffer: foo ---------- ThXs Xs the contents of the buffer before. ---------- Buffer: foo ----------
The translation table table is a string; (aref
table ochar)
gives the translated
character corresponding to ochar. If the length of
table is less than 256, any characters with codes larger
than the length of table are not altered by the
translation.
The return value of translate-region
is the number
of characters that were actually changed by the translation. This
does not count characters that were mapped into themselves in the
translation table.
A register is a sort of variable used in Emacs editing that can hold a variety of different kinds of values. Each register is named by a single character. All ASCII characters and their meta variants (but with the exception of C-g) can be used to name registers. Thus, there are 255 possible registers. A register is designated in Emacs Lisp by the character that is its name.
(name .
contents)
. Normally, there is one element for
each Emacs register that has been used. The object name is a character (an integer) identifying the register.
The contents of a register can have several possible types:
insert-register
finds a number in the register, it converts the number to
decimal.
(window-configuration
position)
(frame-configuration
position)
The functions in this section return unpredictable values unless otherwise stated.
nil
if it
has no contents.
Normally, this command puts point before the inserted text, and
the mark after it. However, if the optional second argument
beforep is non-nil
, it puts the mark before
and point after. You can pass a non-nil
second
argument beforep to this function interactively by
supplying any prefix argument.
If the register contains a rectangle, then the rectangle is inserted with its upper left corner at point. This means that text is inserted in the current line and underneath it on successive lines.
If the register contains something other than saved text (a string) or a rectangle (a list), currently useless things happen. This may be changed in the future.
This subroutine is used by the transposition commands.
Normally, transpose-regions
relocates markers with
the transposed text; a marker previously positioned within one of
the two transposed portions moves along with that portion, thus
remaining between the same two characters in their new position.
However, if leave-markers is non-nil
,
transpose-regions
does not do this--it leaves all
markers unrelocated.
These hook variables let you arrange to take notice of all changes in all buffers (or in a particular buffer, if you make them buffer-local). See also section Properties with Special Meanings, for how to detect changes to specific parts of the text.
The functions you use in these hooks should save and restore the match data if they do anything that uses regular expressions; otherwise, they will interfere in bizarre ways with the editing operations that call them.
The length of the old text is the difference between the buffer positions before and after that text as it was before the change. As for the changed text, its length is simply the difference between the first two arguments.
If a program makes several text changes in the same area of the
buffer, using the macro combine-after-change-calls
around that part of the program can make it run considerably faster
when after-change hooks are in use. When the after-change hooks are
ultimately called, the arguments specify a portion of the buffer
including all of the changes made within the
combine-after-change-calls
body.
Warning: You must not alter the values of
after-change-functions
and
after-change-function
within the body of a
combine-after-change-calls
form.
Note: If the changes you combine occur in widely scattered parts of the buffer, this will still work, but it is not advisable, because it may lead to inefficient behavior for some change hook functions.
nil
for no function). It is called just like the
functions in before-change-functions
.
nil
for no function). It is called just like the
functions in after-change-functions
.
The four variables above are temporarily bound to
nil
during the time that any of these functions is
running. This means that if one of these functions changes the
buffer, that change won't run these functions. If you do want a
hook function to make changes that run these functions, make it
bind these variables back to their usual values.
One inconvenient result of this protective feature is that you
cannot have a function in after-change-functions
or
before-change-functions
which changes the value of
that variable. But that's not a real limitation. If you want those
functions to change the list of functions to run, simply add one
fixed function to the hook, and code that function to look in
another variable for other functions to call. Here is an
example:
(setq my-own-after-change-functions nil) (defun indirect-after-change-function (beg end len) (let ((list my-own-after-change-functions)) (while list (funcall (car list) beg end len) (setq list (cdr list))))) (add-hooks 'after-change-functions 'indirect-after-change-function)