\input ../texinfo
XEmacs Manual
July 1994 (General Public License upgraded, January 1991)
Richard Stallman
and
Lucid, Inc. Copyright (C) 1985, 1986, 1988 Richard M. Stallman.
Copyright (C) 1991, 1992, 1993, 1994 Lucid, Inc.
Copyright (C) 1993, 1994 Sun Microsystems, Inc.
Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies.
Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided also that the sections entitled "The GNU Manifesto", "Distribution" and "GNU General Public License" are included exactly as in the original, and provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one.
Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that the sections entitled "The GNU Manifesto", "Distribution" and "GNU General Public License" may be included in a translation approved by the author instead of in the original English.
This manual documents the use and simple customization of the Emacs editor. The reader is not expected to be a programmer to use this editor, and simple customizations do not require programming skills either. Users who are not interested in customizing Emacs can ignore the scattered customization hints.
This document is primarily a reference manual, but it can also be used as a primer. However, if you are new to Emacs, consider using the on-line, learn-by-doing tutorial, which you get by running Emacs and typing C-h t. With it, you learn Emacs by using Emacs on a specially designed file which describes commands, tells you when to try them, and then explains the results you see. Using the tutorial gives a more vivid introduction than the printed manual.
On first reading, just skim chapters one and two, which describe the notational conventions of the manual and the general appearance of the Emacs display screen. Note which questions are answered in these chapters, so you can refer back later. After reading chapter four you should practice the commands there. The next few chapters describe fundamental techniques and concepts that are used constantly. You need to understand them thoroughly, experimenting with them if necessary.
To find the documentation on a particular command, look in the index. Keys (character commands) and command names have separate indexes. There is also a glossary, with a cross reference for each term.
This manual comes in two forms: the published form and the Info form. The Info form is for on-line perusal with the INFO program; it is distributed along with GNU Emacs. Both forms contain substantially the same text and are generated from a common source file, which is also distributed along with GNU Emacs.
GNU Emacs is a member of the Emacs editor family. There are many Emacs editors, all sharing common principles of organization. For information on the underlying philosophy of Emacs and the lessons learned from its development, write for a copy of AI memo 519a, "Emacs, the Extensible, Customizable Self-Documenting Display Editor", to Publications Department, Artificial Intelligence Lab, 545 Tech Square, Cambridge, MA 02139, USA. At last report they charge $2.25 per copy. Another useful publication is LCS TM-165, "A Cookbook for an Emacs", by Craig Finseth, available from Publications Department, Laboratory for Computer Science, 545 Tech Square, Cambridge, MA 02139, USA. The price today is $3.
This manual is for GNU Emacs installed on UNIX systems. GNU Emacs can also be used on VMS systems, which have different file name syntax and don't support all GNU Emacs features. A VMS edition of this manual may appear in the future.
Copyright (C) 1989 Free Software Foundation, Inc. 675 Mass Ave, Cambridge, MA 02139, USA Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.
The license agreements of most software companies try to keep users at the mercy of those companies. By contrast, our General Public License is intended to guarantee your freedom to share and change free software--to make sure the software is free for all its users. The General Public License applies to the Free Software Foundation's software and to any other program whose authors commit to using it. You can use it for your programs, too.
When we speak of free software, we are referring to freedom, not price. Specifically, the General Public License is designed to make sure that you have the freedom to give away or sell copies of free software, that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs; and that you know you can do these things.
To protect your rights, we need to make restrictions that forbid anyone to deny you these rights or to ask you to surrender the rights. These restrictions translate to certain responsibilities for you if you distribute copies of the software, or if you modify it.
For example, if you distribute copies of a such a program, whether gratis or for a fee, you must give the recipients all the rights that you have. You must make sure that they, too, receive or can get the source code. And you must tell them their rights.
We protect your rights with two steps: (1) copyright the software, and (2) offer you this license which gives you legal permission to copy, distribute and/or modify the software.
Also, for each author's protection and ours, we want to make certain that everyone understands that there is no warranty for this free software. If the software is modified by someone else and passed on, we want its recipients to know that what they have is not the original, so that any problems introduced by others will not reflect on the original authors' reputations.
The precise terms and conditions for copying, distribution and modification follow.
If you develop a new program, and you want it to be of the greatest possible use to humanity, the best way to achieve this is to make it free software which everyone can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest to attach them to the start of each source file to most effectively convey the exclusion of warranty; and each file should have at least the "copyright" line and a pointer to where the full notice is found.
one line to give the program's name and a brief idea of what it does. Copyright (C) 19yy name of author This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 1, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
Also add information on how to contact you by electronic and paper mail.
If the program is interactive, make it output a short notice like this when it starts in an interactive mode:
Gnomovision version 69, Copyright (C) 19yy name of author Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'. This is free software, and you are welcome to redistribute it under certain conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the appropriate parts of the General Public License. Of course, the commands you use may be called something other than `show w' and `show c'; they could even be mouse-clicks or menu items--whatever suits your program.
You should also get your employer (if you work as a programmer) or your school, if any, to sign a "copyright disclaimer" for the program, if necessary. Here a sample; alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the program `Gnomovision' (a program to direct compilers to make passes at assemblers) written by James Hacker. signature of Ty Coon, 1 April 1989 Ty Coon, President of Vice
That's all there is to it!
XEmacs is free; this means that everyone is free to use it and free to redistribute it on a free basis. XEmacs is not in the public domain; it is copyrighted and there are restrictions on its distribution, but these restrictions are designed to permit everything that a good cooperating citizen would want to do. What is not allowed is to try to prevent others from further sharing any version of GNU Emacs that they might get from you. The precise conditions are found in the GNU General Public License that comes with Emacs and also appears following this section.
The easiest way to get a copy of GNU Emacs is from someone else who has it. You need not ask for permission to do so, or tell any one else; just copy it.
If you have access to the Internet, you can get the latest version of XEmacs from anonymous FTP server, `cs.uiuc.edu'. It can also be found at numerous other archive sites around the world; check the file `etc/DISTRIB' in a XEmacs distribution for the latest known list.
XEmacs, the version of Emacs described in this manual, is not distributed by the Free Software Foundation. You can, however, still get other versions of Emacs from there.
If you have access to the Internet, you can get the latest distribution version of GNU Emacs from host `prep.ai.mit.edu' using anonymous login. See the file `/pub/gnu/GNUinfo/FTP' on that host to find out about your options for copying and which files to use.
You may also receive GNU Emacs when you buy a computer. Computer manufacturers are free to distribute copies on the same terms that apply to everyone else. These terms require them to give you the full sources, including whatever changes they may have made, and to permit you to redistribute the GNU Emacs received from them under the usual terms of the General Public License. In other words, the program must be free for you when you get it, not just free for the manufacturer.
If you cannot get a copy in any of those ways, you can order one from the Free Software Foundation. Though Emacs itself is free, the Free Software Foundation's distribution service is not. An order form is included in the file `etc/DISTRIB' in the Emacs distribution. For further information, write to:
Free Software Foundation 675 Mass Ave. Cambridge, MA 02139 USA
The income from distribution fees goes to support the foundation's purpose: the development of more free software to distribute just like GNU Emacs.
If you find GNU Emacs useful, please send a donation to the Free Software Foundation. This will help support development of the rest of the GNU system, and other useful software beyond that. Your donation is tax deductible.
You are reading about GNU Emacs, the GNU incarnation of the advanced, self-documenting, customizable, extensible real-time display editor Emacs. (The `G' in `GNU' is not silent.)
We say that Emacs is a display editor because normally the text being edited is visible on the screen and is updated automatically as you type. See section The Emacs Screen.
We call Emacs a real-time editor because the display is updated very frequently, usually after each character or pair of characters you type. This minimizes the amount of information you must keep in your head as you edit. See section Basic Editing Commands.
We call Emacs advanced because it provides facilities that go beyond simple insertion and deletion: filling of text; automatic indentation of programs; viewing two or more files at once; and dealing in terms of characters, words, lines, sentences, paragraphs, and pages, as well as expressions and comments in several different programming languages. It is much easier to type one command meaning "go to the end of the paragraph" than to find that spot with simple cursor keys.
Self-documenting means that at any time you can type a special character, Control-h, to find out what your options are. You can also use C-h to find out what a command does, or to find all the commands relevant to a topic. See section Help.
Customizable means you can change the definitions of Emacs commands. For example, if you use a programming language in which comments start with `<**' and end with `**>', you can tell the Emacs comment manipulation commands to use those strings (see section Manipulating Comments). Another sort of customization is rearrangement of the command set. For example, you can set up the four basic cursor motion commands (up, down, left and right) on keys in a diamond pattern on the keyboard if you prefer. See section Customization.
Extensible means you can go beyond simple customization and write entirely new commands, programs in the Lisp language to be run by Emacs's own Lisp interpreter. Emacs is an "on-line extensible" system: it is divided into many functions that call each other. You can redefine any function in the middle of an editing session and replace any part of Emacs without making a separate copy of all of Emacs. Most of the editing commands of Emacs are written in Lisp; the few exceptions could have been written in Lisp but are written in C for efficiency. Only a programmer can write an extension to Emacs, but anybody can use it afterward.
The terminology used here reflects the fact that Emacs was developed on environments that had only one window, which took up the entire screen.
Each Emacs screen displays a variety of information:
You can subdivide the Emacs screen into multiple text windows, and use each window for a different file (see section Multiple Windows). Multiple Emacs windows are tiled vertically on the Emacs screen. The upper Emacs window is separated from the lower window by its mode line.
When there are multiple, tiled Emacs windows on a single Emacs screen, the Emacs window receiving input from the keyboard has the keyboard focus and is called the selected window. The selected window contains the cursor, which indicates the insertion point. If you are working in an environment that permits multiple Emacs screens, and you move the focus from one Emacs screen into another, the selected window is the one that was last selected in that screen.
The same text can be displayed simultaneously in several Emacs windows, which can be in different Emacs screens. If you alter the text in an Emacs buffer by editing it in one Emacs window, the changes are visible in all Emacs windows containing that buffer.
When Emacs is running, the cursor shows the location at which editing commands will take effect. This location is called point. You can use keystrokes or the mouse cursor to move point through the text and edit the text at different places.
While the cursor appears to point at a character, you should think of point as between two characters: it points before the character on which the cursor appears. Sometimes people speak of "the cursor" when they mean "point," or speak of commands that move point as "cursor motion" commands.
Each Emacs screen has only one cursor. When output is in progress, the cursor must appear where the typing is being done. This does not mean that point is moving. It is only that Emacs has no way to show you the location of point except when the terminal is idle.
If you are editing several files in Emacs, each file has its own point location. A file that is not being displayed remembers where point is. Point becomes visible at the correct location when you look at the file again.
When there are multiple text windows, each window has its own point location. The cursor shows the location of point in the selected window. The visible cursor also shows you which window is selected. If the same buffer appears in more than one window, point can be moved in each window independently.
The term `point' comes from the character `.', which was the command in TECO (the language in which the original Emacs was written) for accessing the value now called `point'.
The line at the bottom of the screen (below the mode line) is the echo area. Emacs uses this area to communicate with the user:
Each text window's last line is a mode line which describes what is going on in that window. When there is only one text window, the mode line appears right above the echo area. The mode line is in inverse video if the terminal supports that, starts and ends with dashes, and contains text like `Emacs: something'.
If a mode line has something else in place of `Emacs: something', the window above it is in a special subsystem such as Dired. The mode line then indicates the status of the subsystem.
Normally, the mode line has the following appearance:
--ch-Emacs: buf (major minor)----pos------
This gives information about the buffer being displayed in the window: the buffer's name, what major and minor modes are in use, whether the buffer's text has been changed, and how far down the buffer you are currently looking.
ch contains two stars (`**') if the text in the buffer has been edited (the buffer is "modified"), or two dashes (`--') if the buffer has not been edited. Exception: for a read-only buffer, it is `%%'.
buf is the name of the window's chosen buffer. The chosen buffer in the selected window (the window that the cursor is in) is also Emacs's selected buffer, the buffer in which editing takes place. When we speak of what some command does to "the buffer", we mean the currently selected buffer. See section Using Multiple Buffers.
pos tells you whether there is additional text above the top of the screen or below the bottom. If your file is small and it is completely visible on the screen, pos is `All'. Otherwise, pos is `Top' if you are looking at the beginning of the file, `Bot' if you are looking at the end of the file, or `nn%', where nn is the percentage of the file above the top of the screen.
major is the name of the major mode in effect in the buffer. At any time, each buffer is in one and only one major mode. The available major modes include Fundamental mode (the least specialized), Text mode, Lisp mode, and C mode. See section Major Modes, for details on how the modes differ and how you select one.
minor is a list of some of the minor modes that are turned on
in the window's chosen buffer. For example, `Fill' means that Auto
Fill mode is on. Abbrev
means that Word Abbrev mode is on.
Ovwrt
means that Overwrite mode is on. See section Minor Modes, for more
information. `Narrow' means that the buffer being displayed has
editing restricted to only a portion of its text. This is not really a
minor mode, but is like one. See section Narrowing. Def
means that a
keyboard macro is being defined. See section Keyboard Macros.
Some buffers display additional information after the minor modes. For example, Rmail buffers display the current message number and the total number of messages. Compilation buffers and Shell mode display the status of the subprocess.
If Emacs is currently inside a recursive editing level, square brackets (`[...]') appear around the parentheses that surround the modes. If Emacs is in one recursive editing level within another, double square brackets appear, and so on. Since information on recursive editing applies to Emacs in general and not to any one buffer, the square brackets appear in every mode line on the screen or not in any of them. See section Recursive Editing Levels.
Emacs can optionally display the time and system load in all mode lines. To enable this feature, type M-x display-time. The information added to the mode line usually appears after the file name, before the mode names and their parentheses. It looks like this:
hh:mmpm l.ll [d]
(Some fields may be missing if your operating system cannot support them.) hh and mm are the hour and minute, followed always by `am' or `pm'. l.ll is the average number of running processes in the whole system recently. d is an approximate index of the ratio of disk activity to CPU activity for all users.
The word `Mail' appears after the load level if there is mail for you that you have not read yet.
Customization note: the variable mode-line-inverse-video
controls whether the mode line is displayed in inverse video (assuming
the terminal supports it); nil
means no inverse video. The
default is t
. For X screens, simply set the foreground and
background colors appropriately.
XEmacs can be used with the X Window System and a window manager like MWM or TWM. In that case, the X window manager opens, closes, and resizes Emacs screens. You use the window manager's mouse gestures to perform the operations. Consult your window manager guide or reference manual for information on manipulating X windows.
When you are working under X, each X window (that is, each Emacs screen) has a menu bar for mouse-controlled operations (see section XEmacs Pull-down Menus).
Emacs under X is also a multi-screen Emacs. You can use the New Screen menu item from the File menu to create a new Emacs screen in a new X window from the same process. The different screens will share the same buffer list, but you can look at different buffers in the different screens.
The function find-file-other-screen
is just like find-file
,
but picks a new screen to display the buffer in first. If there is only
one screen, then a new screen is created; otherwise some other screen
is used. This is normally bound to C-x 5 C-f, and is what
the Open File, New Screen menu item does.
The function switch-to-buffer-other-screen
is just
like switch-to-buffer
, but picks a new screen to display the
buffer in first. If there is only one screen, then a new screen is created;
otherwise some other screen is used. This is normally bound to
C-x 5 b.
You can specify a different default screen size other than the one
provided. Use the variable default-screen-alist
, which is an
alist of default values for screen creation other than the first one.
These may be set in your init file, like this:
(setq default-screen-alist '((width . 80) (height . 55)))
For values specific to the first Emacs screen, you must use X resources.
The variable x-screen-defaults
takes an alist of default screen
creation parameters for X window screens. These override what is
specified in `~/.Xdefaults' but are overridden by the arguments to
the particular call to x-create-screen
.
When you create a new screen, the variable create-screen-hook
is called with one argument, the screen just created.
If you want to close one or more of the X windows you created using New Screen, use the Delete Screen menu item from the File menu.
If you are working with multiple screens, some special information applies:
screen-title-format
and
screen-icon-title-format
determine the title of the screen and the
title of the icon that results if you shrink the screen.
auto-lower-screen
and
auto-raise-screen
position a screen. If true,
auto-lower-screen
lowers a screen to the bottom when it is no longer
selected. If true, auto-raise-screen
raises a screen to
the top when it is selected. Under X, most ICCCM-compliant window managers
will have options to do this for you, but these variables are provided in
case you are using a broken window manager.
This chapter discusses the character set Emacs uses for input commands
and inside files. You have already learned that the more frequently
used Emacs commands are bound to keys. For example, Control-f is
bound to forward-char
. The following issues are covered:
You will also learn how to customize existing key bindings and create new ones.
Earlier versions of GNU Emacs used only the ASCII character set, which defines 128 different character codes. Some of these codes are assigned graphic symbols like `a' and `='; the rest are control characters, such as Control-a (also called C-a). C-a means you hold down the CTRL key and then press a.
Keybindings in XEmacs are no longer restricted to the set of keystrokes that can be represented in ASCII. Emacs can now tell the difference between, for example, Control-h, Control-Shift-h, and Backspace.
A keystroke is like a piano chord: you get it by simultaneously striking several keys. To be more precise, a keystroke consists of a possibly empty set of modifiers followed by a single keysym. The set of modifiers is small; it consists of Control, Meta, Super, Hyper, and Shift.
The rest of the keys on your keyboard, along with the mouse buttons, make up the set of keysyms. A keysym is usually what is printed on the keys on your keyboard. Here is a table of some of the symbolic names for keysyms:
Use the variable keyboard-translate-table
only if you are on a
dumb tty, as it cannot handle input that cannot be represented as ASCII.
The value of this variable is a string used as a translate table for
keyboard input or nil
. Each character is looked up in this
string and the contents used instead. If the string is of length
n
, character codes N
and up are untranslated. If you are
running Emacs under X, you should do the translations with the
xmodmap
program instead.
XEmacs represents keystrokes as lists. Each list consists of
an arbitrary combination of modifiers followed by a single keysym at the
end of the list. If the keysym corresponds to an ASCII character, you
can use its character code. (A keystroke may also be represented by an
event object, as returned by the read-key-sequence
function;
non-programmers need not worry about this.)
The following table gives some examples of how to list representations for keystrokes. Each list consists of sets of modifiers followed by keysyms:
A complete key sequence is a sequence of keystrokes that Emacs understands as a unit. Key sequences are significant because you can bind them to commands. Note that not all sequences of keystrokes are possible key sequences. In particular, the initial keystrokes in a key sequence must make up a prefix key sequence.
Emacs represents a key sequence as a vector of keystrokes. Thus, the schematic representation of a complete key sequence is as follows:
[(modifier .. modifer keysym) ... (modifier .. modifier keysym)]
Here are some examples of complete key sequences:
A prefix key sequence is the beginning of a series of longer sequences that are valid key sequences; adding any single keystroke to the end of a prefix results in a valid key sequence. For example, control-x is standardly defined as a prefix. Thus there is a two-character key sequence starting with C-x for each valid keystroke, giving numerous possibilities. Here are some samples:
Adding one character to a prefix key does not have to form a complete key. It could make another, longer prefix. For example, [(control x) (\4)] is itself a prefix that leads to any number of different three-character keys, including [(control x) (\4) (f)], [(control x) (\4) (b)] and so on. It would be possible to define one of those three-character sequences as a prefix, creating a series of four-character keys, but we did not define any of them this way.
By contrast, the two-character sequence [(control f) (control k)] is not a key, because the (control f) is a complete key sequence in itself. You cannot give [(control f (control k)] an independent meaning as a command while (control f) is a complete sequence, because Emacs would understand C-f C-k as two commands.
The predefined prefix key sequences in Emacs are (control c), (control x), (control h), [(control x) (\4)], and escape. You can customize Emacs and could make new prefix keys or eliminate the default key sequences. See section Customizing Key Bindings. For example, if you redefine (control f) as a prefix, [(control f) (control k)] automatically becomes a valid key sequence (complete, unless you define it as a prefix as well). Conversely, if you remove the prefix definition of [(control x) (\4)], [(control x) (\4) (f)] (or [(control x) (\4) anything]) is no longer a valid key sequence.
Note that the above paragraphs uses \4 instead of simply 4, because \4 is the symbol whose name is "4", and plain 4 is the integer 4, which would have been interpreted as the ASCII value. Another way of representing the symbol whose name is "4" is to write ?4, which would be interpreted as the number 52, which is the ASCII code for the character "4". We could therefore actually have written 52 directly, but that is far less clear.
[(control c) (control c)]
[(meta control c)]
If there is no META key, you can still type Meta characters using two-character sequences starting with ESC. To enter M-a, you could type ESC a. To enter C-M-a, you would type ESC C-a. ESC is allowed on terminals with Meta keys, too, in case you have formed a habit of using it.
If you are running under X and do not have a META key, it is possible to reconfigure some other key to be a META key. See section Assignment of the SUPER and HYPER Keys.
Emacs believes the terminal has a META key if the variable
meta-flag
is non-nil
. Normally this is set automatically
according to the termcap entry for your terminal type. However, sometimes
the termcap entry is wrong, and then it is useful to set this variable
yourself. See section Variables, for how to do this.
Note: If you are running under the X window system, the setting of
the meta-flag
variable is irrelevant.
Most keyboards do not, by default, have SUPER or HYPER
modifier keys. Under X, you can simulate the SUPER or
HYPER key if you want to bind keys to sequences using super
and hyper. You can use the xmodmap
program to do this.
For example, to turn your CAPS-LOCK key into a SUPER key, do the following:
Create a file called ~/.xmodmap
. In this file, place the lines
remove Lock = Caps_Lock keysym Caps_Lock = Super_L add Mod2 = Super_L
The first line says that the key that is currently called Caps_Lock
should no longer behave as a "lock" key. The second line says that
this should now be called Super_L
instead. The third line says that
the key called Super_L
should be a modifier key, which produces the
Mod2
modifier.
To create a META or HYPER key instead of a SUPER key,
replace the word Super
above with Meta
or Hyper
.
Just after you start up X, execute the command xmodmap /.xmodmap
.
You can add this command to the appropriate initialization file to have
the command executed automatically.
If you have problems, see the documentation for the xmodmap
program. The X keyboard model is quite complicated, and explaining
it is beyond the scope of this manual. However, we reprint the
following description from the X Protocol document for your convenience:
A list of keysyms is associated with each keycode. If that list
(ignoring trailing NoSymbol
entries) is a single keysym `K',
then the list is treated as if it were the list
"K NoSymbol K NoSymbol"
. If the list (ignoring trailing
NoSymbol
entries) is a pair of keysyms `K1 K2', then the
list is treated as if it were the list "K1 K2 K1 K2"
. If the
list (ignoring trailing NoSymbol
entries) is a triple of keysyms
`K1 K2 K3', then the list is treated as if it were the list
"K1 K2 K3 NoSymbol"
.
The first four elements of the list are split into two groups of keysyms. Group 1 contains the first and second keysyms; Group 2 contains third and fourth keysyms. Within each group, if the second element of the group is NoSymbol, then the group should be treated as if the second element were the same as the first element, except when the first element is an alphabetic keysym `K' for which both lowercase and uppercase forms are defined. In that case, the group should be treated as if the first element were the lowercase form of `K' and the second element were the uppercase form of `K'.
The standard rules for obtaining a keysym from a KeyPress event make use of only the Group 1 and Group 2 keysyms; no interpretation of other keysyms in the list is given here. (That is, the last four keysyms are unused.)
Which group to use is determined by modifier state. Switching between
groups is controlled by the keysym named Mode_switch
. Attach that
keysym to some keycode and attach that keycode to any one of the
modifiers Mod1 through Mod5. This modifier is called the group
modifier. For any keycode, Group 1 is used when the group modifier is
off, and Group 2 is used when the group modifier is on.
Within a group, which keysym to use is also determined by modifier
state. The first keysym is used when the Shift
and Lock
modifiers are off. The second keysym is used when the Shift
modifier is on, or when the Lock
modifier is on and the second
keysym is uppercase alphabetic, or when the Lock
modifier is on
and is interpreted as ShiftLock
. Otherwise, when the Lock
modifier is on and is interpreted as CapsLock
, the state of the
Shift
modifier is applied first to select a keysym,
but if that keysym is lower-case alphabetic, then the corresponding
upper-case keysym is used instead.
In addition to the above information on keysyms, we also provide the following description of modifier mapping from the InterClient Communications Conventions Manual:
X11 supports 8 modifier bits, of which 3 are pre-assigned to
Shift
, Lock
, and Control
. Each modifier bit is
controlled by the state of a set of keys, and these sets are specified
in a table accessed by GetModifierMapping()
and
SetModifierMapping()
.
A client needing to use one of the pre-assigned modifiers should assume
that the modifier table has been set up correctly to control these
modifiers. The Lock
modifier should be interpreted as Caps
Lock
or Shift Lock
according to whether the keycodes in its
controlling set include XK_Caps_Lock
or XK_Shift_Lock
.
Clients should determine the meaning of a modifier bit from the keysyms being used to control it.
A client needing to use an extra modifier, for example Meta
, should:
XK_Meta_L
or XK_Meta_R
, it should use that
modifier bit.
XK_Meta_L
or
XK_Meta_R
, it should select an unused modifier bit (one with
an empty controlling set) and:
XL_Meta_L
in its set of keysyms,
add that keycode to the set for the chosen modifier, and then:
XL_Meta_R
in its set of keysyms,
add that keycode to the set for the chosen modifier, and then:
Meta
.
This means that the Mod1
modifier does not necessarily mean
Meta
, although some applications (such as twm and emacs 18)
assume that. Any of the five unassigned modifier bits could mean
Meta
; what matters is that a modifier bit is generated by a
keycode which is bound to the keysym Meta_L
or Meta_R
.
Therefore, if you want to make a META key, the right way
is to make the keycode in question generate both a Meta
keysym
and some previously-unassigned modifier bit.
This section briefly discusses how characters are represented in Emacs buffers. See section Representing Key Sequences for information on representing key sequences to create key bindings.
ASCII graphic characters in Emacs buffers are displayed with their graphics. LFD is the same as a newline character; it is displayed by starting a new line. TAB is displayed by moving to the next tab stop column (usually every 8 spaces). Other control characters are displayed as a caret (`^') followed by the non-control version of the character; thus, C-a is displayed as `^A'. Non-ASCII characters 128 and up are displayed with octal escape sequences; thus, character code 243 (octal), also called M-# when used as an input character, is displayed as `\243'.
The variable ctl-arrow
may be used to alter this behavior.
See section Variables Controlling Display.
This manual is full of passages that tell you what particular keys do. But Emacs does not assign meanings to keys directly. Instead, Emacs assigns meanings to functions, and then gives keys their meanings by binding them to functions.
A function is a Lisp object that can be executed as a program. Usually
it is a Lisp symbol that has been given a function definition; every
symbol has a name, usually made of a few English words separated by
dashes, such as next-line
or forward-word
. It also has a
definition, which is a Lisp program. Only some functions can be the
bindings of keys; these are functions whose definitions use
interactive
to specify how to call them interactively. Such
functions are called commands, and their names are command
names. More information on this subject will appear in the GNU
Emacs Lisp Manual.
The bindings between keys and functions are recorded in various tables called keymaps. See section Customizing Key Bindings for more information on key sequences you can bind commands to. See section Keymaps for information on creating keymaps.
When we say "C-n moves down vertically one line" we are
glossing over a distinction that is irrelevant in ordinary use but is
vital in understanding how to customize Emacs. The function
next-line
is programmed to move down vertically. C-n
has this effect because it is bound to that function. If you rebind
C-n to the function forward-word
then C-n will move
forward by words instead. Rebinding keys is a common method of
customization.
The rest of this manual usually ignores this subtlety to keep
things simple. To give the customizer the information needed, we often
state the name of the command that really does the work in parentheses
after mentioning the key that runs it. For example, we will say that
"The command C-n (next-line
) moves point vertically
down," meaning that next-line
is a command that moves vertically
down and C-n is a key that is standardly bound to it.
While we are on the subject of information for customization only,
it's a good time to tell you about variables. Often the
description of a command will say, "To change this, set the variable
mumble-foo
." A variable is a name used to remember a value.
Most of the variables documented in this manual exist just to facilitate
customization: some command or other part of Emacs uses the variable
and behaves differently depending on its setting. Until you are interested in
customizing, you can ignore the information about variables. When you
are ready to be interested, read the basic information on variables, and
then the information on individual variables will make sense.
See section Variables.
If you are running XEmacs under X, a menu bar on top of the Emacs screen provides access to pull-down menus of file, edit, and help-related commands. The menus provide convenient shortcuts and an easy interface for novice users. They do not provide additions to the functionality available via key commands; you can still invoke commands from the keyboard as in previous versions of Emacs.
There are two ways of selecting an item from a pull-down menu:
If a command in the pull-down menu is not applicable in a given situation, the command is disabled and its name appears faded. You cannot invoke items that are faded. For example, most commands on the Edit menu appear faded until you select text on which they are to operate; after you select a block of text, edit commands are enabled. See section Selecting Text with the Mouse for information on using the mouse to select text. See section Using X Selections for related information.
There are also M-x equivalents for each menu item. To find the equivalent for any left-button menu item, do the following:
Describe Key
prompt.
Emacs displays the function associated with the menu item in a separate window, usually together with some documentation.
The File menu bar item contains the items New Screen, Open File..., Save Buffer, Save Buffer As..., Revert Buffer, Print Buffer, Delete Screen, Kill Buffer and Exit Emacs on the pull-down menu. If you select a menu item, Emacs executes the equivalent command.
find-file
(C-x
C-f).
insert-file
(C-x i).
save-buffer
(C-x C-s).
write-file
(C-x C-w).
revert-file
(M-x
revert-buffer).
kill-buffer
(C-x k), except that kill-buffer
prompts for the name of a buffer to kill.
print-buffer
(M-x print-buffer).
*scratch*
buffer. This
is like the Open File, New Screen... menu item, except that it does
not prompt for or load a file.
split-window-vertically
(C-x 2).
delete-other-windows
(C-x 1).
delete-window
(C-x 0).
save-buffers-kill-emacs
(C-x C-c). Before killing the
Emacs process, the system asks which unsaved buffers to save by going through
the list of all buffers in that Emacs process.The Edit pull-down menu contains the Undo, Cut, Copy, Paste, and Clear menu items. When you select a menu item, Emacs executes the equivalent command. Most commands on the Edit menu work on a block of text, the X selection. They appear faded until you select a block of text (activate a region) with the mouse. See section Using X Selections, see section Deletion and Killing, and see section Yanking for more information.
undo
(C-x u).
yank
command, because the Emacs kill ring and the X clipboard
selection are not the same thing. You can paste in text you
have placed in the clipboard using Copy or Cut. You can also
use Paste to insert text that was pasted into the clipboard from other
applications.
start-kbd-macro
(C-x ().
end-kbd-macro
(C-x )).
call-last-kbd-macro
(C-x e).
The Options pull-down menu contains the Read Only, Case Sensitive Search, Overstrike, Auto Delete Selection, Teach Extended Commands, Syntax Highlighting, Paren Highlighting, Font, Size, Weight, Buffers Menu Length..., Buffers Sub-Menus and Save Options menu items. When you select a menu item, Emacs executes the equivalent command. For some of the menu items, there are sub-menus which you will need to select.
toggle-read-only
(C-x C-q).
quoted-insert
(C-q).
.emacs
file to include the font-lock
mode so that when you select this item, the comments will be
displayed in one face, strings in another, reserved words in another,
and so on. When Fonts is selected, different parts of the program
will appear in different Fonts. When Colors is selected, then the
program will be displayed in different colors. Selecting None
causes the program to appear in just one Font and Color. Selecting
Less resets the Fonts and Colors to a fast, minimal set of
decorations. Selecting More resets the Fonts and Colors to a larger
set of decorations. For example, if Less is selected (which is the
default setting) then you might have all comments in green color.
Whereas, if More is selected then a function name in the comments
themselves might appear in a different Color or Font.
.emacs
file.
The Help Menu gives you access to Emacs Info and provides a menu equivalent for each of the choices you have when using C-h. See section Help for more information.
The Help menu also gives access to UNIX online manual pages via the UNIX Manual Page option.
You can customize any of the pull-down menus by adding or removing menu items and disabling or enabling existing menu items. The following functions are available:
Use the function add-menu
to add a new menu or submenu.
If a menu or submenu of the given name exists already, it is changed.
menu-path identifies the menu under which the new menu should be
inserted. It is a list of strings; for example, ("File")
names
the top-level File menu. ("File" "Foo")
names a hypothetical
submenu of File. If menu-path is nil
, the menu is
added to the menu bar itself.
menu-name is the string naming the menu to be added.
menu-items is a list of menu item descriptions. Each menu item should be a vector of three elements:
t
or nil
to indicate whether the item is selectable
The optional argument before is the name of the menu before which the new menu or submenu should be added. If the menu is already present, it is not moved.
The function add-menu-item
adds a menu item to the specified
menu, creating the menu first if necessary. If the named item already
exists, the menu remains unchanged.
menu-path identifies the menu into which the new menu item should
be inserted. It is a list of strings; for example, ("File")
names the top-level File menu. ("File" "Foo")
names a
hypothetical submenu of File.
item-name is the string naming the menu item to add.
function is the command to invoke when this menu item is selected.
If it is a symbol, it is invoked with call-interactively
, in the
same way that functions bound to keys are invoked. If it is a list, the
list is simply evaluated.
enabled-p controls whether the item is selectable or not.
It should be t
, nil
, or a form to evaluate to decide.
This form will be evaluated just before the menu is displayed, and
the menu item will be selectable if that form returns non-nil
.
For example, to make the rename-file
command available from the
File menu, use the following code:
(add-menu-item '("File") "Rename File" 'rename-file t)
To add a submenu of file management commands using a File Management item, use the following code:
(add-menu-item '("File" "File Management") "Copy File" 'copy-file t) (add-menu-item '("File" "File Management") "Delete File" 'delete-file t) (add-menu-item '("File" "File Management") "Rename File" 'rename-file t)
The optional before argument is the name of a menu item before which the new item should be added. If the item is already present, it is not moved.
To remove a specified menu item from the menu hierarchy, use
delete-menu-item
.
path is a list of strings that identify the position of the menu
item in the menu hierarchy. ("File" "Save")
means the menu item
called Save under the top level File menu. ("Menu" "Foo"
"Item")
means the menu item called Item under the Foo submenu
of Menu.
To disable a menu item, use disable-menu-item
. The disabled
menu item is grayed and can no longer be selected. To make the
item selectable again, use enable-menu-item
.
disable-menu-item
and enable-menu-item
both have the
argument path.
To change the string of the specified menu item, use
relabel-menu-item
. This function also takes the argument path.
new-name is the string to which the menu item will be changed.
The usual way to invoke Emacs is to type emacs RET at the shell. To invoke XEmacs, type xemacs RET. Emacs clears the screen and then displays an initial advisory message and copyright notice. You can begin typing Emacs commands immediately afterward.
Some operating systems insist on discarding all type-ahead when Emacs starts up; they give Emacs no way to prevent this. Therefore, it is wise to wait until Emacs clears the screen before typing the first editing command.
Before Emacs reads the first command, you have not had a chance to
give a command to specify a file to edit. Since Emacs must always have a
current buffer for editing, it presents a buffer, by default, a buffer named
`*scratch*'. The buffer is in Lisp Interaction
mode; you can use it to type Lisp expressions and evaluate them, or you
can ignore that capability and simply doodle. You can specify a
different major mode for this buffer by setting the variable
initial-major-mode
in your init file. See section The Init File, .emacs.
It is possible to give Emacs arguments in the shell command line to specify files to visit, Lisp files to load, and functions to call.
There are two commands for exiting Emacs because there are two kinds of exiting: suspending Emacs and killing Emacs. Suspending means stopping Emacs temporarily and returning control to its superior (usually the shell), allowing you to resume editing later in the same Emacs job, with the same files, same kill ring, same undo history, and so on. This is the usual way to exit. Killing Emacs means destroying the Emacs job. You can run Emacs again later, but you will get a fresh Emacs; there is no way to resume the same editing session after it has been killed.
suspend-emacs
). If used under the X window system,
shrink the X window containing the Emacs screen to an icon (see below).
save-buffers-kill-emacs
).
If you use XEmacs under the X window system, C-z shrinks the X window containing the Emacs screen to an icon. The Emacs process is stopped temporarily, and control is returned to the window manager. If more than one screen is associated with the Emacs process, only the screen from which you used C-z is retained. The X windows containing the other Emacs screens are closed.
To activate the "suspended" Emacs, use the appropriate window manager mouse gestures. Usually left-clicking on the icon reactivates and reopens the X window containing the Emacs screen, but the window manager you use determines what exactly happens. To actually kill the Emacs process, use C-x C-c or the Exit Emacs item on the File menu.
On systems that do not permit programs to be suspended, C-z runs
an inferior shell that communicates directly with the terminal, and
Emacs waits until you exit the subshell. On these systems, the only way
to return to the shell from which Emacs was started (to log out, for
example) is to kill Emacs. C-d or exit
are typical
commands to exit a subshell.
To kill Emacs, type C-x C-c (save-buffers-kill-emacs
). A
two-character key is used for this to make it harder to type. In
XEmacs, selecting the Exit Emacs option of the File menu is an
alternate way of issuing the command.
Unless a numeric argument is used, this command first offers to save any modified buffers. If you do not save all buffers, you are asked for reconfirmation with yes before killing Emacs, since any changes not saved will be lost. If any subprocesses are still running, C-x C-c asks you to confirm killing them, since killing Emacs kills the subprocesses simultaneously.
In most programs running on Unix, certain characters may instantly suspend or kill the program. (In Berkeley Unix these characters are normally C-z and C-c.) This Unix feature is turned off while you are in Emacs. The meanings of C-z and C-x C-c as keys in Emacs were inspired by the standard Berkeley Unix meanings of C-z and C-c, but that is their only relationship with Unix. You could customize these keys to do anything (see section Keymaps).
GNU Emacs supports command line arguments you can use to request various actions when invoking Emacs. The commands are for compatibility with other editors and for sophisticated activities. If you are using XEmacs under the X window system, you can also use a number of standard Xt command line arguments. Command line arguments are not usually needed for editing with Emacs; new users can skip this section.
Many editors are designed to be started afresh each time you want to edit. You start the editor to edit one file; then exit the editor. The next time you want to edit either another file or the same one, you start the editor again. Under these circumstances, it makes sense to use a command line argument to say which file to edit.
The recommended way to use GNU Emacs is to start it only once, just after you log in, and do all your editing in the same Emacs process. Each time you want to edit a file, you visit it using the existing Emacs. Emacs creates a new buffer for each file, and (unless you kill some of the buffers) Emacs eventually has many files in it ready for editing. Usually you do not kill the Emacs process until you are about to log out. Since you usually read files by typing commands to Emacs, command line arguments for specifying a file when Emacs is started are seldom needed.
Emacs accepts command-line arguments that specify files to visit, functions to call, and other activities and operating modes. If you are running XEmacs under the X window system, a number of standard Xt command line arguments are available as well.
The following subsections list:
Here are the arguments allowed:
find-file
. See section Visiting Files.
find-file
, then go to line number
linenum in it.
load
.
See section Libraries of Lisp Code for Emacs.
% emacs -version GNU Emacs 19.6 Lucid of Thu Apr 1 1993 on thalidomide (berkeley-unix)
stderr
only what would normally be printed
in the echo area under program control.
Batch mode is used for running programs written in Emacs Lisp from shell
scripts, makefiles, and so on. Normally the `-l' switch or
`-f' switch will be used as well, to invoke a Lisp program to do
the batch processing.
`-batch' implies `-q' (do not load an init file). It also
causes Emacs to kill itself after all command switches have been
processed. In addition, auto-saving is not done except in buffers for
which it has been explicitly requested.
Note that the init file can get access to the command line argument
values as the elements of a list in the variable
command-line-args
. (The arguments in the second table above will
already have been processed and will not be in the list.) The init file
can override the normal processing of the other arguments by setting
this variable.
One way to use command switches is to visit many files automatically:
emacs *.c
passes each .c
file as a separate argument to Emacs, so that
Emacs visits each file (see section Visiting Files).
Here is an advanced example that assumes you have a Lisp program file called `hack-c-program.el' which, when loaded, performs some useful operation on the current buffer, expected to be a C program.
emacs -batch foo.c -l hack-c-program -f save-buffer -kill > log
Here Emacs is told to visit `foo.c', load `hack-c-program.el'
(which makes changes in the visited file), save `foo.c' (note that
save-buffer
is the function that C-x C-s is bound to), and
then exit to the shell from which the command was executed. `-batch'
guarantees there will be no problem redirecting output to `log',
because Emacs will not assume that it has a display terminal to work
with.
screen-title-format
variable, which controls the title of the X
window corresponding to the selected screen. This is the same format as
mode-line-format
.
screen-icon-title-format
variable, which controls the title of
the icon corresponding to the selected screen.
In addition, XEmacs allows you to use a number of standard Xt command line arguments.
argv[0]
) as
the resource manager name.
We now give the basics of how to enter text, make corrections, and
save the text in a file. If this material is new to you, you might
learn it more easily by running the Emacs learn-by-doing tutorial. To
do this, type Control-h t (help-with-tutorial
).
To insert printing characters into the text you are editing, just type them. This inserts the characters into the buffer at the cursor (that is, at point; see section Point). The cursor moves forward. Any characters after the cursor move forward too. If the text in the buffer is `FOOBAR', with the cursor before the `B', and you type XX, the result is `FOOXXBAR', with the cursor still before the `B'.
To delete text you have just inserted, use DEL. DEL deletes the character before the cursor (not the one that the cursor is on top of or under; that is the character after the cursor). The cursor and all characters after it move backwards. Therefore, if you type a printing character and then type DEL, they cancel out.
To end a line and start typing a new one, type RET. This inserts a newline character in the buffer. If point is in the middle of a line, RET splits the line. Typing DEL when the cursor is at the beginning of a line rubs out the newline before the line, thus joining the line with the preceding line.
Emacs automatically splits lines when they become too long, if you turn on a special mode called Auto Fill mode. See section Filling Text, for information on using Auto Fill mode.
Customization information: DEL, in most modes, runs the command
delete-backward-char
; RET runs the command newline
,
and self-inserting printing characters run the command
self-insert
, which inserts whatever character was typed to invoke
it. Some major modes rebind DEL to other commands.
Direct insertion works for printing characters and SPC, but
other characters act as editing commands and do not insert themselves.
If you need to insert a control character or a character whose code is
above 200 octal, you must quote it by typing the character
control-q (quoted-insert
) first. There are two ways to use
C-q:
A numeric argument to C-q specifies how many copies of the quoted character should be inserted (see section Numeric Arguments).
If you prefer to have text characters replace (overwrite) existing text instead of moving it to the right, you can enable Overwrite mode, a minor mode. See section Minor Modes.
To do more than insert characters, you have to know how to move point (see section Point). Here are a few of the available commands.
beginning-of-line
).
end-of-line
).
forward-char
).
backward-char
).
forward-word
).
backward-word
).
next-line
). This command attempts to keep the horizontal position unchanged, so if you start in the middle of one line, you end in the middle of the next. When on the last line of text, C-n creates a new line and moves onto it. @item C-p
Move up one line, vertically (previous-line
).
recenter
). Text moves
on the screen to bring point to the center of the window.
move-to-window-line
). Text does not move on the
screen. A numeric argument says how many screen lines down from the
top of the window (zero for the top). A negative argument counts from
the bottom (-1 for the bottom).
transpose-chars
).
beginning-of-buffer
). With
numeric argument n, move to n/10 of the way from the top.
See section Numeric Arguments, for more information on numeric arguments.end-of-buffer
).
set-goal-column
). Henceforth, those
commands always move to this column in each line moved into, or as
close as possible given the contents of the line. This goal column remains
in effect until canceled.
If you set the variable track-eol
to a non-nil
value,
C-n and C-p move to the end of the line when at the end of
the starting line. By default, track-eol
is nil
.
delete-backward-char
).
delete-char
).
kill-line
).
kill-word
).
backward-kill-word
).
In contrast to the DEL key, which deletes the character before the cursor, Control-d deletes the character after the cursor, causing the rest of the text on the line to shift left. If Control-d is typed at the end of a line, that line and the next line are joined.
To erase a larger amount of text, use Control-k, which kills a line at a time. If you use C-k at the beginning or in the middle of a line, it kills all the text up to the end of the line. If you use C-k at the end of a line, it joins that line and the next line.
See section Deletion and Killing, for more flexible ways of killing text.
The commands above are sufficient for creating and altering text in an Emacs buffer. More advanced Emacs commands just make things easier. But to keep any text permanently you must put it in a file. Files are named units of text which are stored by the operating system and which you can retrieve by name. To look at or use the contents of a file in any way, including editing the file with Emacs, you must specify the file name.
Consider a file named `/usr/rms/foo.c'. To begin editing this file from Emacs, type:
C-x C-f /usr/rms/foo.c RET
The file name is given as an argument to the command C-x
C-f (find-file
). The command uses the minibuffer to
read the argument. You have to type RET to terminate the argument
(see section The Minibuffer).
You can also use the Open... menu item from the File menu, then type the name of the file to the prompt.
Emacs obeys the command by visiting the file: it creates a
buffer, copies the contents of the file into the buffer, and then
displays the buffer for you to edit. You can make changes in the
buffer, and then save the file by typing C-x C-s
(save-buffer
) or choosing Save Buffer from the File menu. This
makes the changes permanent by copying the altered contents of the
buffer back into the file `/usr/rms/foo.c'. Until then, the
changes are only inside your Emacs buffer, and the file `foo.c' is
not changed.
To create a file, visit the file with C-x C-f as if it already existed or choose Open... from the File menu and provide the name for the new file in the minibuffer. Emacs will create an empty buffer in which you can insert the text you want to put in the file. When you save the buffer with C-x C-s, or by choosing Save Buffer from the File menu, the file is created.
To learn more about using files, see section File Handling.
If you forget what a key does, you can use the Help character
(C-h) to find out: Type C-h k followed by the key you want to know
about. For example, C-h k C-n tells you what C-n
does. C-h is a prefix key; C-h k is just one of its
subcommands (the command describe-key
). The other subcommands of
C-h provide different kinds of help. Type C-h three times
to get a description of all the help facilities. See section Help.
Here are special commands and techniques for entering and removing blank lines.
open-line
).
delete-blank-lines
).
When you want to insert a new line of text before an existing line,
you just type the new line of text, followed by RET. If you
prefer to create a blank line first and then insert the desired text,
use the key C-o (open-line
), which inserts a newline after
point but leaves point in front of the newline. Then type
the text into the new line. C-o F O O has the same effect as
F O O RET, except for the final location of point.
To create several blank lines, type C-o several times, or give C-o an argument indicating how many blank lines to create. See section Numeric Arguments, for more information.
If you have many blank lines in a row and want to get rid of them, use
C-x C-o (delete-blank-lines
). If point is on a blank
line which is adjacent to at least one other blank line, C-x C-o
deletes all but one of the blank lines.
If point is on a blank line with no other adjacent blank line, the
sole blank line is deleted. If point is on a non-blank
line, C-x C-o deletes any blank lines following that non-blank
line.
If you add too many characters to one line without breaking with a RET, the line grows to occupy two (or more) screen lines, with a curved arrow at the extreme right margin of all but the last line. The curved arrow indicates that the following screen line is not really a distinct line in the text, but just the continuation of a line too long to fit the screen. You can use Auto Fill mode (see section Filling Text) to have Emacs insert newlines automatically when a line gets too long.
Instead of continuation, long lines can be displayed by truncation. This means that all the characters that do not fit in the width of the screen or window do not appear at all. They remain in the buffer, temporarily invisible. Three diagonal dots in the last column (instead of the curved arrow inform you that truncation is in effect.
To turn off continuation for a particular buffer, set the
variable truncate-lines
to non-nil
in that buffer.
Truncation instead of continuation also happens whenever horizontal
scrolling is in use, and optionally whenever side-by-side windows are in
use (see section Multiple Windows). Altering the value of truncate-lines
makes
it local to the current buffer; until that time, the default value is in
effect. The default is initially nil
. See section Local Variables.
If you are accustomed to other display editors, you may be surprised that Emacs does not always display the page number or line number of point in the mode line. In Emacs, this information is only rarely needed, and a number of commands are available to compute and print it. Since text is stored in a way that makes it difficult to compute the information, it is not displayed all the time.
count-lines-region
).
what-cursor-position
).
There are several commands for printing line numbers:
count-lines-region
) prints the number of lines in
the region (see section Selecting Text).
The command C-x = (what-cursor-position
) provides
information about point and about the column the cursor is in.
It prints a line in the echo area that looks like this:
Char: x (0170) point=65986 of 563027(12%) column 44
(In fact, this is the output produced when point is before `column 44' in the example.)
The two values after `Char:' describe the character following point, first by showing it and second by giving its octal character code.
`point=' is followed by the position of point expressed as a character count. The front of the buffer counts as position 1, one character later as 2, and so on. The next, larger number is the total number of characters in the buffer. Afterward in parentheses comes the position expressed as a percentage of the total size.
`column' is followed by the horizontal position of point, in columns from the left edge of the window.
If the buffer has been narrowed, making some of the text at the beginning and the end temporarily invisible, C-x = prints additional text describing the current visible range. For example, it might say:
Char: x (0170) point=65986 of 563025(12%) <65102 - 68533> column 44
where the two extra numbers give the smallest and largest character position that point is allowed to assume. The characters between those two positions are the visible ones. See section Narrowing.
If point is at the end of the buffer (or the end of the visible part), C-x = omits any description of the character after point. The output looks like
point=563026 of 563025(100%) column 0
Any Emacs command can be given a numeric argument. Some commands
interpret the argument as a repetition count. For example, giving an
argument of ten to the key C-f (the command forward-char
, move
forward one character) moves forward ten characters. With these commands,
no argument is equivalent to an argument of one. Negative arguments are
allowed. Often they tell a command to move or act backwards.
If your keyboard has a META key, the easiest way to specify a numeric argument is to type digits and/or a minus sign while holding down the the META key. For example,
M-5 C-nmoves down five lines. The characters Meta-1, Meta-2, and so on, as well as Meta--, do this because they are keys bound to commands (
digit-argument
and negative-argument
) that are
defined to contribute to an argument for the next command.
Another way of specifying an argument is to use the C-u
(universal-argument
) command followed by the digits of the argument.
With C-u, you can type the argument digits without holding
down shift keys. To type a negative argument, start with a minus sign.
Just a minus sign normally means -1. C-u works on all terminals.
C-u followed by a character which is neither a digit nor a minus sign has the special meaning of "multiply by four". It multiplies the argument for the next command by four. C-u twice multiplies it by sixteen. Thus, C-u C-u C-f moves forward sixteen characters. This is a good way to move forward "fast", since it moves about 1/5 of a line in the usual size screen. Other useful combinations are C-u C-n, C-u C-u C-n (move down a good fraction of a screen), C-u C-u C-o (make "a lot" of blank lines), and C-u C-k (kill four lines).
Some commands care only about whether there is an argument and not about
its value. For example, the command M-q (fill-paragraph
) with
no argument fills text; with an argument, it justifies the text as well.
(See section Filling Text, for more information on M-q.) Just C-u is a
handy way of providing an argument for such commands.
Some commands use the value of the argument as a repeat count, but do
something peculiar when there is no argument. For example, the command
C-k (kill-line
) with argument n kills n lines,
including their terminating newlines. But C-k with no argument is
special: it kills the text up to the next newline, or, if point is right at
the end of the line, it kills the newline itself. Thus, two C-k
commands with no arguments can kill a non-blank line, just like C-k
with an argument of one. (See section Deletion and Killing, for more information on
C-k.)
A few commands treat a plain C-u differently from an ordinary argument. A few others may treat an argument of just a minus sign differently from an argument of -1. These unusual cases will be described when they come up; they are always to make the individual command more convenient to use.
Emacs allows you to undo all changes you make to the text of a buffer,
up to a certain amount of change (8000 characters). Each buffer records
changes individually, and the undo command always applies to the
current buffer. Usually each editing command makes a separate entry
in the undo records, but some commands such as query-replace
make many entries, and very simple commands such as self-inserting
characters are often grouped to make undoing less tedious.
undo
).
The command C-x u or C-_ allows you to undo changes. The first time you give this command, it undoes the last change. Point moves to the text affected by the undo, so you can see what was undone.
Consecutive repetitions of the C-_ or C-x u commands undo earlier and earlier changes, back to the limit of what has been recorded. If all recorded changes have already been undone, the undo command prints an error message and does nothing.
Any command other than an undo command breaks the sequence of undo commands. Starting at this moment, the previous undo commands are considered ordinary changes that can themselves be undone. Thus, you can redo changes you have undone by typing C-f or any other command that have no important effect, and then using more undo commands.
If you notice that a buffer has been modified accidentally, the easiest way to recover is to type C-_ repeatedly until the stars disappear from the front of the mode line. When that happens, all the modifications you made have been canceled. If you do not remember whether you changed the buffer deliberately, type C-_ once. When you see Emacs undo the last change you made, you probably remember why you made it. If the change was an accident, leave it undone. If it was deliberate, redo the change as described in the preceding paragraph.
Whenever an undo command makes the stars disappear from the mode line, the buffer contents is the same as it was when the file was last read in or saved.
Not all buffers record undo information. Buffers whose names start with spaces don't; these buffers are used internally by Emacs and its extensions to hold text that users don't normally look at or edit. Minibuffers, help buffers, and documentation buffers also don't record undo information.
Emacs can remember at most 8000 or so characters of deleted or modified text in any one buffer for reinsertion by the undo command. There is also a limit on the number of individual insert, delete, or change actions that Emacs can remember.
There are two keys to run the undo
command, C-x u and
C-_, because on some keyboards, it is not obvious how to type
C-_. C-x u is an alternative you can type in the same
fashion on any terminal.
Emacs commands use the minibuffer to read arguments more complicated than a single number. Minibuffer arguments can be file names, buffer names, Lisp function names, Emacs command names, Lisp expressions, and many other things, depending on the command reading the argument. To edit the argument in the minibuffer, you can use Emacs editing commands.
When the minibuffer is in use, it appears in the echo area, and the cursor moves there. The beginning of the minibuffer line displays a prompt indicating what kind of input you should supply and how it will be used. The prompt is often derived from the name of the command the argument is for. The prompt normally ends with a colon.
Sometimes a default argument appears in parentheses after the colon; it, too, is part of the prompt. The default is used as the argument value if you enter an empty argument (e.g., by just typing RET). For example, commands that read buffer names always show a default, which is the name of the buffer that will be used if you type just RET.
The simplest way to give a minibuffer argument is to type the text you want, terminated by RET to exit the minibuffer. To get out of the minibuffer and cancel the command that it was for, type C-g.
Since the minibuffer uses the screen space of the echo area, it can conflict with other ways Emacs customarily uses the echo area. Here is how Emacs handles such conflicts:
Sometimes the minibuffer starts out with text in it. For example, when you are supposed to give a file name, the minibuffer starts out containing the default directory, which ends with a slash. This informs you in which directory the file will be looked for if you do not specify a different one. For example, the minibuffer might start out with:
Find File: /u2/emacs/src/
where `Find File: ' is the prompt. Typing buffer.c specifies
the file
`/u2/emacs/src/buffer.c'. To find files in nearby
directories, use `..'; thus, if you type ../lisp/simple.el, the
file that you visit will be the one named
`/u2/emacs/lisp/simple.el'.
Alternatively, you can use M-DEL to kill directory names you
don't want (see section Words).
You can also type an absolute file name, one starting with a slash or a tilde, ignoring the default directory. For example, to find the file `/etc/termcap', just type the name, giving:
Find File: /u2/emacs/src//etc/termcap
Two slashes in a row are not normally meaningful in Unix file names, but they are allowed in GNU Emacs. They mean, "ignore everything before the second slash in the pair." Thus, `/u2/emacs/src/' is ignored, and you get the file `/etc/termcap'.
If you set insert-default-directory
to nil
, the default
directory is not inserted in the minibuffer. This way, the minibuffer
starts out empty. But the name you type, if relative, is still
interpreted with respect to the same default directory.
The minibuffer is an Emacs buffer (albeit a peculiar one), and the usual Emacs commands are available for editing the text of an argument you are entering.
Since RET in the minibuffer is defined to exit the minibuffer, you must use C-o or C-q LFD to insert a newline into the minibuffer. (Recall that a newline is really the LFD character.)
The minibuffer has its own window, which always has space on the screen but acts as if it were not there when the minibuffer is not in use. The minibuffer window is just like the others; you can switch to another window with C-x o, edit text in other windows, and perhaps even visit more files before returning to the minibuffer to submit the argument. You can kill text in another window, return to the minibuffer window, and then yank the text to use it in the argument. See section Multiple Windows.
There are, however, some restrictions on the use of the minibuffer window. You cannot switch buffers in it--the minibuffer and its window are permanently attached. You also cannot split or kill the minibuffer window, but you can make it taller with C-x ^.
If you are in the minibuffer and issue a command that displays help text in another window, that window will be scrolled if you type M-C-v while in the minibuffer until you exit the minibuffer. This feature is helpful if a completing minibuffer gives you a long list of possible completions.
If the variable minibuffer-confirm-incomplete
is t
, you
are asked for confirmation if there is no known completion for the text
you typed. For example, if you attempted to visit a non-existent file,
the minibuffer might read:
Find File:chocolate_bar.c [no completions, confirm]If you press Return again, that confirms the filename. Otherwise, you can continue editing it.
Emacs supports recursive use of the minibuffer. However, it is easy to do this by accident (because of autorepeating keyboards, for example) and get confused. Therefore, most Emacs commands that use the minibuffer refuse to operate if the minibuffer window is selected. If the minibuffer is active but you have switched to a different window, recursive use of the minibuffer is allowed--if you know enough to try to do this, you probably will not get confused.
If you set the variable enable-recursive-minibuffers
to be
non-nil
, recursive use of the minibuffer is always allowed.
When appropriate, the minibuffer provides a completion facility. You type the beginning of an argument and one of the completion keys, and Emacs visibly fills in the rest, depending on what you have already typed.
When completion is available, certain keys---TAB, RET, and SPC---are redefined to complete an abbreviation present in the minibuffer into a longer string that it stands for, by matching it against a set of completion alternatives provided by the command reading the argument. ? is defined to display a list of possible completions of what you have inserted.
For example, when the minibuffer is being used by Meta-x to read the name of a command, it is given a list of all available Emacs command names to complete against. The completion keys match the text in the minibuffer against all the command names, find any additional characters of the name that are implied by the ones already present in the minibuffer, and add those characters to the ones you have given.
Case is normally significant in completion because it is significant in most of the names that you can complete (buffer names, file names, and command names). Thus, `fo' will not complete to `Foo'. When you are completing a name in which case does not matter, case may be ignored for completion's sake if specified by program.
When a completion list is displayed, the completions will highlight as you move the mouse over them. Clicking the middle mouse button on any highlighted completion will "select" it just as if you had typed it in and hit RET.
Consider the following example. If you type Meta-x au TAB,
TAB looks for alternatives (in this case, command names) that
start with `au'. There are only two commands: auto-fill-mode
and
auto-save-mode
. They are the same as far as auto-
, so the
`au' in the minibuffer changes to `auto-'.
If you type TAB again immediately, there are multiple possibilities for the very next character--it could be `s' or `f'---so no more characters are added; but a list of all possible completions is displayed in another window.
If you go on to type f TAB, this TAB sees
`auto-f'. The only command name starting this way is
auto-fill-mode
, so completion inserts the rest of that command. You
now have `auto-fill-mode' in the minibuffer after typing just au
TAB f TAB. Note that TAB has this effect because in the
minibuffer it is bound to the function minibuffer-complete
when
completion is supposed to be done.
Here is a list of all the completion commands defined in the minibuffer when completion is available.
minibuffer-complete
).
minibuffer-complete-word
).
minibuffer-complete-and-exit
).
minibuffer-list-completions
).
minibuf-select-highlighted-completion
).
SPC completes in a way that is similar to TAB, but it never
goes beyond the next hyphen or space. If you have `auto-f' in the
minibuffer and type SPC, it finds that the completion is
`auto-fill-mode', but it stops completing after `fill-'.
The result is `auto-fill-'. Another SPC at this point
completes all the way to `auto-fill-mode'. SPC in the
minibuffer runs the function minibuffer-complete-word
when
completion is available.
There are three different ways that RET can work in completing minibuffers, depending on how the argument will be used.
The completion commands display a list of all possible completions in a window whenever there is more than one possibility for the very next character. Typing ? explicitly requests such a list. The list of completions counts as help text, so C-M-v typed in the minibuffer scrolls the list.
When completion is done on file names, certain file names are usually
ignored. The variable completion-ignored-extensions
contains a list
of strings; a file whose name ends in any of those strings is ignored as a
possible completion. The standard value of this variable has several
elements including ".o"
, ".elc"
, ".dvi"
and "~"
.
The effect is that, for example, `foo' completes to `foo.c'
even though `foo.o' exists as well. If the only possible completions
are files that end in "ignored" strings, they are not ignored.
If a completion command finds the next character is undetermined, it
automatically displays a list of all possible completions. If the variable
completion-auto-help
is set to nil
, this does not happen,
and you must type ? to display the possible completions.
If the variable minibuffer-confirm-incomplete
is set to t
,
then in contexts where completing-read
allows answers that are
not valid completions, an extra RET must be typed to confirm the
response. This is helpful for catching typos.
Every command that uses the minibuffer at least once is recorded on a special history list, together with the values of the minibuffer arguments, so that you can repeat the command easily. In particular, every use of Meta-x is recorded, since M-x uses the minibuffer to read the command name.
repeat-complex-command
).
previous-history-element
).
next-history-element
).C-x ESC is used to re-execute a recent command that used the minibuffer. With no argument, it repeats the last command. A numeric argument specifies which command to repeat; 1 means the last one, and larger numbers specify earlier commands.
C-x ESC works by turning the previous command into a Lisp expression and then entering a minibuffer initialized with the text for that expression. If you type just RET, the command is repeated as before. You can also change the command by editing the Lisp expression. The expression you finally submit will be executed. The repeated command is added to the front of the command history unless it is identical to the most recently executed command already there.
Even if you don't understand Lisp syntax, it will probably be obvious which command is displayed for repetition. If you do not change the text, you can be sure the command will repeat exactly as before.
If you are in the minibuffer for C-x ESC and the command shown to you is not the one you want to repeat, you can move around the list of previous commands using M-n and M-p. M-p replaces the contents of the minibuffer with the next earlier recorded command, and M-n replaces it with the next later command. After finding the desired previous command, you can edit its expression and then resubmit it by typing RET. Any editing you have done on the command to be repeated is lost if you use M-n or M-p.
M-n and M-p are specially defined within C-x ESC
to run the commands previous-history-element
and
next-history-element
.
The list of previous commands using the minibuffer is stored as a Lisp
list in the variable command-history
. Each element of the list
is a Lisp expression which describes one command and its arguments.
Lisp programs can reexecute a command by feeding the corresponding
command-history
element to eval
.
The Emacs commands that are used often or that must be quick to type are bound to keys--short sequences of characters--for convenient use. Other Emacs commands that are used more rarely are not bound to keys; to run them, you must refer to them by name.
A command name consists, by convention, of one or more words,
separated by hyphens: for example, auto-fill-mode
or
manual-entry
. The use of English words makes the command name
easier to remember than a key made up of obscure characters, even though
it results in more characters to type. You can run any command by name,
even if it can be run by keys as well.
To run a command by name, start with M-x, then type the command name, and finish with RET. M-x uses the minibuffer to read the command name. RET exits the minibuffer and runs the command.
Emacs uses the minibuffer for reading input for many different purposes; on this occasion, the string `M-x' is displayed at the beginning of the minibuffer as a prompt to remind you that your input should be the name of a command to be run. See section The Minibuffer, for full information on the features of the minibuffer.
You can use completion to enter a command name. For example, to
invoke the command forward-char
, type:
M-x forward-char RETor
M-x fo TAB c RET
After you type in M-x fo TAB
emacs will give you a possible list of
completions from which you can choose. Note that forward-char
is the
same command that you invoke with the key C-f. You can call any
command (interactively callable function) defined in Emacs by its name
using M-x regardless of whether or not any keys are bound to it.
If you type C-g while Emacs reads the command name, you cancel the M-x command and get out of the minibuffer, ending up at top level.
To pass a numeric argument to a command you are invoking with M-x, specify the numeric argument before the M-x. M-x passes the argument along to the function that it calls. The argument value appears in the prompt while the command name is being read.
You can use the command M-x interactive
to specify a way of
parsing arguments for interactive use of a function. For example, write:
(defun foo (arg) "Doc string" (interactive "p") ...use arg...)
to make arg
be the prefix argument when foo
is called as a
command. The call to interactive
is actually a declaration
rather than a function; it tells call-interactively
how to read
arguments to pass to the function. When actually called, interactive
returns nil
.
The argument of interactive is usually a string containing a code
letter followed by a prompt. Some code letters do not use I/O to get
the argument and do not need prompts. To prompt for multiple arguments,
you must provide a code letter, its prompt, a newline, and another code
letter, and so forth. If the argument is not a string, it is evaluated
to get a list of arguments to pass to the function. If you do not provide an
argument to interactive
, no arguments are passed when calling
interactively.
Available code letters are:
a
b
B
c
C
d
D
e
f
F
k
m
n
N
n
p
P
r
s
S
v
user-variable-p
x
X
In addition, if the string begins with `*', an error is signaled if the buffer is read-only. This happens before reading any arguments. If the string begins with `@', the window the mouse is over is selected before anything else is done. You may use both `@' and `*'; they are processed in the order that they appear.
Normally, when describing a command that is run by name, we omit the RET that is needed to terminate the name. Thus we may refer to M-x auto-fill-mode rather than M-x auto-fill-mode RET. We mention the RET only when it it necessary to emphasize its presence, for example, when describing a sequence of input that contains a command name and arguments that follow it.
M-x is defined to run the command execute-extended-command
,
which is responsible for reading the name of another command and invoking
it.
Emacs provides extensive help features which revolve around a single character, C-h. C-h is a prefix key that is used only for documentation-printing commands. The characters you can type after C-h are called help options. One help option is C-h; you use it to ask for help about using C-h.
C-h C-h prints a list of the possible help options, and then asks you to type the desired option. It prompts with the string:
A, B, C, F, I, K, L, M, N, S, T, V, W, C-c, C-d, C-n, C-w or C-h for more help:
You should type one of those characters.
Typing a third C-h displays a description of what the options mean; Emacs still waits for you to type an option. To cancel, type C-g.
Here is a summary of the defined help commands.
command-
apropos
).describe-bindings
).
describe-key-
briefly
).
c is for `character'. For more extensive information on key,
use C-h k.
describe-function
). Note that commands are Lisp functions, so
a command name may be used.
info
).
The complete Emacs manual is available online in Info.
describe-key
).
view-lossage
).
describe-mode
).
view-emacs-news
).
describe-pointer
).
describe-syntax
).
help-with-tutorial
).
describe-
variable
).
where-is
).
The most basic C-h options are C-h c
(describe-key-briefly
) and C-h k
(describe-key
).
C-h c key prints the name of the command that key is
bound to in the echo area. For example, C-h c C-f prints
`forward-char'. Since command names are chosen to describe what
the command does, using this option is a good way to get a somewhat cryptic
description of what key does.
C-h k key is similar to C-h c but gives more information. It displays the documentation string of the function key is bound to as well as its name. key is a string or vector of events. When called interactively, key may also be a menu selection. This information does not usually fit into the echo area, so a window is used for the display.
C-h f (describe-function
) reads the name of a Lisp
function using the minibuffer, then displays that function's
documentation string in a window. Since commands are Lisp functions,
you can use the argument function to get the documentation of a
command that you know by name. For example,
C-h f auto-fill-mode RET
displays the documentation for auto-fill-mode
. Using C-h f
is the only way to see the documentation of a command that is not bound
to any key, that is, a command you would normally call using M-x.
If the variable describe-function-show-arglist
is t
,
describe-function
shows its arglist if the function is not
an autoload function.
C-h f is also useful for Lisp functions you are planning to
use in a Lisp program. For example, if you have just written the code
(make-vector len)
and want to make sure you are using
make-vector
properly, type C-h f make-vector RET. Because
C-h f allows all function names, not just command names, you may find
that some of your favorite abbreviations that work in M-x don't work
in C-h f. An abbreviation may be unique among command names, yet fail
to be unique when other function names are allowed.
If you type RET, leaving the minibuffer empty, C-h f by
default describes the function called by the innermost Lisp expression
in the buffer around point, provided that that is a valid, defined Lisp
function name. For example, if point is located following the text
`(make-vector (car x)', the innermost list containing point is the
one starting with `(make-vector', so the default is to describe
the function make-vector
.
C-h f is often useful just to verify that you have the right spelling for the function name. If C-h f mentions a default in the prompt, you have typed the name of a defined Lisp function. If that is what you wanted to know, just type C-g to cancel the C-h f command and continue editing.
C-h w command RET (where-s
) tells you what
keys are bound to command. It prints a list of the keys in the
echo area. Alternatively, it informs you that a command is not bound to
any keys, which implies that you must use M-x to call the
command.
C-h v (describe-variable
) is like C-h f but
describes Lisp variables instead of Lisp functions. Its default is the
Lisp symbol around or before point, if that is the name of a known Lisp
variable. See section Variables.
command-apropos
).
It is possible to ask a question like, "What are the commands for
working with files?" To do this, type C-h a file RET,
which displays a list of all command names that contain `file',
such as copy-file
, find-file
, and so on. With each
command name a brief description of its use and information on the keys
you can use to invoke it is displayed. For example, you would be
informed that you can invoke find-file
by typing C-x C-f.
The a in C-h a stands for `Apropos'; C-h a runs the
Lisp function command-apropos
.
Because C-h a looks only for functions whose names contain the string you specify, you must use ingenuity in choosing the string. If you are looking for commands for killing backwards and C-h a kill-backwards RET doesn't reveal any commands, don't give up. Try just kill, or just backwards, or just back. Be persistent. Pretend you are playing Adventure. Also note that you can use a regular expression as the argument (see section Syntax of Regular Expressions).
Here is a set of arguments to give to C-h a that covers many
classes of Emacs commands, since there are strong conventions for naming
standard Emacs commands. By giving you a feeling for the naming
conventions, this set of arguments can also help you develop a
technique for picking apropos
strings.
char, line, word, sentence, paragraph, region, page, sexp, list, defun, buffer, screen, window, file, dir, register, mode, beginning, end, forward, backward, next, previous, up, down, search, goto, kill, delete, mark, insert, yank, fill, indent, case, change, set, what, list, find, view, describe.
To list all Lisp symbols that contain a match for a regexp, not just the ones that are defined as commands, use the command M-x apropos instead of C-h a.
C-h i (info
) runs the Info program, which is used for
browsing through structured documentation files. The entire Emacs manual
is available within Info. Eventually all the documentation of the GNU
system will be available. Type h after entering Info to run
a tutorial on using Info.
If something surprising happens, and you are not sure what commands you
typed, use C-h l (view-lossage
). C-h l prints the last
100 command characters you typed. If you see commands you don't
know, use C-h c to find out what they do.
Emacs has several major modes. Each mode redefines a few keys and
makes a few other changes in how editing works. C-h m
(describe-mode
) prints documentation on the current major mode,
which normally describes all the commands that are changed in this mode.
C-h b (describe-bindings
) and C-h s
(describe-syntax
) present information about the current Emacs
mode that is not covered by C-h m. C-h b displays a list of
all key bindings currently in effect, with the local bindings of the current
major mode first, followed by the global bindings (see section Customizing Key Bindings). C-h s displays the contents of the syntax table with
explanations of each character's syntax (see section The Syntax Table).
The other C-h options display various files of useful
information. C-h C-w (describe-no-warranty
) displays
details on the complete absence of warranty for GNU Emacs. C-h n
(view-emacs-news
) displays the file `emacs/etc/NEWS', which
contains documentation on Emacs changes arranged chronologically.
C-h t (help-with-tutorial
) displays the learn-by-doing
Emacs tutorial. C-h C-c (describe-copying
) displays the file
`emacs/etc/COPYING', which tells you the conditions you must obey
in distributing copies of Emacs. C-h C-d
(describe-distribution
) displays another file named
`emacs/etc/DISTRIB', which tells you how you can order a copy of
the latest version of Emacs.
Many Emacs commands operate on an arbitrary contiguous part of the current buffer. You can select text in two ways:
Many commands that insert text, such as C-y (yank
) and
M-x insert-buffer, position the mark at one end of the inserted
text--the opposite end from where point is positioned, so that the region
contains the text just inserted.
Aside from delimiting the region, the mark is useful for marking
a spot that you may want to go back to. To make this feature more useful,
Emacs remembers 16 previous locations of the mark in the mark ring
.
Here are some commands for setting the mark:
set-mark-command
).
exchange-point-and-mark
).
For example, to convert part of the buffer to all
upper-case, you can use the C-x C-u (upcase-region
)
command, which operates on the text in the region. First go to the
beginning of the text you want to capitalize and type C-SPC to
put the mark there, then move to the end, and then type C-x C-u to
capitalize the selected region. You can also set the mark at the end of the
text, move to the beginning, and then type C-x C-u. Most commands
that operate on the text in the region have the word region
in
their names.
The most common way to set the mark is with the C-SPC
command (set-mark-command
). This command sets the mark where
point is. You can then move point away, leaving the mark behind. It is
actually incorrect to speak of the character C-SPC; there is
no such character. When you type SPC while holding down
CTRL, you get the character C-@ on most terminals. This
character is actually bound to set-mark-command
. But unless you are
unlucky enough to have a terminal where typing C-SPC does
not produce C-@, you should think of this character as
C-SPC.
Since terminals have only one cursor, Emacs cannot show you where the
mark is located. Most people use the mark soon after they set it, before
they forget where it is. But you can see where the mark is with the
command C-x C-x (exchange-point-and-mark
) which puts the
mark where point was and point where the mark was. The extent of the
region is unchanged, but the cursor and point are now at the previous
location of the mark.
Another way to set the mark is to push the mark to the beginning of a
buffer while leaving point at its original location. If you supply an
argument to C-< (mark-beginning-of-buffer
), the mark is pushed
n/10 of the way from the true beginning of the buffer. You can
also set the mark at the end of a buffer with C->
(mark-end-of-buffer
). It pushes the mark to the end of the buffer,
leaving point alone. Supplying an argument to the command pushes the mark
n/10 of the way from the true end of the buffer.
If you are using XEmacs under the X window system, you can set
the variable zmacs-regions
to t
. This makes the current
region (defined by point and mark) highlight and makes it available as
the X clipboard selection, which means you can use the menu bar items on
it. See section Active Regions for more information.
C-x C-x is also useful when you are satisfied with the location of
point but want to move the mark; do C-x C-x to put point there and
then you can move it. A second use of C-x C-x, if necessary, puts
the mark at the new location with point back at its original location.
Once you have created an active region, you can do many things to the text in it:
There are commands for placing point and the mark around a textual object such as a word, list, paragraph or page.
mark-word
). This command and
the following one do not move point.
mark-sexp
).
mark-paragraph
).
mark-defun
).
mark-whole-buffer
).
mark-page
).
M-@ (mark-word
) puts the mark at the end of the next word,
while C-M-@ (mark-sexp
) puts it at the end of the next Lisp
expression. These characters sometimes save you some typing.
A number of commands are available that set both point and mark and
thus delimit an object in the buffer. M-h (mark-paragraph
)
moves point to the beginning of the paragraph that surrounds or follows
point, and puts the mark at the end of that paragraph
(see section Paragraphs). You can then indent, case-convert, or kill the
whole paragraph. In the same fashion, C-M-h (mark-defun
)
puts point before and the mark after the current or following defun
(see section Defuns). C-x C-p (mark-page
) puts point before
the current page (or the next or previous, depending on the argument),
and mark at the end (see section Pages). The mark goes after the
terminating page delimiter (to include it), while point goes after the
preceding page delimiter (to exclude it). Finally, C-x h
(mark-whole-buffer
) sets up the entire buffer as the region by
putting point at the beginning and the mark at the end.
Aside from delimiting the region, the mark is also useful for marking
a spot that you may want to go back to. To make this feature more
useful, Emacs remembers 16 previous locations of the mark in the
mark ring. Most commands that set the mark push the old mark onto
this ring. To return to a marked location, use C-u C-SPC
(or C-u C-@); this is the command set-mark-command
given a
numeric argument. The command moves point to where the mark was, and
restores the mark from the ring of former marks. Repeated use of this
command moves point to all the old marks on the ring, one by one.
The marks you have seen go to the end of the ring, so no marks are lost.
Each buffer has its own mark ring. All editing commands use the current buffer's mark ring. In particular, C-u C-SPC always stays in the same buffer.
Many commands that can move long distances, such as M-<
(beginning-of-buffer
), start by setting the mark and saving the
old mark on the mark ring. This makes it easier for you to move back
later. Searches set the mark, unless they do not actually move point.
When a command sets the mark, `Mark Set' is printed in the
echo area.
The variable mark-ring-max
is the maximum number of entries to
keep in the mark ring. If that many entries exist and another entry is
added, the last entry in the list is discarded. Repeating C-u
C-SPC circulates through the entries that are currently in the
ring.
The variable mark-ring
holds the mark ring itself, as a list of
marker objects in the order most recent first. This variable is local
in every buffer.
If you are using Lucid GNU Emacs under X, you can use the mouse cursor to select text. There are two mouse cursor shapes:
xterm
uses.
You can set the value of the variable x-mode-pointer-shape
to
determine the shape of the mouse pointer when it is over the mode line. If
the value is nil
, either the variable
x-nontext-pointer-shape
or x-pointer-shape
is used.
If you want to get fancy, you can set the foreground and background
colors of the mouse pointer with the variables
x-pointer-background-color
and x-pointer-foreground-color
.
There are two ways to select a region of text with the mouse:
To select a word in text, double-click with the left mouse button while the mouse cursor is over the word. The word is highlighted when selected. On monochrome monitors, a stippled background indicates that a region of text has been highlighted. On color monitors, a color background indicates highlighted text. You can triple-click to select whole lines.
To select an arbitrary region of text:
Once a region of text is selected, it becomes the primary X selection (see section Using X Selections) as well as the Emacs selected region. You can paste it into other X applications and use the options from the Edit pull-down menu on it. Since it is also the Emacs region, you can use Emacs region commands on it.
Lucid GNU Emacs also provides the following mouse functions. Most of these are not bound to mouse gestures by default, but they are provided for your customization pleasure. For example, if you wanted shift-left (that is, holding down the Shift key and clicking the left mouse button) to delete the character at which you are pointing, then you could do this:
(global-set-key '(shift button1) 'mouse-del-char)
mouse-track
, but also copy it to the cut buffer.
The M-x mouse-track command should be bound to a mouse button. If you click-and-drag, the selection is set to the region between the point of the initial click and the point at which you release the button. These positions do not need to be ordered.
If you click-and-release without moving the mouse, the point is moved, and the selection is disowned (there will be no selection owner.) The mark will be set to the previous position of point.
If you double-click, the selection will extend by symbols instead of by characters. If you triple-click, the selection will extend by lines.
If you drag the mouse off the top or bottom of the window, you can select pieces of text that are larger than the visible part of the buffer; the buffer will scroll as necessary.
The selected text becomes the current X selection, and is also copied to
the top of the kill ring. Point will be left at the position at
which you released the button and the mark will be left at the initial
click position. Bind a mouse click to
mouse-track-and-copy-to-cutbuffer to copy selections to the cut buffer.
(See also the mouse-track-adjust
command, on Shift-button1.)
The M-x mouse-track-adjust command should be bound to a mouse
button. The selection will be enlarged or shrunk so that the point of
the mouse click is one of its endpoints. This is only meaningful
after the mouse-track
command (button1) has been executed.
The M-x mouse-track-delete-and-insert command is exactly the same
as the mouse-track
command on button1, except that point is
not moved; the selected text is immediately inserted after being
selected; and the text of the selection is deleted.
The M-x mouse-track-insert command is exactly the same as the
mouse-track
command on button1, except that point is not moved;
the selected text is immediately inserted after being selected; and the
selection is immediately disowned afterwards.
Killing means erasing text and copying it into the kill ring, from which it can be retrieved by yanking it. Some other systems that have recently become popular use the terms "cutting" and "pasting" for these operations.
The most common way of moving or copying text with Emacs is to kill it and later yank it in one or more places. This is safe because all the text killed recently is stored in the kill ring, and it is versatile, because you can use the same commands for killing syntactic units and for moving those units. There are other ways of copying text for special purposes.
Emacs has only one kill ring, so you can kill text in one buffer and yank it in another buffer. If you are using Lucid GNU Emacs under X, you can also use the X selection mechanism to copy text from one buffer to another, or between applications. See section Using X Selections.
Most commands that erase text from the buffer save it. You can get the text back if you change your mind, or you can move or copy it to other parts of the buffer. Commands which erase text and save it in the kill ring are known as kill commands. Some other commands erase text but do not save it; they are known as delete commands. (This distinction is made only for erasing text in the buffer.)
The commands' names and individual descriptions use the words
`kill' and `delete' to indicate what they do. If you perform
a kill or delete command by mistake, use the C-x u (undo
)
command to undo it (see section Undoing Changes). The delete commands include C-d
(delete-char
) and DEL (delete-backward-char
), which
delete only one character at a time, and those commands that delete only
spaces or newlines. Commands that can destroy significant amounts of
nontrivial data usually kill.
delete-char
).
delete-backward-char
).
delete-horizontal-space
).
just-one-space
).
delete-blank-lines
).
delete-indentation
).
The most basic delete commands are C-d (delete-char
) and
DEL (delete-backward-char
). C-d deletes the
character after point, the one the cursor is "on top of". Point
doesn't move. DEL deletes the character before the cursor, and
moves point back. You can delete newlines like any other characters in
the buffer; deleting a newline joins two lines. Actually, C-d and
DEL aren't always delete commands; if you give them an argument,
they kill instead, since they can erase more than one character this
way.
The other delete commands delete only formatting characters: spaces,
tabs and newlines. M-\ (delete-horizontal-space
) deletes
all spaces and tab characters before and after point.
M-SPC (just-one-space
) does the same but leaves a
single space after point, regardless of the number of spaces that
existed previously (even zero).
C-x C-o (delete-blank-lines
) deletes all blank lines after
the current line. If the current line is blank, it deletes all blank lines
preceding the current line as well as leaving one blank line, the current
line. M-^ (delete-indentation
) joins the current line and
the previous line, or, if given an argument, joins the current line and
the next line by deleting a newline and all surrounding spaces, possibly
leaving a single space. See section Indentation.
kill-line
).
The simplest kill command is C-k. If given at the beginning of a line, it kills all the text on the line, leaving the line blank. If given on a blank line, the blank line disappears. As a consequence, a line disappears completely if you go to the front of a non-blank line and type C-k twice.
More generally, C-k kills from point up to the end of the line, unless it is at the end of a line. In that case, it kills the newline following the line, thus merging the next line into the current one. Emacs ignores invisible spaces and tabs at the end of the line when deciding which case applies: if point appears to be at the end of the line, you can be sure the newline will be killed.
If you give C-k a positive argument, it kills that many lines and the newlines that follow them (however, text on the current line before point is not killed). With a negative argument, C-k kills back to a number of line beginnings. An argument of -2 means kill back to the second line beginning. If point is at the beginning of a line, that line beginning doesn't count, so C-u - 2 C-k with point at the front of a line kills the two previous lines.
C-k with an argument of zero kills all the text before point on the current line.
kill-region
).
See section Words.
kill-word
).
backward-kill-word
).
backward-kill-sentence
).
See section Sentences.
kill-sentence
).
kill-sexp
). See section Lists and Sexps.
zap-to-char
).
C-w (kill-region
) is a very general kill command; it
kills everything between point and the mark. You can use this command to
kill any contiguous sequence of characters by first setting the mark at
one end of a sequence of characters, then going to the other end and
typing C-w.
A convenient way of killing is combined with searching: M-z
(zap-to-char
) reads a character and kills from point up to (but not
including) the next occurrence of that character in the buffer. If there
is no next occurrence, killing goes to the end of the buffer. A numeric
argument acts as a repeat count. A negative argument means to search
backward and kill text before point.
Other syntactic units can be killed: words, with M-DEL and M-d (see section Words); sexps, with C-M-k (see section Lists and Sexps); and sentences, with C-x DEL and M-k (see section Sentences).
Yanking means getting back text which was killed. Some systems call this "pasting". The usual way to move or copy text is to kill it and then yank it one or more times.
yank
).
yank-pop
).
copy-region-as-kill
).
append-next-kill
).
All killed text is recorded in the kill ring, a list of blocks of text that have been killed. There is only one kill ring, used in all buffers, so you can kill text in one buffer and yank it in another buffer. This is the usual way to move text from one file to another. (See section Accumulating Text, for some other ways.)
If you have two separate Emacs processes, you cannot use the kill ring to move text. If you are using Lucid GNU Emacs under X, however, you can use the X selection mechanism to move text from one to another.
If you are using Lucid GNU Emacs under X and have one Emacs process with multiple screens, they do share the same kill ring. You can kill or copy text in one Emacs screen, then yank it in the other screen belonging to the same process.
The command C-y (yank
) reinserts the text of the most recent
kill. It leaves the cursor at the end of the text and sets the mark at
the beginning of the text. See section Selecting Text.
C-u C-y yanks the text, leaves the cursor in front of the text, and sets the mark after it, if the argument is with just a C-u. Any other argument, including C-u and digits, has different results, described below, under "Yanking Earlier Kills".
To copy a block of text, you can also use M-w
(copy-region-as-kill
), which copies the region into the kill ring
without removing it from the buffer. M-w is similar to C-w
followed by C-y but does not mark the buffer as "modified" and
does not actually cut anything.
Normally, each kill command pushes a new block onto the kill ring. However, two or more kill commands in a row combine their text into a single entry, so that a single C-y yanks it all back. This means you don't have to kill all the text you want to yank in one command; you can kill line after line, or word after word, until you have killed what you want, then get it all back at once using C-y. (Thus we join television in leading people to kill thoughtlessly.)
Commands that kill forward from point add onto the end of the previous killed text. Commands that kill backward from point add onto the beginning. This way, any sequence of mixed forward and backward kill commands puts all the killed text into one entry without rearrangement. Numeric arguments do not break the sequence of appending kills. For example, suppose the buffer contains:
This is the first line of sample text and here is the third.
with point at the beginning of the second line. If you type C-k C-u 2 M-DEL C-k, the first C-k kills the text `line of sample text', C-u 2 M-DEL kills `the first' with the newline that followed it, and the second C-k kills the newline after the second line. The result is that the buffer contains `This is and here is the third.' and a single kill entry contains `the firstRETline of sample textRET'---all the killed text, in its original order.
If a kill command is separated from the last kill command by other
commands (not just numeric arguments), it starts a new entry on the kill
ring. To force a kill command to append, first type the command C-M-w
(append-next-kill
). C-M-w tells the following command,
if it is a kill command, to append the text it kills to the last killed
text, instead of starting a new entry. With C-M-w, you can kill
several separated pieces of text and accumulate them to be yanked back
in one place.
To recover killed text that is no longer the most recent kill, you need
the Meta-y (yank-pop
) command. You can use M-y only
after a C-y or another M-y. It takes the text previously
yanked and replaces it with the text from an earlier kill. To recover
the text of the next-to-the-last kill, first use C-y to recover
the last kill, then M-y to replace it with the previous
kill.
You can think in terms of a "last yank" pointer which points at an item in the kill ring. Each time you kill, the "last yank" pointer moves to the new item at the front of the ring. C-y yanks the item which the "last yank" pointer points to. M-y moves the "last yank" pointer to a different item, and the text in the buffer changes to match. Enough M-y commands can move the pointer to any item in the ring, so you can get any item into the buffer. Eventually the pointer reaches the end of the ring; the next M-y moves it to the first item again.
Yanking moves the "last yank" pointer around the ring, but does not change the order of the entries in the ring, which always runs from the most recent kill at the front to the oldest one still remembered.
Use M-y with a numeric argument to advance the "last yank" pointer by the specified number of items. A negative argument moves the pointer toward the front of the ring; from the front of the ring, it moves to the last entry and starts moving forward from there.
Once the text you are looking for is brought into the buffer, you can stop doing M-y commands and the text will stay there. Since the text is just a copy of the kill ring item, editing it in the buffer does not change what's in the ring. As long you don't kill additional text, the "last yank" pointer remains at the same place in the kill ring: repeating C-y will yank another copy of the same old kill.
If you know how many M-y commands it would take to find the text you want, you can yank that text in one step using C-y with a numeric argument. C-y with an argument greater than one restores the text the specified number of entries back in the kill ring. Thus, C-u 2 C-y gets the next to the last block of killed text. It is equivalent to C-y M-y. C-y with a numeric argument starts counting from the "last yank" pointer, and sets the "last yank" pointer to the entry that it yanks.
The variable kill-ring-max
controls the length of the kill
ring; no more than that many blocks of killed text are saved.
In the X window system, mouse selections provide a simple mechanism for text transfer between different applications. In a typical X application, you can select text by pressing the left mouse button and dragging the cursor over the text you want to copy. The text becomes the primary X selection and is highlighted. The highlighted region is also the Emacs selected region.
There are other kinds of X selections besides the Primary selection; one common one is the Clipboard selection. Some applications prefer to transfer data using this selection in preference to the Primary. One can transfer text from the Primary selection to the Clipboard selection with the Copy command under the Edit menu in the menubar.
Usually, the clipboard selection is not visible. However, if you run the
`xclipboard' application, the text most recently copied to the clipboard
(with the Copy command) is displayed in a window. Any time new text is
thus copied, the `xclipboard' application makes a copy of it and displays
it in its window. The value of the clipboard can survive the lifetime of the
running Emacs process. The xclipboard
man page provides more details.
Warning: If you use the `xclipboard' application, remember that it
maintains a list of all things that have been pasted to the clipboard (that
is, copied with the Copy command). If you don't manually delete elements
from this list by clicking on the Delete button in the xclipboard
window, the clipboard will eventually consume a lot of memory.
In summary, some X applications (such as `xterm') allow one to paste text in them from XEmacs in the following way:
With some other applications (notably, the OpenWindows and Motif tools) you must use this method instead:
X cut buffers are a different, older way of transferring text between applications. XEmacs supports cut buffers for compatibility with older programs, even though selections are now the preferred way of transferring text.
X has a concept of applications "owning" selections. When you select text by clicking and dragging inside an application, the application tells the X server that it owns the selection. When another application asks the X server for the value of the selection, the X server requests the information from the owner. When you use selections, the selection data is not actually transferred unless someone wants it; the act of making a selection doesn't transfer data. Cut buffers are different: when you "own" a cut buffer, the data is actually transferred to the X server immediately, and survives the lifetime of the application.
Any time a region of text becomes the primary selection in Emacs, Emacs also copies that text to the cut buffer. This makes it possible to copy text from a Lucid GNU Emacs buffer and paste it into an older, non-selection-based application (such as Emacs 18).
Note: Older versions of Emacs could not access the X selections, only the X cut buffers.
By default, both the text you select in an Emacs buffer using the click-and-drag mechanism and text you select by setting point and the mark is highlighted. You can use Emacs region commands as well as the Cut and Copy commands on the highlighted region you selected with the mouse.
If you prefer, you can make a distinction between text selected with the
mouse and text selected with point and the mark by setting the variable
zmacs-regions
to nil
. In that case:
Active regions originally come from Zmacs, the Lisp Machine editor. The idea behind them is that commands can only operate on a region when the region is in an "active" state. Put simply, you can only operate on a region that is highlighted.
The variable zmacs-regions
checks whether LISPM-style active
regions should be used. This means that commands that operate on the
region (the area between point and the mark) only work while
the region is in the active state, which is indicated by highlighting.
Most commands causes the region to not be in the active state;
for example, C-w only works immediately after activating the
region.
More specifically:
mark-defun
.
set-mark-command
(C-SPC) pushes a mark and activates the
region. Moving the cursor with normal motion commands (C-n,
C-p, etc.) will cause the region between point and the
recently-pushed mark to be highlighted. It will remain highlighted
until some non-motion comand is executed.
exchange-point-and-mark
(C-x C-x) activates the region.
So if you mark a region and execute a command that operates on it, you
can reactivate the same region with C-x C-x (or perhaps C-x
C-x C-x C-x) to operate on it again.
Generally, commands that push marks as a means of navigation, such as
beginning-of-buffer
(M-<) and end-of-buffer
(M->), do not activate the region. However, commands that push
marks as a means of marking an area of text, such as mark-defun
(M-C-h), mark-word
(M-@), and mark-whole-buffer
(C-x h), do activate the region.
When zmacs-regions
is t
, there is no distinction between
the primary X selection and the active region selected by point and the
mark. To see this, set the mark (C-SPC) and move the cursor
with any cursor-motion command: the region between point and mark is
highlighted, and you can watch it grow and shrink as you move the
cursor.
Any other commands besides cursor-motion commands (such as inserting or deleting text) will cause the region to no longer be active; it will no longer be highlighted, and will no longer be the primary selection. Errors also remove highlighting from a region.
Commands that require a region (such as C-w) signal an error if
the region is not active. Certain commands cause the region to be in
its active state. The most common ones are push-mark
(C-SPC) and exchange-point-and-mark
(C-x C-x).
When zmacs-regions
is t
, programs can be non-intrusive
on the state of the region by setting the variable zmacs-region-stays
to a non-nil
value. If you are writing a new Emacs command that
is conceptually a "motion" command and should not interfere with the
current highlightedness of the region, then you may set this variable.
It is reset to nil
after each user command is executed.
When zmacs-regions
is t
, programs can make the region between
point and mark go into the active (highlighted) state by using the
function zmacs-activate-region
. Only a small number of commands
should ever do this.
When zmacs-regions
is t
, programs can deactivate the region
between point and the mark by using zmacs-deactivate-region
.
Note: you should not have to call this function; the command loop calls
it when appropriate.
Usually you copy or move text by killing it and yanking it, but there are other ways that are useful for copying one block of text in many places, or for copying many scattered blocks of text into one place.
If you like, you can accumulate blocks of text from scattered locations either into a buffer or into a file. The relevant commands are described here. You can also use Emacs registers for storing and accumulating text. See section Registers.
append-to-buffer
).
To accumulate text into a buffer, use the command M-x append-to-buffer, which inserts a copy of the region into the buffer buffername, at the location of point in that buffer. If there is no buffer with the given name, one is created.
If you append text to a buffer that has been used for editing, the
copied text goes to the place where point is. Point in that buffer is
left at the end of the copied text, so successive uses of
append-to-buffer
accumulate the text in the specified buffer in
the same order as they were copied. Strictly speaking, this command does
not always append to the text already in the buffer; but if this command
is the only command used to alter a buffer, it does always append to the
existing text because point is always at the end.
M-x prepend-to-buffer is similar to append-to-buffer
, but
point in the other buffer is left before the copied text, so successive
prependings add text in reverse order. M-x copy-to-buffer is
similar, except that any existing text in the other buffer is deleted,
so the buffer is left containing just the text newly copied into it.
You can retrieve the accumulated text from that buffer with M-x
insert-buffer, which takes buffername as an argument. It inserts
a copy of the text in buffer buffername into the selected buffer.
You could alternatively select the other buffer for editing, perhaps moving
text from it by killing or with append-to-buffer
. See section Using Multiple Buffers, for
background information on buffers.
Instead of accumulating text within Emacs in a buffer, you can append text directly into a file with M-x append-to-file, which takes file-name as an argument. It adds the text of the region to the end of the specified file. The file is changed immediately on disk. This command is normally used with files that are not being visited in Emacs. Using it on a file that Emacs is visiting can produce confusing results, because the file's text inside Emacs does not change while the file itself changes.
The rectangle commands affect rectangular areas of text: all characters between a certain pair of columns, in a certain range of lines. Commands are provided to kill rectangles, yank killed rectangles, clear them out, or delete them. Rectangle commands are useful with text in multicolumnar formats, like code with comments at the right, or for changing text into or out of such formats.
To specify the rectangle a command should work on, put the mark at one corner and point at the opposite corner. The specified rectangle is called the region-rectangle because it is controlled about the same way the region is controlled. Remember that a given combination of point and mark values can be interpreted either as specifying a region or as specifying a rectangle; it is up to the command that uses them to choose the interpretation.
The rectangle operations fall into two classes: commands deleting and moving rectangles, and commands for blank rectangles.
There are two ways to get rid of the text in a rectangle: you can discard the text (delete it) or save it as the "last killed" rectangle. The commands for these two ways are M-x delete-rectangle and M-x kill-rectangle. In either case, the portion of each line that falls inside the rectangle's boundaries is deleted, causing following text (if any) on the line to move left.
Note that "killing" a rectangle is not killing in the usual sense; the rectangle is not stored in the kill ring, but in a special place that only records the most recently killed rectangle (that is, does not append to a killed rectangle). Different yank commands have to be used and only one rectangle is stored, because yanking a rectangle is quite different from yanking linear text and yank-popping commands are difficult to make sense of.
Inserting a rectangle is the opposite of deleting one. You specify where to put the upper left corner by putting point there. The rectangle's first line is inserted at point, the rectangle's second line is inserted at a point one line vertically down, and so on. The number of lines affected is determined by the height of the saved rectangle.
To insert the last killed rectangle, type M-x yank-rectangle. This can be used to convert single-column lists into double-column lists; kill the second half of the list as a rectangle and then yank it beside the first line of the list.
There are two commands for working with blank rectangles: M-x clear-rectangle erases existing text, and M-x open-rectangle inserts a blank rectangle. Clearing a rectangle is equivalent to deleting it and then inserting a blank rectangle of the same size.
Rectangles can also be copied into and out of registers. See section Saving Rectangles in Registers.
Emacs registers are places in which you can save text or positions for later use. Text saved in a register can be copied into the buffer once or many times; a position saved in a register is used by moving point to that position. Rectangles can also be copied into and out of registers (see section Rectangles).
Each register has a name, which is a single character. A register can store either a piece of text, a position, or a rectangle, but only one thing at any given time. Whatever you store in a register remains there until you store something else in that register.
M-x view-register reads a register name as an argument and then displays the contents of the specified register.
Saving a position records a spot in a buffer so you can move back there later. Moving to a saved position re-selects the buffer and moves point to the spot.
point-to-register
).
register-to-point
).
To save the current location of point in a register, choose a name r and type C-x r SPC r. The register r retains the location thus saved until you store something else in that register.
The command C-x r j r moves point to the location recorded in register r. The register is not affected; it continues to record the same location. You can jump to the same position using the same register as often as you want.
When you want to insert a copy of the same piece of text many times, it
can be impractical to use the kill ring, since each subsequent kill moves
the piece of text further down on the ring. It becomes hard to keep
track of the argument needed to retrieve the same text with C-y. An
alternative is to store the text in a register with C-x r s
(copy-to-register
) and then retrieve it with C-x r g
(insert-register
).
copy-to-register
).
insert-register
).
C-x r s r stores a copy of the text of the region into the register named r. Given a numeric argument, C-x r s deletes the text from the buffer as well.
C-x r g r inserts the text from register r in the buffer. By default it leaves point before the text and places the mark after it. With a numeric argument, it puts point after the text and the mark before it.
A register can contain a rectangle instead of lines of text. The rectangle is represented as a list of strings. See section Rectangles, for basic information on rectangles and how to specify rectangles in a buffer.
copy-rectangle-to-register
).
With a numeric argument, delete it as well.
insert-register
).
The C-x r g command inserts linear text if the register contains that, or inserts a rectangle if the register contains one.
Since only part of a large buffer fits in the window, Emacs tries to show the part that is likely to be interesting. The display control commands allow you to specify which part of the text you want to see.
recenter
).
scroll-up
).
scroll-down
).
recenter
).
scroll-left
).
scroll-right
).
set-selective-display
).
If a buffer contains text that is too large to fit entirely within the window that is displaying the buffer, Emacs shows a contiguous section of the text. The section shown always contains point.
Scrolling means moving text up or down in the window so that different parts of the text are visible. Scrolling forward means that text moves up, and new text appears at the bottom. Scrolling backward moves text down and new text appears at the top.
Scrolling happens automatically if you move point past the bottom or top of the window. You can also explicitly request scrolling with the commands in this section.
The most basic scrolling command is C-l (recenter
) with no
argument. It clears the entire screen and redisplays all windows. In
addition, it scrolls the selected window so that point is halfway down
from the top of the window.
The scrolling commands C-v and M-v let you move all the text
in the window up or down a few lines. C-v (scroll-up
) with an
argument shows you that many more lines at the bottom of the window, moving
the text and point up together as C-l might. C-v with a
negative argument shows you more lines at the top of the window.
Meta-v (scroll-down
) is like C-v, but moves in the
opposite direction.
To read the buffer a windowful at a time, use C-v with no
argument. C-v takes the last two lines at the bottom of the
window and puts them at the top, followed by nearly a whole windowful of
lines not previously visible. Point moves to the new top of the window
if it was in the text scrolled off the top. M-v with no argument
moves backward with similar overlap. The number of lines of overlap
across a C-v or M-v is controlled by the variable
next-screen-context-lines
; by default, it is two.
Another way to scroll is using C-l with a numeric argument. C-l does not clear the screen when given an argument; it only scrolls the selected window. With a positive argument n, C-l repositions text to put point n lines down from the top. An argument of zero puts point on the very top line. Point does not move with respect to the text; rather, the text and point move rigidly on the screen. C-l with a negative argument puts point that many lines from the bottom of the window. For example, C-u - 1 C-l puts point on the bottom line, and C-u - 5 C-l puts it five lines from the bottom. Just C-u as argument, as in C-u C-l, scrolls point to the center of the screen.
Scrolling happens automatically if point has moved out of the visible
portion of the text when it is time to display. Usually scrolling is
done to put point vertically centered within the window. However, if
the variable scroll-step
has a non-zero value, an attempt is made to
scroll the buffer by that many lines; if that is enough to bring point back
into visibility, that is what happens.
The text in a window can also be scrolled horizontally. This means that each line of text is shifted sideways in the window, and one or more characters at the beginning of each line are not displayed at all. When a window has been scrolled horizontally in this way, text lines are truncated rather than continued (see section Continuation Lines), with a `$' appearing in the first column when there is text truncated to the left, and in the last column when there is text truncated to the right.
The command C-x < (scroll-left
) scrolls the selected
window to the left by n columns with argument n. With no
argument, it scrolls by almost the full width of the window (two columns
less, to be precise). C-x > (scroll-right
) scrolls
similarly to the right. The window cannot be scrolled any farther to
the right once it is displaying normally (with each line starting at the
window's left margin); attempting to do so has no effect.
Emacs can hide lines indented more than a certain number of columns (you specify how many columns). This allows you to get an overview of a part of a program.
To hide lines, type C-x $ (set-selective-display
) with a
numeric argument n. (See section Numeric Arguments, for information on giving
the argument.) Lines with at least n columns of indentation
disappear from the screen. The only indication of their presence are
three dots (`...'), which appear at the end of each visible
line that is followed by one or more invisible ones.
The invisible lines are still present in the buffer, and most editing commands see them as usual, so it is very easy to put point in the middle of invisible text. When this happens, the cursor appears at the end of the previous line, after the three dots. If point is at the end of the visible line, before the newline that ends it, the cursor appears before the three dots.
The commands C-n and C-p move across the invisible lines as if they were not there.
To make everything visible again, type C-x $ with no argument.
This section contains information for customization only. Beginning users should skip it.
The variable mode-line-inverse-video
controls whether the mode
line is displayed in inverse video (assuming the terminal supports it);
nil
means don't do so. See section The Mode Line.
If the variable inverse-video
is non-nil
, Emacs attempts
to invert all the lines of the display from what they normally are.
When you reenter Emacs after suspending, Emacs normally clears the
screen and redraws the entire display. On some terminals with more than
one page of memory, it is possible to arrange the termcap entry so that
the `ti' and `te' strings (output to the terminal when Emacs
is entered and exited, respectively) switch between pages of memory so
as to use one page for Emacs and another page for other output. In that
case, you might want to set the variable no-redraw-on-reenter
to
non-nil
so that Emacs will assume, when resumed, that the screen
page it is using still contains what Emacs last wrote there.
The variable echo-keystrokes
controls the echoing of multi-character
keys; its value is the number of seconds of pause required to cause echoing
to start, or zero, meaning don't echo at all. See section The Echo Area.
If the variable ctl-arrow
is nil
, control characters in the
buffer are displayed with octal escape sequences, all except newline and
tab. If its value is t
, then control characters will be printed
with an up-arrow, for example ^A.
If its value is not t
and not nil
, then characters whose
code is greater than 160 (that is, the space character (32) with its
high bit set) will be assumed to be printable, and will be displayed
without alteration. This is the default when running under X Windows,
since XEmacs assumes an ISO/8859-1 character set (also known as
"Latin1"). The ctl-arrow
variable may also be set to an
integer, in which case all characters whose codes are greater than or
equal to that value will be assumed to be printable.
Altering the value of ctl-arrow
makes it local to the current
buffer; until that time, the default value is in effect. See section Local Variables.
Normally, a tab character in the buffer is displayed as whitespace which
extends to the next display tab stop position, and display tab stops come
at intervals equal to eight spaces. The number of spaces per tab is
controlled by the variable tab-width
, which is made local by
changing it, just like ctl-arrow
. Note that how the tab character
in the buffer is displayed has nothing to do with the definition of
TAB as a command.
If you set the variable selective-display-ellipses
to nil
,
the three dots at the end of a line that precedes invisible
lines do not appear. There is no visible indication of the invisible lines.
This variable becomes local automatically when set.
Like other editors, Emacs has commands for searching for occurrences of a string. The principal search command is unusual in that it is incremental: it begins to search before you have finished typing the search string. There are also non-incremental search commands more like those of other editors.
Besides the usual replace-string
command that finds all
occurrences of one string and replaces them with another, Emacs has a fancy
replacement command called query-replace
which asks interactively
which occurrences to replace.
An incremental search begins searching as soon as you type the first character of the search string. As you type in the search string, Emacs shows you where the string (as you have typed it so far) is found. When you have typed enough characters to identify the place you want, you can stop. Depending on what you do next, you may or may not need to terminate the search explicitly with a RET.
isearch-forward
).
isearch-backward
).
C-s starts an incremental search. C-s reads characters from the keyboard and positions the cursor at the first occurrence of the characters that you have typed. If you type C-s and then F, the cursor moves right after the first `F'. Type an O, and see the cursor move to after the first `FO'. After another O, the cursor is after the first `FOO' after the place where you started the search. Meanwhile, the search string `FOO' has been echoed in the echo area.
The echo area display ends with three dots when actual searching is going on. When search is waiting for more input, the three dots are removed. (On slow terminals, the three dots are not displayed.)
If you make a mistake in typing the search string, you can erase characters with DEL. Each DEL cancels the last character of the search string. This does not happen until Emacs is ready to read another input character; first it must either find, or fail to find, the character you want to erase. If you do not want to wait for this to happen, use C-g as described below.
When you are satisfied with the place you have reached, you can type RET (or C-m), which stops searching, leaving the cursor where the search brought it. Any command not specially meaningful in searches also stops the search and is then executed. Thus, typing C-a exits the search and then moves to the beginning of the line. RET is necessary only if the next command you want to type is a printing character, DEL, ESC, or another control character that is special within searches (C-q, C-w, C-r, C-s, or C-y).
Sometimes you search for `FOO' and find it, but were actually looking for a different occurance of it. To move to the next occurrence of the search string, type another C-s. Do this as often as necessary. If you overshoot, you can cancel some C-s characters with DEL.
After you exit a search, you can search for the same string again by typing just C-s C-s: the first C-s is the key that invokes incremental search, and the second C-s means "search again".
If the specified string is not found at all, the echo area displays the text `Failing I-Search'. The cursor is after the place where Emacs found as much of your string as it could. Thus, if you search for `FOOT', and there is no `FOOT', the cursor may be after the `FOO' in `FOOL'. At this point there are several things you can do. If you mistyped the search string, correct it. If you like the place you have found, you can type RET or some other Emacs command to "accept what the search offered". Or you can type C-g, which removes from the search string the characters that could not be found (the `T' in `FOOT'), leaving those that were found (the `FOO' in `FOOT'). A second C-g at that point cancels the search entirely, returning point to where it was when the search started.
If a search is failing and you ask to repeat it by typing another C-s, it starts again from the beginning of the buffer. Repeating a failing backward search with C-r starts again from the end. This is called wrapping around. `Wrapped' appears in the search prompt once this has happened.
The C-g "quit" character does special things during searches; just what it does depends on the status of the search. If the search has found what you specified and is waiting for input, C-g cancels the entire search. The cursor moves back to where you started the search. If C-g is typed when there are characters in the search string that have not been found--because Emacs is still searching for them, or because it has failed to find them--then the search string characters which have not been found are discarded from the search string. The search is now successful and waiting for more input, so a second C-g cancels the entire search.
To search for a control character such as C-s or DEL or ESC, you must quote it by typing C-q first. This function of C-q is analogous to its meaning as an Emacs command: it causes the following character to be treated the way a graphic character would normally be treated in the same context.
To search backwards, you can use C-r instead of C-s to
start the search; C-r is the key that runs the command
(isearch-backward
) to search backward. You can also use
C-r to change from searching forward to searching backwards. Do
this if a search fails because the place you started was too far down in the
file. Repeated C-r keeps looking for more occurrences backwards.
C-s starts going forward again. You can cancel C-r in a
search with DEL.
The characters C-y and C-w can be used in incremental search to grab text from the buffer into the search string. This makes it convenient to search for another occurrence of text at point. C-w copies the word after point as part of the search string, advancing point over that word. Another C-s to repeat the search will then search for a string including that word. C-y is similar to C-w but copies the rest of the current line into the search string.
The characters M-p and M-n can be used in an incremental search to recall things which you have searched for in the past. A list of the last 16 things you have searched for is retained, and M-p and M-n let you cycle through that ring.
The character M-TAB does completion on the elements in
the search history ring. For example, if you know that you have
recently searched for the string POTATOE
, you could type
C-s P O M-TAB. If you had searched for other strings
beginning with PO
then you would be shown a list of them, and
would need to type more to select one.
You can change any of the special characters in incremental search via
the normal keybinding mechanism: simply add a binding to the
isearch-mode-map
. For example, to make the character
C-b mean "search backwards" while in isearch-mode, do this:
(define-key isearch-mode-map "\C-b" 'isearch-repeat-backward)
These are the default bindings of isearch-mode:
isearch-delete-char
).
isearch-exit
).
isearch-quote-char
).
isearch-repeat-forward
).
isearch-repeat-backward
).
isearch-yank-line
).
isearch-yank-word
).
isearch-abort
).
isearch-ring-retreat
).
isearch-ring-advance
).
isearch-complete
).
Any other character which is normally inserted into a buffer when typed is automatically added to the search string in isearch-mode.
Incremental search on a slow terminal uses a modified style of display that is designed to take less time. Instead of redisplaying the buffer at each place the search gets to, it creates a new single-line window and uses that to display the line the search has found. The single-line window appears as soon as point gets outside of the text that is already on the screen.
When the search is terminated, the single-line window is removed. Only at this time the window in which the search was done is redisplayed to show its new value of point.
The three dots at the end of the search string, normally used to indicate that searching is going on, are not displayed in slow style display.
The slow terminal style of display is used when the terminal baud rate is
less than or equal to the value of the variable search-slow-speed
,
initially 1200.
The number of lines to use in slow terminal search display is controlled
by the variable search-slow-window-lines
. Its normal value is 1.
Emacs also has conventional non-incremental search commands, which require you type the entire search string before searching begins.
To do a non-incremental search, first type C-s RET (or C-s C-m). This enters the minibuffer to read the search string. Terminate the string with RET to start the search. If the string is not found, the search command gets an error.
By default, C-s invokes incremental search, but if you give it an empty argument, which would otherwise be useless, it invokes non-incremental search. Therefore, C-s RET invokes non-incremental search. C-r RET also works this way.
Forward and backward non-incremental searches are implemented by the
commands search-forward
and search-backward
. You can bind
these commands to keys. The reason that incremental
search is programmed to invoke them as well is that C-s RET
is the traditional sequence of characters used in Emacs to invoke
non-incremental search.
Non-incremental searches performed using C-s RET do
not call search-forward
right away. They first check
if the next character is C-w, which requests a word search.
Word search looks for a sequence of words without regard to how the words are separated. More precisely, you type a string of many words, using single spaces to separate them, and the string is found even if there are multiple spaces, newlines or other punctuation between the words.
Word search is useful in editing documents formatted by text formatters. If you edit while looking at the printed, formatted version, you can't tell where the line breaks are in the source file. Word search, allows you to search without having to know the line breaks.
Word search is a special case of non-incremental search. It is invoked with C-s RET C-w followed by the search string, which must always be terminated with another RET. Being non-incremental, this search does not start until the argument is terminated. It works by constructing a regular expression and searching for that. See section Regular Expression Search.
You can do a backward word search with C-r RET C-w.
Forward and backward word searches are implemented by the commands
word-search-forward
and word-search-backward
. You can
bind these commands to keys. The reason that incremental
search is programmed to invoke them as well is that C-s RET C-w
is the traditional Emacs sequence of keys for word search.
A regular expression (regexp, for short) is a pattern that denotes a set of strings, possibly an infinite set. Searching for matches for a regexp is a powerful operation that editors on Unix systems have traditionally offered. In GNU Emacs, you can search for the next match for a regexp either incrementally or not.
Incremental search for a regexp is done by typing M-C-s
(isearch-forward-regexp
). This command reads a search string
incrementally just like C-s, but it treats the search string as a
regexp rather than looking for an exact match against the text in the
buffer. Each time you add text to the search string, you make the regexp
longer, and the new regexp is searched for. A reverse regexp search command
isearch-backward-regexp
also exists, but no key runs it.
All of the control characters that do special things within an ordinary incremental search have the same functionality in incremental regexp search. Typing C-s or C-r immediately after starting a search retrieves the last incremental search regexp used: incremental regexp and non-regexp searches have independent defaults.
Non-incremental search for a regexp is done by the functions
re-search-forward
and re-search-backward
. You can invoke
them with M-x or bind them to keys. You can also call
re-search-forward
by way of incremental regexp search with
M-C-s RET.
Regular expressions have a syntax in which a few characters are special constructs and the rest are ordinary. An ordinary character is a simple regular expression which matches that character and nothing else. The special characters are `$', `^', `.', `*', `+', `?', `[', `]' and `\'; no new special characters will be defined. Any other character appearing in a regular expression is ordinary, unless a `\' precedes it.
For example, `f' is not a special character, so it is ordinary, and therefore `f' is a regular expression that matches the string `f' and no other string. (It does not match the string `ff'.) Likewise, `o' is a regular expression that matches only `o'.
Any two regular expressions a and b can be concatenated. The result is a regular expression which matches a string if a matches some amount of the beginning of that string and b matches the rest of the string.
As a simple example, you can concatenate the regular expressions `f' and `o' to get the regular expression `fo', which matches only the string `fo'. To do something nontrivial, you need to use one of the following special characters:
Note: for historical compatibility, special characters are treated as ordinary ones if they are in contexts where their special meanings make no sense. For example, `*foo' treats `*' as ordinary since there is no preceding expression on which the `*' can act. It is poor practice to depend on this behavior; better to quote the special character anyway, regardless of where is appears.
Usually, `\' followed by any character matches only that character. However, there are several exceptions: characters which, when preceded by `\', are special constructs. Such characters are always ordinary when encountered on their own. Here is a table of `\' constructs.
Here is a complicated regexp used by Emacs to recognize the end of a sentence together with any whitespace that follows. It is given in Lisp syntax to enable you to distinguish the spaces from the tab characters. In Lisp syntax, the string constant begins and ends with a double-quote. `\"' stands for a double-quote as part of the regexp, `\\' for a backslash as part of the regexp, `\t' for a tab and `\n' for a newline.
"[.?!][]\"')]*\\($\\|\t\\| \\)[ \t\n]*"
This regexp contains four parts: a character set matching period, `?' or `!'; a character set matching close-brackets, quotes or parentheses, repeated any number of times; an alternative in backslash-parentheses that matches end-of-line, a tab or two spaces; and a character set matching whitespace characters, repeated any number of times.
All searches in Emacs normally ignore the case of the text they are searching through; if you specify searching for `FOO', `Foo' and `foo' are also considered a match. Regexps, and in particular character sets, are included: `[aB]' matches `a' or `A' or `b' or `B'.
If you want a case-sensitive search, set the variable
case-fold-search
to nil
. Then all letters must match
exactly, including case. case-fold-search
is a per-buffer
variable; altering it affects only the current buffer, but
there is a default value which you can change as well. See section Local Variables.
You can also use Case Sensitive Search from the Options menu
on your screen.
Global search-and-replace operations are not needed as often in Emacs as
they are in other editors, but they are available. In addition to the
simple replace-string
command which is like that found in most
editors, there is a query-replace
command which asks you, for each
occurrence of a pattern, whether to replace it.
The replace commands all replace one string (or regexp) with one
replacement string. It is possible to perform several replacements in
parallel using the command expand-region-abbrevs
. See section Controlling Abbrev Expansion.
To replace every instance of `foo' after point with `bar', use the command M-x replace-string with the two arguments `foo' and `bar'. Replacement occurs only after point: if you want to cover the whole buffer you must go to the beginning first. By default, all occurrences up to the end of the buffer are replaced. To limit replacement to part of the buffer, narrow to that part of the buffer before doing the replacement (see section Narrowing).
When replace-string
exits, point is left at the last occurrence
replaced. The value of point when the replace-string
command was
issued is remembered on the mark ring; C-u C-SPC moves back
there.
A numeric argument restricts replacement to matches that are surrounded by word boundaries.
replace-string
replaces exact matches for a single string. The
similar command replace-regexp
replaces any match for a specified
pattern.
In replace-regexp
, the newstring need not be constant. It
can refer to all or part of what is matched by the regexp. `\&'
in newstring stands for the entire text being replaced.
`\d' in newstring, where d is a digit, stands for
whatever matched the d'th parenthesized grouping in regexp.
For example,
M-x replace-regexp RET c[ad]+r RET \&-safe RET
would replace (for example) `cadr' with `cadr-safe' and `cddr' with `cddr-safe'.
M-x replace-regexp RET \(c[ad]+r\)-safe RET \1 RET
would perform exactly the opposite replacements. To include a `\' in the text to replace with, you must give `\\'.
If the arguments to a replace command are in lower case, the command preserves case when it makes a replacement. Thus, the following command:
M-x replace-string RET foo RET bar RET
replaces a lower-case `foo' with a lower case `bar', `FOO'
with `BAR', and `Foo' with `Bar'. If upper-case letters are
used in the second argument, they remain upper-case every time that
argument is inserted. If upper-case letters are used in the first
argument, the second argument is always substituted exactly as given, with
no case conversion. Likewise, if the variable case-replace
is set
to nil
, replacement is done without case conversion. If
case-fold-search
is set to nil
, case is significant in
matching occurrences of `foo' to replace; also, case conversion of the
replacement string is not done.
If you want to change only some of the occurrences of `foo' to
`bar', not all of them, you can use query-replace
instead of
M-%. This command finds occurrences of `foo' one by one,
displays each occurrence, and asks you whether to replace it. A numeric
argument to query-replace
tells it to consider only occurrences
that are bounded by word-delimiter characters.
Aside from querying, query-replace
works just like
replace-string
, and query-replace-regexp
works
just like replace-regexp
.
The things you can type when you are shown an occurrence of string or a match for regexp are:
replace-string
, provided case-replace
is non-nil
,
as it normally is.
query-replace
. If you want to do further
replacement you must use C-x ESC to restart (see section Repeating Minibuffer Commands).
query-replace
.
If you type any other character, Emacs exits the query-replace
, and
executes the character as a command. To restart the query-replace
,
use C-x ESC, which repeats the query-replace
because it
used the minibuffer to read its arguments. See section Repeating Minibuffer Commands.
Here are some other commands that find matches for a regular expression. They all operate from point to the end of the buffer.
This chapter describes commands that are especially useful when you catch a mistake in your text just after you have made it, or when you change your mind while composing text on line.
delete-backward-char
).
backward-kill-word
).
backward-kill-sentence
).
The DEL character (delete-backward-char
) is the most
important correction command. When used among graphic (self-inserting)
characters, it can be thought of as canceling the last character typed.
When your mistake is longer than a couple of characters, it might be more convenient to use M-DEL or C-x DEL. M-DEL kills back to the start of the last word, and C-x DEL kills back to the start of the last sentence. C-x DEL is particularly useful when you are thinking of what to write as you type it, in case you change your mind about phrasing. M-DEL and C-x DEL save the killed text for C-y and M-y to retrieve. See section Yanking.
M-DEL is often useful even when you have typed only a few characters wrong, if you know you are confused in your typing and aren't sure exactly what you typed. At such a time, you cannot correct with DEL except by looking at the screen to see what you did. It requires less thought to kill the whole word and start over.
transpose-chars
).
transpose-words
).
transpose-sexps
).
transpose-lines
).
The common error of transposing two adjacent characters can be fixed
with the C-t command (transpose-chars
). Normally,
C-t transposes the two characters on either side of point. When
given at the end of a line, C-t transposes the last two characters
on the line, rather than transposing the last character of the line with
the newline, which would be useless. If you catch a
transposition error right away, you can fix it with just C-t.
If you catch the error later, move the cursor back to between
the two transposed characters. If you transposed a space with the last
character of the word before it, the word motion commands are a good way
of getting there. Otherwise, a reverse search (C-r) is often the
best way. See section Searching and Replacement.
Meta-t (transpose-words
) transposes the word before point
with the word after point. It moves point forward over a word, dragging
the word preceding or containing point forward as well. The punctuation
characters between the words do not move. For example, `FOO, BAR'
transposes into `BAR, FOO' rather than `BAR FOO,'.
C-M-t (transpose-sexps
) is a similar command for transposing
two expressions (see section Lists and Sexps), and C-x C-t (transpose-lines
)
exchanges lines. It works like M-t but in determines the
division of the text into syntactic units differently.
A numeric argument to a transpose command serves as a repeat count: it tells the transpose command to move the character (word, sexp, line) before or containing point across several other characters (words, sexps, lines). For example, C-u 3 C-t moves the character before point forward across three other characters. This is equivalent to repeating C-t three times. C-u - 4 M-t moves the word before point backward across four words. C-u - C-M-t would cancel the effect of plain C-M-t.
A numeric argument of zero transposes the character (word, sexp, line) ending after point with the one ending after the mark (otherwise a command with a repeat count of zero would do nothing).
A common error is to type words in the wrong case. Because of this, the word case-conversion commands M-l, M-u, and M-c do not move the cursor when used with a negative argument. As soon as you see you have mistyped the last word, you can simply case-convert it and continue typing. See section Case Conversion Commands.
spell-word
).
To check the spelling of the word before point, and optionally correct
it, use the command M-$ (spell-word
). This command runs an
inferior process containing the spell
program to see whether the
word is correct English. If it is not, it asks you to edit the word (in
the minibuffer) into a corrected spelling, and then performs a
query-replace
to substitute the corrected spelling for the old
one throughout the buffer.
If you exit the minibuffer without altering the original spelling, it
means you do not want to do anything to that word. In that case, the
query-replace
is not done.
M-x spell-buffer checks each word in the buffer the same way that
spell-word
does, doing a query-replace
for
every incorrect word if appropriate.
M-x spell-region is similar to spell-buffer
but operates
only on the region, not the entire buffer.
M-x spell-string reads a string as an argument and checks whether that is a correctly spelled English word. It prints a message giving the answer in the echo area.
The basic unit of stored data in Unix is the file. To edit a file, you must tell Emacs to examine the file and prepare a buffer containing a copy of the file's text. This is called visiting the file. Editing commands apply directly to text in the buffer; that is, to the copy inside Emacs. Your changes appear in the file itself only when you save the buffer back into the file.
In addition to visiting and saving files, Emacs can delete, copy, rename, and append to files, and operate on file directories.
Most Emacs commands that operate on a file require you to specify the file name. (Saving and reverting are exceptions; the buffer knows which file name to use for them.) File names are specified in the minibuffer (see section The Minibuffer). Completion is available, to make it easier to specify long file names. See section Completion.
There is always a default file name which is used if you enter an empty argument by typing just RET. Normally the default file name is the name of the file visited in the current buffer; this makes it easy to operate on that file with any of the Emacs file commands.
Each buffer has a default directory, normally the same as the
directory of the file visited in that buffer. When Emacs reads a file
name, the default directory is used if you do not specify a directory.
If you specify a directory in a relative fashion, with a name that does
not start with a slash, it is interpreted with respect to the default
directory. The default directory of the current buffer is kept in the
variable default-directory
, which has a separate value in every
buffer. The value of the variable should end with a slash.
For example, if the default file name is `/u/rms/gnu/gnu.tasks' then the default directory is `/u/rms/gnu/'. If you type just `foo', which does not specify a directory, it is short for `/u/rms/gnu/foo'. `../.login' would stand for `/u/rms/.login'. `new/foo' would stand for the filename `/u/rms/gnu/new/foo'.
The variable default-directory-alist
takes an alist of major
modes and their opinions on default-directory
as a Lisp
expression to evaluate. A resulting value of nil
is ignored in
favor of default-directory
.
You can create a new directory with the function make-directory
,
which takes as an argument a file name string. The current directory is
displayed in the minibuffer when the function is called; you can delete
the old directory name and supply a new directory name. For example, if
the current directory is `/u/rms/gnu', you can delete `gnu'
and type `oryx' and RET to create `/u/rms/oryx'.
Removing a directory is similar to creating one. To remove a directory,
use remove-directory
; it takes one argument, a file name string.
The command M-x pwd prints the current buffer's default directory,
and the command M-x cd sets it (to a value read using the
minibuffer). A buffer's default directory changes only when the cd
command is used. A file-visiting buffer's default directory is initialized
to the directory of the file that is visited there. If a buffer is created
with C-x b, its default directory is copied from that of the
buffer that was current at the time.
The default directory name actually appears in the minibuffer when the
minibuffer becomes active to read a file name. This serves two
purposes: it shows you what the default is, so that you can type a
relative file name and know with certainty what it will mean, and it
allows you to edit the default to specify a different directory. To
inhibit the insertion of the default directory, set the variable
insert-default-directory
to nil
.
Note that it is legitimate to type an absolute file name after you enter the minibuffer, ignoring the presence of the default directory name. The final minibuffer contents may look invalid, but that is not so. See section Minibuffers for File Names.
`$' in a file name is used to substitute environment variables. For example, if you have used the shell command `setenv FOO rms/hacks' to set up an environment variable named `FOO', then you can use `/u/$FOO/test.c' or `/u/${FOO}/test.c' as an abbreviation for `/u/rms/hacks/test.c'. The environment variable name consists of all the alphanumeric characters after the `$'; alternatively, it may be enclosed in braces after the `$'. Note that the `setenv' command affects Emacs only if done before Emacs is started.
To access a file with `$' in its name, type `$$'. This pair
is converted to a single `$' at the same time variable substitution
is performed for single `$'. The Lisp function that performs the
substitution is called substitute-in-file-name
. The substitution
is performed only on filenames read as such using the minibuffer.
find-file
).
find-alternate-file
).
find-file-other-window
). Don't
change this window.
find-file-other-screen
). Don't
change this window or screen.
Visiting a file means copying its contents into an Emacs buffer so you can edit it. Emacs creates a new buffer for each file you visit. We say that the buffer is visiting the file that it was created to hold. Emacs constructs the buffer name from the file name by throwing away the directory and keeping just the file name. For example, a file named `/usr/rms/emacs.tex' is displayed in a buffer named `emacs.tex'. If a buffer with that name exists, a unique name is constructed by appending `<2>', `<3>',and so on, using the lowest number that makes a name that is not already in use.
Each window's mode line shows the name of the buffer that is being displayed in that window, so you can always tell what buffer you are editing.
The changes you make with Emacs are made in the Emacs buffer. They do not take effect in the file that you visit, or any other permanent place, until you save the buffer. Saving the buffer means that Emacs writes the current contents of the buffer into its visited file. See section Saving Files.
If a buffer contains changes that have not been saved, the buffer is said to be modified. This is important because it implies that some changes will be lost if the buffer is not saved. The mode line displays two stars near the left margin if the buffer is modified.
To visit a file, use the command C-x C-f (find-file
). Follow
the command with the name of the file you wish to visit, terminated by a
RET. If you are using Lucid GNU Emacs under X, you can also use the
Open... command from the File menu bar item.
The file name is read using the minibuffer (see section The Minibuffer), with defaulting and completion in the standard manner (see section File Names). While in the minibuffer, you can abort C-x C-f by typing C-g.
C-x C-f has completed successfully when text appears on the screen and a new buffer name appears in the mode line. If the specified file does not exist and could not be created or cannot be read, an error results. The error message is printed in the echo area, and includes the name of the file that Emacs was trying to visit.
If you visit a file that is already in Emacs, C-x C-f does not make another copy. It selects the existing buffer containing that file. However, before doing so, it checks that the file itself has not changed since you visited or saved it last. If the file has changed, Emacs prints a warning message. See section Protection Against Simultaneous Editing.
You can switch to a specific file called out in the current buffer by
calling the function find-this-file
. By providing a prefix
argument, this function calls filename-at-point
and switches to a
buffer visiting the file filename. It creates one if none already
exists. You can use this function to edit the file mentioned in the
buffer you are working in or to test if the file exists. You can do that
by using the minibuffer completion after snatching the all or part of
the filename.
If the variable find-file-use-truenames
's value is
non-nil
, a buffer's visited filename will always be traced back
to the real file. The filename will never be a symbolic link, and there
will never be a symbolic link anywhere in its directory path. In other
words, the buffer-file-name
and buffer-file-truename
will
be equal.
If the variable find-file-compare-truenames
value is
non-nil
, the find-file
command will check the
buffer-file-truename
of all visited files when deciding whether a
given file is already in a buffer, instead of just
buffer-file-name
. If you attempt to visit another file which is
a hard-link or symbolic-link to a file that is already in a buffer, the
existing buffer will be found instead of a newly created one.
If you want to create a file, just visit it. Emacs prints `(New File)' in the echo area, but in other respects behaves as if you had visited an existing empty file. If you make any changes and save them, the file is created.
If you visit a nonexistent file unintentionally (because you typed the
wrong file name), use the C-x C-v (find-alternate-file
)
command to visit the file you wanted. C-x C-v is similar to C-x
C-f, but it kills the current buffer (after first offering to save it if
it is modified). C-x C-v is allowed even if the current buffer
is not visiting a file.
If the file you specify is actually a directory, Dired is called on
that directory (see section Dired, the Directory Editor). To inhibit this, set the variable
find-file-run-dired
to nil
; then it is an error to try to
visit a directory.
C-x 4 f (find-file-other-window
) is like C-x C-f
except that the buffer containing the specified file is selected in another
window. The window that was selected before C-x 4 f continues to
show the same buffer it was already showing. If you use this command when
only one window is being displayed, that window is split in two, with one
window showing the same buffer as before, and the other one showing the
newly requested file. See section Multiple Windows.
C-x 5 C-f (find-file-other-screen
) is like C-x C-f
except that it creates a new screen in which the file is displayed.
Use the function find-this-file-other-window
to edit a file
mentioned in the buffer you are editing or to test if that file exists.
To do do this, use the minibuffer completion after snatching the part or
all of the filename. By providing a prefix argument, the function calls
filename-at-point
and switches you to a buffer visiting the file
filename in another window. The function creates a buffer if none
already exists. This function is similar to find-file-other-window
.
There are two hook variables that allow extensions to modify the
operation of visiting files. Visiting a file that does not exist runs the
functions in the list find-file-not-found-hooks
; the value of this
variable is expected to be a list of functions which are
called one by one until one of them returns non-nil
. Any visiting
of a file, whether extant or not, expects find-file-hooks
to
contain list of functions and calls them all, one by one. In both cases
the functions receive no arguments. Visiting a nonexistent file
runs the find-file-not-found-hooks
first.
Saving a buffer in Emacs means writing its contents back into the file that was visited in the buffer.
save-buffer
).
save-some-buffers
).
not-modified
).
write-file
).
To save a file and make your changes permanent, type
C-x C-s (save-buffer
). After saving is finished, C-x C-s
prints a message such as:
Wrote /u/rms/gnu/gnu.tasks
If the selected buffer is not modified (no changes have been made in it since the buffer was created or last saved), Emacs does not save it because it would have no effect. Instead, C-x C-s prints a message in the echo area saying:
(No changes need to be saved)
The command C-x s (save-some-buffers
) can save any or all
modified buffers. First it asks, for each modified buffer, whether to
save it. The questions should be answered with y or n.
C-x C-c, the key that kills Emacs, invokes
save-some-buffers
and therefore asks the same questions.
If you have changed a buffer and do not want the changes to be saved,
you should take some action to prevent it. Otherwise, you are liable to
save it by mistake each time you use save-some-buffers
or a
related command. One thing you can do is type M-~
(not-modified
), which removes the indication that the buffer
is modified. If you do this, none of the save commands will believe
that the buffer needs to be saved. (`~' is often used as a
mathematical symbol for `not'; thus Meta-~ is `not', metafied.)
You could also use set-visited-file-name
(see below) to mark the
buffer as visiting a different file name, not in use for
anything important.
You can also undo all the changes made since the file was visited or saved, by reading the text from the file again. This is called reverting. See section Reverting a Buffer. Alternatively, you can undo all the changes by repeating the undo command C-x u; but this only works if you have not made more changes than the undo mechanism can remember.
M-x set-visited-file-name alters the name of the file that the
current buffer is visiting. It prompts you for the new file name in the
minibuffer. You can also use set-visited-file-name
on a buffer
that is not visiting a file. The buffer's name is changed to correspond
to the file it is now visiting unless the new name is already used by a
different buffer; in that case, the buffer name is not changed.
set-visited-file-name
does not save the buffer in the newly
visited file; it just alters the records inside Emacs so that it will
save the buffer in that file. It also marks the buffer as "modified"
so that C-x C-s will save.
If you wish to mark a buffer as visiting a different file and save it
right away, use C-x C-w (write-file
). It is precisely
equivalent to set-visited-file-name
followed by C-x C-s.
C-x C-s used on a buffer that is not visiting a file has the
same effect as C-x C-w; that is, it reads a file name, marks the
buffer as visiting that file, and saves it there. The default file name in
a buffer that is not visiting a file is made by combining the buffer name
with the buffer's default directory.
If Emacs is about to save a file and sees that the date of the latest version on disk does not match what Emacs last read or wrote, Emacs notifies you of this fact, because it probably indicates a problem caused by simultaneous editing and requires your immediate attention. See section Protection Against Simultaneous Editing.
If the variable require-final-newline
is non-nil
, Emacs
puts a newline at the end of any file that doesn't already end in one,
every time a file is saved or written.
Use the hook variable write-file-hooks
to implement other ways
to write files, and specify things to be done before files are written. The
value of this variable should be a list of Lisp functions. When a file
is to be written, the functions in the list are called, one by one, with
no arguments. If one of them returns a non-nil
value, Emacs
takes this to mean that the file has been written in some suitable
fashion; the rest of the functions are not called, and normal writing is
not done. Use the hook variable after-save-hook
to list
all the functions to be called after writing out a buffer to a file.
Because Unix does not provide version numbers in file names, rewriting a
file in Unix automatically destroys all record of what the file used to
contain. Thus, saving a file from Emacs throws away the old contents of
the file--or it would, except that Emacs carefully copies the old contents
to another file, called the backup file, before actually saving.
(Make sure that the variable make-backup-files
is non-nil
.
Backup files are not written if this variable is nil
).
At your option, Emacs can keep either a single backup file or a series of numbered backup files for each file you edit.
Emacs makes a backup for a file only the first time a file is saved from one buffer. No matter how many times you save a file, its backup file continues to contain the contents from before the file was visited. Normally this means that the backup file contains the contents from before the current editing session; however, if you kill the buffer and then visit the file again, a new backup file is made by the next save.
If you choose to have a single backup file (the default), the backup file's name is constructed by appending `~' to the file name being edited; thus, the backup file for `eval.c' is `eval.c~'.
If you choose to have a series of numbered backup files, backup file names are made by appending `.~', the number, and another `~' to the original file name. Thus, the backup files of `eval.c' would be called `eval.c.~1~', `eval.c.~2~', and so on, through names like `eval.c.~259~' and beyond.
If protection stops you from writing backup files under the usual names, the backup file is written as `%backup%~' in your home directory. Only one such file can exist, so only the most recently made backup is available.
The choice of single backup or numbered backups is controlled by the
variable version-control
. Its possible values are:
t
nil
never
version-control
may be set locally in an individual buffer to
control the making of backups for that buffer's file. For example,
Rmail mode locally sets version-control
to never
to make sure
that there is only one backup for an Rmail file. See section Local Variables.
To prevent unlimited consumption of disk space, Emacs can delete numbered
backup versions automatically. Generally Emacs keeps the first few backups
and the latest few backups, deleting any in between. This happens every
time a new backup is made. The two variables that control the deletion are
kept-old-versions
and kept-new-versions
. Their values are, respectively
the number of oldest (lowest-numbered) backups to keep and the number of
newest (highest-numbered) ones to keep, each time a new backup is made.
The values are used just after a new backup version is made;
that newly made backup is included in the count in kept-new-versions
.
By default, both variables are 2.
If trim-versions-without-asking
is non-nil
, excess
middle versions are deleted without notification. If it is nil
, the
default, you are asked whether the excess middle versions should
really be deleted.
You can also use Dired's . (Period) command to delete old versions. See section Dired, the Directory Editor.
You can make backup files by copying the old file or by renaming it. This makes a difference when the old file has multiple names. If you rename the old file into the backup file, the alternate names become names for the backup file. If you copy the old file instead, the alternate names remain names for the file that you are editing, and the contents accessed by those names will be the new contents.
How you make a backup file may also affect the file's owner and group. If you use copying, they do not change. If renaming is used, you become the file's owner, and the file's group becomes the default (different operating systems have different defaults for the group).
Having the owner change is usually a good idea, because then the owner
is always the person who last edited the file. Occasionally there is a
file whose owner should not change. Since most files should change
owners, it is a good idea to use local variable lists to set
backup-by-copying-when-mismatch
for the special cases where the
owner should not change (see section Local Variables in Files).
Three variables control the choice of renaming or copying.
Normally, renaming is done. If the variable backup-by-copying
is
non-nil
, copying is used. Otherwise, if the variable
backup-by-copying-when-linked
is non-nil
, copying is
done for files that have multiple names, but renaming may still be done when
the file being edited has only one name. If the variable
backup-by-copying-when-mismatch
is non-nil
, copying is
done if renaming would cause the file's owner or group to change.
Simultaneous editing occurs when two users visit the same file, both make changes, and both save their changes. If no one was informed that this was happening, and you saved first, you would later find that your changes were lost. On some systems, Emacs notices immediately when the second user starts to change a file already being edited, and issues a warning. When this is not possible, or if the second user has started to change the file despite the warning, Emacs checks when the file is saved, and issues a second warning when a user is about to overwrite a file containing another user's changes. If you are the user editing the file, you can take corrective action at this point and prevent actual loss of work.
When you make the first modification in an Emacs buffer that is visiting
a file, Emacs records that you have locked the file. (It does this by
writing another file in a directory reserved for this purpose.) The lock
is removed when you save the changes. The idea is that the file is locked
whenever the buffer is modified. If you begin to modify the buffer while
the visited file is locked by someone else, this constitutes a collision,
and Emacs asks you what to do. It does this by calling the Lisp function
ask-user-about-lock
, which you can redefine to customize what it
does. The standard definition of this function asks you a
question and accepts three possible answers:
file-locked
) and the modification you
were trying to make in the buffer does not actually take place.
Note that locking works on the basis of a file name; if a file has multiple names, Emacs does not realize that the two names are the same file and cannot prevent two users from editing it simultaneously under different names. However, basing locking on names means that Emacs can interlock the editing of new files that do not really exist until they are saved.
Some systems are not configured to allow Emacs to make locks. On these systems, Emacs cannot detect trouble in advance, but it can still detect it in time to prevent you from overwriting someone else's changes.
Every time Emacs saves a buffer, it first checks the last-modification date of the existing file on disk to see that it has not changed since the file was last visited or saved. If the date does not match, it implies that changes were made in the file in some other way, and these changes are about to be lost if Emacs actually does save. To prevent this, Emacs prints a warning message and asks for confirmation before saving. Occasionally you will know why the file was changed and know that it does not matter; then you can answer yes and proceed. Otherwise, you should cancel the save with C-g and investigate the situation.
The first thing you should do when notified that simultaneous editing
has already taken place is to list the directory with C-u C-x C-d
(see section Listing a File Directory). This will show the file's current
author. You should attempt to contact that person and ask him not to
continue editing. Often the next step is to save the contents of your
Emacs buffer under a different name, and use diff
to compare the
two files.
Simultaneous editing checks are also made when you visit a file that is already visited with C-x C-f and when you start to modify a file. This is not strictly necessary, but it is useful to find out about such a problem as early as possible, when corrective action takes less work.
Another way to protect your file is to set the read, write, and
executable permissions for the file. Use the function
set-default-file-modes
to set the UNIX umask
value to the
nmask argument. The umask
value is the default protection
mode for new files.
If you have made extensive changes to a file and then change your mind about them, you can get rid of all changes by reading in the previous version of the file. To do this, use M-x revert-buffer, which operates on the current buffer. Since reverting a buffer can result in very extensive changes, you must confirm it with yes.
If the current buffer has been auto-saved more recently than it has been
saved explicitly, revert-buffer
offers to read the auto save file
instead of the visited file (see section Auto-Saving: Protection Against Disasters). Emacs asks you about
the auto-save file before the request for confirmation of the
revert-buffer operation, and demands y or n
as an answer. If you have started to type yes for confirmation
without realizing that the auto-save question was going to be asked, the
y will answer that question, but the es will not be valid
confirmation. This gives you a chance to cancel the operation with
C-g and try again with the answers you really intend.
revert-buffer
keeps point at the same distance (measured in
characters) from the beginning of the file. If the file was edited only
slightly, you will be at approximately the same piece of text after
reverting as before. If you have made more extensive changes, the value of
point in the old file may bring you to a totally different piece of text
than your last editing point.
A buffer reverted from its visited file is marked "not modified" until you make a change.
Some kinds of buffers whose contents reflect data bases other than files,
such as Dired buffers, can also be reverted. For them, reverting means
recalculating their contents from the appropriate data. Buffers
created randomly with C-x b cannot be reverted; revert-buffer
reports an error when asked to do so.
Emacs saves all the visited files from time to time (based on counting your keystrokes) without being asked. This is called auto-saving. It prevents you from losing more than a limited amount of work if the system crashes.
When Emacs determines it is time for auto-saving, each buffer is considered and is auto-saved if auto-saving is turned on for it and it has changed since the last time it was auto-saved. If any auto-saving is done, the message `Auto-saving...' is displayed in the echo area until auto-saving is finished. Errors occurring during auto-saving are caught so that they do not interfere with the execution of commands you have been typing.
Auto-saving does not normally write to the files you visited, because it can be undesirable to save a program that is in an inconsistent state when you have made only half of a planned change. Instead, auto-saving is done in a different file called the auto-save file, and the visited file is changed only when you save explicitly, for example, with C-x C-s.
Normally, the name of the auto-save file is generated by appending
`#' to the front and back of the visited file name. Thus, a buffer
visiting file `foo.c' would be auto-saved in a file `#foo.c#'.
Most buffers that are not visiting files are auto-saved only if you
request it explicitly; when they are auto-saved, the auto-save file name
is generated by appending `#%' to the front and `#' to the
back of buffer name. For example, the `*mail*' buffer in which you
compose messages to be sent is auto-saved in a file named
`#%*mail*#'. Names of auto-save files are generated this way
unless you customize the functions make-auto-save-file-name
and
auto-save-file-name-p
to do something different. The file name
to be used for auto-saving a buffer is calculated at the time auto-saving is
turned on in that buffer.
If you want auto-saving to be done in the visited file, set the variable
auto-save-visited-file-name
to be non-nil
. In this mode,
there is really no difference between auto-saving and explicit saving.
Emacs deletes a buffer's auto-save file when you explicitly save the
buffer. To inhibit the deletion, set the variable
delete-auto-save-files
to nil
. Changing the visited file
name with C-x C-w or set-visited-file-name
renames any
auto-save file to correspond to the new visited name.
Each time you visit a file, auto-saving is turned on for that file's
buffer if the variable auto-save-default
is non-nil
(but
not in batch mode; see section Entering and Exiting Emacs). The default for this
variable is t
, so Emacs auto-saves buffers that visit files by
default. You can use the command M-x auto-save-mode to turn
auto-saving for a buffer on or off. Like other minor mode commands,
M-x auto-save-mode turns auto-saving on with a positive argument,
off with a zero or negative argument; with no argument, it toggles.
Emacs performs auto-saving periodically based on counting how many
characters you have typed since the last time auto-saving happened. The
variable auto-save-interval
specifies the number of characters
between auto-saves. By default, it is 300. Emacs also auto-saves
whenever you call the function do-auto-save
.
Emacs also does auto-saving whenever it gets a fatal error. This
includes killing the Emacs job with a shell command such as kill
-emacs
, or disconnecting a phone line or network connection.
You can set the number of seconds of idle time before an auto-save is
done. Setting the value of the variable auto-save-timeout
to zero or
nil
will disable auto-saving due to idleness.
The actual amount of idle time between auto-saves is logarithmically related to the size of the current buffer. This variable is the number of seconds after which an auto-save will happen when the current buffer is 50k or less; the timeout will be 2 1/4 times this in a 200k buffer, 3 3/4 times this in a 1000k buffer, and 4 1/2 times this in a 2000k buffer.
For this variable to have any effect, you must do (require 'timer)
.
If you want to use the contents of an auto-save file to recover from a loss of data, use the command M-x recover-file RET file RET. Emacs visits file and then (after your confirmation) restores the contents from the auto-save file `#file#'. You can then save the file with C-x C-s to put the recovered text into file itself. For example, to recover file `foo.c' from its auto-save file `#foo.c#', do:
M-x recover-file RET foo.c RET C-x C-s
Before asking for confirmation, M-x recover-file displays a directory listing describing the specified file and the auto-save file, so you can compare their sizes and dates. If the auto-save file is older, M-x recover-file does not offer to read it.
Auto-saving is disabled by M-x recover-file because using this command implies that the auto-save file contains valuable data from a past session. If you save the data in the visited file and then go on to make new changes, turn auto-saving back on with M-x auto-save-mode.
Version control systems are packages that can record multiple versions of a source file, usually storing the unchanged parts of the file just once. Version control systems also record history information such as the creation time of each version, who created it, and a description of what was changed in that version.
The GNU project recommends the version control system known as RCS, which is free software and available from the Free Software Foundation. Emacs supports use of either RCS or SCCS (a proprietary, but widely used, version control system that is not quite as powerful as RCS) through a facility called VC. The same Emacs commands work with either RCS or SCCS, so you hardly have to know which one of them you are using.
When a file is under version control, we also say that it is registered in the version control system. Each registered file has a corresponding master file which represents the file's present state plus its change history, so that you can reconstruct from it either the current version or any specified earlier version. Usually the master file also records a log entry for each version describing what was changed in that version.
The file that is maintained under version control is sometimes called the work file corresponding to its master file.
To examine a file, you check it out. This extracts a version of the source file (typically, the most recent) from the master file. If you want to edit the file, you must check it out locked. Only one user can do this at a time for any given source file. (This kind of locking is completely unrelated to the locking that Emacs uses to detect simultaneous editing of a file.)
When you are done with your editing, you must check in the new version. This records the new version in the master file, and unlocks the source file so that other people can lock it and thus modify it.
Checkin and checkout are the basic operations of version control. You
can do both of them with a single Emacs command: C-x C-q
(vc-toggle-read-only
).
A snapshot is a coherent collection of versions of the various files that make up a program. See section Snapshots.
When you visit a file that is maintained using version control, the mode line displays `RCS' or `SCCS' to inform you that version control is in use, and also (in case you care) which low-level system the file is actually stored in. Normally, such a source file is read-only, and the mode line indicates this with `%%'. With RCS, the mode line also indicates the number of the head version, which is normally also the version you are looking at.
These are the commands for editing a file maintained with version control:
(C-x v is the prefix key for version control commands; all of these commands except for C-x C-q start with C-x v.)
When you want to modify a file maintained with version control, type
C-x C-q (vc-toggle-read-only
). This checks out the
file, and tells RCS or SCCS to lock the file. This means making the
file writable for you (but not for anyone else).
When you are finished editing the file, type C-x C-q again. When used on a file that is checked out, this command checks the file in. But check-in does not start immediately; first, you must enter the log entry---a description of the changes in the new version. C-x C-q pops up a buffer for you to enter this in. When you are finished typing in the log entry, type C-c C-c to terminate it; this is when actual check-in takes place.
Once you have checked in your changes, the file is unlocked, so that other users can lock it and modify it.
Emacs does not save backup files for source files that are maintained
with version control. If you want to make backup files despite version
control, set the variable vc-make-backup-files
to a
non-nil
value.
Normally the work file exists all the time, whether it is locked or
not. If you set vc-keep-workfiles
to nil
, then checking
in a new version with C-x C-q deletes the work file; but any
attempt to visit the file with Emacs creates it again.
It is not impossible to lock a file that someone else has locked. If you try to check out a file that is locked, C-x C-q asks you whether you want to "steal the lock." If you say yes, the file becomes locked by you, but a message is sent to the person who had formerly locked the file, to inform him of what has happened. The mode line indicates that a file is locked by someone else by displaying the login name of that person, before the version number.
If you want to discard your current set of changes and revert to the
last version checked in, use C-x v u (vc-revert-buffer
).
This cancels your last check-out, leaving the file unlocked. If you want
to make a different set of changes, you must first check the file out
again. C-x v u requires confirmation, unless it sees that
you haven't made any changes since the last checked-in version.
C-x v u is also the command to use if you lock a file and then don't actually change it.
You can cancel a change after checking it in, with C-x v c
(vc-cancel-version
). This command discards all record of the
most recent checked in version, so be careful about using it. It
requires confirmation with yes. By default, C-x v c reverts
your workfile and buffer to the previous version (the one that precedes
the version that is deleted), but you can prevent the reversion by
giving the command a prefix argument. Then the buffer does not change.
This command with a prefix argument is useful when you have checked in a change and then discover a trivial error in it; you can cancel the erroneous check-in, fix the error, and repeat the check-in.
Be careful when invoking C-x v c, as it is easy to throw away a lot of work with it. To help you be careful, this command always requires confirmation with `yes'.
You can register the visited file for version control using
C-x v i (vc-register
). If the variable
vc-default-back-end
is non-nil
, it specifies which
version control system to use; otherwise, this uses RCS if it is
installed on your system and SCCS if not. After C-x v i,
the file is unlocked and read-only. Type C-x C-q if you wish to
edit it.
By default, the initial version number is 1.1. If you want to use a different number, give C-x v i a prefix argument; then it reads the initial version number using the minibuffer.
If vc-initial-comment
is non-nil
, C-x v i reads
an initial comment (much like a log entry) to describe the purpose of
this source file.
To specify the version number for a subsequent checkin, use the
command C-u C-x v v. C-x v v (vc-next-action
) is the
command that C-x C-q uses to do the "real work" when the visited
file uses version control. When used for checkin, and given a prefix
argument, it reads the version number with the minibuffer.
If vc-suppress-confirm
is non-nil
, then C-x C-q
and C-x v i can save the current buffer without asking, and
C-x v u also operates without asking for confirmation.
(This variable does not affect C-x v c; that is so drastic
that it should always ask for confirmation.)
VC mode does much of its work by running the shell commands for RCS
and SCCS. If vc-command-messages
is non-nil
, VC displays
messages to indicate which shell commands it runs, and additional
messages when the commands finish.
Normally, VC assumes that it can deduce the locked/unlocked state of files by looking at the file permissions of the work file; this is fast. However, if the `RCS' or `SCCS' subdirectory is actually a symbolic link, then VC does not trust the file permissions to reflect this status.
You can specify the criterion for whether to trust the file permissions
by setting the variable vc-mistrust-permissions
. Its value may
be t
(always mistrust the file permissions and check the master
file), nil
(always trust the file permissions), or a function of
one argument which makes the decision. The argument is the directory
name of the `RCS' or `SCCS' subdirectory. A non-nil
value from the function says to mistrust the file permissions.
If you find that the file permissions of work files are changed
erroneously, set vc-mistrust-permissions
to t
. Then VC
always checks the master file to determine the file's status.
You can specify additional directories to search for version control
programs by setting the variable vc-path
. These directories
are searched before the usual search path. The proper result usually
happens automatically.
When you're editing an initial comment or log entry for inclusion in a master file, finish your entry by typing C-c C-c.
vc-finish-logentry
).
This finishes check-in.
To abort check-in, just don't type C-c C-c in that buffer. You can switch buffers and do other editing. As long as you don't try to check in another file, the entry you were editing remains in its buffer, and you can go back to that buffer at any time to complete the check-in.
If you change several source files for the same reason, it is often convenient to specify the same log entry for many of the files. To do this, use the history of previous log entries. The commands M-n, M-p, M-s and M-r for doing this work just like the minibuffer history commands (except that these versions are used outside the minibuffer).
Each time you check in a file, the log entry buffer is put into VC Log
mode, which involves running two hooks: text-mode-hook
and
vc-log-mode-hook
.
If you use RCS for a program and also maintain a change log file for it (see section Change Logs), you can generate change log entries automatically from the version control log entries:
vc-update-change-log
).
This command works with RCS only; it does not work with SCCS.
For example, suppose the first line of `ChangeLog' is dated 10 April 1992, and that the only check-in since then was by Nathaniel Bowditch to `rcs2log' on 8 May 1992 with log text `Ignore log messages that start with `#'.'. Then C-x v a visits `ChangeLog' and inserts text like this:
Fri May 8 21:45:00 1992 Nathaniel Bowditch (nat@apn.org) * rcs2log: Ignore log messages that start with `#'.
You can then edit the new change log entry further as you wish.
Normally, the log entry for file `foo' is displayed as `* foo: text of log entry'. The `:' after `foo' is omitted if the text of the log entry starts with `(functionname): '. For example, if the log entry for `vc.el' is `(vc-do-command): Check call-process status.', then the text in `ChangeLog' looks like this:
Wed May 6 10:53:00 1992 Nathaniel Bowditch (nat@apn.org) * vc.el (vc-do-command): Check call-process status.
When C-x v a adds several change log entries at once, it groups related log entries together if they all are checked in by the same author at nearly the same time. If the log entries for several such files all have the same text, it coalesces them into a single entry. For example, suppose the most recent checkins have the following log entries:
For `vc.texinfo': Fix expansion typos. For `vc.el': Don't call expand-file-name. For `vc-hooks.el': Don't call expand-file-name.
They appear like this in `ChangeLog':
Wed Apr 1 08:57:59 1992 Nathaniel Bowditch (nat@apn.org) * vc.texinfo: Fix expansion typos. * vc.el, vc-hooks.el: Don't call expand-file-name.
Normally, C-x v a separates log entries by a blank line, but you can mark several related log entries to be clumped together (without an intervening blank line) by starting the text of each related log entry with a label of the form `{clumpname} '. The label itself is not copied to `ChangeLog'. For example, suppose the log entries are:
For `vc.texinfo': {expand} Fix expansion typos. For `vc.el': {expand} Don't call expand-file-name. For `vc-hooks.el': {expand} Don't call expand-file-name.
Then the text in `ChangeLog' looks like this:
Wed Apr 1 08:57:59 1992 Nathaniel Bowditch (nat@apn.org) * vc.texinfo: Fix expansion typos. * vc.el, vc-hooks.el: Don't call expand-file-name.
A log entry whose text begins with `#' is not copied to `ChangeLog'. For example, if you merely fix some misspellings in comments, you can log the change with an entry beginning with `#' to avoid putting such trivia into `ChangeLog'.
vc-version-other-window
).
You can examine any version of a file by first visiting it, and then
using C-x v ~ version RET
(vc-version-other-window
). This puts the text of version
version in a file named `filename.~version~',
then visits it in a separate window.
To compare two versions of a file, use the command C-x v =
(vc-diff
).
Plain C-x v = compares the current buffer contents (saving them in the file if necessary) with the last checked-in version of the file. With a prefix argument, C-x v = reads a file name and two version numbers, then compares those versions of the specified file.
If you supply a directory name instead of the name of a work file, this command compares the two specified versions of all registered files in that directory and its subdirectories. You can also specify a snapshot name (see section Snapshots) instead of one or both version numbers.
You can specify a checked-in version by its number; you can specify the most recent checked-in version with an empty version number.
This command works by running the diff
utility, getting the
options from the variable diff-switches
. It displays the output
in a special buffer in another window. Unlike the M-x diff
command, C-x v = does not try to find the changes in the old and
new versions. This is because one or both versions normally do not
exist as files. They exist only in the records of the master file.
See section Comparing Files, for more information about M-x diff.
To view the detailed version control status and history of a file,
type C-x v l (vc-print-log
). It displays the history of
changes to the current file, including the text of the log entries. The
output appears in a separate window.
When you are working on a large program, it's often useful to find all
the files that are currently locked, or all the files maintained in
version control at all. You can use C-x v d (vc-directory
)
to show all the locked files in or beneath the current directory. This
includes all files that are locked by any user. C-u C-x v d lists
all files in or beneath the current directory that are maintained with
version control.
The list of files is displayed as a buffer that uses an augmented Dired mode. The names of the users locking various files are shown (in parentheses) in place of the owner and group. All the normal Dired commands work in this buffer. Most interactive VC commands work also, and apply to the file name on the current line.
The C-x v v command (vc-next-action
), when used in the
augmented Dired buffer, operates on all the marked files (or the file on
the current line). If it operates on more than one file, it handles
each file according to its current state; thus, it may check out one
file and check in another (because it is already checked out). If it
has to check in any files, it reads a single log entry, then uses that
text for all the files being checked in. This can be convenient for
registering or checking in several files at once, as part of the same
change.
When you rename a registered file, you must also rename its master
file correspondingly to get proper results. Use vc-rename-file
to rename the source file as you specify, and rename its master file
accordingly. It also updates any snapshots (see section Snapshots) that
mention the file, so that they use the new name; despite this, the
snapshot thus modified may not completely work (see section Snapshot Caveats).
You cannot use vc-rename-file
on a file that is locked by
someone else.
A snapshot is a named set of file versions (one for each registered file) that you can treat as a unit. One important kind of snapshot is a release, a (theoretically) stable version of the system that is ready for distribution to users.
There are two basic commands for snapshots; one makes a snapshot with a given name, the other retrieves a named snapshot.
C-x v s name RET
vc-create-snapshot
).
C-x v r name RET
vc-retrieve-snapshot
).
This command reports an error if any files are locked at or below the
current directory, without changing anything; this is to avoid
overwriting work in progress.
A snapshot uses a very small amount of resources--just enough to record the list of file names and which version belongs to the snapshot. Thus, you need not hesitate to create snapshots whenever they are useful.
You can give a snapshot name as an argument to C-x v = or C-x v ~ (see section Examining And Comparing Old Versions). Thus, you can use it to compare a snapshot against the current files, or two snapshots against each other, or a snapshot against a named version.
VC's snapshot facilities are modeled on RCS's named-configuration support. They use RCS's native facilities for this, so under VC snapshots made using RCS are visible even when you bypass VC.
For SCCS, VC implements snapshots itself. The files it uses contain name/file/version-number triples. These snapshots are visible only through VC.
A snapshot is a set of checked-in versions. So make sure that all the files are checked in and not locked when you make a snapshot.
File renaming and deletion can create some difficulties with snapshots. This is not a VC-specific problem, but a general design issue in version control systems that no one has solved very well yet.
If you rename a registered file, you need to rename its master along
with it (the command vc-rename-file
does this automatically). If
you are using SCCS, you must also update the records of the snapshot, to
mention the file by its new name (vc-rename-file
does this,
too). An old snapshot that refers to a master file that no longer
exists under the recorded name is invalid; VC can no longer retrieve
it. It would be beyond the scope of this manual to explain enough about
RCS and SCCS to explain how to update the snapshots by hand.
Using vc-rename-file
makes the snapshot remain valid for
retrieval, but it does not solve all problems. For example, some of the
files in the program probably refer to others by name. At the very
least, the makefile probably mentions the file that you renamed. If you
retrieve an old snapshot, the renamed file is retrieved under its new
name, which is not the name that the makefile expects. So the program
won't really work as retrieved.
Sometimes it is convenient to put version identification strings directly into working files. Certain special strings called version headers are replaced in each successive version by the number of that version.
You can use the C-x v h command (vc-insert-headers
) to
insert a suitable header string.
The default header string is `\$Id\$' for RCS and `\%W\%'
for SCCS. (The actual strings inserted do not have the backslashes
in them. They were placed in the Info source file so that the
strings don't get interpreted as version-control headers when the
Info source files are maintained under version control.) You can
specify other headers to insert by setting the variable
vc-header-alist
. Its value is a list of elements of the form
(program . string)
where program is RCS
or SCCS
and string is the string to use.
Instead of a single string, you can specify a list of strings; then each string in the list is inserted as a separate header on a line of its own.
It is often necessary to use "superfluous" backslashes when writing the strings that you put in this variable. This is to prevent the string in the constant from being interpreted as a header itself if the Emacs Lisp file containing it is maintained with version control.
Each header is inserted surrounded by tabs, inside comment delimiters,
on a new line at the start of the buffer. Normally the ordinary comment
start and comment end strings of the current mode are used, but for
certain modes, there are special comment delimiters for this purpose;
the variable vc-comment-alist
specifies them. Each element of
this list has the form (mode starter ender)
.
The variable vc-static-header-alist
specifies further strings
to add based on the name of the buffer. Its value should be a list of
elements of the form (regexp . format)
. Whenever
regexp matches the buffer name, format is inserted as part
of the header. A header line is inserted for each element that matches
the buffer name, and for each string specified by
vc-header-alist
. The header line is made by processing the
string from vc-header-alist
with the format taken from the
element. The default value for vc-static-header-alist
is:
(("\\.c$" . "\n#ifndef lint\nstatic char vcid[] = \"\%s\";\n\ #endif /* lint */\n"))
which specifies insertion of a string of this form:
#ifndef lint static char vcid[] = "string"; #endif /* lint */
Files are organized by Unix into directories. A directory listing is a list of all the files in a directory. Emacs provides directory listings in brief format (file names only) and verbose format (sizes, dates, and authors included).
list-directory
).
To print a directory listing, use C-x C-d
(list-directory
). This command prompts in the minibuffer for a
file name which is either a directory to be listed or pattern
containing wildcards for the files to be listed. For example,
C-x C-d /u2/emacs/etc RET
lists all the files in directory `/u2/emacs/etc'. An example of specifying a file name pattern is:
C-x C-d /u2/emacs/src/*.c RET
Normally, C-x C-d prints a brief directory listing containing just
file names. A numeric argument (regardless of value) tells it to print a
verbose listing (like ls -l
).
Emacs obtains the text of a directory listing by running ls
in
an inferior process. Two Emacs variables control the switches passed to
ls
: list-directory-brief-switches
is a string giving the
switches to use in brief listings ("-CF"
by default).
list-directory-verbose-switches
is a string giving the switches
to use in a verbose listing ("-l"
by default).
The variable directory-abbrev-alist
is an alist of abbreviations
for file directories. The list consists of elements of the form
(FROM . TO)
, each meaning to replace FROM
with TO
when it appears in a directory name. This replacement is done when
setting up the default directory of a newly visited file. Every FROM
string should start with ``^''.
Use this feature when you have directories which you normally refer to
via absolute symbolic links. Make TO
the name of the link, and
FROM
the name it is linked to.
The command M-x diff compares two files, displaying the
differences in an Emacs buffer named `*Diff*'. It works by running
the diff
program, using options taken from the variable
diff-switches
, whose value should be a string.
The buffer `*Diff*' has Compilation mode as its major mode, so you can use C-x ` to visit successive changed locations in the two source files. You can also move to a particular hunk of changes and type C-c C-c to find the corresponding source location. You can also use the other special commands of Compilation mode: SPC and DEL for scrolling, and M-p and M-n for cursor motion. See section Running `make', or Compilers Generally.
The command M-x diff-backup compares a specified file with its most
recent backup. If you specify the name of a backup file,
diff-backup
compares it with the source file that it is a backup
of.
The command M-x compare-windows compares the text in the current window with that in the next window. Comparison starts at point in each window. Point moves forward in each window, a character at a time in each window, until the next characters in the two windows are different. Then the command is finished. For more information about windows in Emacs, section Multiple Windows.
With a numeric argument, compare-windows
ignores changes in
whitespace. If the variable compare-ignore-case
is
non-nil
, it ignores differences in case as well.
Dired makes it easy to delete or visit many of the files in a single directory at once. It creates an Emacs buffer containing a listing of the directory. You can use the normal Emacs commands to move around in this buffer and special Dired commands to operate on the files.
To invoke dired, type C-x d or M-x dired. The command reads a
directory name or wildcard file name pattern as a minibuffer argument just
like the list-directory
command, C-x C-d. Where dired
differs from list-directory
is in naming the buffer after the
directory name or the wildcard pattern used for the listing, and putting
the buffer into Dired mode so that the special commands of Dired are
available in it. The variable dired-listing-switches
is a string
used as an argument to ls
in making the directory; this string
must contain `-l'.
To display the Dired buffer in another window rather than in the selected
window, use C-x 4 d (dired-other-window)
instead of C-x d.
Once the Dired buffer exists, you can switch freely between it and other Emacs buffers. Whenever the Dired buffer is selected, certain special commands are provided that operate on files that are listed. The Dired buffer is "read-only", and inserting text in it is not useful, so ordinary printing characters such as d and x are used for Dired commands. Most Dired commands operate on the file described by the line that point is on. Some commands perform operations immediately; others "flag" a file to be operated on later.
Most Dired commands that operate on the current line's file also treat a numeric argument as a repeat count, meaning to act on the files of the next few lines. A negative argument means to operate on the files of the preceding lines, and leave point on the first of those lines.
All the usual Emacs cursor motion commands are available in Dired buffers. Some special purpose commands are also provided. The keys C-n and C-p are redefined so that they try to position the cursor at the beginning of the filename on the line, rather than at the beginning of the line.
For extra convenience, SPC and n in Dired are equivalent to C-n. p is equivalent to C-p. Moving by lines is done so often in Dired that it deserves to be easy to type. DEL (move up and unflag) is often useful simply for moving up.
The g command in Dired runs revert-buffer
to reinitialize
the buffer from the actual disk directory and show any changes made in the
directory by programs other than Dired. All deletion flags in the Dired
buffer are lost when this is done.
The primary use of Dired is to flag files for deletion and then delete them.
You can flag a file for deletion by moving to the line describing the file and typing d or C-d. The deletion flag is visible as a `D' at the beginning of the line. Point is moved to the beginning of the next line, so that repeated d commands flag successive files.
The files are flagged for deletion rather than deleted immediately to avoid the danger of deleting a file accidentally. Until you direct Dired to delete the flagged files, you can remove deletion flags using the commands u and DEL. u works just like d, but removes flags rather than making flags. DEL moves upward, removing flags; it is like u with numeric argument automatically negated.
To delete the flagged files, type x. This command first displays a list of all the file names flagged for deletion, and requests confirmation with yes. Once you confirm, all the flagged files are deleted, and their lines are deleted from the text of the Dired buffer. The shortened Dired buffer remains selected. If you answer no or quit with C-g, you return immediately to Dired, with the deletion flags still present and no files actually deleted.
The #, ~, and . commands flag many files for deletion, based on their names. These commands are useful precisely because they do not actually delete any files; you can remove the deletion flags from any flagged files that you really wish to keep.
# flags for deletion all files that appear to have been made by auto-saving (that is, files whose names begin and end with `#'). ~ flags for deletion all files that appear to have been made as backups for files that were edited (that is, files whose names end with `~').
. (Period) flags just some of the backup files for deletion: only
numeric backups that are not among the oldest few nor the newest few
backups of any one file. Normally dired-kept-versions
(not
kept-new-versions
; that applies only when saving) specifies the
number of newest versions of each file to keep, and
kept-old-versions
specifies the number of oldest versions to keep.
Period with a positive numeric argument, as in C-u 3 ., specifies the
number of newest versions to keep, overriding dired-kept-versions
.
A negative numeric argument overrides kept-old-versions
, using minus
the value of the argument to specify the number of oldest versions of each
file to keep.
Some file operations in Dired take place immediately when they are requested.
Emacs has commands for performing many other operations on files. All operate on one file; they do not accept wildcard file names.
You can use the command M-x add-name-to-file to add a name to an existing file without removing the old name. The new name must belong on the file system that the file is on.
M-x append-to-file adds the text of the region to the end of the specified file.
M-x copy-file reads the file old and writes a new file named new with the same contents. Confirmation is required if a file named new already exists, because copying overwrites the old contents of the file new.
M-x delete-file deletes a specified file, like the rm
command in the shell. If you are deleting many files in one directory, it
may be more convenient to use Dired (see section Dired, the Directory Editor).
M-x insert-file inserts a copy of the contents of a specified file into the current buffer at point, leaving point unchanged before the contents and the mark after them. See section Selecting Text.
M-x make-symbolic-link reads two file names old and linkname, and then creates a symbolic link named linkname and pointing at old. Future attempts to open file linkname will then refer to the file named old at the time the opening is done, or will result in an error if the name old is not in use at that time. Confirmation is required if you create the link while linkname is in use. Note that not all systems support symbolic links.
M-x rename-file reads two file names old and new using the minibuffer, then renames file old as new. If a file named new already exists, you must confirm with yes or renaming is not done; this is because renaming causes the previous meaning of the name new to be lost. If old and new are on different file systems, the file old is copied and deleted.
M-x view-file allows you to scan or read a file by sequential
screenfuls. It reads a file name argument using the minibuffer. After
reading the file into an Emacs buffer, view-file
reads and displays
one windowful. You can then type SPC to scroll forward one window,
or DEL to scroll backward. Various other commands are provided for
moving around in the file, but none for changing it; type C-h while
viewing a file for a list of them. Most commands are the default Emacs
cursor motion commands. To exit from viewing, type C-c.
Text you are editing in Emacs resides in an object called a buffer. Each time you visit a file, Emacs creates a buffer to hold the file's text. Each time you invoke Dired, Emacs creates a buffer to hold the directory listing. If you send a message with C-x m, a buffer named `*mail*' is used to hold the text of the message. When you ask for a command's documentation, it appears in a buffer called `*Help*'.
At any time, one and only one buffer is selected. It is also called the current buffer. Saying a command operates on "the buffer" really means that the command operates on the selected buffer, as most commands do.
When Emacs creates multiple windows, each window has a chosen buffer which is displayed there, but at any time only one of the windows is selected and its chosen buffer is the selected buffer. Each window's mode line displays the name of the buffer the window is displaying (see section Multiple Windows).
Each buffer has a name which can be of any length but is case-sensitive. You can select a buffer using its name. Most buffers are created when you visit files; their names are derived from the files' names. You can also create an empty buffer with any name you want. A newly started Emacs has a buffer named `*scratch*' which you can use for evaluating Lisp expressions in Emacs.
Each buffer records what file it is visiting, whether it is modified, and what major mode and minor modes are in effect in it (see section Major Modes). Any Emacs variable can be made local to a particular buffer, meaning its value in that buffer can be different from the value in other buffers. See section Local Variables.
switch-to-buffer
).
switch-to-buffer-other-window
).
To select a buffer named bufname, type C-x b bufname
RET. This is the command switch-to-buffer
with argument
bufname. You can use completion on an abbreviation for the buffer
name you want (see section Completion). An empty argument to C-x b
specifies the most recently selected buffer that is not displayed in any
window.
Most buffers are created when you visit files, or use Emacs commands
that display text. You can also create a buffer explicitly by typing
C-x b bufname RET, which creates a new, empty buffer
that is not visiting any file, and selects it for editing. The new
buffer's major mode is determined by the value of
default-major-mode
(see section Major Modes). Buffers not visiting
files are usually used for making notes to yourself. If you try to save
one, you are asked for the file name to use.
The function switch-to-buffer-other-screen
is similar to
switch-to-buffer
except that it creates a new screen in which to
display the selected buffer.
Use M-x switch-to-other-buffer to visit the previous buffer. If you supply a positive integer n, the nth most recent buffer is displayed. If you supply an argument of 0, the current buffer is moved to the bottom of the buffer stack.
Note that you can also use C-x C-f and any other command for visiting a file to switch buffers. See section Visiting Files.
list-buffers
).
To print a list of all existing buffers, type C-x C-b. Each line in the list shows one buffer's name, major mode, and visited file. A `*' at the beginning of a line indicates the buffer has been "modified". If several buffers are modified, it may be time to save some with C-x s (see section Saving Files). A `%' indicates a read-only buffer. A `.' marks the selected buffer. Here is an example of a buffer list:
MR Buffer Size Mode File -- ------ ---- ---- ---- .* emacs.tex 383402 Texinfo /u2/emacs/man/emacs.tex *Help* 1287 Fundamental files.el 23076 Emacs-Lisp /u2/emacs/lisp/files.el % RMAIL 64042 RMAIL /u/rms/RMAIL *% man 747 Dired /u2/emacs/man/ net.emacs 343885 Fundamental /u/rms/net.emacs fileio.c 27691 C /u2/emacs/src/fileio.c NEWS 67340 Text /u2/emacs/etc/NEWS *scratch* 0 Lisp Interaction
Note that the buffer `*Help*' was made by a help request; it is not
visiting any file. The buffer man
was made by Dired on the
directory `/u2/emacs/man/'.
As you move the mouse over the `*Buffer List*' buffer, the lines
are highlighted. This visual cue indicates that clicking the right
mouse button (button3
) will pop up a menu of commands on the
buffer represented by this line. This menu duplicates most of those
commands which are bound to keys in the `*Buffer List*' buffer.
toggle-read-only
).
A buffer can be read-only, which means that commands to change
its text are not allowed. Normally, read-only buffers are created by
subsystems such as Dired and Rmail that have special commands to operate
on the text. Emacs also creates a read-only buffer if you
visit a file that is protected. To make changes in a read-only buffer,
use the command C-x C-q (toggle-read-only
). It makes a
read-only buffer writable, and makes a writable buffer read-only. This
works by setting the variable buffer-read-only
, which has a local
value in each buffer and makes a buffer read-only if its value is
non-nil
.
M-x rename-buffer changes the name of the current buffer, prompting for the new name in the minibuffer. There is no default. If you specify a name that is used by a different buffer, an error is signalled and renaming is not done.
M-x view-buffer is similar to M-x view-file (see section Miscellaneous File Operations), but it examines an already existing Emacs buffer. View mode provides convenient commands for scrolling through the buffer but not for changing it. When you exit View mode, the resulting value of point remains in effect.
To copy text from one buffer to another, use the commands M-x append-to-buffer and M-x insert-buffer. See section Accumulating Text.
After using Emacs for a while, you may accumulate a large number of buffers and may want to eliminate the ones you no longer need. There are several commands for doing this.
kill-buffer
).
C-x k (kill-buffer
) kills one buffer, whose name you
specify in the minibuffer. If you type just RET in the
minibuffer, the default, killing the current buffer, is used. If the
current buffer is killed, the buffer that has been selected recently but
does not appear in any window now is selected. If the buffer being
killed contains unsaved changes, you are asked to confirm with yes
before the buffer is killed.
The command M-x kill-some-buffers asks about each buffer, one by
one. An answer of y means to kill the buffer. Killing the current
buffer or a buffer containing unsaved changes selects a new buffer or asks
for confirmation just like kill-buffer
.
The buffer-menu facility is like a "Dired for buffers"; it allows you to request operations on various Emacs buffers by editing a buffer containing a list of them. You can save buffers, kill them (here called deleting them, for consistency with Dired), or display them.
The command buffer-menu
writes a list of all Emacs buffers into
the buffer `*Buffer List*', and selects that buffer in Buffer Menu
mode. The buffer is read-only. You can only change it using the special
commands described in this section. Most of the commands are graphic
characters. You can use Emacs cursor motion commands in the
`*Buffer List*' buffer. If the cursor is on a line describing a
buffer, the following special commands apply to that buffer:
All commands that add or remove flags to request later operations also move down a line. They accept a numeric argument as a repeat count, unless otherwise specified.
There are also special commands to use the buffer list to select another buffer, and to specify one or more other buffers for display in additional windows.
Going back between a buffer-menu
buffer and other Emacs buffers is
easy. You can, for example, switch from the `*Buffer List*'
buffer to another Emacs buffer, and edit there. You can then reselect the
buffer-menu
buffer and perform operations already
requested, or you can kill that buffer or pay no further attention to it.
All that buffer-menu
does directly is create and select a
suitable buffer, and turn on Buffer Menu mode. All the other
capabilities of the buffer menu are implemented by special commands
provided in Buffer Menu mode.
The only difference between buffer-menu
and list-buffers
is
that buffer-menu
selects the `*Buffer List*' buffer and
list-buffers
does not. If you run list-buffers
(that is,
type C-x C-b) and select the buffer list manually, you can use all
the commands described here.
Emacs can split the screen into two or many windows, which can display parts of different buffers or different parts of one buffer. If you are running Lucid GNU Emacs under X, that means you can have the X window that contains the Emacs screen have multiple subwindows.
When Emacs displays multiple windows, each window has one Emacs buffer designated for display. The same buffer may appear in more than one window; if it does, any changes in its text are displayed in all the windows that display it. Windows showing the same buffer can show different parts of it, because each window has its own value of point.
At any time, one window is the selected window; the buffer displayed by that window is the current buffer. The cursor shows the location of point in that window. Each other window has a location of point as well, but since the terminal has only one cursor, it cannot show the location of point in the other windows.
Commands to move point affect the value of point for the selected Emacs
window only. They do not change the value of point in any other Emacs
window, including those showing the same buffer. The same is true for commands
such as C-x b to change the selected buffer in the selected window;
they do not affect other windows at all. However, there are other commands
such as C-x 4 b that select a different window and switch buffers in
it. Also, all commands that display information in a window, including
(for example) C-h f (describe-function
) and C-x C-b
(list-buffers
), work by switching buffers in a non-selected window
without affecting the selected window.
Each window has its own mode line, which displays the buffer name, modification status, and major and minor modes of the buffer that is displayed in the window. See section The Mode Line, for details on the mode line.
split-window-vertically
).
split-window-horizontally
).
The command C-x 2 (split-window-vertically
) breaks the
selected window into two windows, one above the other. Both windows
start out displaying the same buffer, with the same value of point. By
default each of the two windows gets half the height of the window that
was split. A numeric argument specifies how many lines to give to the
top window.
C-x 3 (split-window-horizontally
) breaks the selected
window into two side-by-side windows. A numeric argument specifies how
many columns to give the one on the left. A line of vertical bars
separates the two windows. Windows that are not the full width of the
screen have truncated mode lines which do not always appear in inverse
video, because Emacs display routines cannot display a region of inverse
video that is only part of a line on the screen.
When a window is less than the full width, many text lines are too
long to fit. Continuing all those lines might be confusing. Set the
variable truncate-partial-width-windows
to non-nil
to
force truncation in all windows less than the full width of the screen,
independent of the buffer and its value for truncate-lines
.
See section Continuation Lines.
Horizontal scrolling is often used in side-by-side windows. See section Controlling the Display.
You can resize a window and store that configuration in a register by
supplying a register argument to window-configuration-to-register
(C-x 6). To return to the window configuration established with
window-configuration-to-register
, use jump-to-register
(C-x j).
other-window
). That is the letter `o', not zero.
scroll-other-window
).
To select a different window, use C-x o (other-window
).
That is an `o', for `other', not a zero. When there are more than
two windows, the command moves through all the windows in a cyclic
order, generally top to bottom and left to right. From the rightmost
and bottommost window, it goes back to the one at the upper left corner.
A numeric argument, n, moves several steps in the cyclic order of
windows. A negative numeric argument moves around the cycle in the
opposite order. If the optional second argument all_screens is
non-nil
, the function cycles through all screens. When the
minibuffer is active, the minibuffer is the last window in the cycle;
you can switch from the minibuffer window to one of the other windows,
and later switch back and finish supplying the minibuffer argument that
is requested. See section Editing in the Minibuffer.
The command M-x other-window-any-screen also selects the window
n steps away in the cyclic order. However, unlike other-window
,
this command selects a window on the next or previous screen instead of
wrapping around to the top or bottom of the current screen, when there
are no more windows.
The usual scrolling commands (see section Controlling the Display) apply to the selected
window only. M-C-v (scroll-other-window
) scrolls the
window that C-x o would select. Like C-v, it takes positive
and negative arguments.
The command M-x compare-windows compares the text in the current window with the text in the next window. Comparison starts at point in each window. Point moves forward in each window, a character at a time, until the next set of characters in the two windows are different. Then the command is finished.
A prefix argument ignore-whitespace means ignore changes in
whitespace. The variable compare-windows-whitespace
controls how
whitespace is skipped.
If compare-ignore-case
is non-nil
, changes in case are
also ignored.
C-x 4 is a prefix key for commands that select another window (splitting the window if there is only one) and select a buffer in that window. Different C-x 4 commands have different ways of finding the buffer to select.
switch-to-buffer-other-window
.
find-file-other-window
. See section Visiting Files.
dired-other-window
. See section Dired, the Directory Editor.
mail-other-window
, and its same-window version is C-x m
(see section Sending Mail).
find-tag-other-window
, the multiple-window variant of M-.
(see section Tag Tables).
If the variable display-buffer-function
is non-nil
, its value is
the function to call to handle display-buffer
. It receives two
arguments, the buffer and a flag that if non-nil
means that the
currently selected window is not acceptable. Commands such as
switch-to-buffer-other-window
and find-file-other-window
work using this function.
delete-window
). That is a zero.
If there is more than one Emacs screen, deleting the sole remaining
window on that screen deletes the screen as well. If the current screen
is the only screen, it is not deleted.
delete-other-windows
).
enlarge-window
).
enlarge-window-horizontally
).
To delete a window, type C-x 0 (delete-window
). (That is a
zero.) The space occupied by the deleted window is distributed among the
other active windows (but not the minibuffer window, even if that is active
at the time). Once a window is deleted, its attributes are forgotten;
there is no automatic way to make another window of the same shape or
showing the same buffer. The buffer continues to exist, and you can
select it in any window with C-x b.
C-x 1 (delete-other-windows
) is more powerful than C-x 0;
it deletes all the windows except the selected one (and the minibuffer).
The selected window expands to use the whole screen except for the echo
area.
To readjust the division of space among existing windows, use C-x
^ (enlarge-window
). It makes the currently selected window
longer by one line or as many lines as a numeric argument specifies.
With a negative argument, it makes the selected window smaller.
C-x } (enlarge-window-horizontally
) makes the selected
window wider by the specified number of columns. The extra screen space
given to a window comes from one of its neighbors, if that is possible;
otherwise, all the competing windows are shrunk in the same proportion.
If this makes some windows too small, those windows are deleted and their
space is divided up. Minimum window size is specified by the variables
window-min-height
and window-min-width
.
You can also resize windows within a screen by clicking the left mouse button on a modeline, and dragging.
Clicking the right button on a mode line pops up a menu of common window manager operations. This menu contains the following options:
Emacs has many different major modes, each of which customizes Emacs for editing text of a particular sort. The major modes are mutually exclusive; at any time, each buffer has one major mode. The mode line normally contains the name of the current major mode in parentheses. See section The Mode Line.
The least specialized major mode is called Fundamental mode. This mode has no mode-specific redefinitions or variable settings. Each Emacs command behaves in its most general manner, and each option is in its default state. For editing any specific type of text, such as Lisp code or English text, you should switch to the appropriate major mode, such as Lisp mode or Text mode.
Selecting a major mode changes the meanings of a few keys to become more specifically adapted to the language being edited. TAB, DEL, and LFD are changed frequently. In addition, commands which handle comments use the mode to determine how to delimit comments. Many major modes redefine the syntactical properties of characters appearing in the buffer. See section The Syntax Table.
The major modes fall into three major groups. Lisp mode (which has several variants), C mode, and Muddle mode are for specific programming languages. Text mode, Nroff mode, TeX mode, and Outline mode are for editing English text. The remaining major modes are not intended for use on users' files; they are used in buffers created by Emacs for specific purposes and include Dired mode for buffers made by Dired (see section Dired, the Directory Editor), Mail mode for buffers made by C-x m (see section Sending Mail), and Shell mode for buffers used for communicating with an inferior shell process (see section Interactive Inferior Shell).
Most programming language major modes specify that only blank lines separate paragraphs. This is so that the paragraph commands remain useful. See section Paragraphs. They also cause Auto Fill mode to use the definition of TAB to indent the new lines it creates. This is because most lines in a program are usually indented. See section Indentation.
You can select a major mode explicitly for the current buffer, but most of the time Emacs determines which mode to use based on the file name or some text in the file.
Use a M-x command to explicitly select a new major mode. Add
-mode
to the name of a major mode to get the name of a command to
select that mode. For example, to enter Lisp mode, execute M-x
lisp-mode.
When you visit a file, Emacs usually chooses the right major mode
based on the file's name. For example, files whose names end in
.c
are edited in C mode. The variable auto-mode-alist
controls the correspondence between file names and major mode. Its value
is a list in which each element has the form:
(regexp . mode-function)
For example, one element normally found in the list has the form
("\\.c$" . c-mode)
. It is responsible for selecting C mode
for files whose names end in `.c'. (Note that `\\' is needed in
Lisp syntax to include a `\' in the string, which is needed to
suppress the special meaning of `.' in regexps.) The only practical
way to change this variable is with Lisp code.
You can specify which major mode should be used for editing a certain file by a special sort of text in the first non-blank line of the file. The mode name should appear in this line both preceded and followed by `-*-'. Other text may appear on the line as well. For example,
;-*-Lisp-*-
tells Emacs to use Lisp mode. Note how the semicolon is used to make Lisp treat this line as a comment. Such an explicit specification overrides any default mode based on the file name.
Another format of mode specification is:
-*-Mode: modename;-*-
which allows other things besides the major mode name to be specified. However, Emacs does not look for anything except the mode name.
The major mode can also be specified in a local variables list. See section Local Variables in Files.
When you visit a file that does not specify a major mode to use, or
when you create a new buffer with C-x b, Emacs uses the major mode
specified by the variable default-major-mode
. Normally this
value is the symbol fundamental-mode
, which specifies Fundamental
mode. If default-major-mode
is nil
, the major mode is
taken from the previously selected buffer.
newline-and-indent
).
delete-indentation
). This would cancel out
the effect of LFD.
split-line
).
back-to-indentation
).
indent-region
).
indent-rigidly
).
tab-to-tab-stop
).
Most programming languages have some indentation convention. For Lisp code, lines are indented according to their nesting in parentheses. The same general idea is used for C code, though details differ.
Use the TAB command to indent a line whatever the language. Each major mode defines this command to perform indentation appropriate for the particular language. In Lisp mode, TAB aligns a line according to its depth in parentheses. No matter where in the line you are when you type TAB, it aligns the line as a whole. In C mode, TAB implements a subtle and sophisticated indentation style that knows about many aspects of C syntax.
In Text mode, TAB runs the command tab-to-tab-stop
, which
indents to the next tab stop column. You can set the tab stops with
M-x edit-tab-stops.
If you just want to insert a tab character in the buffer, you can type C-q TAB.
To move over the indentation on a line, type Meta-m
(back-to-indentation
). This command, given anywhere on a line,
positions point at the first non-blank character on the line.
To insert an indented line before the current line, type C-a C-o TAB. To make an indented line after the current line, use C-e LFD.
C-M-o (split-line
) moves the text from point to the end of
the line vertically down, so that the current line becomes two lines.
C-M-o first moves point forward over any spaces and tabs. Then it
inserts after point a newline and enough indentation to reach the same
column point is on. Point remains before the inserted newline; in this
regard, C-M-o resembles C-o.
To join two lines cleanly, use the Meta-^
(delete-indentation
) command to delete the indentation at the
front of the current line, and the line boundary as well. Empty spaces
are replaced by a single space, or by no space if at the beginning of a
line, before a close parenthesis, or after an open parenthesis.
To delete just the indentation of a line, go to the beginning of the
line and use Meta-\ (delete-horizontal-space
), which
deletes all spaces and tabs around the cursor.
There are also commands for changing the indentation of several lines at
once. Control-Meta-\ (indent-region
) gives each line which
begins in the region the "usual" indentation by invoking TAB at the
beginning of the line. A numeric argument specifies the column to indent
to. Each line is shifted left or right so that its first non-blank
character appears in that column. C-x TAB
(indent-rigidly
) moves all the lines in the region right by its
argument (left, for negative arguments). The whole group of lines moves
rigidly sideways, which is how the command gets its name.
M-x indent-relative indents at point based on the previous line
(actually, the last non-empty line.) It inserts whitespace at point, moving
point, until it is underneath an indentation point in the previous line.
An indentation point is the end of a sequence of whitespace or the end of
the line. If point is farther right than any indentation point in the
previous line, the whitespace before point is deleted and the first
indentation point then applicable is used. If no indentation point is
applicable even then, tab-to-tab-stop
is run (see next section).
indent-relative
is the definition of TAB in Indented Text
mode. See section Commands for Human Languages.
For typing in tables, you can use Text mode's definition of TAB,
tab-to-tab-stop
. This command inserts indentation before point,
enough to reach the next tab stop column. Even if you are not in Text mode,
this function is associated with M-i anyway.
You can arbitrarily set the tab stops used by M-i. They are
stored as a list of column-numbers in increasing order in the variable
tab-stop-list
.
The convenient way to set the tab stops is using M-x edit-tab-stops,
which creates and selects a buffer containing a description of the tab stop
settings. You can edit this buffer to specify different tab stops, and
then type C-c C-c to make those new tab stops take effect. In the
tab stop buffer, C-c C-c runs the function
edit-tab-stops-note-changes
rather than the default
save-buffer
. edit-tab-stops
records which buffer was current
when you invoked it, and stores the tab stops in that buffer. Normally
all buffers share the same tab stops and changing them in one buffer
affects all. If you make tab-stop-list
local in one
buffer edit-tab-stops
in that buffer edits only the local
settings.
Below is the text representing ordinary tab stops every eight columns:
: : : : : : 0 1 2 3 4 0123456789012345678901234567890123456789012345678 To install changes, type C-c C-c
The first line contains a colon at each tab stop. The remaining lines help you see where the colons are and tell you what to do.
Note that the tab stops that control tab-to-tab-stop
have nothing
to do with displaying tab characters in the buffer. See section Variables Controlling Display,
for more information on that.
Emacs normally uses both tabs and spaces to indent lines. If you prefer,
all indentation can be made from spaces only. To request this, set
indent-tabs-mode
to nil
. This is a per-buffer variable;
altering the variable affects only the current buffer, but there is a
default value which you can change as well. See section Local Variables.
There are also commands to convert tabs to spaces or vice versa, always preserving the columns of all non-blank text. M-x tabify scans the region for sequences of spaces, and converts sequences of at least three spaces to tabs if that is possible without changing indentation. M-x untabify changes all tabs in the region to corresponding numbers of spaces.
The term text has two widespread meanings in our area of the computer field. One is data that is a sequence of characters. In this sense of the word any file that you edit with Emacs is text. The other meaning is more restrictive: a sequence of characters in a human language for humans to read (possibly after processing by a text formatter), as opposed to a program or commands for a program.
Human languages have syntactic and stylistic conventions that editor commands should support or use to advantage: conventions involving words, sentences, paragraphs, and capital letters. This chapter describes Emacs commands for all these things. There are also commands for filling, or rearranging paragraphs into lines of approximately equal length. The commands for moving over and killing words, sentences, and paragraphs, while intended primarily for editing text, are also often useful for editing programs.
Emacs has several major modes for editing human language text. If a file contains plain text, use Text mode, which customizes Emacs in small ways for the syntactic conventions of text. For text which contains embedded commands for text formatters, Emacs has other major modes, each for a particular text formatter. Thus, for input to TeX, you can use TeX mode; for input to nroff, Nroff mode.
You should use Text mode--rather than Fundamental or Lisp mode--to
edit files of text in a human language. Invoke M-x text-mode to
enter Text mode. In Text mode, TAB runs the function
tab-to-tab-stop
, which allows you to use arbitrary tab stops set
with M-x edit-tab-stops (see section Tab Stops). Features concerned
with comments in programs are turned off unless they are explicitly invoked.
The syntax table is changed so that periods are not considered part of a
word, while apostrophes, backspaces and underlines are.
A similar variant mode is Indented Text mode, intended for editing
text in which most lines are indented. This mode defines TAB to
run indent-relative
(see section Indentation), and makes Auto Fill
indent the lines it creates. As a result, a line made by Auto Filling,
or by LFD, is normally indented just like the previous line. Use
M-x indented-text-mode to select this mode.
Entering Text mode or Indented Text mode calls the value of the
variable text-mode-hook
with no arguments, if that value exists
and is not nil
. This value is also called when modes related to
Text mode are entered; this includes Nroff mode, TeX mode, Outline
mode, and Mail mode. Your hook can look at the value of
major-mode
to see which of these modes is actually being entered.
Two modes similar to Text mode are of use for editing text that is to be passed through a text formatter before achieving its final readable form.
Nroff mode is a mode like Text mode but modified to handle nroff commands present in the text. Invoke M-x nroff-mode to enter this mode. Nroff mode differs from Text mode in only a few ways. All nroff command lines are considered paragraph separators, so that filling never garbles the nroff commands. Pages are separated by `.bp' commands. Comments start with backslash-doublequote. There are also three special commands that are not available in Text mode:
forward-text-line
). An argument is a repeat count.
backward-text-line
).
count-text-lines
).
The other feature of Nroff mode is Electric Nroff newline mode. This is a minor mode that you can turn on or off with M-x electric-nroff-mode (see section Minor Modes). When the mode is on and you use RET to end a line containing an nroff command that opens a kind of grouping, Emacs automatically inserts the matching nroff command to close that grouping on the following line. For example, if you are at the beginning of a line and type .(b RET, the matching command `.)b' will be inserted on a new line following point.
Entering Nroff mode calls the value of the variable
text-mode-hook
with no arguments, if that value exists and is not
nil
; then it does the same with the variable
nroff-mode-hook
.
TeX is a powerful text formatter written by Donald Knuth; like GNU Emacs, it is free. LaTeX is a simplified input format for TeX, implemented by TeX macros. It is part of TeX.
Emacs has a special TeX mode for editing TeX input files. It provides facilities for checking the balance of delimiters and for invoking TeX on all or part of the file.
TeX mode has two variants, Plain TeX mode and LaTeX mode,
which are two distinct major modes that differ only slightly. These
modes are designed for editing the two different input formats. The
command M-x tex-mode looks at the contents of a buffer to
determine whether it appears to be LaTeX input or not; it then
selects the appropriate mode. If it can't tell which is right (e.g.,
the buffer is empty), the variable tex-default-mode
controls
which mode is used.
The commands M-x plain-tex-mode and M-x latex-mode explicitly select one of the variants of TeX mode. Use these commands when M-x tex-mode does not guess right.
TeX for Unix systems can be obtained from the University of Washington for a distribution fee.
To order a full distribution, send $140.00 for a 1/2 inch 9-track tape, $165.00 for two 4-track 1/4 inch cartridge tapes (foreign sites $150.00, for 1/2 inch, $175.00 for 1/4 inch, to cover the extra postage) payable to the University of Washington to:
The Director Northwest Computer Support Group, DW-10 University of Washington Seattle, Washington 98195
Purchase orders are acceptable, but there is an extra charge of $10.00 to pay for processing charges. (The total cost comes to $150 for domestic sites, $175 for foreign sites).
The normal distribution is a tar tape, blocked 20, 1600 bpi, on an industry standard 2400 foot half-inch reel. The physical format for the 1/4 inch streamer cartridges uses QIC-11, 8000 bpi, 4-track serpentine recording for the SUN. Also, SystemV tapes can be written in cpio format, blocked 5120 bytes, ASCII headers.
Here are the special commands provided in TeX mode for editing the text of the file.
TeX-insert-quote
).
tex-terminate-
paragraph
).
tex-insert-braces
).
up-list
).
tex-close-latex-block
).
In TeX, the character `"' is not normally used; you use `"'
to start a quotation and `"' to end one. TeX mode defines the key
" to insert `"' after whitespace or an open brace, `"'
after a backslash, or `"' otherwise. This is done by the command
tex-insert-quote
. If you need the character `"' itself in
unusual contexts, use C-q to insert it. Also, " with a
numeric argument always inserts that number of `"' characters.
In TeX mode, `$' has a special syntax code which attempts to understand the way TeX math mode delimiters match. When you insert a `$' that is meant to exit math mode, the position of the matching `$' that entered math mode is displayed for a second. This is the same feature that displays the open brace that matches a close brace that is inserted. However, there is no way to tell whether a `$' enters math mode or leaves it; so when you insert a `$' that enters math mode, the previous `$' position is shown as if it were a match, even though they are actually unrelated.
If you prefer to keep braces balanced at all times, you can use C-c {
(tex-insert-braces
) to insert a pair of braces. It leaves point
between the two braces so you can insert the text that belongs inside.
Afterward, use the command C-c } (up-list
) to move forward
past the close brace.
There are two commands for checking the matching of braces. LFD
(tex-terminate-paragraph
) checks the paragraph before point, and
inserts two newlines to start a new paragraph. It prints a message in the
echo area if any mismatch is found. M-x validate-tex-buffer checks
the entire buffer, paragraph by paragraph. When it finds a paragraph that
contains a mismatch, it displays point at the beginning of the paragraph
for a few seconds and pushes a mark at that spot. Scanning continues
until the whole buffer has been checked or until you type another key.
The positions of the last several paragraphs with mismatches can be
found in the mark ring (see section The Mark Ring).
Note that square brackets and parentheses, not just braces, are matched in TeX mode. This is wrong if you want to check TeX syntax. However, parentheses and square brackets are likely to be used in text as matching delimiters and it is useful for the various motion commands and automatic match display to work with them.
In LaTeX input, `\begin' and `\end' commands must balance.
After you insert a `\begin', use C-c C-f
(tex-close-latex-block
) to insert automatically a matching
`\end' (on a new line following the `\begin'). A blank line is
inserted between the two, and point is left there.
You can invoke TeX as an inferior of Emacs on either the entire contents of the buffer or just a region at a time. Running TeX in this way on just one chapter is a good way to see what your changes look like without taking the time to format the entire file.
tex-region
).
tex-buffer
).
tex-recenter-output-buffer
).
tex-kill-job
).
tex-print
).
tex-show-print-queue
).
You can pass the current buffer through an inferior TeX using
C-c C-b (tex-buffer
). The formatted output appears in a file
in `/tmp'; to print it, type C-c C-p (tex-print
).
Afterward use C-c C-q (tex-show-print-queue
) to view the
progress of your output towards being printed.
The console output from TeX, including any error messages, appears in a buffer called `*TeX-shell*'. If TeX gets an error, you can switch to this buffer and feed it input (this works as in Shell mode; see section Interactive Inferior Shell). Without switching to this buffer, you can scroll it so that its last line is visible by typing C-c C-l.
Type C-c C-k (tex-kill-job
) to kill the TeX process if
you see that its output is no longer useful. Using C-c C-b or
C-c C-r also kills any TeX process still running.
You can pass an arbitrary region through an inferior TeX by typing
C-c C-r (tex-region
). This is tricky, however, because
most files of TeX input contain commands at the beginning to set
parameters and define macros. Without them, no later part of the file
will format correctly. To solve this problem, C-c C-r allows you
to designate a part of the file as containing essential commands; it is
included before the specified region as part of the input to TeX.
The designated part of the file is called the header.
To indicate the bounds of the header in Plain TeX mode, insert two special strings in the file: `%**start of header' before the header, and `%**end of header' after it. Each string must appear entirely on one line, but there may be other text on the line before or after. The lines containing the two strings are included in the header. If `%**start of header' does not appear within the first 100 lines of the buffer, C-c C-r assumes there is no header.
In LaTeX mode, the header begins with `\documentstyle' and ends
with
`\begin{document}'. These are commands that LaTeX requires
you to use, so you don't need to do anything special to identify the
header.
When you enter either kind of TeX mode, Emacs calls with no
arguments the value of the variable text-mode-hook
, if that value
exists and is not nil
. Emacs then calls the variable
TeX-mode-hook
and either plain-TeX-mode-hook
or
LaTeX-mode-hook
under the same conditions.
Outline mode is a major mode similar to Text mode but intended for editing outlines. It allows you to make parts of the text temporarily invisible so that you can see just the overall structure of the outline. Type M-x outline-mode to turn on Outline mode in the current buffer.
When you enter Outline mode, Emacs calls with no arguments the value
of the variable text-mode-hook
, if that value exists and is not
nil
; then it does the same with the variable
outline-mode-hook
.
When a line is invisible in outline mode, it does not appear on the screen. The screen appears exactly as if the invisible line were deleted, except that an ellipsis (three periods in a row) appears at the end of the previous visible line (only one ellipsis no matter how many invisible lines follow).
All editing commands treat the text of the invisible line as part of the previous visible line. For example, C-n moves onto the next visible line. Killing an entire visible line, including its terminating newline, really kills all the following invisible lines as well; yanking everything back yanks the invisible lines and they remain invisible.
Outline mode assumes that the lines in the buffer are of two types: heading lines and body lines. A heading line represents a topic in the outline. Heading lines start with one or more stars; the number of stars determines the depth of the heading in the outline structure. Thus, a heading line with one star is a major topic; all the heading lines with two stars between it and the next one-star heading are its subtopics; and so on. Any line that is not a heading line is a body line. Body lines belong to the preceding heading line. Here is an example:
* Food This is the body, which says something about the topic of food. ** Delicious Food This is the body of the second-level header. ** Distasteful Food This could have a body too, with several lines. *** Dormitory Food * Shelter A second first-level topic with its header line.
A heading line together with all following body lines is called collectively an entry. A heading line together with all following deeper heading lines and their body lines is called a subtree.
You can customize the criterion for distinguishing heading lines by
setting the variable outline-regexp
. Any line whose beginning
has a match for this regexp is considered a heading line. Matches that
start within a line (not at the beginning) do not count. The length of
the matching text determines the level of the heading; longer matches
make a more deeply nested level. Thus, for example, if a text formatter
has commands `@chapter', `@section' and `@subsection'
to divide the document into chapters and sections, you can make those
lines count as heading lines by setting outline-regexp
to
`"@chap\\|@\\(sub\\)*section"'. Note the trick: the two words
`chapter' and `section' are the same length, but by defining
the regexp to match only `chap' we ensure that the length of the
text matched on a chapter heading is shorter, so that Outline mode will
know that sections are contained in chapters. This works as long as no
other command starts with `@chap'.
Outline mode makes a line invisible by changing the newline before it into an ASCII Control-M (code 015). Most editing commands that work on lines treat an invisible line as part of the previous line because, strictly speaking, it is part of that line, since there is no longer a newline in between. When you save the file in Outline mode, Control-M characters are saved as newlines, so the invisible lines become ordinary lines in the file. Saving does not change the visibility status of a line inside Emacs.
Some special commands in Outline mode move backward and forward to heading lines.
outline-next-visible-heading
).
outline-previous-visible-heading
).
outline-forward-same-level
).
outline-backward-same-level
).
outline-up-heading
).
C-c C-n (next-visible-heading
) moves down to the next
heading line. C-c C-p (previous-visible-heading
) moves
similarly backward. Both accept numeric arguments as repeat counts. The
names emphasize that invisible headings are skipped, but this is not really
a special feature. All editing commands that look for lines ignore the
invisible lines automatically.
More advanced motion commands understand the levels of headings.
The commands C-c C-f (outline-forward-same-level
) and
C-c C-b (outline-backward-same-level
) move from one
heading line to another visible heading at the same depth in
the outline. C-c C-u (outline-up-heading
) moves
backward to another heading that is less deeply nested.
The other special commands of outline mode are used to make lines visible
or invisible. Their names all start with hide
or show
.
Most of them exist as pairs of opposites. They are not undoable; instead,
you can undo right past them. Making lines visible or invisible is simply
not recorded by the undo mechanism.
hide-subtree
).
show-subtree
).
show-children
).
Two commands that are exact opposites are M-x hide-entry and M-x show-entry. They are used with point on a heading line, and apply only to the body lines of that heading. The subtopics and their bodies are not affected.
Two more powerful opposites are C-c C-h (hide-subtree
) and
C-c C-s (show-subtree
). Both should be used when point is
on a heading line, and both apply to all the lines of that heading's
subtree: its body, all its subheadings, both direct and indirect, and
all of their bodies. In other words, the subtree contains everything
following this heading line, up to and not including the next heading of
the same or higher rank.
Intermediate between a visible subtree and an invisible one is having all the subheadings visible but none of the body. There are two commands for doing this, one that hides the bodies and one that makes the subheadings visible. They are M-x hide-leaves and M-x show-branches.
A little weaker than show-branches
is C-c C-i
(show-children
). It makes just the direct subheadings
visible--those one level down. Deeper subheadings remain
invisible.
Two commands have a blanket effect on the whole file. M-x hide-body makes all body lines invisible, so that you see just the outline structure. M-x show-all makes all lines visible. You can think of these commands as a pair of opposites even though M-x show-all applies to more than just body lines.
You can turn off the use of ellipses at the ends of visible lines by
setting selective-display-ellipses
to nil
. The result is
no visible indication of the presence of invisible lines.
Emacs has commands for moving over or operating on words. By convention, the keys for them are all Meta- characters.
forward-word
).
backward-word
).
kill-word
).
backward-kill-word
).
mark-word
).
transpose-words
).
Notice how these keys form a series that parallels the character-based C-f, C-b, C-d, C-t and DEL. M-@ is related to C-@, which is an alias for C-SPC.
The commands Meta-f (forward-word
) and Meta-b
(backward-word
) move forward and backward over words. They are
analogous to Control-f and Control-b, which move over single
characters. Like their Control- analogues, Meta-f and
Meta-b move several words if given an argument. Meta-f with a
negative argument moves backward, and Meta-b with a negative argument
moves forward. Forward motion stops after the last letter of the
word, while backward motion stops before the first letter.
Meta-d (kill-word
) kills the word after point. To be
precise, it kills everything from point to the place Meta-f would
move to. Thus, if point is in the middle of a word, Meta-d kills
just the part after point. If some punctuation comes between point and the
next word, it is killed along with the word. (To kill only the
next word but not the punctuation before it, simply type Meta-f to get
to the end and kill the word backwards with Meta-DEL.)
Meta-d takes arguments just like Meta-f.
Meta-DEL (backward-kill-word
) kills the word before
point. It kills everything from point back to where Meta-b would
move to. If point is after the space in `FOO, BAR', then
`FOO, ' is killed. To kill just `FOO', type
Meta-b Meta-d instead of Meta-DEL.
Meta-t (transpose-words
) exchanges the word before or
containing point with the following word. The delimiter characters
between the words do not move. For example, transposing `FOO,
BAR' results in `BAR, FOO' rather than `BAR FOO,'.
See section Transposing Text, for more on transposition and on arguments to
transposition commands.
To operate on the next n words with an operation which applies
between point and mark, you can either set the mark at point and then move
over the words, or you can use the command Meta-@ (mark-word
)
which does not move point but sets the mark where Meta-f would move
to. It can be given arguments just like Meta-f.
The word commands' understanding of syntax is completely controlled by the syntax table. For example, any character can be declared to be a word delimiter. See section The Syntax Table.
The Emacs commands for manipulating sentences and paragraphs are mostly on Meta- keys, and therefore are like the word-handling commands.
backward-sentence
).
forward-sentence
).
kill-sentence
).
backward-kill-sentence
).
The commands Meta-a and Meta-e (backward-sentence
and forward-sentence
) move to the beginning and end of the
current sentence, respectively. They resemble Control-a and
Control-e, which move to the beginning and end of a line. Unlike
their counterparts, Meta-a and Meta-e move over successive
sentences if repeated or given numeric arguments. Emacs assumes
the typist's convention is followed, and thus considers a sentence to
end wherever there is a `.', `?', or `!' followed by the
end of a line or two spaces, with any number of `)', `]',
`'', or `"' characters allowed in between. A sentence also
begins or ends wherever a paragraph begins or ends.
Neither M-a nor M-e moves past the newline or spaces beyond the sentence edge at which it is stopping.
M-a and M-e have a corresponding kill command, just like
C-a and C-e have C-k. The command is M-k
(kill-sentence
) which kills from point to the end of the
sentence. With minus one as an argument it kills back to the beginning
of the sentence. Larger arguments serve as repeat counts.
There is a special command, C-x DEL
(backward-kill-sentence
), for killing back to the beginning of a
sentence, which is useful when you change your mind in the middle of
composing text.
The variable sentence-end
controls recognition of the end of a
sentence. It is a regexp that matches the last few characters of a
sentence, together with the whitespace following the sentence. Its
normal value is:
"[.?!][]\"')]*\\($\\|\t\\| \\)[ \t\n]*"
This example is explained in the section on regexps. See section Syntax of Regular Expressions.
The Emacs commands for manipulating paragraphs are also Meta- keys.
backward-paragraph
).
forward-paragraph
).
mark-paragraph
).
Meta-[ moves to the beginning of the current or previous paragraph, while Meta-] moves to the end of the current or next paragraph. Blank lines and text formatter command lines separate paragraphs and are not part of any paragraph. An indented line starts a new paragraph.
In major modes for programs (as opposed to Text mode), paragraphs begin and end only at blank lines. As a result, the paragraph commands continue to be useful even though there are no paragraphs per se.
When there is a fill prefix, paragraphs are delimited by all lines which don't start with the fill prefix. See section Filling Text.
To operate on a paragraph, you can use the command
Meta-h (mark-paragraph
) to set the region around it. This
command puts point at the beginning and mark at the end of the paragraph
point was in. If point is between paragraphs (in a run of blank lines or
at a boundary), the paragraph following point is surrounded by point and
mark. If there are blank lines preceding the first line of the paragraph,
one of the blank lines is included in the region. Thus, for example,
M-h C-w kills the paragraph around or after point.
The precise definition of a paragraph boundary is controlled by the
variables paragraph-separate
and paragraph-start
. The value
of paragraph-start
is a regexp that matches any line that
either starts or separates paragraphs. The value of
paragraph-separate
is another regexp that matches only lines
that separate paragraphs without being part of any paragraph. Lines that
start a new paragraph and are contained in it must match both regexps. For
example, normally paragraph-start
is "^[ \t\n\f]"
and paragraph-separate
is "^[ \t\f]*$"
.
Normally it is desirable for page boundaries to separate paragraphs. The default values of these variables recognize the usual separator for pages.
Files are often thought of as divided into pages by the formfeed character (ASCII Control-L, octal code 014). For example, if a file is printed on a line printer, each "page" of the file starts on a new page of paper. Emacs treats a page-separator character just like any other character. It can be inserted with C-q C-l or deleted with DEL. You are free to paginate your file or not. However, since pages are often meaningful divisions of the file, commands are provided to move over them and operate on them.
backward-page
).
forward-page
).
mark-page
).
count-lines-page
).
The C-x [ (backward-page
) command moves point to
immediately after the previous page delimiter. If point is already
right after a page delimiter, the command skips that one and stops at
the previous one. A numeric argument serves as a repeat count. The
C-x ] (forward-page
) command moves forward past the next
page delimiter.
The C-x C-p command (mark-page
) puts point at the beginning
of the current page and the mark at the end. The page delimiter at the end
is included (the mark follows it). The page delimiter at the front is
excluded (point follows it). You can follow this command by C-w to
kill a page you want to move elsewhere. If you insert the page after a page
delimiter, at a place where C-x ] or C-x [ would take you,
the page will be properly delimited before and after once again.
A numeric argument to C-x C-p is used to specify which page to go to, relative to the current one. Zero means the current page. One means the next page, and -1 means the previous one.
The C-x l command (count-lines-page
) can help you decide
where to break a page in two. It prints the total number of lines in
the current page in the echo area, then divides the lines into those
preceding the current line and those following it, for example
Page has 96 (72+25) lines
Notice that the sum is off by one; this is correct if point is not at the beginning of a line.
The variable page-delimiter
should have as its value a regexp that
matches the beginning of a line that separates pages. This defines
where pages begin. The normal value of this variable is "^\f"
,
which matches a formfeed character at the beginning of a line.
If you use Auto Fill mode, Emacs fills text (breaks it up into lines that fit in a specified width) as you insert it. When you alter existing text it is often no longer be properly filled afterwards and you can use explicit commands for filling.
Auto Fill mode is a minor mode in which lines are broken automatically when they become too wide. Breaking happens only when you type a SPC or RET.
M-x auto-fill-mode turns Auto Fill mode on if it was off, or off if it was on. With a positive numeric argument the command always turns Auto Fill mode on, and with a negative argument it always turns it off. The presence of the word `Fill' in the mode line, inside the parentheses, indicates that Auto Fill mode is in effect. Auto Fill mode is a minor mode; you can turn it on or off for each buffer individually. See section Minor Modes.
In Auto Fill mode, lines are broken automatically at spaces when they get longer than desired. Line breaking and rearrangement takes place only when you type SPC or RET. To insert a space or newline without permitting line-breaking, type C-q SPC or C-q LFD (recall that a newline is really a linefeed). C-o inserts a newline without line breaking.
Auto Fill mode works well with Lisp mode: when it makes a new line in
Lisp mode, it indents that line with TAB. If a line ending in a
Lisp comment gets too long, the text of the comment is split into two
comment lines. Optionally, new comment delimiters are inserted at the
end of the first line and the beginning of the second, so that each line
is a separate comment. The variable comment-multi-line
controls
the choice (see section Manipulating Comments).
Auto Fill mode does not refill entire paragraphs. It can break lines but cannot merge lines. Editing in the middle of a paragraph can result in a paragraph that is not correctly filled. The easiest way to make the paragraph properly filled again is using an explicit fill commands.
Many users like Auto Fill mode and want to use it in all text files. The section on init files explains how you can arrange this permanently for yourself. See section The Init File, .emacs.
fill-paragraph
).
fill-region
).
set-fill-column
).
To refill a paragraph, use the command Meta-q
(fill-paragraph
). It causes the paragraph containing point, or
the one after point if point is between paragraphs, to be refilled. All
line breaks are removed, and new ones are inserted where necessary.
M-q can be undone with C-_. See section Undoing Changes.
To refill many paragraphs, use M-g (fill-region
), which
divides the region into paragraphs and fills each of them.
Meta-q and Meta-g use the same criteria as Meta-h for finding paragraph boundaries (see section Paragraphs). For more control, you can use M-x fill-region-as-paragraph, which refills everything between point and mark. This command recognizes only blank lines as paragraph separators.
A numeric argument to M-g or M-q causes it to justify the text as well as filling it. Extra spaces are inserted to make the right margin line up exactly at the fill column. To remove the extra spaces, use M-q or M-g with no argument.
The variable auto-fill-inhibit-regexp
takes as a value a regexp to
match lines that should not be auto-filled.
The command Meta-s (center-line
) centers the current line
within the current fill column. With an argument, it centers several lines
individually and moves past them.
The maximum line width for filling is in the variable
fill-column
. Altering the value of fill-column
makes it
local to the current buffer; until then, the default value--initially
70--is in effect. See section Local Variables.
The easiest way to set fill-column
is to use the command C-x
f (set-fill-column
). With no argument, it sets fill-column
to the current horizontal position of point. With a numeric argument, it
uses that number as the new fill column.
To fill a paragraph in which each line starts with a special marker (which might be a few spaces, giving an indented paragraph), use the fill prefix feature. The fill prefix is a string which is not included in filling. Emacs expects every line to start with a fill prefix.
set-fill-prefix
).
fill-paragraph
).
To specify a fill prefix, move to a line that starts with the desired
prefix, put point at the end of the prefix, and give the command
C-x . (set-fill-prefix
). That's a period after the
C-x. To turn off the fill prefix, specify an empty prefix: type
C-x . with point at the beginning of a line.
When a fill prefix is in effect, the fill commands remove the fill prefix from each line before filling and insert it on each line after filling. Auto Fill mode also inserts the fill prefix inserted on new lines it creates. Lines that do not start with the fill prefix are considered to start paragraphs, both in M-q and the paragraph commands; this is just right if you are using paragraphs with hanging indentation (every line indented except the first one). Lines which are blank or indented once the prefix is removed also separate or start paragraphs; this is what you want if you are writing multi-paragraph comments with a comment delimiter on each line.
The fill prefix is stored in the variable fill-prefix
. Its value
is a string, or nil
when there is no fill prefix. This is a
per-buffer variable; altering the variable affects only the current buffer,
but there is a default value which you can change as well. See section Local Variables.
Another way to use fill prefixes is through M-x fill-individual-paragraphs. This function divides the region into groups of consecutive lines with the same amount and kind of indentation and fills each group as a paragraph, using its indentation as a fill prefix.
Emacs has commands for converting either a single word or any arbitrary range of text to upper case or to lower case.
downcase-word
).
upcase-word
).
capitalize-word
).
downcase-region
).
upcase-region
).
The word conversion commands are used most frequently. Meta-l
(downcase-word
) converts the word after point to lower case,
moving past it. Thus, repeating Meta-l converts successive words.
Meta-u (upcase-word
) converts to all capitals instead,
while Meta-c (capitalize-word
) puts the first letter of the
word into upper case and the rest into lower case. The word conversion
commands convert several words at once if given an argument. They are
especially convenient for converting a large amount of text from all
upper case to mixed case: you can move through the text using
M-l, M-u, or M-c on each word as appropriate,
occasionally using M-f instead to skip a word.
When given a negative argument, the word case conversion commands apply to the appropriate number of words before point, but do not move point. This is convenient when you have just typed a word in the wrong case: you can give the case conversion command and continue typing.
If a word case conversion command is given in the middle of a word, it
applies only to the part of the word which follows point. This is just
like what Meta-d (kill-word
) does. With a negative argument,
case conversion applies only to the part of the word before point.
The other case conversion commands are C-x C-u
(upcase-region
) and C-x C-l (downcase-region
), which
convert everything between point and mark to the specified case. Point and
mark do not move.
Emacs has many commands designed to understand the syntax of programming languages such as Lisp and C. These commands can:
The commands available for words, sentences, and paragraphs are useful in editing code even though their canonical application is for editing human language text. Most symbols contain words (see section Words); sentences can be found in strings and comments (see section Sentences). Paragraphs per se are not present in code, but the paragraph commands are useful anyway, because Lisp mode and C mode define paragraphs to begin and end at blank lines (see section Paragraphs). Judicious use of blank lines to make the program clearer also provides interesting chunks of text for the paragraph commands to work on.
The selective display feature is useful for looking at the overall structure of a function (see section Selective Display). This feature causes only the lines that are indented less than a specified amount to appear on the screen.
Emacs has several major modes for the programming languages Lisp, Scheme (a variant of Lisp), C, Fortran, and Muddle. Ideally, a major mode should be implemented for each programming language you might want to edit with Emacs; but often the mode for one language can serve for other syntactically similar languages. The language modes that exist are those that someone decided to take the trouble to write.
There are several variants of Lisp mode, which differ in the way they interface to Lisp execution. See section Major Modes for Lisp.
Each of the programming language modes defines the TAB key to run
an indentation function that knows the indentation conventions of that
language and updates the current line's indentation accordingly. For
example, in C mode TAB is bound to c-indent-line
. LFD
is normally defined to do RET followed by TAB; thus it, too,
indents in a mode-specific fashion.
In most programming languages, indentation is likely to vary from line to
line. So the major modes for those languages rebind DEL to treat a
tab as if it were the equivalent number of spaces (using the command
backward-delete-char-untabify
). This makes it possible to rub out
indentation one column at a time without worrying whether it is made up of
spaces or tabs. In these modes, use C-b C-d to delete a tab
character before point.
Programming language modes define paragraphs to be separated only by blank lines, so that the paragraph commands remain useful. Auto Fill mode, if enabled in a programming language major mode, indents the new lines which it creates.
Turning on a major mode calls a user-supplied function called the
mode hook, which is the value of a Lisp variable. For example,
turning on C mode calls the value of the variable c-mode-hook
if
that value exists and is non-nil
. Mode hook variables for other
programming language modes include lisp-mode-hook
,
emacs-lisp-mode-hook
, lisp-interaction-mode-hook
,
scheme-mode-hook
, and muddle-mode-hook
. The mode hook
function receives no arguments.
By convention, Emacs keys for dealing with balanced expressions are usually Control-Meta- characters. They tend to be analogous in function to their Control- and Meta- equivalents. These commands are usually thought of as pertaining to expressions in programming languages, but can be useful with any language in which some sort of parentheses exist (including English).
The commands fall into two classes. Some commands deal only with lists (parenthetical groupings). They see nothing except parentheses, brackets, braces (depending on what must balance in the language you are working with), and escape characters that might be used to quote those.
The other commands deal with expressions or sexps. The word `sexp' is derived from s-expression, the term for a symbolic expression in Lisp. In Emacs, the notion of `sexp' is not limited to Lisp. It refers to an expression in the language your program is written in. Each programming language has its own major mode, which customizes the syntax tables so that expressions in that language count as sexps.
Sexps typically include symbols, numbers, and string constants, as well as anything contained in parentheses, brackets, or braces.
In languages that use prefix and infix operators, such as C, it is not possible for all expressions to be sexps. For example, C mode does not recognize `foo + bar' as an sexp, even though it is a C expression; it recognizes `foo' as one sexp and `bar' as another, with the `+' as punctuation between them. This is a fundamental ambiguity: both `foo + bar' and `foo' are legitimate choices for the sexp to move over if point is at the `f'. Note that `(foo + bar)' is a sexp in C mode.
Some languages have obscure forms of syntax for expressions that nobody has bothered to make Emacs understand properly.
forward-sexp
).
backward-sexp
).
kill-sexp
).
backward-up-list
).
down-list
).
forward-list
).
backward-list
).
transpose-sexps
).
mark-sexp
).
To move forward over an sexp, use C-M-f (forward-sexp
). If
the first significant character after point is an opening delimiter
(`(' in Lisp; `(', `[', or `{' in C), C-M-f
moves past the matching closing delimiter. If the character begins a
symbol, string, or number, C-M-f moves over that. If the character
after point is a closing delimiter, C-M-f just moves past it. (This
last is not really moving across an sexp; it is an exception which is
included in the definition of C-M-f because it is as useful a
behavior as anyone can think of for that situation.)
The command C-M-b (backward-sexp
) moves backward over a
sexp. The detailed rules are like those above for C-M-f, but with
directions reversed. If there are any prefix characters (single quote,
back quote, and comma, in Lisp) preceding the sexp, C-M-b moves back
over them as well.
C-M-f or C-M-b with an argument repeats that operation the specified number of times; with a negative argument, it moves in the opposite direction.
In languages such as C where the comment-terminator can be recognized, the sexp commands move across comments as if they were whitespace. In Lisp and other languages where comments run until the end of a line, it is very difficult to ignore comments when parsing backwards; therefore, in such languages the sexp commands treat the text of comments as if it were code.
Killing an sexp at a time can be done with C-M-k (kill-sexp
).
C-M-k kills the characters that C-M-f would move over.
The list commands, C-M-n (forward-list
) and
C-M-p (backward-list
), move over lists like the sexp
commands but skip over any number of other kinds of sexps (symbols,
strings, etc). In some situations, these commands are useful because
they usually ignore comments, since the comments usually do not contain
any lists.
C-M-n and C-M-p stay at the same level in parentheses, when
that is possible. To move up one (or n) levels, use C-M-u
(backward-up-list
).
C-M-u moves backward up past one unmatched opening delimiter. A
positive argument serves as a repeat count; a negative argument reverses
direction of motion and also requests repetition, so it moves forward and
up one or more levels.
To move down in list structure, use C-M-d
(down-list
). In Lisp mode, where `(' is the only opening
delimiter, this is nearly the same as searching for a `('. An
argument specifies the number of levels of parentheses to go down.
C-M-t (transpose-sexps
) drags the previous sexp across
the next one. An argument serves as a repeat count, and a negative
argument drags backwards (thus canceling out the effect of C-M-t with
a positive argument). An argument of zero, rather than doing nothing,
transposes the sexps ending after point and the mark.
To make the region be the next sexp in the buffer, use C-M-@
(mark-sexp
) which sets the mark at the same place that
C-M-f would move to. C-M-@ takes arguments like
C-M-f. In particular, a negative argument is useful for putting
the mark at the beginning of the previous sexp.
The list and sexp commands' understanding of syntax is completely controlled by the syntax table. Any character can, for example, be declared to be an opening delimiter and act like an open parenthesis. See section The Syntax Table.
In Emacs, a parenthetical grouping at the top level in the buffer is
called a defun. The name derives from the fact that most
top-level lists in Lisp are instances of the special form
defun
, but Emacs calls any top-level parenthetical
grouping counts a defun regardless of its contents or
the programming language. For example, in C, the body of a
function definition is a defun.
beginning-of-defun
).
end-of-defun
).
mark-defun
).
The commands to move to the beginning and end of the current defun are
C-M-a (beginning-of-defun
) and C-M-e (end-of-defun
).
To operate on the current defun, use C-M-h (mark-defun
)
which puts point at the beginning and the mark at the end of the current
or next defun. This is the easiest way to prepare for moving the defun
to a different place. In C mode, C-M-h runs the function
mark-c-function
, which is almost the same as mark-defun
,
but which backs up over the argument declarations, function name, and
returned data type so that the entire C function is inside the region.
To compile and evaluate the current defun, use M-x compile-defun. This function prints the results in the minibuffer. If you include an argument, it inserts the value in the current buffer after the defun.
Emacs assumes that any open-parenthesis found in the leftmost column is the start of a defun. Therefore, never put an open-parenthesis at the left margin in a Lisp file unless it is the start of a top level list. Never put an open-brace or other opening delimiter at the beginning of a line of C code unless it starts the body of a function. The most likely problem case is when you want an opening delimiter at the start of a line inside a string. To avoid trouble, put an escape character (`\' in C and Emacs Lisp, `/' in some other Lisp dialects) before the opening delimiter. It will not affect the contents of the string.
The original Emacs found defuns by moving upward a level of parentheses until there were no more levels to go up. This required scanning back to the beginning of the buffer for every function. To speed this up, Emacs was changed to assume that any `(' (or other character assigned the syntactic class of opening-delimiter) at the left margin is the start of a defun. This heuristic is nearly always right; however, it mandates the convention described above.
The best way to keep a program properly indented ("ground") is to use Emacs to re-indent it as you change the program. Emacs has commands to indent properly either a single line, a specified number of lines, or all of the lines inside a single parenthetical grouping.
newline-and-indent
).
The basic indentation command is TAB, which gives the current
line the correct indentation as determined from the previous lines. The
function that TAB runs depends on the major mode; it is
lisp-indent-line
in Lisp mode, c-indent-line
in C mode,
etc. These functions understand different syntaxes for different
languages, but they all do about the same thing. TAB in any
programming language major mode inserts or deletes whitespace at the
beginning of the current line, independent of where point is in the
line. If point is inside the whitespace at the beginning of the line,
TAB leaves it at the end of that whitespace; otherwise, TAB
leaves point fixed with respect to the characters around it.
Use C-q TAB to insert a tab at point.
When entering a large amount of new code, use LFD
(newline-and-indent
), which is equivalent to a RET followed
by a TAB. LFD creates a blank line, then gives it the
appropriate indentation.
TAB indents the second and following lines of the body of a parenthetical grouping each under the preceding one; therefore, if you alter one line's indentation to be nonstandard, the lines below tend to follow it. This is the right behavior in cases where the standard result of TAB does not look good.
Remember that Emacs assumes that an open-parenthesis, open-brace, or other opening delimiter at the left margin (including the indentation routines) is the start of a function. You should therefore never have an opening delimiter in column zero that is not the beginning of a function, not even inside a string. This restriction is vital for making the indentation commands fast. See section Defuns, for more information on this behavior.
Several commands are available to re-indent several lines of code which have been altered or moved to a different level in a list structure.
indent-sexp
).
indent-region
).
To re-indent the contents of a single list, position point before the
beginning of it and type C-M-q. This key is bound to
indent-sexp
in Lisp mode, indent-c-exp
in C mode, and
bound to other suitable functions in other modes. The indentation of
the line the sexp starts on is not changed; therefore, only the relative
indentation within the list, and not its position, is changed. To
correct the position as well, type a TAB before C-M-q.
If the relative indentation within a list is correct but the indentation of its beginning is not, go to the line on which the list begins and type C-u TAB. When you give TAB a numeric argument, it moves all the lines in the group, starting on the current line, sideways the same amount that the current line moves. The command does not move lines that start inside strings, or C preprocessor lines when in C mode.
Another way to specify a range to be re-indented is with point and
mark. The command C-M-\ (indent-region
) applies TAB
to every line whose first character is between point and mark.
The indentation pattern for a Lisp expression can depend on the function called by the expression. For each Lisp function, you can choose among several predefined patterns of indentation, or define an arbitrary one with a Lisp program.
The standard pattern of indentation is as follows: the second line of the expression is indented under the first argument, if that is on the same line as the beginning of the expression; otherwise, the second line is indented underneath the function name. Each following line is indented under the previous line whose nesting depth is the same.
If the variable lisp-indent-offset
is non-nil
, it overrides
the usual indentation pattern for the second line of an expression, so that
such lines are always indented lisp-indent-offset
more columns than
the containing list.
Certain functions override the standard pattern. Functions
whose names start with def
always indent the second line by
lisp-body-indention
extra columns beyond the open-parenthesis
starting the expression.
Individual functions can override the standard pattern in various
ways, according to the lisp-indent-function
property of the
function name. (Note: lisp-indent-function
was formerly called
lisp-indent-hook
). There are four possibilities for this
property:
nil
defun
def
is used for
this function also.
lisp-body-indent
more columns than the open-parenthesis starting the containing
expression. If the argument is distinguished and is either the first
or second argument, it is indented twice that many extra columns.
If the argument is distinguished and not the first or second argument,
the standard pattern is followed for that line.
parse-partial-sexp
(a Lisp primitive for
indentation and nesting computation) when it parses up to the
beginning of this line.
Two variables control which commands perform C indentation and when.
If c-auto-newline
is non-nil
, newlines are inserted both
before and after braces that you insert and after colons and semicolons.
Correct C indentation is done on all the lines that are made this way.
If c-tab-always-indent
is non-nil
, the TAB command
in C mode does indentation only if point is at the left margin or within
the line's indentation. If there is non-whitespace to the left of point,
TAB just inserts a tab character in the buffer. Normally,
this variable is nil
, and TAB always reindents the current line.
C does not have anything analogous to particular function names for which special forms of indentation are desirable. However, it has a different need for customization facilities: many different styles of C indentation are in common use.
There are six variables you can set to control the style that Emacs C mode will use.
c-indent-level
c-continued-statement-offset
if
or body of a while
.
c-brace-offset
c-brace-imaginary-offset
c-argdecl-indent
c-label-offset
The variable c-indent-level
controls the indentation for C
statements with respect to the surrounding block. In the example:
{ foo ();
the difference in indentation between the lines is c-indent-level
.
Its standard value is 2.
If the open-brace beginning the compound statement is not at the beginning
of its line, the c-indent-level
is added to the indentation of the
line, not the column of the open-brace. For example,
if (losing) { do_this ();
One popular indentation style is that which results from setting
c-indent-level
to 8 and putting open-braces at the end of a line
in this way. Another popular style prefers to put the open-brace on a
separate line.
In fact, the value of the variable c-brace-imaginary-offset
is
also added to the indentation of such a statement. Normally this variable
is zero. Think of this variable as the imaginary position of the open
brace, relative to the first non-blank character on the line. By setting
the variable to 4 and c-indent-level
to 0, you can get this style:
if (x == y) { do_it (); }
When c-indent-level
is zero, the statements inside most braces
line up exactly under the open brace. An exception are braces in column
zero, like those surrounding a function's body. The statements inside
those braces are not placed at column zero. Instead,
c-brace-offset
and c-continued-statement-offset
(see
below) are added to produce a typical offset between brace levels, and
the statements are indented that far.
c-continued-statement-offset
controls the extra indentation for
a line that starts within a statement (but not within parentheses or
brackets). These lines are usually statements inside other statements,
like the then-clauses of if
statements and the bodies of
while
statements. The c-continued-statement-offset
parameter determines the difference in indentation between the two lines in:
if (x == y) do_it ();
The default value for c-continued-statement-offset
is 2. Some
popular indentation styles correspond to a value of zero for
c-continued-statement-offset
.
c-brace-offset
is the extra indentation given to a line that
starts with an open-brace. Its standard value is zero;
compare:
if (x == y) {
with:
if (x == y) do_it ();
If you set c-brace-offset
to 4, the first example becomes:
if (x == y) {
c-argdecl-indent
controls the indentation of declarations of the
arguments of a C function. It is absolute: argument declarations receive
exactly c-argdecl-indent
spaces. The standard value is 5 and
results in code like this:
char * index (string, char) char *string; int char;
c-label-offset
is the extra indentation given to a line that
contains a label, a case statement, or a default:
statement. Its
standard value is -2 and results in code like this:
switch (c) { case 'x':
If c-label-offset
were zero, the same code would be indented as:
switch (c) { case 'x':
This example assumes that the other variables above also have their default values.
Using the indentation style produced by the default settings of the variables just discussed and putting open braces on separate lines produces clear and readable files. For an example, look at any of the C source files of GNU Emacs.
The Emacs parenthesis-matching feature shows you automatically how parentheses match in the text. Whenever a self-inserting character that is a closing delimiter is typed, the cursor moves momentarily to the location of the matching opening delimiter, provided that is visible on the screen. If it is not on the screen, some text starting with that opening delimiter is displayed in the echo area. Either way, you see the grouping you are closing off.
In Lisp, automatic matching applies only to parentheses. In C, it also applies to braces and brackets. Emacs knows which characters to regard as matching delimiters based on the syntax table set by the major mode. See section The Syntax Table.
If the opening delimiter and closing delimiter are mismatched--as in `[x)'---the echo area displays a warning message. The correct matches are specified in the syntax table.
Two variables control parenthesis matching displays.
blink-matching-paren
turns the feature on or off. The default is
t
(match display is on); nil
turns it off.
blink-matching-paren-distance
specifies how many characters back
Emacs searches to find a matching opening delimiter. If the match is
not found in the specified region, scanning stops, and nothing is
displayed. This prevents wasting lots of time scanning when there is no
match. The default is 4000.
The comment commands insert, kill and align comments.
indent-for-comment
).
set-comment-column
).
kill-comment
).
indent-new-comment-line
).
The command that creates a comment is Meta-;
(indent-for-comment
). If there is no comment already on the
line, a new comment is created and aligned at a specific column called
the comment column. Emacs creates the comment by inserting the
string at the value of comment-start
; see below. Point is left
after that string. If the text of the line extends past the comment
column, indentation is done to a suitable boundary (usually, at least
one space is inserted). If the major mode has specified a string to
terminate comments, that string is inserted after point, to keep the
syntax valid.
You can also use Meta-; to align an existing comment. If a line already contains the string that starts comments, M-; just moves point after it and re-indents it to the conventional place. Exception: comments starting in column 0 are not moved.
Some major modes have special rules for indenting certain kinds of comments in certain contexts. For example, in Lisp code, comments which start with two semicolons are indented as if they were lines of code, instead of at the comment column. Comments which start with three semicolons are supposed to start at the left margin. Emacs understands these conventions by indenting a double-semicolon comment using TAB and by not changing the indentation of a triple-semicolon comment at all.
;; This function is just an example. ;;; Here either two or three semicolons are appropriate. (defun foo (x) ;;; And now, the first part of the function: ;; The following line adds one. (1+ x)) ; This line adds one.
In C code, a comment preceded on its line by nothing but whitespace is indented like a line of code.
Even when an existing comment is properly aligned, M-; is still useful for moving directly to the start of the comment.
C-u - C-x ; (kill-comment
) kills the comment on the
current line, if there is one. The indentation before the start of the
comment is killed as well. If there does not appear to be a comment in
the line, nothing happens. To reinsert the comment on another line,
move to the end of that line, type first C-y, and then M-;
to realign the comment. Note that C-u - C-x ; is not a distinct
key; it is C-x ; (set-comment-column
) with a negative
argument. That command is programmed to call kill-comment
when
called with a negative argument. However, kill-comment
is a
valid command which you could bind directly to a key if you wanted to.
If you are typing a comment and want to continue it on another line,
use the command Meta-LFD (indent-new-comment-line
),
which terminates the comment you are typing, creates a new blank line
afterward, and begins a new comment indented under the old one. If
Auto Fill mode is on and you go past the fill column while typing, the
comment is continued in just this fashion. If point is
not at the end of the line when you type M-LFD, the text on
the rest of the line becomes part of the new comment line.
The comment column is stored in the variable comment-column
. You
can explicitly set it to a number. Alternatively, the command C-x ;
(set-comment-column
) sets the comment column to the column point is
at. C-u C-x ; sets the comment column to match the last comment
before point in the buffer, and then calls Meta-; to align the
current line's comment under the previous one. Note that C-u - C-x ;
runs the function kill-comment
as described above.
comment-column
is a per-buffer variable; altering the variable
affects only the current buffer. You can also change the default value.
See section Local Variables. Many major modes initialize this variable
for the current buffer.
The comment commands recognize comments based on the regular expression
that is the value of the variable comment-start-skip
. This regexp
should not match the null string. It may match more than the comment
starting delimiter in the strictest sense of the word; for example, in C
mode the value of the variable is "/\\*+ *"
, which matches extra
stars and spaces after the `/*' itself. (Note that `\\' is
needed in Lisp syntax to include a `\' in the string, which is needed
to deny the first star its special meaning in regexp syntax. See section Syntax of Regular Expressions.)
When a comment command makes a new comment, it inserts the value of
comment-start
to begin it. The value of comment-end
is
inserted after point and will follow the text you will insert
into the comment. In C mode, comment-start
has the value
"/* "
and comment-end
has the value " */"
.
comment-multi-line
controls how M-LFD
(indent-new-comment-line
) behaves when used inside a comment. If
comment-multi-line
is nil
, as it normally is, then
M-LFD terminates the comment on the starting line and starts
a new comment on the new following line. If comment-multi-line
is not nil
, then M-LFD sets up the new following line
as part of the same comment that was found on the starting line. This
is done by not inserting a terminator on the old line and not inserting
a starter on the new line. In languages where multi-line comments are legal,
the value you choose for this variable is a matter of taste.
The variable comment-indent-hook
should contain a function that
is called to compute the indentation for a newly inserted comment or for
aligning an existing comment. Major modes set this variable differently.
The function is called with no arguments, but with point at the
beginning of the comment, or at the end of a line if a new comment is to
be inserted. The function should return the column in which the comment
ought to start. For example, in Lisp mode, the indent hook function
bases its decision on the number of semicolons that begin an existing
comment and on the code in the preceding lines.
insert-parentheses
).
move-over-close-and-reindent
).
The commands M-( (insert-parentheses
) and M-)
(move-over-close-
) are designed to facilitate a style of
editing which keeps parentheses balanced at all times. M-( inserts a
pair of parentheses, either together as in `()', or, if given an
argument, around the next several sexps, and leaves point after the open
parenthesis. Instead of typing ( F O O ), you can type M-( F O
O, which has the same effect except for leaving the cursor before the
close parenthesis. You can then type M-), which moves past the
close parenthesis, deletes any indentation preceding it (in this example
there is none), and indents with LFD after it.
and-reindent
Completion usually happens in the minibuffer. An exception is completion for Lisp symbol names, which is available in all buffers.
The command M-TAB (lisp-complete-symbol
) takes the
partial Lisp symbol before point to be an abbreviation, and compares it
against all non-trivial Lisp symbols currently known to Emacs. Any
additional characters that they all have in common are inserted at point.
Non-trivial symbols are those that have function definitions, values, or
properties.
If there is an open-parenthesis immediately before the beginning of the partial symbol, only symbols with function definitions are considered as completions.
If the partial name in the buffer has more than one possible completion and they have no additional characters in common, a list of all possible completions is displayed in another window.
As you edit Lisp code to be run in Emacs, you can use the commands
C-h f (describe-function
) and C-h v
(describe-variable
) to print documentation of functions and
variables you want to call. These commands use the minibuffer to
read the name of a function or variable to document, and display the
documentation in a window.
For extra convenience, these commands provide default arguments based on the code in the neighborhood of point. C-h f sets the default to the function called in the innermost list containing point. C-h v uses the symbol name around or adjacent to point as its default.
The M-x manual-entry command gives you access to documentation
on Unix commands, system calls, and libraries. The command reads a
topic as an argument, and displays the Unix manual page for that topic.
manual-entry
always searches all 8 sections of the
manual and concatenates all the entries it finds. For example,
the topic `termcap' finds the description of the termcap library
from section 3, followed by the description of the termcap data base
from section 5.
The Emacs command M-x add-change-log-entry helps you keep a record of when and why you have changed a program. It assumes that you have a file in which you write a chronological sequence of entries describing individual changes. The default is to store the change entries in a file called `ChangeLog' in the same directory as the file you are editing. The same `ChangeLog' file therefore records changes for all the files in a directory.
A change log entry starts with a header line that contains your name and the current date. Except for these header lines, every line in the change log starts with a tab. One entry can describe several changes; each change starts with a line starting with a tab and a star. M-x add-change-log-entry visits the change log file and creates a new entry unless the most recent entry is for today's date and your name. In either case, it adds a new line to start the description of another change just after the header line of the entry. When M-x add-change-log-entry is finished, all is prepared for you to edit in the description of what you changed and how. You must then save the change log file yourself.
The change log file is always visited in Indented Text mode, which means that LFD and auto-filling indent each new line like the previous line. This is convenient for entering the contents of an entry, which must be indented. See section Text Mode.
Here is an example of the formatting conventions used in the change log for Emacs:
Wed Jun 26 19:29:32 1985 Richard M. Stallman (rms at mit-prep) * xdisp.c (try_window_id): If C-k is done at end of next-to-last line, this fn updates window_end_vpos and cannot leave window_end_pos nonnegative (it is zero, in fact). If display is preempted before lines are output, this is inconsistent. Fix by setting blank_end_of_window to nonzero. Tue Jun 25 05:25:33 1985 Richard M. Stallman (rms at mit-prep) * cmds.c (Fnewline): Call the auto fill hook if appropriate. * xdisp.c (try_window_id): If point is found by compute_motion after xp, record that permanently. If display_text_line sets point position wrong (case where line is killed, point is at eob and that line is not displayed), set it again in final compute_motion.
A tag table is a description of how a multi-file program is broken up into files. It lists the names of the component files and the names and positions of the functions in each file. Grouping the related files makes it possible to search or replace through all the files with one command. Recording the function names and positions makes it possible to use the Meta-. command, which finds the definition of a function without asking for information on the file it is in.
Tag tables are stored in files called tag table files. The conventional name for a tag table file is `TAGS'.
Each entry in the tag table records the name of one tag, the name of the file that the tag is defined in (implicitly), and the position in that file of the tag's definition.
The programming language of a file determines what names are recorded in the tag table depends on. Normally, Emacs includes all functions and subroutines, and may also include global variables, data types, and anything else convenient. Each recorded name is called a tag.
In Lisp code, any function defined with defun
, any variable
defined with defvar
or defconst
, and the first argument of
any expression that starts with `(def' in column zero, is a tag.
In C code, any C function is a tag, and so is any typedef if -t
is
specified when the tag table is constructed.
In Fortran code, functions and subroutines are tags.
In LaTeX text, the argument of any of the commands \chapter
,
\section
, \subsection
, \subsubsection
, \eqno
,
\label
, \ref
, \cite
, \bibitem
, and
\typeout
is a tag.
The etags
program is used to create a tag table file. It knows
the syntax of C, Fortran, LaTeX, Scheme, and Emacs Lisp/Common Lisp. To
use etags
, use it as a shell command:
etags inputfiles...
The program reads the specified files and writes a tag table
named `TAGS' in the current working directory. etags
recognizes the language used in an input file based on the name and
contents of the file; there are no switches for specifying the language.
The -t
switch tells etags
to record typedefs in C code as
tags.
If the tag table data become outdated due to changes in the files described in the table, you can update the tag table by running the program from the shell again. It is not necessary to do this often.
If the tag table fails to record a tag, or records it for the wrong file, Emacs cannot find its definition. However, if the position recorded in the tag table becomes a little bit wrong (due to some editing in the file that the tag definition is in), the only consequence is to slow down finding the tag slightly. Even if the stored position is very wrong, Emacs will still find the tag, but it must search the entire file for it.
You should update a tag table when you define new tags you want to have listed, when you move tag definitions from one file to another, or when changes become substantial. You don't have to update the tag table after each edit, or even every day.
At any time Emacs has one selected tag table, and all the commands
for working with tag tables use the selected one. To select a tag table,
use the variable tag-table-alist
.
The value of tag-table-alist
is a list that determines which
TAGS
files should be active for a given buffer. This is not
really an association list, in that all elements are checked. The car
of each element of this list is a pattern against which the buffers file
name is compared; if it matches, then the cdr of the list should be the
name of the tags table to use. If more than one element of this list
matches the buffers file name, all of the associated tags tables are
used. Earlier ones are searched first.
If the car of elements of this list are strings, they are treated
as regular-expressions against which the file is compared (like the
auto-mode-alist
). If they are not strings, they are evaluated.
If they evaluate to non-nil
, the current buffer is considered to
match.
If the cdr of the elements of this list are strings, they are assumed to name a tags file. If they name a directory, the string `tags' is appended to them to get the file name. If they are not strings, they are evaluated and must return an appropriate string.
For example:
(setq tag-table-alist '(("/usr/src/public/perl/" . "/usr/src/public/perl/perl-3.0/") ("\\.el$" . "/usr/local/emacs/src/") ("/jbw/gnu/" . "/usr15/degree/stud/jbw/gnu/") ("" . "/usr/local/emacs/src/") ))
The example defines the tag table alist in the following way:
If you had a file called `/usr/jbw/foo.el', it would use both
`TAGS' files,
`/usr/local/emacs/src/TAGS' and
`/usr15/degree/stud/jbw/gnu/TAGS' (in that order), because it
matches both patterns.
If the buffer-local variable buffer-tag-table
is set, it names a
tags table that is searched before all others when find-tag
is
executed from this buffer.
If there is a file called `TAGS' in the same directory as the file
in question, then that tags file will always be used as well (after the
buffer-tag-table
but before the tables specified by this list).
If the variable tags-file-name
is set, the `TAGS' file it names
will apply to all buffers (for backwards compatibility.) It is searched
first.
If the value of the variable tags-always-build-completion-table
is t
, the tags file will always be added to the completion table
without asking first, regardless of the size of the tags file.
The function M-x visit-tags-table, is largely made obsolete by
the variable tag-table-alist
, tells tags commands to use the tags
table file file first. The file should be the name of a
file created with the etags
program. A directory name is also
acceptable; it means the file `TAGS' in that directory. The
function only stores the file name you provide in the variable
tags-file-name
. Emacs does not actually read in the tag table
contents until you try to use them. You can set the variable explicitly
instead of using visit-tags-table
. The value of the variable
tags-file-name
is the name of the tags table used by all buffers.
This is for backward compatibility, and is largely supplanted by the
variable tag-table-alist
.
The most important thing that a tag table enables you to do is to find the definition of a specific tag.
find-tag
).
find-tag-other-window
).
M-. (find-tag
) is the command to find the definition of
a specified tag. It searches through the tag table for that tag, as a
string, then uses the tag table information to determine the file in
which the definition is used and the approximate character position of
the definition in the file. Then find-tag
visits the file,
moves point to the approximate character position, and starts searching
ever-increasing distances away for the text that should appear at
the beginning of the definition.
If an empty argument is given (by typing RET), the sexp in the buffer before or around point is used as the name of the tag to find. See section Lists and Sexps, for information on sexps.
The argument to find-tag
need not be the whole tag name; it can
be a substring of a tag name. However, there can be many tag names
containing the substring you specify. Since find-tag
works by
searching the text of the tag table, it finds the first tag in the table
that the specified substring appears in. To find other tags that match
the substring, give find-tag
a numeric argument, as in C-u
M-.. This does not read a tag name, but continues searching the tag
table's text for another tag containing the same substring last used.
If your keyboard has a real META key, M-0 M-. is an easier
alternative to C-u M-..
If the optional second argument other-window is non-nil
, it uses
another window to display the tag.
Multiple active tags tables and completion are supported.
Variables of note include the following:
Like most commands that can switch buffers, find-tag
has another
similar command that displays the new buffer in another window. C-x 4
. invokes the function find-tag-other-window
. (This key sequence
ends with a period.)
Emacs comes with a tag table file `TAGS' (in the directory
containing Lisp libraries) that includes all the Lisp libraries and all
the C sources of Emacs. By specifying this file with visit-tags-table
and then using M-. you can quickly look at the source of any Emacs
function.
The commands in this section visit and search all the files listed in the selected tag table, one by one. For these commands, the tag table serves only to specify a sequence of files to search. A related command is M-x grep (see section Running `make', or Compilers Generally).
query-replace
on each file in the selected tag table.
tags-loop-continue
).
M-x tags-search reads a regexp using the minibuffer, then visits
the files of the selected tag table one by one, and searches through each
file for that regexp. It displays the name of the file being searched so
you can follow its progress. As soon as an occurrence is found,
tags-search
returns.
After you have found one match, you probably want to find all the rest.
To find one more match, type M-, (tags-loop-continue
) to
resume the tags-search
. This searches the rest of the current
buffer, followed by the remaining files of the tag table.
M-x tags-query-replace performs a single query-replace
through all the files in the tag table. It reads a string to search for
and a string to replace with, just like ordinary M-x query-replace.
It searches much like M-x tags-search but repeatedly, processing
matches according to your input. See section Replacement Commands, for more information on
query-replace
.
It is possible to get through all the files in the tag table with a single invocation of M-x tags-query-replace. But since any unrecognized character causes the command to exit, you may need to continue from where you left off. You can use M-, to do this. It resumes the last tags search or replace command that you did.
It may have struck you that tags-search
is a lot like grep
.
You can also run grep
itself as an inferior of Emacs and have Emacs
show you the matching lines one by one. This works mostly the same as
running a compilation and having Emacs show you where the errors were.
See section Running `make', or Compilers Generally.
If you wish to process all the files in a selected tag table, but M-x tags-search and M-x tags-query-replace are not giving you the desired result, you can use M-x next-file.
M-x list-tags reads the name of one of the files described by the selected tag table, and displays a list of all the tags defined in that file. The "file name" argument is really just a string to compare against the names recorded in the tag table; it is read as a string rather than a file name. Therefore, completion and defaulting are not available, and you must enter the string the same way it appears in the tag table. Do not include a directory as part of the file name unless the file name recorded in the tag table contains that directory.
M-x tags-apropos is like apropos
for tags. It reads a regexp,
then finds all the tags in the selected tag table whose entries match that
regexp, and displays the tag names found.
Fortran mode provides special motion commands for Fortran statements and subprograms, and indentation commands that understand Fortran conventions of nesting, line numbers, and continuation statements.
Special commands for comments are provided because Fortran comments are unlike those of other languages.
Built-in abbrevs optionally save typing when you insert Fortran keywords.
Use M-x fortran-mode to switch to this major mode. Doing so calls
the value of fortran-mode-hook
as a function of no arguments if
that variable has a non-nil
value.
Fortran mode was contributed by Michael Prange.
Fortran mode provides special commands to move by subprograms (functions and subroutines) and by statements. There is also a command to put the region around one subprogram, which is convenient for killing it or moving it.
beginning-of-fortran-subprogram
).
end-of-fortran-subprogram
).
mark-fortran-subprogram
).
fortran-next-
statement
).
fortran-
previous-statement
).
Special commands and features are available for indenting Fortran code. They make sure various syntactic entities (line numbers, comment line indicators, and continuation line flags) appear in the columns that are required for standard Fortran.
fortran-indent-line
).
fortran-indent-subprogram
).
TAB is redefined by Fortran mode to reindent the current line for
Fortran (fortran-indent-line
). Line numbers and continuation
markers are indented to their required columns, and the body of the
statement is independently indented, based on its nesting in the program.
The key C-M-q is redefined as fortran-indent-subprogram
, a
command that reindents all the lines of the Fortran subprogram (function or
subroutine) containing point.
The key M-LFD is redefined as fortran-split-line
, a
command to split a line in the appropriate fashion for Fortran. In a
non-comment line, the second half becomes a continuation line and is
indented accordingly. In a comment line, both halves become separate
comment lines.
If a number is the first non-whitespace in the line, it is assumed to be a line number and is moved to columns 0 through 4. (Columns are always counted from 0 in GNU Emacs.) If the text on the line starts with the conventional Fortran continuation marker `$', it is moved to column 5. If the text begins with any non whitespace character in column 5, it is assumed to be an unconventional continuation marker and remains in column 5.
Line numbers of four digits or less are normally indented one space.
This amount is controlled by the variable fortran-line-number-indent
,
which is the maximum indentation a line number can have. Line numbers
are indented to right-justify them to end in column 4 unless that would
require more than the maximum indentation. The default value of the
variable is 1.
Simply inserting a line number is enough to indent it according to these
rules. As each digit is inserted, the indentation is recomputed. To turn
off this feature, set the variable fortran-electric-line-number
to
nil
. Then inserting line numbers is like inserting anything else.
Fortran mode assumes that you follow certain conventions that simplify the task of understanding a Fortran program well enough to indent it properly:
fortran-continuation-char
.
By default, this character is `$'.
If you fail to follow these conventions, the indentation commands may indent some lines unaesthetically. However, a correct Fortran program will retain its meaning when reindented even if the conventions are not followed.
Several additional variables control how Fortran indentation works.
fortran-do-indent
fortran-if-indent
fortran-continuation-indent
fortran-check-all-num-for-matching-do
nil
, indentation assumes that each `do'
statement ends on a `continue' statement. Therefore, when
computing indentation for a statement other than `continue', it
can save time by not checking for a `do' statement ending there.
If this is non-nil
, indenting any numbered statement must check
for a `do' that ends there. The default is nil
.
fortran-minimum-statement-indent
The usual Emacs comment commands assume that a comment can follow a line of code. In Fortran, the standard comment syntax requires an entire line to be just a comment. Therefore, Fortran mode replaces the standard Emacs comment commands and defines some new variables.
Fortran mode can also handle a non-standard comment syntax where comments
start with `!' and can follow other text. Because only some Fortran
compilers accept this syntax, Fortran mode will not insert such comments
unless you have specified to do so in advance by setting the variable
comment-start
to `"!"' (see section Variables).
fortran-comment-indent
).
fortran-comment-region
).
M-; in Fortran mode is redefined as the command
fortran-comment-indent
. Like the usual M-; command,
it recognizes an existing comment and aligns its text appropriately.
If there is no existing comment, a comment is inserted and aligned.
Inserting and aligning comments is not the same in Fortran mode as in other modes. When a new comment must be inserted, a full-line comment is inserted if the current line is blank. On a non-blank line, a non-standard `!' comment is inserted if you previously specified you wanted to use them. Otherwise a full-line comment is inserted on a new line before the current line.
Non-standard `!' comments are aligned like comments in other
languages, but full-line comments are aligned differently. In a
standard full-line comment, the comment delimiter itself must always
appear in column zero. What can be aligned is the text within the
comment. You can choose from three styles of alignment by setting the
variable fortran-comment-indent-style
to one of these values:
fixed
fortran-comment-line-column
. This is the default.
relative
fortran-comment-line-column
columns of indentation.
nil
You can also specify the character to be used to indent within
full-line comments by setting the variable fortran-comment-indent-char
to the character you want to use.
Fortran mode introduces two variables comment-line-start
and
comment-line-start-skip
, which do for full-line comments what
comment-start
and comment-start-skip
do for
ordinary text-following comments. Normally these are set properly by
Fortran mode, so you do not need to change them.
The normal Emacs comment command C-x ; has not been redefined. It can therefore be used if you use `!' comments, but is useless in Fortran mode otherwise.
The command C-c ; (fortran-comment-region
) turns all the
lines of the region into comments by inserting the string `C$$$' at
the front of each one. With a numeric arg, the region is turned back into
live code by deleting `C$$$' from the front of each line. You can
control the string used for the comments by setting the variable
fortran-comment-region
. Note that here we have an example of a
command and a variable with the same name; the two uses of the name never
conflict because in Lisp and in Emacs it is always clear from the context
which one is referred to.
fortran-column-ruler
).
fortran-window-create
).
The command C-c C-r (fortran-column-ruler
) shows a column
ruler above the current line. The comment ruler consists of two lines
of text that show you the locations of columns with special significance
in Fortran programs. Square brackets show the limits of the columns for
line numbers, and curly brackets show the limits of the columns for the
statement body. Column numbers appear above them.
Note that the column numbers count from zero, as always in GNU Emacs. As a result, the numbers may not be those you are familiar with; but the actual positions in the line are standard Fortran.
The text used to display the column ruler is the value of the variable
fortran-comment-ruler
. By changing this variable, you can change
the display.
For even more help, use C-c C-w (fortran-window-create
), a
command which splits the current window horizontally, resulting in a window 72
columns wide. When you edit in this window, you can immediately see
when a line gets too wide to be correct Fortran.
Fortran mode provides many built-in abbrevs for common keywords and declarations. These are the same sort of abbrevs that you can define yourself. To use them, you must turn on Abbrev mode. see section Abbrevs.
The built-in abbrevs are unusual in one way: they all start with a semicolon. You cannot normally use semicolon in an abbrev, but Fortran mode makes this possible by changing the syntax of semicolon to "word constituent".
For example, one built-in Fortran abbrev is `;c' for `continue'. If you insert `;c' and then insert a punctuation character such as a space or a newline, the `;c' changes automatically to `continue', provided Abbrev mode is enabled.
Type `;?' or `;C-h' to display a list of all built-in Fortran abbrevs and what they stand for.
The previous chapter discusses the Emacs commands that are useful for making changes in programs. This chapter deals with commands that assist in the larger process of developing and maintaining programs.
Emacs can run compilers for non-interactive languages like C and Fortran as inferior processes, feeding the error log into an Emacs buffer. It can also parse the error messages and visit the files in which errors are found, moving point to the line where the error occurred.
grep
asynchronously under Emacs, with matching lines
listed in the buffer named `*compilation*'.
M-x compile
command.
grep
subprocess.
grep
match.
To run make
or another compiler, type M-x compile. This
command reads a shell command line using the minibuffer, then executes
the specified command line in an inferior shell with output going to the
buffer named `*compilation*'. By default, the current buffer's
default directory is used as the working directory for the execution of
the command; therefore, the makefile comes from this directory.
When the shell command line is read, the minibuffer appears containing a
default command line (the command you used the last time you typed
M-x compile). If you type just RET, the same command line is used
again. The first M-x compile provides make -k
as the default.
The default is taken from the variable compile-command
; if the
appropriate compilation command for a file is something other than
make -k
, it can be useful to have the file specify a local value for
compile-command
(see section Local Variables in Files).
When you start a compilation, the buffer `*compilation*' is displayed in another window but not selected. Its mode line displays the word `run' or `exit' in the parentheses to tell you whether compilation is finished. You do not have to keep this buffer visible; compilation continues in any case.
To kill the compilation process, type M-x-kill-compilation. The mode line of the `*compilation*' buffer changes to say `signal' instead of `run'. Starting a new compilation also kills any running compilation, as only one can occur at any time. Starting a new compilation prompts for confirmation before actually killing a compilation that is running.
To parse the compiler error messages, type C-x `
(next-error
). The character following C-x is the grave
accent, not the single quote. The command displays the buffer
`*compilation*' in one window and the buffer in which the next
error occurred in another window. Point in that buffer is moved to the
line where the error was found. The corresponding error message is
scrolled to the top of the window in which `*compilation*' is
displayed.
The first time you use C-x ` after the start of a compilation, it parses all the error messages, visits all the files that have error messages, and creates markers pointing at the lines the error messages refer to. It then moves to the first error message location. Subsequent uses of C-x ` advance down the data set up by the first use. When the preparsed error messages are exhausted, the next C-x ` checks for any more error messages that have come in; this is useful if you start editing compiler errors while compilation is still going on. If no additional error messages have come in, C-x ` reports an error.
C-u C-x ` discards the preparsed error message data and parses the `*compilation*' buffer again, then displays the first error. This way, you can process the same set of errors again.
Instead of running a compiler, you can run grep
and see the
lines on which matches were found. To do this, type M-x grep with
an argument line that contains the same arguments you would give to
grep
: a grep
-style regexp (usually in single quotes to
quote the shell's special characters) followed by filenames, which may
use wildcard characters. The output from grep
goes in the
`*compilation*' buffer. You can use C-x ` to find the lines that
match as if they were compilation errors.
Note: a shell is used to run the compile command, but the shell is not
run in interactive mode. In particular, this means that the shell starts
up with no prompt. If you find your usual shell prompt making an
unsightly appearance in the `*compilation*' buffer, it means you
have made a mistake in your shell's initialization file (`.cshrc'
or `.shrc' or ...) by setting the prompt unconditionally. The
shell initialization file should set the prompt only if there already is
a prompt. Here's how to do it in csh
:
if ($?prompt) set prompt = ...
Emacs has four different major modes for Lisp. They are the same in terms of editing commands, but differ in the commands for executing Lisp expressions.
Lisp code for Emacs editing commands is stored in files whose names conventionally end in `.el'. This ending tells Emacs to edit them in Emacs-Lisp mode (see section Major Modes for Lisp).
To execute a file of Emacs Lisp, use M-x load-file. This command reads the file name you provide in the minibuffer, then executes the contents of that file as Lisp code. It is not necessary to visit the file first; in fact, this command reads the file as found on disk, not the text in an Emacs buffer.
Once a file of Lisp code is installed in the Emacs Lisp library
directories, users can load it using M-x load-library. Programs can
load it by calling load-library
, or with load
, a more primitive
function that is similar but accepts some additional arguments.
M-x load-library differs from M-x load-file in that it searches a sequence of directories and tries three file names in each directory. The three names are: first, the specified name with `.elc' appended; second, the name with `.el' appended; third, the specified name alone. A `.elc' file would be the result of compiling the Lisp file into byte code; if possible, it is loaded in preference to the Lisp file itself because the compiled file loads and runs faster.
Because the argument to load-library
is usually not in itself
a valid file name, file name completion is not available. In fact, when
using this command, you usually do not know exactly what file name
will be used.
The sequence of directories searched by M-x load-library is
specified by the variable load-path
, a list of strings that are
directory names. The elements of this list may not begin with "`~'",
so you must call expand-file-name
on them before adding them to
the list. The default value of the list contains the directory where
the Lisp code for Emacs itself is stored. If you have libraries of your
own, put them in a single directory and add that directory to
load-path
. nil
in this list stands for the current
default directory, but it is probably not a good idea to put nil
in the list. If you start wishing that nil
were in the list, you
should probably use M-x load-file for this case.
The variable is initialized by the EMACSLOADPATH environment variable. If no value is specified, the variable takes the default value specified in the file `paths.h' when Emacs was built. If a path isn't specified in `paths.h', a default value is obtained from the file system, near the directory in which the Emacs executable resides.
Like M-x load-library, M-x locate-library searches the
directories in load-path
to find the file that M-x load-library
would load. If the optional second argument nosuffix is
non-nil
, the suffixes `.elc' or `.el' are not added to
the specified name library (like calling load
instead of
load-library
).
You often do not have to give any command to load a library, because the
commands defined in the library are set up to autoload that library.
Running any of those commands causes load
to be called to load the
library; this replaces the autoload definitions with the real ones from the
library.
If autoloading a file does not finish, either because of an error or
because of a C-g quit, all function definitions made by the file
are undone automatically. So are any calls to provide
. As a
consequence, the entire file is loaded a second time if you use one of
the autoloadable commands again. This prevents problems when the
command is no longer autoloading but is working incorrectly because the file
was only partially loaded. Function definitions are undone only for
autoloading; explicit calls to load
do not undo anything if
loading is not completed.
The variable after-load-alist
takes an alist of expressions to be
evaluated when particular files are loaded. Each element has the form
(filename forms...)
. When load
is run and the filename
argument is filename, the forms in the corresponding element are
executed at the end of loading.
filename must match exactly. Normally filename is the
name of a library, with no directory specified, since that is how load
is normally called. An error in forms
does not undo the load, but
it does prevent execution of the rest of the forms
.
Emacs Lisp code can be compiled into byte-code which loads faster, takes up less space when loaded, and executes faster.
byte-compile-file creates a byte-code compiled file from an
Emacs-Lisp source file. The default argument for this function is the
file visited in the current buffer. The function reads the specified
file, compiles it into byte code, and writes an output file whose name
is made by appending `c' to the input file name. Thus, the file
`rmail.el' would be compiled into `rmail.elc'. To compile a
file of Lisp code named filename into a file of byte code and
then load it, use byte-compile-and-load-file
. To compile and
evaluate Lisp code in a given buffer, use byte-compile-buffer
.
To recompile all changed Lisp files in a directory, use M-x byte-recompile-directory. Specify just the directory name as an argument. Each `.el' file that has been byte-compiled before is byte-compiled again if it has changed since the previous compilation. A numeric argument to this command tells it to offer to compile each `.el' file that has not been compiled yet. You must answer y or n to each offer.
You can use the function batch-byte-compile
to invoke Emacs
non-interactively from the shell to do byte compilation. When you use
this function, the files to be compiled are specified with command-line
arguments. Use a shell command of the form:
emacs -batch -f batch-byte-compile files...
Directory names may also be given as arguments; in that case,
byte-recompile-directory
is invoked on each such directory.
batch-byte-compile
uses all remaining command-line arguments as
file or directory names, then kills the Emacs process.
M-x disassemble explains the result of byte compilation. Its argument is a function name. It displays the byte-compiled code in a help window in symbolic form, one instruction per line. If the instruction refers to a variable or constant, that is shown, too.
GNU Emacs can run Mocklisp files by converting them to Emacs Lisp first. To convert a Mocklisp file, visit it and then type M-x convert-mocklisp-buffer. Then save the resulting buffer of Lisp file in a file whose name ends in `.el' and use the new file as a Lisp library.
You cannot currently byte-compile converted Mocklisp code. The reason is that converted Mocklisp code uses some special Lisp features to deal with Mocklisp's incompatible ideas of how arguments are evaluated and which values signify "true" or "false".
Lisp programs intended to be run in Emacs should be edited in Emacs-Lisp mode; this will happen automatically for file names ending in `.el'. By contrast, Lisp mode itself should be used for editing Lisp programs intended for other Lisp systems. Emacs-Lisp mode can be selected with the command M-x emacs-lisp-mode.
For testing of Lisp programs to run in Emacs, it is useful to be able to evaluate part of the program as it is found in the Emacs buffer. For example, if you change the text of a Lisp function definition and then evaluate the definition, Emacs installs the change for future calls to the function. Evaluation of Lisp expressions is also useful in any kind of editing task for invoking non-interactive functions (functions that are not commands).
eval-expression
).
eval-last-sexp
).
eval-defun
).
M-ESC (eval-expression
) is the most basic command
for evaluating a Lisp expression interactively. It reads the expression
using the minibuffer, so you can execute any expression on a buffer
regardless of what the buffer contains. When evaluation is complete,
the current buffer is once again the buffer that was current when
M-ESC was typed.
M-ESC can easily confuse users, especially on keyboards
with autorepeat, where it can result from holding down the ESC key
for too long. Therefore, eval-expression
is normally a disabled
command. Attempting to use this command asks for confirmation and gives
you the option of enabling it; once you enable the command, you are no
longer required to confirm. See section Disabling Commands.
In Emacs-Lisp mode, the key C-M-x is bound to the function
eval-defun
, which parses the defun containing point or following point
as a Lisp expression and evaluates it. The value is printed in the echo
area. This command is convenient for installing in the Lisp environment
changes that you have just made in the text of a function definition.
The command C-x C-e (eval-last-sexp
) performs a similar job
but is available in all major modes, not just Emacs-Lisp mode. It finds
the sexp before point, reads it as a Lisp expression, evaluates it, and
prints the value in the echo area. It is sometimes useful to type in an
expression and then, with point still after it, type C-x C-e.
If C-M-x or C-x C-e are given a numeric argument, they print the value by inserting it into the current buffer at point, rather than in the echo area. The argument value does not matter.
The most general command for evaluating Lisp expressions from a buffer
is eval-region
. M-x eval-region parses the text of the
region as one or more Lisp expressions, evaluating them one by one.
M-x eval-current-buffer is similar, but it evaluates the entire
buffer. This is a reasonable way to install the contents of a file of
Lisp code that you are just ready to test. After finding and fixing a
bug, use C-M-x on each function that you change, to keep the Lisp
world in step with the source file.
GNU Emacs contains a debugger for Lisp programs executing inside it.
This debugger is normally not used; many commands frequently get Lisp
errors when invoked in inappropriate contexts (such as C-f at the
end of the buffer) and it would be unpleasant to enter a special
debugging mode in this case. When you want to make Lisp errors invoke
the debugger, you must set the variable debug-on-error
to
non-nil
. Quitting with C-g is not considered an error, and
debug-on-error
has no effect on the handling of C-g.
However, if you set debug-on-quit
to be non-nil
, C-g will
invoke the debugger. This can be useful for debugging an infinite loop;
type C-g once the loop has had time to reach its steady state.
debug-on-quit
has no effect on errors.
You can make Emacs enter the debugger when a specified function
is called or at a particular place in Lisp code. Use M-x
debug-on-entry with argument fun-name to have Emacs enter the
debugger as soon as fun-name is called. Use
M-x cancel-debug-on-entry to make the function stop entering the
debugger when called. (Redefining the function also does this.) To enter
the debugger from some other place in Lisp code, you must insert the
expression (debug)
there and install the changed code with
C-M-x. See section Evaluating Emacs-Lisp Expressions.
When the debugger is entered, it displays the previously selected buffer in one window and a buffer named `*Backtrace*' in another window. The backtrace buffer contains one line for each level of Lisp function execution currently going on. At the beginning of the buffer is a message describing the reason that the debugger was invoked, for example, an error message if it was invoked due to an error.
The backtrace buffer is read-only and is in Backtrace mode, a special major mode in which letters are defined as debugger commands. The usual Emacs editing commands are available; you can switch windows to examine the buffer that was being edited at the time of the error, and you can switch buffers, visit files, and perform any other editing operations. However, the debugger is a recursive editing level (see section Recursive Editing Levels); it is a good idea to return to the backtrace buffer and explictly exit the debugger when you don't want to use it any more. Exiting the debugger kills the backtrace buffer.
The contents of the backtrace buffer show you the functions that are executing and the arguments that were given to them. It also allows you to specify a stack frame by moving point to the line describing that frame. The frame whose line point is on is considered the current frame. Some of the debugger commands operate on the current frame. Debugger commands are mainly used for stepping through code one expression at a time. Here is a list of them:
wrong-type-argument
errors will use the debugger's return value
instead of the invalid argument; no-catch
errors will use the
debugger value as a throw tag instead of the tag that was not found.
If an error was signaled by calling the Lisp function signal
,
the debugger's return value is returned as the value of signal
.
The buffer `*scratch*', which is selected when Emacs starts up, is provided for evaluating Lisp expressions interactively inside Emacs. Both the expressions you evaluate and their output goes in the buffer.
The `*scratch*' buffer's major mode is Lisp Interaction mode, which
is the same as Emacs-Lisp mode except for one command, LFD. In
Emacs-Lisp mode, LFD is an indentation command. In Lisp
Interaction mode, LFD is bound to eval-print-last-sexp
. This
function reads the Lisp expression before point, evaluates it, and inserts
the value in printed representation before point.
The way to use the `*scratch*' buffer is to insert Lisp expressions at the end, ending each one with LFD so that it will be evaluated. The result is a complete typescript of the expressions you have evaluated and their values.
The rationale for this feature is that Emacs must have a buffer when it starts up, but that buffer is not useful for editing files since a new buffer is made for every file that you visit. The Lisp interpreter typescript is the most useful thing I can think of for the initial buffer to do. M-x lisp-interaction-mode will put any buffer in Lisp Interaction mode.
Emacs has facilities for running programs in other Lisp systems. You can run a Lisp process as an inferior of Emacs, and pass expressions to it to be evaluated. You can also pass changed function definitions directly from the Emacs buffers in which you edit the Lisp programs to the inferior Lisp process.
To run an inferior Lisp process, type M-x run-lisp. This runs the
program named lisp
, the same program you would run by typing
lisp
as a shell command, with both input and output going through an
Emacs buffer named `*lisp*'. In other words, any "terminal output"
from Lisp will go into the buffer, advancing point, and any "terminal
input" for Lisp comes from text in the buffer. To give input to Lisp, go
to the end of the buffer and type the input, terminated by RET. The
`*lisp*' buffer is in Inferior Lisp mode, which has all the
special characteristics of Lisp mode and Shell mode (see section Shell Mode).
Use Lisp mode to run the source files of programs in external Lisps. You can select this mode with M-x lisp-mode. It is used automatically for files whose names end in `.l' or `.lisp', as most Lisp systems usually expect.
When you edit a function in a Lisp program you are running, the easiest
way to send the changed definition to the inferior Lisp process is the key
C-M-x. In Lisp mode, this key runs the function lisp-send-defun
,
which finds the defun around or following point and sends it as input to
the Lisp process. (Emacs can send input to any inferior process regardless
of what buffer is current.)
Contrast the meanings of C-M-x in Lisp mode (for editing programs to be run in another Lisp system) and Emacs-Lisp mode (for editing Lisp programs to be run in Emacs): in both modes it has the effect of installing the function definition that point is in, but the way of doing so is different according to where the relevant Lisp environment is found. See section Major Modes for Lisp.
An abbrev is a word which expands into some different text. Abbrevs are defined by the user to expand in specific ways. For example, you might define `foo' as an abbrev expanding to `find outer otter'. With this abbrev defined, you would be able to get `find outer otter ' into the buffer by typing f o o SPC.
Abbrevs expand only when Abbrev mode (a minor mode) is enabled.
Disabling Abbrev mode does not cause abbrev definitions to be discarded,
but they do not expand until Abbrev mode is enabled again. The command
M-x abbrev-mode toggles Abbrev mode; with a numeric argument, it
turns Abbrev mode on if the argument is positive, off otherwise.
See section Minor Modes. abbrev-mode
is also a variable; Abbrev mode is
on when the variable is non-nil
. The variable abbrev-mode
automatically becomes local to the current buffer when it is set.
Abbrev definitions can be mode-specific---active only in one major mode. Abbrevs can also have global definitions that are active in all major modes. The same abbrev can have a global definition and various mode-specific definitions for different major modes. A mode-specific definition for the current major mode overrides a global definition.
You can define Abbrevs interactively during an editing session. You can also save lists of abbrev definitions in files and reload them in later sessions. Some users keep extensive lists of abbrevs that they load in every session.
A second kind of abbreviation facility is called the dynamic expansion. Dynamic abbrev expansion happens only when you give an explicit command and the result of the expansion depends only on the current contents of the buffer. See section Dynamic Abbrev Expansion.
add-global-abbrev
).
add-mode-abbrev
).
inverse-add-global-abbrev
).
inverse-add-mode-abbrev
).
The usual way to define an abbrev is to enter the text you want the
abbrev to expand to, position point after it, and type C-x a g
(add-global-abbrev
). This reads the abbrev itself using the
minibuffer, and then defines it as an abbrev for one or more words
before point. Use a numeric argument to say how many words before point
should be taken as the expansion. For example, to define the abbrev
`foo' as in the example above, insert the text `find outer
otter', then type
C-u 3 C-x a g f o o RET.
An argument of zero to C-x a g means to use the contents of the region as the expansion of the abbrev being defined.
The command C-x a l (add-mode-abbrev
) is similar, but
defines a mode-specific abbrev. Mode-specific abbrevs are active only in a
particular major mode. C-x a l defines an abbrev for the major mode
in effect at the time C-x a l is typed. The arguments work the
same way they do for C-x a g.
If the text of an abbrev you want is already in the buffer instead of
the expansion, use command C-x a i g (inverse-add-global-abbrev
)
instead of C-x a g, or use C-x a i l
(inverse-add-mode-abbrev
) instead of C-x a l. These commands
are called "inverse" because they invert the meaning of the argument
found in the buffer and the argument read using the minibuffer.
To change the definition of an abbrev, just add the new definition. You will be asked to confirm if the abbrev has a prior definition. To remove an abbrev definition, give a negative argument to C-x a g or C-x a l. You must choose the command to specify whether to kill a global definition or a mode-specific definition for the current mode, since those two definitions are independent for one abbrev.
M-x kill-all-abbrevs removes all existing abbrev definitions.
An abbrev expands whenever it is in a buffer just before point and you type a self-inserting punctuation character (SPC, comma, etc.). Most often an abbrev is used by inserting the abbrev followed by punctuation.
Abbrev expansion preserves case; thus, `foo' expands into `find
outer otter', `Foo' into `Find outer otter', and `FOO' into
`FIND OUTER OTTER' or `Find Outer Otter' according to the
variable abbrev-all-caps
(a non-nil
value chooses the first
of the two expansions).
Two commands are available to control abbrev expansion:
abbrev-prefix-mark
).
expand-abbrev
).
This is effective even when Abbrev mode is not enabled.
You may wish to expand an abbrev with a prefix attached. For example,
if `cnst' expands into `construction', you may want to use it
to enter `reconstruction'. It does not work to type recnst,
because that is not necessarily a defined abbrev. Instead, you can use
the command M-' (abbrev-prefix-mark
) between the prefix
`re' and the abbrev `cnst'. First, insert `re'. Then
type M-'; this inserts a minus sign in the buffer to indicate that
it has done its work. Then insert the abbrev `cnst'. The buffer
now contains `re-cnst'. Now insert a punctuation character to
expand the abbrev `cnst' into `construction'. The minus sign
is deleted at this point by M-'. The resulting text is the
desired `reconstruction'.
If you actually want the text of the abbrev in the buffer, rather than its expansion, insert the following punctuation with C-q. Thus, foo C-q - leaves `foo-' in the buffer.
If you expand an abbrev by mistake, you can undo the expansion (replace
the expansion by the original abbrev text) with M-x unexpand-abbrev.
You can also use C-_ (undo
) to undo the expansion; but that
will first undo the insertion of the punctuation character.
M-x expand-region-abbrevs searches through the region for defined abbrevs, and offers to replace each one it finds with its expansion. This command is useful if you have typed text using abbrevs but forgot to turn on Abbrev mode first. It may also be useful together with a special set of abbrev definitions for making several global replacements at once. The command is effective even if Abbrev mode is not enabled.
The output from M-x list-abbrevs looks like this:
(lisp-mode-abbrev-table) "dk" 0 "define-key" (global-abbrev-table) "dfn" 0 "definition"
(Some blank lines of no semantic significance, and some other abbrev tables, have been omitted.)
A line containing a name in parentheses is the header for abbrevs in a
particular abbrev table; global-abbrev-table
contains all the global
abbrevs, and the other abbrev tables that are named after major modes
contain the mode-specific abbrevs.
Within each abbrev table, each non-blank line defines one abbrev. The word at the beginning is the abbrev. The number that appears is the number of times the abbrev has been expanded. Emacs keeps track of this to help you see which abbrevs you actually use, in case you want to eliminate those that you don't use often. The string at the end of the line is the expansion.
M-x edit-abbrevs allows you to add, change or kill abbrev
definitions by editing a list of them in an Emacs buffer. The list has
the format described above. The buffer of abbrevs is called
`*Abbrevs*', and is in Edit-Abbrevs mode. This mode redefines the
key C-c C-c to install the abbrev definitions as specified in the
buffer. The edit-abbrevs-redefine
command does this.
Any abbrevs not described in the buffer are eliminated when this is
done.
edit-abbrevs
is actually the same as list-abbrevs
, except
that it selects the buffer `*Abbrevs*' whereas list-abbrevs
merely displays it in another window.
These commands allow you to keep abbrev definitions between editing sessions.
Use M-x write-abbrev-file to save abbrev definitions for use in a later session. The command reads a file name using the minibuffer and writes a description of all current abbrev definitions into the specified file. The text stored in the file looks like the output of M-x list-abbrevs.
M-x read-abbrev-file prompts for a file name using the
minibuffer and reads the specified file, defining abbrevs according to
its contents. M-x quietly-read-abbrev-file is the same but does
not display a message in the echo area; it is actually useful primarily
in the `.emacs' file. If you give an empty argument to either of
these functions, the file name Emacs uses is the value of the variable
abbrev-file-name
, which is by default "~/.abbrev_defs"
.
Emacs offers to save abbrevs automatically if you have changed any of
them, whenever it offers to save all files (for C-x s or C-x
C-c). Set the variable save-abbrevs
to nil
to inhibit
this feature.
The commands M-x insert-abbrevs and M-x define-abbrevs are similar to the previous commands but work on text in an Emacs buffer. M-x insert-abbrevs inserts text into the current buffer before point, describing all current abbrev definitions; M-x define-abbrevs parses the entire current buffer and defines abbrevs accordingly.
The abbrev facility described above operates automatically as you insert text, but all abbrevs must be defined explicitly. By contrast, dynamic abbrevs allow the meanings of abbrevs to be determined automatically from the contents of the buffer, but dynamic abbrev expansion happens only when you request it explicitly.
dabbrev-expand
).
For example, if the buffer contains `does this follow ' and you type f o M-/, the effect is to insert `follow' because that is the last word in the buffer that starts with `fo'. A numeric argument to M-/ says to take the second, third, etc. distinct expansion found looking backward from point. Repeating M-/ searches for an alternative expansion by looking farther back. After the entire buffer before point has been considered, the buffer after point is searched.
Dynamic abbrev expansion is completely independent of Abbrev mode; the expansion of a word with M-/ is completely independent of whether it has a definition as an ordinary abbrev.
If you want to create a picture made out of text characters (for example,
a picture of the division of a register into fields, as a comment in a
program), use the command edit-picture
to enter Picture mode.
In Picture mode, editing is based on the quarter-plane model of text. In this model, the text characters lie studded on an area that stretches infinitely far to the right and downward. The concept of the end of a line does not exist in this model; the most you can say is where the last non-blank character on the line is found.
Of course, Emacs really always considers text as a sequence of characters, and lines really do have ends. But in Picture mode most frequently-used keys are rebound to commands that simulate the quarter-plane model of text. They do this by inserting spaces or by converting tabs to spaces.
Most of the basic editing commands of Emacs are redefined by Picture mode to do essentially the same thing but in a quarter-plane way. In addition, Picture mode defines various keys starting with the C-c prefix to run special picture editing commands.
One of these keys, C-c C-c, is pretty important. Often a picture
is part of a larger file that is usually edited in some other major mode.
M-x edit-picture records the name of the previous major mode.
You can then use the C-c C-c command (picture-mode-exit
) to
restore that mode. C-c C-c also deletes spaces from the ends of
lines, unless you give it a numeric argument.
The commands used in Picture mode all work in other modes (provided the `picture' library is loaded), but are only bound to keys in Picture mode. Note that the descriptions below talk of moving "one column" and so on, but all the picture mode commands handle numeric arguments as their normal equivalents do.
Turning on Picture mode calls the value of the variable
picture-mode-hook
as a function, with no arguments, if that value
exists and is non-nil
.
Most keys do the same thing in Picture mode that they usually do, but do
it in a quarter-plane style. For example, C-f is rebound to run
picture-forward-column
, which moves point one column to
the right, by inserting a space if necessary, so that the actual end of the
line makes no difference. C-b is rebound to run
picture-backward-column
, which always moves point left one column,
converting a tab to multiple spaces if necessary. C-n and C-p
are rebound to run picture-move-down
and picture-move-up
,
which can either insert spaces or convert tabs as necessary to make sure
that point stays in exactly the same column. C-e runs
picture-end-of-line
, which moves to after the last non-blank
character on the line. There was no need to change C-a, as the choice
of screen model does not affect beginnings of lines.
Insertion of text is adapted to the quarter-plane screen model through
the use of Overwrite mode (see section Minor Modes). Self-inserting characters
replace existing text, column by column, rather than pushing existing text
to the right. RET runs picture-newline
, which just moves to
the beginning of the following line so that new text will replace that
line.
Text is erased instead of deleted and killed. DEL
(picture-backward-clear-column
) replaces the preceding character
with a space rather than removing it. C-d
(picture-clear-column
) does the same in a forward direction.
C-k (picture-clear-line
) really kills the contents of lines,
but never removes the newlines from a buffer.
To do actual insertion, you must use special commands. C-o
(picture-open-line
) creates a blank line, but does so after
the current line; it never splits a line. C-M-o, split-line
,
makes sense in Picture mode, so it remains unchanged. LFD
(picture-duplicate-line
) inserts another line
with the same contents below the current line.
To actually delete parts of the picture, use C-w, or with
C-c C-d (which is defined as delete-char
, as C-d is
in other modes), or with one of the picture rectangle commands
(see section Picture Mode Rectangle Commands).
Since "self-inserting" characters just overwrite and move point in Picture mode, there is no essential restriction on how point should be moved. Normally point moves right, but you can specify any of the eight orthogonal or diagonal directions for motion after a "self-inserting" character. This is useful for drawing lines in the buffer.
picture-movement-left
).
picture-movement-right
).
picture-movement-up
).
picture-movement-down
).
picture-movement-nw
).
picture-movement-ne
).
picture-movement-sw
).
picture-movement-se
).
Two motion commands move based on the current Picture insertion
direction. The command C-c C-f (picture-motion
) moves in the
same direction as motion after "insertion" currently does, while C-c
C-b (picture-motion-reverse
) moves in the opposite direction.
picture-tab-search
). With no argument, it moves to a point
underneath the next "interesting" character that follows whitespace in
the previous non-blank line. "Next" here means "appearing at a
horizontal position greater than the one point starts out at". With an
argument, as in C-u M-TAB, the command moves to the next such
interesting character in the current line. M-TAB does not
change the text; it only moves point. "Interesting" characters are
defined by the variable picture-tab-chars
, which contains a string
of characters considered interesting. Its default value is
"!-~"
.
TAB itself runs picture-tab
, which operates based on the
current tab stop settings; it is the Picture mode equivalent of
tab-to-tab-stop
. Without arguments it just moves point, but with
a numeric argument it clears the text that it moves over.
The context-based and tab-stop-based forms of tabbing are brought
together by the command C-c TAB (picture-set-tab-stops
.)
This command sets the tab stops to the positions which M-TAB
would consider significant in the current line. If you use this command
with TAB, you can get the effect of context-based tabbing. But
M-TAB is more convenient in the cases where it is sufficient.
Picture mode defines commands for working on rectangular pieces of the text in ways that fit with the quarter-plane model. The standard rectangle commands may also be useful (see section Rectangles).
picture-clear-rectangle
). With
argument, kill it.
picture-clear-rectangle-to-register
).
picture-yank-rectangle
). With argument,
insert instead.
picture-yank-rectangle-from-register
).
The picture rectangle commands C-c C-k
(picture-clear-rectangle
) and C-c C-w
(picture-clear-rectangle-to-register
) differ from the standard
rectangle commands in that they normally clear the rectangle instead of
deleting it; this is analogous with the way C-d is changed in Picture
mode.
However, deletion of rectangles can be useful in Picture mode, so these commands delete the rectangle if given a numeric argument.
The Picture mode commands for yanking rectangles differ from the standard
ones in overwriting instead of inserting. This is the same way that
Picture mode insertion of other text is different from other modes.
C-c C-y (picture-yank-rectangle
) inserts (by overwriting) the
rectangle that was most recently killed, while C-c C-x
(picture-yank-rectangle-from-register
) does for the
rectangle found in a specified register.
Since most region commands in Picture mode operate on rectangles, when you select a region of text with the mouse in Picture mode, it is highlighted as a rectangle.
To send a message in Emacs, start by typing the command (C-x m) to select and initialize the `*mail*' buffer. You can then edit the text and headers of the message in the mail buffer, and type the command (C-c C-c) to send the message.
mail
).
mail-other-window
).
mail-send-and-exit
).
The command C-x m (mail
) selects a buffer named
`*mail*' and initializes it with the skeleton of an outgoing message.
C-x 4 m (mail-other-window
) selects the `*mail*' buffer
in a different window, leaving the previous current buffer visible.
Because the buffer for mail composition is an ordinary Emacs buffer, you can switch to other buffers while in the middle of composing mail, and switch back later (or never). If you use the C-x m command again when you have been composing another message but have not sent it, a new mail buffer will be created; in this way, you can compose multiple messages at once. You can switch back to and complete an unsent message by using the normal buffer selection mechanisms.
C-u C-x m is another way to switch back to a message in progress: it will search for an existing, unsent mail message buffer and select it.
In addition to the text or contents, a message has header fields, which say who sent it, when, to whom, why, and so on. Some header fields, such as the date and sender, are created automatically after the message is sent. Others, such as the recipient names, must be specified by you in order to send the message properly.
Mail mode provides a few commands to help you edit some header fields, and some are preinitialized in the buffer automatically at times. You can insert or edit any header fields using ordinary editing commands.
The line in the buffer that says:
--text follows this line--
is a special delimiter that separates the headers you have specified from
the text. Whatever follows this line is the text of the message; the
headers precede it. The delimiter line itself does not appear in the
message actually sent. The text used for the delimiter line is controlled
by the variable mail-header-separator
.
Here is an example of what the headers and text in the `*mail*' buffer might look like.
To: rms@mc CC: mly@mc, rg@oz Subject: The Emacs Manual --Text follows this line-- Please ignore this message.
There are several header fields you can use in the `*mail*' buffer. Each header field starts with a field name at the beginning of a line, terminated by a colon. It does not matter whether you use upper or lower case in the field name. After the colon and optional whitespace comes the contents of the field.
The `To', `CC', `BCC' and `FCC' fields can appear any number of times, to specify many places to send the message.
The `To', `CC', and `BCC', fields can have continuation lines. All the lines starting with whitespace, following the line on which the field starts, are considered part of the field. For example,
To: foo@here, this@there, me@gnu.cambridge.mass.usa.earth.spiral3281
If you have a `~/.mailrc' file, Emacs scans it for mail aliases the
first time you try to send mail in an Emacs session. Emacs expands
aliases found in the `To', `CC', and `BCC' fields where
appropriate. You can set the variable mail-abbrev-mailrc-file
to
the name of the file with mail aliases. If nil
, `~/.mailrc'
is used.
Your `.mailrc' file ensures that word-abbrevs are defined for each of your mail aliases when point is in a `To', `CC', `BCC', or `From' field. The aliases are defined in your `.mailrc' file or in a file specified by the MAILRC environment variable if it exists. Your mail aliases expand any time you type a word-delimiter at the end of an abbreviation.
In this version of Emacs, what you see is what you get: in contrast to some other versions, no abbreviations are expanded after you have sent the mail. This means you don't suffer the annoyance of having the system do things behind your back -- if the system rewrites an address you typed, you know it immediately, instead of after the mail has been sent and it's too late to do anything about it. For example, you will never again be in trouble because you forgot to delete an old alias from your `.mailrc' and a new local user is given a userid which conflicts with one of your aliases.
Your mail alias abbrevs are in effect only when point is in an
appropriate header field. The mail aliases will not expand in the body
of the message, or in other header fields. The default mode-specific
abbrev table mail-mode-abbrev-table
is used instead if defined.
That means if you have been using mail-mode specific abbrevs, this code
will not adversely affect you. You can control which header fields the
abbrevs are used in by changing the variable mail-abbrev-mode-regexp
.
If auto-fill mode is on, abbrevs wrap at commas instead of at word boundaries, and header continuation lines will be properly indented.
You can also insert a mail alias with mail-interactive-insert-alias
.
This function, which is bound to C-c C-a, prompts you for an alias
(with completion) and inserts its expansion at point.
In this version of Emacs, it is possible to have lines like the following in your `.mailrc' file:
alias someone "John Doe <doe@quux.com>"
That is, if you want an address to have embedded spaces, simply surround it with double-quotes. The quotes are necessary because the format of the `.mailrc' file uses spaces as address delimiters.
Aliases in the `.mailrc' file may be nested. For example, assume you define aliases like:
alias group1 fred ethel alias group2 larry curly moe alias everybody group1 group2
When you now type `everybody' on the `To' line, it will expand to:
fred, ethyl, larry, curly, moe
Aliases may contain forward references; the alias of `everybody' in the example above can preceed the aliases of `group1' and `group2'.
In this version of Emacs, you can use the source
`.mailrc' command
for reading aliases from some other file as well.
Aliases may contain hyphens, as in "alias foo-bar foo@bar"
, even
though word-abbrevs normally cannot contain hyphens.
To read in the contents of another `.mailrc'-type file from Emacs, use the
command M-x merge-mail-aliases
. The rebuild-mail-aliases
command is similar, but deletes existing aliases first.
If you want multiple addresses separated by a string other than `,'
(a comma), then set the variable mail-alias-seperator-string
to
it. This has to be a comma bracketed by whitespace if you want any kind
of reasonable behavior.
If the variable mail-archive-file-name
is non-nil
, it
should be a string naming a file. Each time you start to edit a message
to send, an `FCC' field is entered for that file. Unless you
remove the `FCC' field, every message is written into that
file when it is sent.
The major mode used in the `*mail*' buffer is Mail mode. Mail mode is similar to Text mode, but several commands are provided on the C-c prefix. These commands all deal specifically with editing or sending the message.
mail-send
).
mail-send-and-exit
).
mail-to
).
mail-subject
).
mail-cc
).
mail-signature
).
mail-yank-original
).
This command does nothing unless your command to start sending a
message was issued with Rmail.
mail-fill-yanked-message
).
There are two ways to send a message. C-c C-c
(mail-send-and-exit
) is the usual way to send the message. It
sends the message and then deletes the window (if there is another
window) or switches to another buffer. It puts the `*mail*' buffer
at the lowest priority for automatic reselection, since you are finished
with using it. C-c C-s (mail-send
) sends the
message and marks the `*mail*' buffer unmodified, but leaves that
buffer selected so that you can modify the message (perhaps with new
recipients) and send it again.
Mail mode provides some other special commands that are useful for
editing the headers and text of the message before you send it. There are
three commands defined to move point to particular header fields, all based
on the prefix C-c C-f (`C-f' is for "field"). They are
C-c C-f C-t (mail-to
) to move to the `To' field, C-c
C-f C-s (mail-subject
) for the `Subject' field, and C-c
C-f C-c (mail-cc
) for the `CC' field. These fields have
special motion commands because they are edited most frequently.
C-c C-w (mail-signature
) adds a standard piece of text at
the end of the message to say more about who you are. The text comes
from the file `.signature' in your home directory.
When you use an Rmail command to send mail from the Rmail mail reader,
you can use C-c C-y mail-yank-original
inside the
`*mail*' buffer to insert the
text of the message you are replying to. Normally Rmail indents each line
of that message four spaces and eliminates most header fields. A
numeric argument specifies the number of spaces to indent. An argument
of just C-u says not to indent at all and not to eliminate
anything. C-c C-y always uses the current message from the
`RMAIL' buffer, so you can insert several old messages by selecting
one in `RMAIL', switching to `*mail*' and yanking it, then
switching back to `RMAIL' to select another.
After using C-c C-y, you can use the command C-c C-q
(mail-fill-yanked-message
) to fill the paragraphs of the yanked
old message or messages. One use of C-c C-q fills all such
paragraphs, each one separately.
Clicking the right mouse button in a mail buffer pops up a menu of the above commands, for easy access.
Turning on Mail mode (which C-x m does automatically) calls the
value of text-mode-hook
, if it is not void or nil
, and
then calls the value of mail-mode-hook
if that is not void or
nil
.
Rmail is an Emacs subsystem for reading and disposing of mail that you receive. Rmail stores mail messages in files called Rmail files. You read the messages in an Rmail file in a special major mode, Rmail mode, which redefines most letters to run commands for managing mail. To enter Rmail, type M-x rmail. This reads your primary mail file, merges new mail in from your inboxes, displays the first new message, and lets you begin reading.
Using Rmail in the simplest fashion, you have one Rmail file, `~/RMAIL', in which all of your mail is saved. It is called your primary mail file. You can also copy messages into other Rmail files and then edit those files with Rmail.
Rmail displays only one message at a time. It is called the current message. Rmail mode's special commands can move to another message, delete the message, copy the message into another file, or send a reply.
Within the Rmail file, messages are arranged sequentially in order of receipt. They are also assigned consecutive integers as their message numbers. The number of the current message is displayed in Rmail's mode line, followed by the total number of messages in the file. You can move to a message by specifying its message number using the j key (see section Moving Among Messages).
Following the usual conventions of Emacs, changes in an Rmail file become
permanent only when the file is saved. You can do this with s
(rmail-save
), which also expunges deleted messages from the file
first (see section Deleting Messages). To save the file without expunging, use
C-x C-s. Rmail saves the Rmail file automatically when moving new
mail from an inbox file (see section Rmail Files and Inboxes).
You can exit Rmail with q (rmail-quit
); this expunges and
saves the Rmail file and then switches to another buffer. However, there is
no need to `exit' formally. If you switch from Rmail to editing in
other buffers, and never happen to switch back, you have exited. Just
make sure to save the Rmail file eventually (like any other file you
have changed). C-x s is a good enough way to do this
(see section Saving Files).
When Rmail displays a message that does not fit on the screen, you have to scroll through it. You could use C-v, M-v, and M-<, but scrolling is so frequent in Rmail that it deserves to be easier to type.
scroll-up
).
scroll-down
).
rmail-beginning-of-message
).
Since the most common thing to do while reading a message is to scroll
through it by screenfuls, Rmail makes SPC and DEL synonyms of
C-v (scroll-up
) and M-v (scroll-down
)
The command . (rmail-beginning-of-message
) scrolls back to the
beginning of a selected message. This is not quite the same as M-<:
first, it does not set the mark; secondly, it resets the buffer
boundaries to the current message if you have changed them.
The most basic thing to do with a message is to read it. The way to do this in Rmail is to make the message current. You can make any message current, given its message number, by using the j command, but people most often move sequentially through the file, since this is the order of receipt of messages. When you enter Rmail, you are positioned at the first new message (new messages are those received after you last used Rmail), or at the last message if there are no new messages this time. Move forward to see other new messages if there are any; move backward to re-examine old messages.
rmail-next-undeleted-message
).
rmail-previous-undeleted-message
).
rmail-next-message
).
rmail-previous-message
).
rmail-show-message
).
rmail-last-message
).
rmail-search
). If regexp is empty, the last regexp used is
used again.
To move among messages in Rmail, you can use n and p.
These keys move through the messages sequentially but skip over deleted
messages, which is usually what you want to do. Their command
definitions are named rmail-next-undeleted-message
and
rmail-previous-undeleted-message
. If you do not want to skip
deleted messages--for example, if you want to move to a message to
undelete it--use the variants M-n (rmail-next-message
) and
M-p (rmail-previous-message
). A numeric argument to any of
these commands serves as a repeat count.
In Rmail, you can specify a numeric argument by just typing the digits. It is not necessary to type C-u first.
The M-s (rmail-search
) command is Rmail's version of
search. The usual incremental search command C-s works in Rmail,
but searches only within the current message. The purpose of M-s
is to search for another message. It reads a regular expression
(see section Syntax of Regular Expressions) non-incrementally, then starts searching at the
beginning of the following message for a match. The message containing
the match is selected.
To search backward in the file for another message, give M-s a negative argument. In Rmail you can do this with - M-s.
It is also possible to search for a message based on labels. See section Labels.
To move to a message specified by absolute message number, use j
(rmail-show-message
) with the message number as argument. With
no argument, j selects the first message. >
(rmail-last-message
) selects the last message.
When you no longer need to keep a message, you can delete it. This flags it as ignorable, and some Rmail commands will pretend it is no longer present, but it still has its place in the Rmail file and still has its message number.
Expunging the Rmail file actually removes the deleted messages. The remaining messages are renumbered consecutively. Expunging is the only action that changes the message number of any message, except for undigestifying (see section Digest Messages).
rmail-delete-forward
).
rmail-delete-backward
).
rmail-undelete-previous-message
).
rmail-expunge
). These two
commands are synonyms.
There are two Rmail commands for deleting messages. Both delete the
current message and select another message. d
(rmail-delete-forward
) moves to the following message, skipping
messages already deleted, while C-d (rmail-delete-backward
)
moves to the previous non-deleted message. If there is no non-deleted
message to move to in the specified direction, the message that was just
deleted remains current.
To make all deleted messages disappear from the Rmail file, type
e (rmail-expunge
). Until you do this, you can still
undelete the deleted messages.
To undelete, type
u (rmail-undelete-previous-message
), which cancels the
effect of a d command (usually). It undeletes the current message
if the current message is deleted. Otherwise it moves backward to previous
messages until a deleted message is found, and undeletes that message.
You can usually undo a d with a u because the u moves back to and undeletes the message that the d deleted. This does not work when the d skips a few already-deleted messages that follow the message being deleted; in that case the u command undeletes the last of the messages that were skipped. There is no clean way to avoid this problem. However, by repeating the u command, you can eventually get back to the message you intended to undelete. You can also reach that message with M-p commands and then type u.
A deleted message has the `deleted' attribute, and as a result `deleted' appears in the mode line when the current message is deleted. In fact, deleting or undeleting a message is nothing more than adding or removing this attribute. See section Labels.
Unix places your incoming mail in a file called your inbox. When you start up Rmail, it copies the new messages from your inbox into your primary mail file, an Rmail file which also contains other messages saved from previous Rmail sessions. In this file, you actually read the mail with Rmail. The operation is called getting new mail. You can repeat it at any time using the g key in Rmail. The inbox file name is `/usr/spool/mail/username' in Berkeley Unix, `/usr/mail/username' in system V.
There are two reason for having separate Rmail files and inboxes.
When getting new mail, Rmail first copies the new mail from the inbox file to the Rmail file and saves the Rmail file. It then deletes the inbox file. This way a system crash may cause duplication of mail between the inbox and the Rmail file, but it cannot lose mail.
Copying mail from an inbox in the system's mailer directory actually puts it in an intermediate file, `~/.newmail'. This is because the interlocking is done by a C program that copies to another file. `~/.newmail' is deleted after mail merging is successful. If there is a crash at the wrong time, this file will continue to exist and will be used as an inbox the next time you get new mail.
Rmail operates by default on your primary mail file, which is named `~/RMAIL' and which receives your incoming mail from your system inbox file. You can also have other mail files and edit them with Rmail. These files can receive mail through their own inboxes, or you can move messages into them by explicit command in Rmail (see section Copying Messages Out to Files).
rmail-input
).
rmail-get-new-mail
).
To run Rmail on a file other than your primary mail file, you may use
the i (rmail-input
) command in Rmail. This visits the
file, puts it in Rmail mode, and then gets new mail from the file's
inboxes if any. You can also use M-x rmail-input even when not in
Rmail.
The file you read with i does not have to be in Rmail file format. It could also be Unix mail format, mmdf format, or it could be a mixture of all three, as long as each message has one of the three formats. Rmail recognizes all three and converts all the messages to proper Rmail format before showing you the file.
Each Rmail file can contain a list of inbox file names; you can specify this list with M-x set-rmail-inbox-list RET files RET. The argument can contain any number of file names, separated by commas. It can also be empty, which specifies that this file should have no inboxes. Once a list of inboxes is specified, the Rmail file remembers it permanently until it is explicitly changed.
If an Rmail file has inboxes, new mail is merged in from the inboxes
when you bring the Rmail file into Rmail, and when you use the g
(rmail-get-new-mail
) command. If the Rmail file
specifies no inboxes, then no new mail is merged in at these times. A
special exception is made for your primary mail file: Rmail uses the
standard system inbox for it if it does not specify an inbox.
To merge mail from a file that is not the usual inbox, give the g key a numeric argument, as in C-u g. Rmail prompts you for a file name and merges mail from that file. The inbox file is not deleted or changed at all when you use g with an argument, so this is a general way of merging one file of messages into another.
rmail-output-to-rmail-file
).
rmail-output
).
If an Rmail file has no inboxes, use explicit o commands to write Rmail files.
o (rmail-output-to-rmail-file
) appends the current
message in Rmail format to the end of a specified file. This is the
best command to use to move messages between Rmail files. If you are
currently visiting the other Rmail file, copying is done into the other
file's Emacs buffer instead. You should eventually save the buffer on
disk.
The C-o (rmail-output
) command in Rmail appends a copy of
the current message to a specified file, in Unix mail file format. This
is useful for moving messages into files to be read by other mail
processors that do not understand Rmail format.
Copying a message with o or C-o gives the original copy of the message the `filed' attribute. `filed' appears in the mode line when such a message is current.
Normally you should use only o to output messages to other Rmail files, never C-o. But it is also safe if you always use C-o, never o. When a file is visited in Rmail, the last message is checked, and if it is in Unix format, the entire file is scanned and all Unix-format messages are converted to Rmail format. (The reason for checking the last message is that scanning the file is slow and most Rmail files have only Rmail format messages.) If you use C-o consistently, the last message is guaranteed to be in Unix format, so Rmail will convert all messages properly.
When you and other users want to append mail to the same file, you probably always want to use C-o instead of o. Other mail processors may not know Rmail format but will know Unix format.
In any case, always use o to add to an Rmail file that is being visited in Rmail. Adding messages with C-o to the actual disk file will trigger a "simultaneous editing" warning when you ask to save the Emacs buffer, and the messages will be lost if you do save.
Each message can have various labels assigned to it as a means of classification. A label has a name; different names mean different labels. Any given label is either present or absent on a particular message. A few label names have standard meanings and are given to messages automatically by Rmail when appropriate; these special labels are called attributes. All other labels are assigned by the user.
rmail-add-label
).
rmail-kill-label
).
rmail-next-labeled-message
).
rmail-previous-labeled-message
).
rmail-summary-by-labels
).
Specifying an empty string for one these commands means to use the last label specified for any of these commands.
The a (rmail-add-label
) and k
(rmail-kill-label
) commands allow you to assign or remove any
label on the current message. If the label argument is empty, it
means to assign or remove the label most recently assigned or
removed.
Once you have given messages labels to classify them as you wish, there are two ways to use the labels: in moving and in summaries.
The command C-M-n labels RET
(rmail-next-labeled-message
) moves to the next message that has one
of the labels labels. labels is one or more label names,
separated by commas. C-M-p (rmail-previous-labeled-message
)
is similar, but moves backwards to previous messages. A preceding numeric
argument to either one serves as a repeat count.
The command C-M-l labels RET
(rmail-summary-by-labels
) displays a summary containing only the
messages that have at least one of a specified set of messages. The
argument labels is one or more label names, separated by commas.
See section Summaries, for information on summaries.
If the labels argument to C-M-n, C-M-p or C-M-l is empty, it means to use the last set of labels specified for any of these commands.
Some labels such as `deleted' and `filed' have built-in meanings and are assigned to or removed from messages automatically at appropriate times; these labels are called attributes. Here is a list of Rmail attributes:
rmail-reply
). See section Sending Replies.
rmail-forward
). See section Sending Replies.
All other labels are assigned or removed only by the user, and it is up to the user to decide what they mean.
A summary is a buffer Rmail creates and displays to give you an overview of the mail in an Rmail file. It contains one line per message; each line shows the message number, the sender, the labels, and the subject. When you select the summary buffer, you can use a number of commands to select messages by moving in the summary buffer, or to delete or undelete messages.
A summary buffer applies to a single Rmail file only; if you are editing multiple Rmail files, they have separate summary buffers. The summary buffer name is generated by appending `-summary' to the Rmail buffer's name. Only one summary buffer is displayed at a time unless you make several windows and select the summary buffers by hand.
Here are the commands to create a summary for the current Rmail file. Summaries do not update automatically; to make an updated summary, you must use one of the commands again.
rmail-summary
).
rmail-summary-by-labels
).
rmail-summary-by-recipients
).
The h or C-M-h (rmail-summary
) command fills the
summary buffer for the current Rmail file with a summary of all the
messages in the file. It then displays and selects the summary buffer
in another window.
The l or C-M-l labels RET
(rmail-summary-by-labels
) makes
a partial summary mentioning only the messages that have one or more of the
labels labels. labels should contain label names separated by
commas.
C-M-r rcpts RET (rmail-summary-by-recipients
)
makes a partial summary mentioning only the messages that have one or more
of the recipients rcpts. rcpts should contain mailing
addresses separated by commas.
Note that there is only one summary buffer for any Rmail file; making one kind of summary discards any previously made summary.
Summary buffers are given the major mode Rmail Summary mode, which provides the following special commands:
rmail-summary-goto-msg
).
rmail-summary-next-all
).
rmail-summary-
previous-all
).
rmail-summary-next-msg
).
rmail-summary-previous-msg
).
rmail-summary-delete-forward
).
rmail-summary-undelete
).
rmail-summary-scroll-msg-up
).
rmail-summary-scroll-msg-down
).
rmail-summary-exit
).
rmail-summary-quit
).
The keys C-n(rmail-summary-next-all
) and C-p
(rmail-summary-previous-all
) are modified in Rmail Summary mode.
In addition to moving point in the summary buffer, they also cause the
line's message to become current in the associated Rmail buffer. That
buffer is also made visible in another window if it is not currently
visible.
n and p are similar to C-n and C-p, but skip lines that say `message deleted'. They are like the n and p keys of Rmail itself. Note, however, that in a partial summary these commands move only among the message listed in the summary.
The other Emacs cursor motion commands are not changed in Rmail
Summary mode, so it is easy to get the point on a line whose message is
not selected in Rmail. This can also happen if you switch to the Rmail
window and switch messages there. To get the Rmail buffer back in sync
with the summary, use the j (rmail-summary-goto-msg
)
command, which selects the message of the current summary line in Rmail.
Deletion and undeletion can also be done from the summary buffer.
They always work based on where point is located in the summary buffer,
ignoring which message is selected in Rmail. d
(rmail-summary-delete-forward
) deletes the current line's
message, then moves to the next line whose message is not deleted and
selects that message. The inverse is u
(rmail-summary-undelete
), which moves back (if necessary) to a
line whose message is deleted, undeletes that message, and selects it in
Rmail.
When moving through messages with the summary buffer, it is convenient
to be able to scroll the message while remaining in the summary window.
The commands SPC (rmail-summary-scroll-msg-up
) and
DEL (rmail-summary-scroll-msg-down
) do this. They scroll
the message just as they do when the Rmail buffer is selected.
When you are finished using the summary, type x
(rmail-summary-exit
) to kill the summary buffer's window.
You can also exit Rmail while in the summary. q
(rmail-summary-quit
) kills the summary window, then saves the
Rmail file and switches to another buffer.
Rmail has several commands that use Mail mode to send mail. See section Sending Mail, for information on using Mail mode. Only the special commands of Rmail for entering Mail mode are documented here. Note that the usual keys for sending mail, C-x m and C-x 4 m, are available in Rmail mode and work just as they usually do.
rmail-mail
).
rmail-continue
).
rmail-reply
).
rmail-forward
).
To reply to a the message you are reading in Rmail, type r
(rmail-reply
). This displays the `*mail*' buffer in another
window, much like C-x 4 m, but pre-initializes the `Subject',
`To', `CC', and `In-reply-to' header fields based on the
message you reply to. The `To' field is given the sender of
that message, and the `CC' gets all the recipients of that message.
Recipients that match elements of the list
rmail-dont-reply-to
are omitted; by default, this list contains
your own mailing address.
Once you have initialized the `*mail*' buffer this way, sending the mail goes as usual (see section Sending Mail). You can edit the presupplied header fields if they are not what you want.
One additional Mail mode command is available when you invoke mail
from Rmail: C-c C-y (mail-yank-original
) inserts into the
outgoing message a copy of the current Rmail message. Normally this is
the message you are replying to, but you can also switch to the Rmail
buffer, select a different message, switch back, and yank the new current
message. Normally the yanked message is indented four spaces and has
most header fields deleted from it; an argument to C-c C-y
specifies the amount to indent. C-u C-c C-y neither indents
the message nor deletes any header fields.
Another frequent reason to send mail in Rmail is to forward the current
message to other users. f (rmail-forward
) makes this easy by
preinitializing the `*mail*' buffer with the current message as the
text and a subject designating a forwarded message. All you have to do is
fill in the recipients and send.
You can use the m (rmail-mail
) command to start editing an
outgoing message that is not a reply. It leaves the header fields empty.
Its only difference from C-x 4 m is that it makes the Rmail buffer
accessible for C-c y, just as r does. Thus m can be
used to reply to or forward a message; it can do anything r or f
can do.
The c (rmail-continue
) command resumes editing the
`*mail*' buffer, to finish editing an outgoing message you were
already composing, or to alter a message you have sent.
Rmail mode provides a few special commands for moving within and editing the current message. In addition, the usual Emacs commands are available (except for a few, such as C-M-n and C-M-h, that are redefined by Rmail for other purposes). However, the Rmail buffer is normally read-only, and to alter it you must use the Rmail command w described below.
rmail-toggle-headers
).
rmail-edit-current-message
).
Rmail reformats the header of each message before displaying it.
Normally this involves deleting most header fields, on the grounds that
they are not interesting. The variable rmail-ignored-headers
should contain a regexp that matches the header fields to discard in
this way. The original headers are saved permanently; to see what they
look like, use the t (rmail-toggle-headers
) command. This
discards the reformatted headers of the current message and displays it
with the original headers. Repeating t reformats the message
again. Selecting the message again also reformats.
The Rmail buffer is normally read-only, and most of the characters you
would type to modify it (including most letters) are redefined as Rmail
commands. This is usually not a problem since people rarely want to
change the text of a message. When you do want to do this, type w
(rmail-edit-current-message
), which changes from Rmail mode to
Rmail Edit mode, another major mode which is nearly the same as Text
mode. The mode line indicates this change.
In Rmail Edit mode, letters insert themselves as usual and the Rmail commands are not available. When you are finished editing the message and are ready to go back to Rmail, type C-c C-c, which switches back to Rmail mode. To return to Rmail mode but cancel all the editing you have done, type C-c C-].
Entering Rmail Edit mode calls the value of the variable
text-mode-hook
with no arguments, if that value exists and is not
nil
. It then does the same with the variable
rmail-edit-mode-hook
and finally adds the attribute `edited'
to the message.
A digest message is a message which exists to contain and carry several other messages. Digests are used on moderated mailing lists. All messages that arrive for the list during a period of time, such as one day, are put inside a single digest which is then sent to the subscribers. Transmitting the single digest uses much less computer time than transmitting the individual messages even though the total size is the same, because the per-message overhead in network mail transmission is considerable.
When you receive a digest message, the most convenient way to read it is to undigestify it: to turn it back into many individual messages. You can then read and delete the individual messages as it suits you.
To undigestify a message, select it and then type M-x undigestify-rmail-message. This copies each submessage as a separate Rmail message and inserts them all following the digest. The digest message itself is flagged as deleted.
This chapter contains several brief topics that do not fit anywhere else.
Emacs provides several commands for sorting text in a buffer. All operate on the contents of the region (the text between point and the mark). They divide the text of the region into many sort records, identify a sort key for each record, and then reorder the records using the order determined by the sort keys. The records are ordered so that their keys are in alphabetical order, or, for numerical sorting, in numerical order. In alphabetical sorting, all upper-case letters `A' through `Z' come before lower-case `a', in accordance with the ASCII character sequence.
The sort commands differ in how they divide the text into sort records and in which part of each record they use as the sort key. Most of the commands make each line a separate sort record, but some commands use paragraphs or pages as sort records. Most of the sort commands use each entire sort record as its own sort key, but some use only a portion of the record as the sort key.
For example, if the buffer contains:
On systems where clash detection (locking of files being edited) is implemented, Emacs also checks the first time you modify a buffer whether the file has changed on disk since it was last visited or saved. If it has, you are asked to confirm that you want to change the buffer.
then if you apply M-x sort-lines to the entire buffer you get:
On systems where clash detection (locking of files being edited) is implemented, Emacs also checks the first time you modify a buffer saved. If it has, you are asked to confirm that you want to change the buffer. whether the file has changed on disk since it was last visited or
where the upper case `O' comes before all lower case letters. If you apply instead C-u 2 M-x sort-fields you get:
implemented, Emacs also checks the first time you modify a buffer saved. If it has, you are asked to confirm that you want to change the buffer. On systems where clash detection (locking of files being edited) is whether the file has changed on disk since it was last visited or
where the sort keys were `Emacs', `If', `buffer', `systems', and `the'.
M-x sort-columns requires more explanation. You specify the columns by putting point at one of the columns and the mark at the other column. Because this means you cannot put point or the mark at the beginning of the first line to sort, this command uses an unusual definition of `region': all of the line point is in is considered part of the region, and so is all of the line the mark is in.
For example, to sort a table by information found in columns 10 to 15, you could put the mark on column 10 in the first line of the table, and point on column 15 in the last line of the table, and then use this command. Or you could put the mark on column 15 in the first line and point on column 10 in the last line.
This can be thought of as sorting the rectangle specified by point and the mark, except that the text on each line to the left or right of the rectangle moves along with the text inside the rectangle. See section Rectangles.
Emacs has commands for passing single command lines to inferior shell processes; it can also run a shell interactively with input and output to an Emacs buffer `*shell*'.
shell-command
).
shell-command-on-region
).
M-! (shell-command
) reads a line of text using the
minibuffer and creates an inferior shell to execute the line as a command.
Standard input from the command comes from the null device. If the shell
command produces any output, the output goes to an Emacs buffer named
`*Shell Command Output*', which is displayed in another window but not
selected. A numeric argument, as in M-1 M-!, directs this command to
insert any output into the current buffer. In that case, point is left
before the output and the mark is set after the output.
M-| (shell-command-on-region
) is like M-! but passes
the contents of the region as input to the shell command, instead of no
input. If a numeric argument is used to direct output to the current
buffer, then the old region is deleted first and the output replaces it as
the contents of the region.
Both M-! and M-| use shell-file-name
to specify the
shell to use. This variable is initialized based on your SHELL
environment variable when you start Emacs. If the file name does not
specify a directory, the directories in the list exec-path
are
searched; this list is initialized based on the PATH
environment
variable when you start Emacs. You can override either or both of these
default initializations in your `.emacs' file.
When you use M-! and M-|, Emacs has to wait until the shell command completes. You can quit with C-g; that terminates the shell command.
To run a subshell interactively with its typescript in an Emacs buffer, use M-x shell. This creates (or reuses) a buffer named `*shell*' and runs a subshell with input coming from and output going to that buffer. That is to say, any "terminal output" from the subshell will go into the buffer, advancing point, and any "terminal input" for the subshell comes from text in the buffer. To give input to the subshell, go to the end of the buffer and type the input, terminated by RET.
Emacs does not wait for the subshell to do anything. You can switch windows or buffers and edit them while the shell is waiting, or while it is running a command. Output from the subshell waits until Emacs has time to process it; this happens whenever Emacs is waiting for keyboard input or for time to elapse.
To get multiple subshells, change the name of buffer `*shell*' to something different by using M-x rename-buffer. The next use of M-x shell creates a new buffer `*shell*' with its own subshell. By renaming this buffer as well you can create a third one, and so on. All the subshells run independently and in parallel.
The file name used to load the subshell is the value of the variable
explicit-shell-file-name
, if that is non-nil
. Otherwise, the
environment variable ESHELL
is used, or the environment variable
SHELL
if there is no ESHELL
. If the file name specified
is relative, the directories in the list exec-path
are searched
(see section Single Shell Commands).
As soon as the subshell is started, it is sent as input the contents of
the file `~/.emacs_shellname', if that file exists, where
shellname is the name of the file that the shell was loaded from.
For example, if you use csh
, the file sent to it is
`~/.emacs_csh'.
cd
, pushd
, and popd
commands given to the inferior
shell are watched by Emacs so it can keep the `*shell*' buffer's
default directory the same as the shell's working directory. These
commands are recognized syntactically by examining lines of input that are
sent. If you use aliases for these commands, you can tell Emacs to
recognize them also. For example, if the value of the variable
shell-pushd-regexp
matches the beginning of a shell command line,
that line is regarded as a pushd
command. Change this variable when
you add aliases for `pushd'. Likewise, shell-popd-regexp
and
shell-cd-regexp
are used to recognize commands with the meaning of
`popd' and `cd'.
M-x shell-resync-dirs queries the shell and resynchronizes Emacs' idea of what the current directory stack is. M-x shell-dirtrack-toggle turns directory tracking on and off.
Emacs keeps a history of the most recent commands you have typed in the
`*shell*' buffer. If you are at the beginning of a shell command
line and type M-p, the previous shell input is inserted into the
buffer before point. Immediately typing M-p again deletes that
input and inserts the one before it. By repeating M-p you can
move backward through your commands until you find one you want to
repeat. You may then edit the command before typing RET if you
wish. M-n moves forward through the command history, in case you
moved backward past the one you wanted while using M-p. If you
type the first few characters of a previous command and then type
M-p, the most recent shell input starting with those characters is
inserted. This can be very convenient when you are repeating a sequence
of shell commands. The variable input-ring-size
controls how
many commands are saved in your input history. The default is 30.
The shell buffer uses Shell mode, which defines several special keys attached to the C-c prefix. They are chosen to resemble the usual editing and job control characters present in shells that are not under Emacs, except that you must type C-c first. Here is a list of the special key bindings of Shell mode:
send-shell-input
). When a line is copied, any
text at the beginning of the line that matches the variable
shell-prompt-pattern
is left out; this variable's value should be a
regexp string that matches the prompts that you use in your subshell.
shell-send-eof
).
comint-delchar-or-maybe-eof
).
kill-shell-input
).
backward-kill-word
).
interrupt-shell-subjob
).
stop-shell-subjob
).
quit-shell-subjob
).
kill-output-from-shell
).
show-output-from-shell
).
copy-last-shell-input
). No final newline
is inserted, and the input copied is not resubmitted until you type
RET.
comint-previous-input
).
comint-next-input
).
comint-dynamic-complete
).
Narrowing means focusing in on some portion of the buffer, making the rest temporarily invisible and inaccessible. Cancelling the narrowing and making the entire buffer once again visible is called widening. The amount of narrowing in effect in a buffer at any time is called the buffer's restriction.
narrow-to-region
).
widen
).
Narrowing sometimes makes it easier to concentrate on a single subroutine or paragraph by eliminating clutter. It can also be used to restrict the range of operation of a replace command or repeating keyboard macro. The word `Narrow' appears in the mode line whenever narrowing is in effect. When you have narrowed to a part of the buffer, that part appears to be all there is. You can't see the rest, can't move into it (motion commands won't go outside the visible part), and can't change it in any way. However, the invisible text is not gone; if you save the file, it will be saved.
The primary narrowing command is C-x n n (narrow-to-region
).
It sets the current buffer's restrictions so that the text in the current
region remains visible but all text before the region or after the region
is invisible. Point and mark do not change.
Because narrowing can easily confuse users who do not understand it,
narrow-to-region
is normally a disabled command. Attempting to use
this command asks for confirmation and gives you the option of enabling it;
once you enable the command, confirmation will no longer be required. See section Disabling Commands.
To undo narrowing, use C-x n w (widen
). This makes all
text in the buffer accessible again.
Use the C-x = command to get information on what part of the buffer you narrowed down. See section Cursor Position Information.
The Emacs commands for making hardcopy derive their names from the Unix commands `print' and `lpr'.
print-buffer
, but prints only the current region.
lpr-buffer
, but prints only the current region.
All the hardcopy commands pass extra switches to the lpr
program
based on the value of the variable lpr-switches
. Its value should
be a list of strings, each string a switch starting with `-'. For
example, the value could be ("-Pfoo")
to print on printer
`foo'.
A recursive edit is a situation in which you are using Emacs
commands to perform arbitrary editing while in the middle of another
Emacs command. For example, when you type C-r inside a
query-replace
, you enter a recursive edit in which you can change
the current buffer. When you exit from the recursive edit, you go back to
the query-replace
.
Exiting a recursive edit means returning to the unfinished
command, which continues execution. For example, exiting the recursive
edit requested by C-r in query-replace
causes query replacing
to resume. Exiting is done with C-M-c (exit-recursive-edit
).
You can also abort a recursive edit. This is like exiting, but
also quits the unfinished command immediately. Use the command C-]
(abort-recursive-edit
) for this. See section Quitting and Aborting.
The mode line shows you when you are in a recursive edit by displaying square brackets around the parentheses that always surround the major and minor mode names. Every window's mode line shows the square brackets, since Emacs as a whole, rather than any particular buffer, is in a recursive edit.
It is possible to be in recursive edits within recursive edits. For
example, after typing C-r in a query-replace
, you might
type a command that entered the debugger. In such a case, two or
more sets of square brackets appear in the mode line(s). Exiting the
inner recursive edit (here with the debugger c command)
resumes the query-replace command where it called the debugger. After
the end of the query-replace command, you would be able to exit the
first recursive edit. Aborting exits only one level of recursive edit;
it returns to the command level of the previous recursive edit. You can
then abort that one as well.
The command M-x top-level aborts all levels of recursive edits, returning immediately to the top level command reader.
The text you edit inside the recursive edit need not be the same text that you were editing at top level. If the command that invokes the recursive edit selects a different buffer first, that is the buffer you will edit recursively. You can switch buffers within the recursive edit in the normal manner (as long as the buffer-switching keys have not been rebound). While you could theoretically do the rest of your editing inside the recursive edit, including visiting files, this could have surprising effects (such as stack overflow) from time to time. It is best if you always exit or abort a recursive edit when you no longer need it.
In general, GNU Emacs tries to avoid using recursive edits. It is usually preferable to allow users to switch among the possible editing modes in any order they like. With recursive edits, the only way to get to another state is to go "back" to the state that the recursive edit was invoked from.
M-x dissociated-press is a command for scrambling a file of text either word by word or character by character. Starting from a buffer of straight English, it produces extremely amusing output. The input comes from the current Emacs buffer. Dissociated Press writes its output in a buffer named `*Dissociation*', and redisplays that buffer after every couple of lines (approximately) to facilitate reading it.
dissociated-press
asks every so often whether to continue
operating. Answer n to stop it. You can also stop at any time by
typing C-g. The dissociation output remains in the `*Dissociation*'
buffer for you to copy elsewhere if you wish.
Dissociated Press operates by jumping at random from one point in the buffer to another. In order to produce plausible output rather than gibberish, it insists on a certain amount of overlap between the end of one run of consecutive words or characters and the start of the next. That is, if it has just printed out `president' and then decides to jump to a different point in the file, it might spot the `ent' in `pentagon' and continue from there, producing `presidentagon'. Long sample texts produce the best results.
A positive argument to M-x dissociated-press tells it to operate character by character, and specifies the number of overlap characters. A negative argument tells it to operate word by word and specifies the number of overlap words. In this mode, whole words are treated as the elements to be permuted, rather than characters. No argument is equivalent to an argument of two. For your againformation, the output goes only into the buffer `*Dissociation*'. The buffer you start with is not changed.
Dissociated Press produces nearly the same results as a Markov chain based on a frequency table constructed from the sample text. It is, however, an independent, ignoriginal invention. Dissociated Press techniquitously copies several consecutive characters from the sample between random choices, whereas a Markov chain would choose randomly for each word or character. This makes for more plausible sounding results and runs faster.
It is a mustatement that too much use of Dissociated Press can be a developediment to your real work. Sometimes to the point of outragedy. And keep dissociwords out of your documentation, if you want it to be well userenced and properbose. Have fun. Your buggestions are welcome.
Besides producing a file of scrambled text with Dissociated Press, you can generate random sentences by using CONX.
*conx*
buffer.
conx
database.
conx
database that has been previously saved with
M-x conx-save
.
conx
database.
conx
database to a file for future retrieval.
Copy text from a buffer using M-x conx-buffer or M-x conx-region
and then type M-x conx. Output is continuously generated until you
type ^G. You can save the conx
database to a file with
M-x conx-save, which you can retrieve with M-x conx-load
.
To clear the database, use M-x conx-init
.
If you are a little bit bored, you can try M-x hanoi. If you are considerably bored, give it a numeric argument. If you are very, very bored, try an argument of 9. Sit back and watch.
When you are frustrated, try the famous Eliza program. Just do M-x doctor. End each input by typing RET twice.
When you are feeling strange, type M-x yow.
GNU Emacs can be programmed to emulate (more or less) most other editors. Standard facilities can emulate these:
M-x evi
.
If you want be in evi mode whenever you bring up Emacs, include this
line in your `.emacs' file:
(setq term-setup-hook 'evi)See section Using evi Mode for more information on evi Mode.
vi-mode
command.
vip-mode
as
it is with vi-mode
because terminating insert mode does
not use it.
For full information, see the long comment at the beginning of the
source file, which is `lisp/vip.el' in the Emacs distribution.
Warning: loading more than one vi emulator at once may cause name conflicts; no one has checked.
M-x evi
.
If you want be in evi mode whenever you bring up Emacs, include this
line in your `.emacs' file:
(setq term-setup-hook 'evi)
You can find a customization file for evi-mode in `~/.evirc'. This file has to contain Lisp code, just like the `.emacs' file, and is loaded whenever you invoke evi mode. The file allows you to rebind keys in evi mode, just as you can in other Emacs modes.
Note that evi also loads a file of vi commands from `.exrc', just like vi.
By default, all Emacs commands are disabled in evi mode. This leaves you with only vi commands. You may customize evi mode to make certain keybindings accessible. For example, to enable all Emacs command sequences that begin with Control-x or with Meta, include the following lines in your `.evirc' file:
(evi-define-key evi-all-keymaps "\C-x" ctl-x-map) (setq evi-meta-prefix-char ?\C-a) (evi-define-key evi-all-keymaps "\C-a" esc-map)
When you are in evi mode, typing Control-z stops vi emulation,
leaving you in Emacs. To get back into evi mode, use Meta-x evi
again. To exit Emacs, use Control-x Control-c.
The file management commands used by vi have been adapted to Emacs. They have slightly different meanings than the vi commands themselves:
:e
:e
overrides any complaints about the current buffer being modified
and discards all modifications. With a filename argument, it edits that
file in the current window, using the copy already in the editor if it
was previously read in. There is no difference between :e!
filename
and :e filename
. As a shorthand for editing the most
recently accessed buffer not in the window, use :e#
. :E
:e
, but edits the file in another window, creating that
window if necessary. If used with no filename, this command splits the
current buffer into two windows. :n
:N
:n
, but switches to another window or creates another window
and puts the next file into it.
All ex
commands that accept filenames as arguments perform file
completion using SPC or TAB. Completion begins after the
space that separates the command from the filename.
Many of the ex
commands are not implemented. The following
commands are implemented:
cd, chdir, copy, delete, edit, file, global, map, move, next print, put, quit, read, set, source, substitute, tag, write, wq, yank, !, <, >
The following ex
options are implemented:
autoindent, ignorecase, magic, notimeout, shiftwidth, showmatch, tabstop, wrapscan
This chapter talks about various topics relevant to adapting the behavior of Emacs in minor ways.
All kinds of customization affect only the particular Emacs job that you do them in. They are completely lost when you kill the Emacs job, and have no effect on other Emacs jobs you may run at the same time or later. The only way an Emacs job can affect anything outside of it is by writing a file; in particular, the only way to make a customization `permanent' is to put something in your `.emacs' file or other appropriate file to do the customization in each session. See section The Init File, .emacs.
Minor modes are options which you can use or not. For example, Auto Fill mode is a minor mode in which SPC breaks lines between words as you type. All the minor modes are independent of each other and of the selected major mode. Most minor modes inform you in the mode line when they are on; for example, `Fill' in the mode line means that Auto Fill mode is on.
Append -mode
to the name of a minor mode to get the name of a
command function that turns the mode on or off. Thus, the command to
enable or disable Auto Fill mode is called M-x auto-fill-mode. These
commands are usually invoked with M-x, but you can bind keys to them
if you wish. With no argument, the function turns the mode on if it was
off and off if it was on. This is known as toggling. A positive
argument always turns the mode on, and an explicit zero argument or a
negative argument always turns it off.
Auto Fill mode allows you to enter filled text without breaking lines explicitly. Emacs inserts newlines as necessary to prevent lines from becoming too long. See section Filling Text.
Overwrite mode causes ordinary printing characters to replace existing text instead of moving it to the right. For example, if point is in front of the `B' in `FOOBAR', and you type a G in Overwrite mode, it changes to `FOOGAR', instead of `FOOGBAR'.
Abbrev mode allows you to define abbreviations that automatically expand as you type them. For example, `amd' might expand to `abbrev mode'. See section Abbrevs, for full information.
A variable is a Lisp symbol which has a value. Variable names can contain any characters, but by convention they are words separated by hyphens. A variable can also have a documentation string, which describes what kind of value it should have and how the value will be used.
Lisp allows any variable to have any kind of value, but most variables
that Emacs uses require a value of a certain type. Often the value has
to be a string or a number. Sometimes we say that a certain feature is
turned on if a variable is "non-nil
," meaning that if the
variable's value is nil
, the feature is off, but the feature is
on for any other value. The conventional value to turn on the
feature--since you have to pick one particular value when you set the
variable--is t
.
Emacs uses many Lisp variables for internal recordkeeping, as any Lisp program must, but the most interesting variables for you are the ones that exist for the sake of customization. Emacs does not (usually) change the values of these variables; instead, you set the values, and thereby alter and control the behavior of certain Emacs commands. These variables are called options. Most options are documented in this manual and appear in the Variable Index (see section Variable Index).
One example of a variable which is an option is fill-column
, which
specifies the position of the right margin (as a number of characters from
the left margin) to be used by the fill commands (see section Filling Text).
To examine the value of a single variable, use C-h v
(describe-variable
), which reads a variable name using the
minibuffer, with completion. It prints both the value and the
documentation of the variable.
C-h v fill-column RET
prints something like:
fill-column's value is 75 Documentation: *Column beyond which automatic line-wrapping should happen. Automatically becomes local when set in any fashion.
The star at the beginning of the documentation indicates that this variable is an option. C-h v is not restricted to options; it allows any variable name.
If you know which option you want to set, you can use M-x set-variable to set it. This prompts for the variable name in the minibuffer (with completion), and then prompts for a Lisp expression for the new value using the minibuffer a second time. For example,
M-x set-variable RET fill-column RET 75 RET
sets fill-column
to 75, as if you had executed the Lisp expression
(setq fill-column 75)
.
Setting variables in this way, like all means of customizing Emacs except where explicitly stated, affects only the current Emacs session.
M-x list-options displays a list of all Emacs option variables in an Emacs buffer named `*List Options*'. Each option is shown with its documentation and its current value. Here is what a portion of it might look like:
;; exec-path: ("." "/usr/local/bin" "/usr/ucb" "/bin" "/usr/bin" "/u2/emacs/etc") *List of directories to search programs to run in subprocesses. Each element is a string (directory name) or nil (try the default directory). ;; ;; fill-column: 75 *Column beyond which automatic line-wrapping should happen. Automatically becomes local when set in any fashion. ;;
M-x edit-options goes one step further and immediately selects the `*List Options*' buffer; this buffer uses the major mode Options mode, which provides commands that allow you to point at an option and change its value:
nil
,
it becomes t
; otherwise it becomes nil
.
t
.
nil
.
You can make any variable local to a specific Emacs buffer. This means that the variable's value in that buffer is independent of its value in other buffers. A few variables are always local in every buffer. All other Emacs variables have a global value which is in effect in all buffers that have not made the variable local.
Major modes always make the variables they set local to the buffer. This is why changing major modes in one buffer has no effect on other buffers.
M-x make-local-variable reads the name of a variable and makes it local to the current buffer. Further changes in this buffer will not affect others, and changes in the global value will not affect this buffer.
M-x make-variable-buffer-local reads the name of a variable and
changes the future behavior of the variable so that it automatically
becomes local when it is set. More precisely, once you have marked a
variable in this way, the usual ways of setting the
variable will automatically invoke make-local-variable
first. We
call such variables per-buffer variables.
Some important variables have been marked per-buffer already. They
include abbrev-mode
, auto-fill-function
,
case-fold-search
, comment-column
, ctl-arrow
,
fill-column
, fill-prefix
, indent-tabs-mode
,
left-margin
, mode-line-format
, overwrite-mode
,
selective-display-ellipses
, selective-display
,
tab-width
, and truncate-lines
. Some other variables are
always local in every buffer, but they are used for internal
purposes.
Note: the variable auto-fill-function
was formerly named
auto-fill-hook
.
If you want a variable to cease to be local to the current buffer, call M-x kill-local-variable and provide the name of a variable to the prompt. The global value of the variable is again in effect in this buffer. Setting the major mode kills all the local variables of the buffer.
To set the global value of a variable, regardless of whether the
variable has a local value in the current buffer, you can use the
Lisp function setq-default
. It works like setq
.
If there is a local value in the current buffer, the local value is
not affected by setq-default
; thus, the new global value may
not be visible until you switch to another buffer, as in the case of:
(setq-default fill-column 75)
setq-default
is the only way to set the global value of a variable
that has been marked with make-variable-buffer-local
.
Programs can look at a variable's default value with default-value
.
This function takes a symbol as an argument and returns its default value.
The argument is evaluated; usually you must quote it explicitly, as in
the case of:
(default-value 'fill-column)
A file can contain a local variables list, which specifies the values to use for certain Emacs variables when that file is edited. Visiting the file checks for a local variables list and makes each variable in the list local to the buffer in which the file is visited, with the value specified in the file.
A local variables list goes near the end of the file, in the last page. (It is often best to put it on a page by itself.) The local variables list starts with a line containing the string `Local Variables:', and ends with a line containing the string `End:'. In between come the variable names and values, one set per line, as `variable: value'. The values are not evaluated; they are used literally.
The line which starts the local variables list does not have to say just `Local Variables:'. If there is other text before `Local Variables:', that text is called the prefix, and if there is other text after, that is called the suffix. If a prefix or suffix are present, each entry in the local variables list should have the prefix before it and the suffix after it. This includes the `End:' line. The prefix and suffix are included to disguise the local variables list as a comment so the compiler or text formatter will ignore it. If you do not need to disguise the local variables list as a comment in this way, there is no need to include a prefix or a suffix.
Two "variable" names are special in a local variables list: a value
for the variable mode
sets the major mode, and a value for the
variable eval
is simply evaluated as an expression and the value
is ignored. These are not real variables; setting them in any other
context does not have the same effect. If mode
is used in a
local variables list, it should be the first entry in the list.
Here is an example of a local variables list:
;;; Local Variables: *** ;;; mode:lisp *** ;;; comment-column:0 *** ;;; comment-start: ";;; " *** ;;; comment-end:"***" *** ;;; End: ***
Note that the prefix is `;;; ' and the suffix is ` ***'. Note also that comments in the file begin with and end with the same strings. Presumably the file contains code in a language which is enough like Lisp for Lisp mode to be useful but in which comments start and end differently. The prefix and suffix are used in the local variables list to make the list look like several lines of comments when the compiler or interpreter for that language reads the file.
The start of the local variables list must be no more than 3000 characters from the end of the file, and must be in the last page if the file is divided into pages. Otherwise, Emacs will not notice it is there. The purpose is twofold: a stray `Local Variables:' not in the last page does not confuse Emacs, and Emacs never needs to search a long file that contains no page markers and has no local variables list.
You may be tempted to turn on Auto Fill mode with a local variable list. That is inappropriate. Whether you use Auto Fill mode or not is a matter of personal taste, not a matter of the contents of particular files. If you want to use Auto Fill, set up major mode hooks with your `.emacs' file to turn it on (when appropriate) for you alone (see section The Init File, .emacs). Don't try to use a local variable list that would impose your taste on everyone working with the file.
XEmacs allows you to specify local variables in the first line
of a file, in addition to specifying them in the Local Variables
section at the end of a file.
If the first line of a file contains two occurences of `-*-'
, Emacs
uses the information between them to determine what the major mode and
variable settings should be. For example, these are all legal:
;;; -*- mode: emacs-lisp -*- ;;; -*- mode: postscript; version-control: never -*- ;;; -*- tags-file-name: "/foo/bar/TAGS" -*-
For historical reasons, the syntax `-*- modename -*-'
is allowed
as well; for example, you can use:
;;; -*- emacs-lisp -*-
The variable enable-local-variables
controls the use of local
variables lists in files you visit. The value can be t
,
nil
, or something else. A value of t
means local variables
lists are obeyed; nil
means they are ignored; anything else means
query.
The command M-x normal-mode
always obeys local variables lists
and ignores this variable.
A keyboard macro is a command defined by the user to abbreviate a sequence of keys. For example, if you discover that you are about to type C-n C-d forty times, you can speed your work by defining a keyboard macro to invoke C-n C-d and calling it with a repeat count of forty.
start-kbd-macro
).
end-kbd-macro
).
call-last-kbd-macro
).
kbd-macro-query
).
Keyboard macros differ from other Emacs commands in that they are written in the Emacs command language rather than in Lisp. This makes it easier for the novice to write them and makes them more convenient as temporary hacks. However, the Emacs command language is not powerful enough as a programming language to be useful for writing anything general or complex. For such things, Lisp must be used.
You define a keyboard macro by executing the commands which are its definition. Put differently, as you are defining a keyboard macro, the definition is being executed for the first time. This way, you see what the effects of your commands are, and don't have to figure them out in your head. When you are finished, the keyboard macro is defined and also has been executed once. You can then execute the same set of commands again by invoking the macro.
To start defining a keyboard macro, type C-x (
(start-kbd-macro
). From then on, anything you type continues to be
executed, but also becomes part of the definition of the macro. `Def'
appears in the mode line to remind you of what is going on. When you are
finished, the C-x ) command (end-kbd-macro
) terminates the
definition, without becoming part of it.
For example,
C-x ( M-f foo C-x )
defines a macro to move forward a word and then insert `foo'.
You can give C-x ) a repeat count as an argument, in which case it repeats the macro that many times right after defining it, but defining the macro counts as the first repetition (since it is executed as you define it). If you give C-x ) an argument of 4, it executes the macro immediately 3 additional times. An argument of zero to C-x e or C-x ) means repeat the macro indefinitely (until it gets an error or you type C-g).
Once you have defined a macro, you can invoke it again with the
C-x e command (call-last-kbd-macro
). You can give the
command a repeat count numeric argument to execute the macro many times.
To repeat an operation at regularly spaced places in the text, define a macro and include as part of the macro the commands to move to the next place you want to use it. For example, if you want to change each line, you should position point at the start of a line, and define a macro to change that line and leave point at the start of the next line. Repeating the macro will then operate on successive lines.
After you have terminated the definition of a keyboard macro, you can add to the end of its definition by typing C-u C-x (. This is equivalent to plain C-x ( followed by retyping the whole definition so far. As a consequence it re-executes the macro as previously defined.
To save a keyboard macro for longer than until you define the
next one, you must give it a name using M-x name-last-kbd-macro.
This reads a name as an argument using the minibuffer and defines that name
to execute the macro. The macro name is a Lisp symbol, and defining it in
this way makes it a valid command name for calling with M-x or for
binding a key to with global-set-key
(see section Keymaps). If you
specify a name that has a prior definition other than another keyboard
macro, Emacs prints an error message and nothing is changed.
Once a macro has a command name, you can save its definition in a file. You can then use it in another editing session. First visit the file you want to save the definition in. Then use the command:
M-x insert-kbd-macro RET macroname RET
This inserts some Lisp code that, when executed later, will define the same
macro with the same definition it has now. You need not understand Lisp
code to do this, because insert-kbd-macro
writes the Lisp code for you.
Then save the file. You can load the file with load-file
(see section Libraries of Lisp Code for Emacs). If the file you save in is your initialization file
`~/.emacs' (see section The Init File, .emacs), then the macro will be defined each
time you run Emacs.
If you give insert-kbd-macro
a prefix argument, it creates
additional Lisp code to record the keys (if any) that you have bound to the
keyboard macro, so that the macro is reassigned the same keys when you
load the file.
You can use C-x q (kbd-macro-query
), to get an effect similar
to that of query-replace
. The macro asks you each time
whether to make a change. When you are defining the macro, type C-x
q at the point where you want the query to occur. During macro
definition, the C-x q does nothing, but when you invoke the macro,
C-x q reads a character from the terminal to decide whether to
continue.
The special answers to a C-x q query are SPC, DEL, C-d, C-l, and C-r. Any other character terminates execution of the keyboard macro and is then read as a command. SPC means to continue. DEL means to skip the remainder of this repetition of the macro, starting again from the beginning in the next repetition. C-d means to skip the remainder of this repetition and cancel further repetition. C-l redraws the screen and asks you again for a character to specify what to do. C-r enters a recursive editing level, in which you can perform editing that is not part of the macro. When you exit the recursive edit using C-M-c, you are asked again how to continue with the keyboard macro. If you type a SPC at this time, the rest of the macro definition is executed. It is up to you to leave point and the text in a state such that the rest of the macro will do what you want.
C-u C-x q, which is C-x q with a numeric argument, performs a different function. It enters a recursive edit reading input from the keyboard, both when you type it during the definition of the macro and when it is executed from the macro. During definition, the editing you do inside the recursive edit does not become part of the macro. During macro execution, the recursive edit gives you a chance to do some particularized editing. See section Recursive Editing Levels.
This section deals with the keymaps that define the bindings between keys and functions, and shows how you can customize these bindings.
A command is a Lisp function whose definition provides for interactive use. Like every Lisp function, a command has a function name, which is a Lisp symbol whose name usually consists of lower case letters and hyphens.
The bindings between characters and command functions are recorded in
data structures called keymaps. Emacs has many of these. One, the
global keymap, defines the meanings of the single-character keys that
are defined regardless of major mode. It is the value of the variable
global-map
.
Each major mode has another keymap, its local keymap, which
contains overriding definitions for the single-character keys that are
redefined in that mode. Each buffer records which local keymap is
installed for it at any time, and the current buffer's local keymap is
the only one that directly affects command execution. The local keymaps
for Lisp mode, C mode, and many other major modes always exist even when
not in use. They are the values of the variables lisp-mode-map
,
c-mode-map
, and so on. For less frequently used major modes, the
local keymap is sometimes constructed only when the mode is used for the
first time in a session, to save space.
There are local keymaps for the minibuffer, too; they contain various completion and exit commands.
minibuffer-local-map
is used for ordinary input (no completion).
minibuffer-local-ns-map
is similar, except that SPC exits
just like RET. This is used mainly for Mocklisp compatibility.
minibuffer-local-completion-map
is for permissive completion.
minibuffer-local-must-match-map
is for strict completion and
for cautious completion.
repeat-complex-command-map
is for use in C-x ESC.
isearch-mode-map
contains the bindings of the special keys which
are bound in the pseudo-mode entered with C-s and C-r.
Finally, each prefix key has a keymap which defines the key sequences
that start with it. For example, ctl-x-map
is the keymap used for
characters following a C-x.
ctl-x-map
is the variable name for the map used for characters that
follow C-x.
help-map
is used for characters that follow C-h.
esc-map
is for characters that follow ESC. All Meta
characters are actually defined by this map.
ctl-x-4-map
is for characters that follow C-x 4.
mode-specific-map
is for characters that follow C-c.
The definition of a prefix key is the keymap to use for looking up
the following character. Sometimes the definition is actually a Lisp
symbol whose function definition is the following character keymap. The
effect is the same, but it provides a command name for the prefix key that
you can use as a description of what the prefix key is for. Thus the
binding of C-x is the symbol Ctl-X-Prefix
, whose function
definition is the keymap for C-x commands, the value of
ctl-x-map
.
Prefix key definitions can appear in either the global map or a local map. The definitions of C-c, C-x, C-h, and ESC as prefix keys appear in the global map, so these prefix keys are always available. Major modes can locally redefine a key as a prefix by putting a prefix key definition for it in the local map.
A mode can also put a prefix definition of a global prefix character such
as C-x into its local map. This is how major modes override the
definitions of certain keys that start with C-x. This case is
special, because the local definition does not entirely replace the global
one. When both the global and local definitions of a key are other
keymaps, the next character is looked up in both keymaps, with the local
definition overriding the global one. The character after the
C-x is looked up in both the major mode's own keymap for redefined
C-x commands and in ctl-x-map
. If the major mode's own keymap
for C-x commands contains nil
, the definition from the global
keymap for C-x commands is used.
You can redefine an Emacs key by changing its entry in a keymap. You can change the global keymap, in which case the change is effective in all major modes except those that have their own overriding local definitions for the same key. Or you can change the current buffer's local map, which affects all buffers using the same major mode.
cmd is a symbol naming an interactively-callable function.
When called interactively, key is the next complete key sequence
that you type. When called as a function, key is a string, a
vector of events, or a vector of key-description lists as described in
the define-key
function description. The binding goes in
the current buffer's local map, which is shared with other buffers in
the same major mode.
The following example:
M-x global-set-key RET C-f next-line RET
redefines C-f to move down a line. The fact that cmd is read second makes it serve as a kind of confirmation for key.
These functions offer no way to specify a particular prefix keymap as the one to redefine in, but that is not necessary, as you can include prefixes in key. key is read by reading characters one by one until they amount to a complete key (that is, not a prefix key). Thus, if you type C-f for key, Emacs enters the minibuffer immediately to read cmd. But if you type C-x, another character is read; if that character is 4, another character is read, and so on. For example,
M-x global-set-key RET C-x 4 $ spell-other-window RET
redefines C-x 4 $ to run the (fictitious) command
spell-other-window
.
The most general way to modify a keymap is the function
define-key
, used in Lisp code (such as your `.emacs' file).
define-key
takes three arguments: the keymap, the key to modify
in it, and the new definition. See section The Init File, .emacs, for an example.
substitute-key-definition
is used similarly; it takes three
arguments, an old definition, a new definition, and a keymap, and
redefines in that keymap all keys that were previously defined with the
old definition to have the new definition instead.
You can use the functions global-set-key
and define-key
to rebind keys under program control.
(global-set-key keys cmd)
(define-key keymap keys def)
keys is the sequence of keystrokes to bind.
def is anything that can be a key's definition:
nil
, meaning key is undefined in this keymap
(string . defn)
, meaning that defn is the definition
(defn should be a valid definition in its own right)
(keymap . char)
, meaning use the definition of
char in map keymap
For backward compatibility, XEmacs allows you to specify key sequences as strings. However, the preferred method is to use the representations of key sequences as vectors of keystrokes. See section Keystrokes, Key Sequences, and Key Bindings, for more information about the rules for constructing key sequences.
Emacs allows you to abbreviate representations for key sequences in most places where there is no ambiguity. Here are some rules for abbreviation:
f1
is equivalent to (f1)
.
(control a)
is equivalent to [(control a)]
.
65
is equivalent to A
. (This is not so much an
abbreviation as an alternate representation.)
Here are some examples of programmatically binding keys:
;;; Bindmy-command
to f1 (global-set-key 'f1 'my-command) ;;; Bindmy-command
to Shift-f1 (global-set-key '(shift f1) 'my-command) ;;; Bindmy-command
to C-c Shift-f1 (global-set-key '[(control c) (shift f1)] 'my-command) ;;; Bindmy-command
to the middle mouse button. (global-set-key 'button2 'my-command) ;;; Bindmy-command
to META CTL Right Mouse Button ;;; in the keymap that is in force when you are runningdired
. (define-key dired-mode-map '(meta control button3) 'my-command)
For backward compatibility, you can still use strings to represent key sequences. Thus you can use comands like the following:
;;; Bind end-of-line
to C-f
(global-set-key "\C-f" 'end-of-line)
Note, however, that in some cases you may be binding more than one key sequence by using a single command. This situation can arise because in ASCII, C-i and TAB have the same representation. Therefore, when Emacs sees:
(global-set-key "\C-i" 'end-of-line)
it is unclear whether the user intended to bind C-i or TAB. The solution XEmacs adopts is to bind both of these key sequences.
After binding a command to two key sequences with a form like:
(define-key global-map "\^X\^I" 'command-1)
it is possible to redefine only one of those sequences like so:
(define-key global-map [(control x) (control i)] 'command-2) (define-key global-map [(control x) tab] 'command-3)
This applies only when running under a window system. If you are talking to Emacs through an ASCII-only channel, you do not get any of these features.
Here is a table of pairs of key sequences that behave in a similar fashion:
control h backspace control l clear control i tab control m return control j linefeed control [ escape control @ control space
Disabling a command marks it as requiring confirmation before it can be executed. The purpose of disabling a command is to prevent beginning users from executing it by accident and being confused.
The direct mechanism for disabling a command is to have a non-nil
disabled
property on the Lisp symbol for the command. These
properties are normally set by the user's `.emacs' file with
Lisp expressions such as:
(put 'delete-region 'disabled t)
If the value of the disabled
property is a string, that string
is included in the message printed when the command is used:
(put 'delete-region 'disabled "Text deleted this way cannot be yanked back!\n")
You can disable a command either by editing the `.emacs' file directly or with the command M-x disable-command, which edits the `.emacs' file for you. See section The Init File, .emacs.
When you attempt to invoke a disabled command interactively in Emacs, a window is displayed containing the command's name, its documentation, and some instructions on what to do next; then Emacs asks for input saying whether to execute the command as requested, enable it and execute, or cancel it. If you decide to enable the command, you are asked whether to do this permanently or just for the current session. Enabling permanently works by automatically editing your `.emacs' file. You can use M-x enable-command at any time to enable any command permanently.
Whether a command is disabled is independent of what key is used to invoke it; it also applies if the command is invoked using M-x. Disabling a command has no effect on calling it as a function from Lisp programs.
All the Emacs commands which parse words or balance parentheses are controlled by the syntax table. The syntax table specifies which characters are opening delimiters, which are parts of words, which are string quotes, and so on. Actually, each major mode has its own syntax table (though sometimes related major modes use the same one) which it installs in each buffer that uses that major mode. The syntax table installed in the current buffer is the one that all commands use, so we call it "the" syntax table. A syntax table is a Lisp object, a vector of length 256 whose elements are numbers.
The syntax table entry for a character is a number that encodes six pieces of information:
The syntactic classes are stored internally as small integers, but are usually described to or by the user with characters. For example, `(' is used to specify the syntactic class of opening delimiters. Here is a table of syntactic classes, with the characters that specify them.
The characters flagged as part of two-character comment delimiters can
have other syntactic functions most of the time. For example, `/' and
`*' in C code, when found separately, have nothing to do with
comments. The comment-delimiter significance overrides when the pair of
characters occur together in the proper order. Only the list and sexp
commands use the syntax table to find comments; the commands specifically
for comments have other variables that tell them where to find comments.
Moreover, the list and sexp commands notice comments only if
parse-sexp-ignore-comments
is non-nil
. This variable is set
to nil
in modes where comment-terminator sequences are liable to
appear where there is no comment, for example, in Lisp mode where the
comment terminator is a newline but not every newline ends a comment.
It is possible to alter a character's syntax table entry by storing a new number in the appropriate element of the syntax table, but it would be hard to determine what number to use. Emacs therefore provides a command that allows you to specify the syntactic properties of a character in a convenient way.
M-x modify-syntax-entry is the command to change a character's syntax. It can be used interactively and is also used by major modes to initialize their own syntax tables. Its first argument is the character to change. The second argument is a string that specifies the new syntax. When called from Lisp code, there is a third, optional argument, which specifies the syntax table in which to make the change. If not supplied, or if this command is called interactively, the third argument defaults to the current buffer's syntax table.
Use C-h s (describe-syntax
) to display a description of
the contents of the current syntax table. The description of each
character includes both the string you have to pass to
modify-syntax-entry
to set up that character's current syntax,
and some English to explain that string if necessary.
When you start Emacs, it normally loads the file `.emacs' in your home directory. This file, if it exists, should contain Lisp code. It is called your initialization file or init file. Use the command line switches `-q' and `-u' to tell Emacs whether to load an init file (see section Entering and Exiting Emacs).
When the `.emacs' file is read, the variable init-file-user
says which user's init file it is. The value may be the null string or a
string containing a user's name. If the value is a null string, it means
that the init file was taken from the user that originally logged in.
In all cases, (concat "~" init-file-user "/")
evaluates to the
directory name of the directory where the `.emacs' file was looked
for.
At some sites there is a default init file, which is the
library named `default.el', found via the standard search path for
libraries. The Emacs distribution contains no such library; your site
may create one for local customizations. If this library exists, it is
loaded whenever you start Emacs. But your init file, if any, is loaded
first; if it sets inhibit-default-init
non-nil
, then
`default' is not loaded.
If you have a large amount of code in your `.emacs' file, you
should move it into another file named `something.el',
byte-compile it (see section Libraries of Lisp Code for Emacs), and load that file from your
`.emacs' file using load
.
The `.emacs' file contains one or more Lisp function call
expressions. Each consists of a function name followed by
arguments, all surrounded by parentheses. For example, (setq
fill-column 60)
represents a call to the function setq
which is
used to set the variable fill-column
(see section Filling Text) to 60.
The second argument to setq
is an expression for the new value
of the variable. This can be a constant, a variable, or a function call
expression. In `.emacs', constants are used most of the time.
They can be:
?x
, ?\n
, ?\"
, ?\)
. Note that
strings and characters are not interchangeable in Lisp; some contexts
require one and some contexts require the other.
t
stands for `true'.
nil
stands for `false'.
Here are some examples of doing certain commonly desired things with Lisp expressions:
(setq c-tab-always-indent nil)Here we have a variable whose value is normally
t
for `true'
and the alternative is nil
for `false'.
(setq-default case-fold-search nil)This sets the default value, which is effective in all buffers that do not have local values for the variable. Setting
case-fold-search
with setq
affects only the current buffer's local value, which
is probably not what you want to do in an init file.
(setq default-major-mode 'text-mode)Note that
text-mode
is used because it is the command for entering
the mode we want. A single-quote is written before it to make a symbol
constant; otherwise, text-mode
would be treated as a variable name.
(setq text-mode-hook '(lambda () (auto-fill-mode 1)))Here we have a variable whose value should be a Lisp function. The function we supply is a list starting with
lambda
, and a single
quote is written in front of it to make it (for the purpose of this
setq
) a list constant rather than an expression. Lisp functions
are not explained here; for mode hooks it is enough to know that
(auto-fill-mode 1)
is an expression that will be executed when
Text mode is entered. You could replace it with any other expression
that you like, or with several expressions in a row.
(setq text-mode-hook 'turn-on-auto-fill)This is another way to accomplish the same result.
turn-on-auto-fill
is a symbol whose function definition is
(lambda () (auto-fill-mode 1))
.
(load "foo")When the argument to
load
is a relative pathname, not starting
with `/' or `~', load
searches the directories in
load-path
(see section Loading Libraries).
(load "~/foo.elc")Here an absolute file name is used, so no searching is done.
make-symbolic-link
.
(global-set-key "\C-xl" 'make-symbolic-link)or
(define-key global-map "\C-xl" 'make-symbolic-link)Note once again the single-quote used to refer to the symbol
make-symbolic-link
instead of its value as a variable.
(define-key c-mode-map "\C-xl" 'make-symbolic-link)
(define-key c-mode-map 'f1 'make-symbolic-link)
(define-key c-mode-map '(shift f1) 'make-symbolic-link)
next-line
in Fundamental mode
to run forward-line
instead.
(substitute-key-definition 'next-line 'forward-line global-map)
(global-unset-key "\C-x\C-v")One reason to undefine a key is so that you can make it a prefix. Simply defining C-x C-v anything would make C-x C-v a prefix, but C-x C-v must be freed of any non-prefix definition first.
(modify-syntax-entry ?\$ "." text-mode-syntax-table)
eval-expression
without confirmation.
(put 'eval-expression 'disabled nil)
Each terminal type can have a Lisp library to be loaded into Emacs when
it is run on that type of terminal. For a terminal type named
termtype, the library is called `term/termtype' and it is
found by searching the directories load-path
as usual and trying the
suffixes `.elc' and `.el'. Normally it appears in the
subdirectory `term' of the directory where most Emacs libraries are
kept.
The usual purpose of the terminal-specific library is to define the escape sequences used by the terminal's function keys using the library `keypad.el'. See the file `term/vt100.el' for an example of how this is done.
When the terminal type contains a hyphen, only the part of the name
before the first hyphen is significant in choosing the library name.
Thus, terminal types `aaa-48' and `aaa-30-rv' both use
the library `term/aaa'. The code in the library can use
(getenv "TERM")
to find the full terminal type name.
The library's name is constructed by concatenating the value of the
variable term-file-prefix
and the terminal type. Your `.emacs'
file can prevent the loading of the terminal-specific library by setting
term-file-prefix
to nil
.
The value of the variable term-setup-hook
, if not nil
, is
called as a function of no arguments at the end of Emacs initialization,
after both your `.emacs' file and any terminal-specific library have
been read. You can set the value in the `.emacs' file to override
part of any of the terminal-specific libraries and to define
initializations for terminals that do not have a library.
You can now change how the audible bell sounds using the variable
sound-alist
.
sound-alist
's value is an list associating symbols with, among
other things, strings of audio-data. When ding
is called with
one of the symbols, the associated sound data is played instead of the
standard beep. This only works if you are logged in on the console of a
machine with audio hardware. To listen to a sound of the provided type,
call the function play-sound
with the argument sound. You
can also set the volume of the sound with the optional argument
volume.
Each element of sound-alist
is a list describing a sound.
The first element of the list is the name of the sound being defined.
Subsequent elements of the list are alternating keyword/value pairs:
sound
t
here means use the default X beep.
volume
bell-volume
.
pitch
duration
For compatibility, elements of `sound-alist' may also be of the form:
( sound-name . <sound> ) ( sound-name <volume> <sound> )
You should probably add things to this list by calling the function
load-sound-file
.
Note that you can only play audio data if running on the console screen of a machine with audio hardware which emacs understands, which at this time means a Sun SparcStation, SGI, or HP9000s700.
Also note that the pitch, duration, and volume options are available everywhere, but most X servers ignore the `pitch' option.
The variable bell-volume
should be an integer from 0 to 100,
with 100 being loudest, which controls how loud the sounds emacs makes
should be. Elements of the sound-alist
may override this value.
This variable applies to the standard X bell sound as well as sound files.
If the symbol t
is in place of a sound-string, Emacs uses the
default X beep. This allows you to define beep-types of
different volumes even when not running on the console.
You can add things to this list by calling the function
load-sound-file
, which reads in an audio-file and adds its data to
the sound-alist. You can specify the sound with the sound-name
argument and the file into which the sounds are loaded with the
filename argument. The optional volume argument sets the
volume.
load-sound-file (filename sound-name &optional volume)
To load and install some sound files as beep-types, use the function
load-default-sounds
(note that this only works if you are on
display 0 of a machine with audio hardware).
The following beep-types are used by Emacs itself. Other Lisp packages may use other beep types, but these are the ones that the C kernel of Emacs uses.
auto-save-error
command-error
undefined-key
undefined-click
no-completion
y-or-n-p
y
or n
yes-or-no-p
yes
or no
XEmacs has objects called extents and faces. An extent is a region of text and a face is a collection of textual attributes, such as fonts and colors. Every extent is displayed in some face; therefore, changing the properties of a face immediately updates the display of all associated extents. Faces can be screen-local: you can have a region of text that displays with completely different attributes when its buffer is viewed from a different X window.
The display attributes of faces may be specified either in Lisp or through the X resource manager.
You can change the face of an extent with the functions in this section. All the functions prompt for a face as an argument; use completion for a list of possible values.
nil
if this is not possible.
nil
if not possible.
nil
if not possible.
nil
if not possible.
nil
if not possible.
nil
if not possible.
nil
if not possible.
You can exchange the foreground and background color of the selected
face with the function invert-face
. If the face does not
specify both foreground and background, then its foreground and
background are set to the background and foreground of the default face.
When calling this from a program, you can supply the optional argument
screen to specify which screen is affected; otherwise, all screens
are affected.
You can set the background color of the specified face with the
function set-face-background
. The argument color
should
be a string, the name of a color. When called from a program, if the
optional screen argument is provided, the face is changed only
in that screen; otherwise, it is changed in all screens.
You can set the background pixmap of the specified face with the
function set-face-background-pixmap
. The pixmap argument
name should be a string, the name of a file of pixmap data. The
directories listed in the x-bitmap-file-path
variable are
searched. The bitmap may also be a list of the form (width
height data
), where width and height are the size in
pixels, and data is a string containing the raw bits of the
bitmap. If the optional screen argument is provided, the face is
changed only in that screen; otherwise, it is changed in all screens.
The variable x-bitmap-file-path
takes as a value a list of the
directories in which X bitmap files may be found. If the value is
nil
, the list is initialized from the *bitmapFilePath
resource.
If the environment variable XBMLANGPATH is set, then it is consulted
before the x-bitmap-file-path
variable.
You can set the font of the specified face with the function
set-face-font
. The font argument should be a string, the
name of a font. When called from a program, if the
optional screen argument is provided, the face is changed only
in that screen; otherwise, it is changed in all screens.
You can set the foreground color of the specified face with the
function set-face-foreground
. The argument color should be
a string, the name of a color. If the optional screen argument is
provided, the face is changed only in that screen; otherwise, it is
changed in all screens.
You can set underline the specified face with the function
set-face-underline-p
. The argument underline-p can be used
to make underlining an attribute of the face or not. If the optional
screen argument is provided, the face is changed only in that
screen; otherwise, it is changed in all screens.
The Emacs resources are generally set per-screen. Each Emacs screen can have
its own name or the same name as another, depending on the name passed to the
x-create-screen
function.
You can specify resources for all screens with the syntax:
Emacs*parameter: value
or
Emacs*EmacsScreen*parameter:value
You can specify resources for a particular screen with the syntax:
Emacs*SCREEN-NAME*parameter: value
To make the default size of all Emacs screens 80 columns by 55 lines, set the resource this way:
Emacs*EmacsScreen*geometry: 80x55
As a special case, this geometry specification also works:
Emacs.geometry: 80x55
since that is the syntax used with most other applications (since most
other applications have only one top-level window, unlike Emacs). In
general, however, the top-level shell does not have any interesting
resources on it, and you should set the resources on the screens
instead. To set the geometry of a particular screen named fred
, do
this: Emacs*fred.geometry: 80x55
.
Important! Do not use the following syntax:
Emacs*geometry: 80x55
You should never use *geometry
with any X application. It does
not set the Emacs screen to be 80 columns by 55 lines. It actually sets
all Emacs screens and windows to be 80 by 35 in whatever units they care
to measure in. In particular, it tells the Emacs text pane to be 80 by
55 in characters and tells the menu bar pane to be 80 x 55 pixels, which
is probably not what you want.
The -geometry
command-line argument sets only the geometry of the
initial screen created by Emacs.
Emacs screens accept the following resources:
iconic
(class Iconic
): boolean
internalBorderWidth
(class InternalBorderWidth
): int
interline
(class Interline
): int
cursorColor
(class CursorColor
): color-name
textPointer
(class Cursor
): cursor-name
x-pointer-shape
.
selectionPointer
(class Cursor
): cursor-name
x-selection-pointer-shape
.
spacePointer
(class Cursor
): cursor-name
x-nontext-pointer-shape
.
modeLinePointer
(class Cursor
): cursor-name
x-mode-pointer-shape
.
gcPointer
(class Cursor
): cursor-name
x-gc-pointer-shape
.
scrollbarPointer
(class Cursor
): cursor-name
x-scrollbar-pointer-shape
.
pointerColor
(class Foreground
): color-name
pointerBackground
(class Background
): color-name
x-pointer-foreground-color
and x-pointer-background-color
.
The attributes of faces are also per-screen. They can be specified as:
Emacs*FACE_NAME.parameter: value
or
Emacs*SCREEN_NAME*FACE_NAME.parameter: value
Faces accept the following resources:
attributeFont
(class AttributeFont
): font-name
attributeForeground
(class AttributeForeground
): color-name
attributeBackground
(class AttributeBackground
): color-name
attributeBackgroundPixmap
(class AttributeBackgroundPixmap
): file-name
attributeUnderline
(class AttributeUnderline
): boolean
All text is displayed in some face, defaulting to the face named
default
. To set the font of normal text, use
Emacs*default.attributeFont
. To set it in the screen named
fred
, use Emacs*fred*default.attributeFont
.
These are the names of the predefined faces:
default
bold
italic
bold-italic
modeline
highlight
left-margin
right-margin
primary-selection
isearch
info-node
bold-italic
.
info-xref
bold
. (Note that, when the mouse passes over a
cross-reference, the cross-reference's face is determined from a
combination of the info-xref
and highlight
faces.)
Other packages might define their own faces; to see a list of all faces,
use any of the interactive face-manipulation commands such as
set-face-font
and type `?' when you are prompted for the
name of a face.
If you type an Emacs command you did not intend, the results are often mysterious. This chapter discusses how you can undo your mistake or recover from a mysterious situation. Emacs bugs and system crashes are also considered.
abort-recursive-edit
).
undo
).
There are two ways of cancelling commands which are not finished executing: quitting with C-g, and aborting with C-] or M-x top-level. Quitting is cancelling a partially typed command or one which is already running. Aborting is getting out of a recursive editing level and cancelling the command that invoked the recursive edit.
Quitting with C-g is used for getting rid of a partially typed
command or a numeric argument that you don't want. It also stops a
running command in the middle in a relatively safe way, so you can use
it if you accidentally start executing a command that takes a long
time. In particular, it is safe to quit out of killing; either your
text will all still be there, or it will all be in the kill
ring (or maybe both). Quitting an incremental search does special
things documented under searching; in general, it may take two
successive C-g characters to get out of a search. C-g works
by setting the variable quit-flag
to t
the instant
C-g is typed; Emacs Lisp checks this variable frequently and quits
if it is non-nil
. C-g is only actually executed as a
command if it is typed while Emacs is waiting for input.
If you quit twice in a row before the first C-g is recognized, you activate the "emergency escape" feature and return to the shell. See section Emergency Escape.
You can use C-] (abort-recursive-edit
) to get out
of a recursive editing level and cancel the command which invoked it.
Quitting with C-g does not do this, and could not do this because it
is used to cancel a partially typed command within the recursive
editing level. Both operations are useful. For example, if you are in the
Emacs debugger (see section The Emacs-Lisp Debugger) and have typed C-u 8 to enter a
numeric argument, you can cancel that argument with C-g and remain in
the debugger.
The command M-x top-level is equivalent to "enough" C-] commands to get you out of all the levels of recursive edits that you are in. C-] only gets you out one level at a time, but M-x top-level goes out all levels at once. Both C-] and M-x top-level are like all other commands and unlike C-g in that they are effective only when Emacs is ready for a command. C-] is an ordinary key and has its meaning only because of its binding in the keymap. See section Recursive Editing Levels.
C-x u (undo
) is not strictly speaking a way of cancelling a
command, but you can think of it as cancelling a command already finished
executing. See section Undoing Changes.
This section describes various conditions in which Emacs fails to work, and how to recognize them and correct them.
Recursive editing levels are important and useful features of Emacs, but they can seem like malfunctions to the user who does not understand them.
If the mode line has square brackets `[...]' around the parentheses that contain the names of the major and minor modes, you have entered a recursive editing level. If you did not do this on purpose, or if you don't understand what that means, you should just get out of the recursive editing level. To do so, type M-x top-level. This is called getting back to top level. See section Recursive Editing Levels.
If the data on the screen looks wrong, the first thing to do is see whether the text is actually wrong. Type C-l, to redisplay the entire screen. If the text appears correct after this, the problem was entirely in the previous screen update.
Display updating problems often result from an incorrect termcap entry for the terminal you are using. The file `etc/TERMS' in the Emacs distribution gives the fixes for known problems of this sort. `INSTALL' contains general advice for these problems in one of its sections. Very likely there is simply insufficient padding for certain display operations. To investigate the possibility that you have this sort of problem, try Emacs on another terminal made by a different manufacturer. If problems happen frequently on one kind of terminal but not another kind, the real problem is likely to be a bad termcap entry, though it could also be due to a bug in Emacs that appears for terminals that have or lack specific features.
If C-l shows that the text is wrong, try undoing the changes to it using C-x u until it gets back to a state you consider correct. Also try C-h l to find out what command you typed to produce the observed results.
If a large portion of text appears to be missing at the beginning or end of the buffer, check for the word `Narrow' in the mode line. If it appears, the text is still present, but marked off-limits. To make it visible again, type C-x n w. See section Narrowing.
If Emacs spontaneously displays `I-search:' at the bottom of the screen, it means that the terminal is sending C-s and C-q according to the poorly designed xon/xoff "flow control" protocol. You should try to prevent this by putting the terminal in a mode where it will not use flow control, or by giving it enough padding that it will never send a C-s. If that cannot be done, you must tell Emacs to expect flow control to be used, until you can get a properly designed terminal.
Information on how to do these things can be found in the file `INSTALL' in the Emacs distribution.
Because at times there have been bugs causing Emacs to loop without
checking quit-flag
, a special feature causes Emacs to be suspended
immediately if you type a second C-g while the flag is already set,
so you can always get out of GNU Emacs. Normally Emacs recognizes and
clears quit-flag
(and quits!) quickly enough to prevent this from
happening.
When you resume Emacs after a suspension caused by multiple C-g, it asks two questions before going back to what it had been doing:
Auto-save? (y or n) Abort (and dump core)? (y or n)
Answer each one with y or n followed by RET.
Saying y to `Auto-save?' causes immediate auto-saving of all modified buffers in which auto-saving is enabled.
Saying y to `Abort (and dump core)?' causes an illegal
instruction to be executed, dumping core. This is to enable a wizard to
figure out why Emacs was failing to quit in the first place. Execution
does not continue after a core dump. If you answer n, execution
does continue. With luck, Emacs will ultimately check
quit-flag
and quit normally. If not, and you type another
C-g, it is suspended again.
If Emacs is not really hung, but is just being slow, you may invoke the double C-g feature without really meaning to. In that case, simply resume and answer n to both questions, and you will arrive at your former state. Presumably the quit you requested will happen soon.
The double-C-g feature may be turned off when Emacs is running under a window system, since the window system always enables you to kill Emacs or to create another window and run another program.
If using Emacs (or something else) becomes terribly frustrating and none of the techniques described above solve the problem, Emacs can still help you.
First, if the Emacs you are using is not responding to commands, type C-g C-g to get out of it and then start a new one.
Second, type M-x doctor RET.
The doctor will make you feel better. Each time you say something to the doctor, you must end it by typing RET RET. This lets the doctor know you are finished.
Sometimes you will encounter a bug in Emacs. Although we cannot promise we can or will fix the bug, and we might not even agree that it is a bug, we want to hear about bugs you encounter in case we do want to fix them.
To make it possible for us to fix a bug, you must report it. In order to do so effectively, you must know when and how to do it.
If Emacs executes an illegal instruction, or dies with an operating system error message that indicates a problem in the program (as opposed to something like "disk full"), then it is certainly a bug.
If Emacs updates the display in a way that does not correspond to what is in the buffer, then it is certainly a bug. If a command seems to do the wrong thing but the problem corrects itself if you type C-l, it is a case of incorrect display updating.
Taking forever to complete a command can be a bug, but you must make certain that it was really Emacs's fault. Some commands simply take a long time. Type C-g and then C-h l to see whether the input Emacs received was what you intended to type; if the input was such that you know it should have been processed quickly, report a bug. If you don't know whether the command should take a long time, find out by looking in the manual or by asking for assistance.
If a command you are familiar with causes an Emacs error message in a case where its usual definition ought to be reasonable, it is probably a bug.
If a command does the wrong thing, that is a bug. But be sure you know for certain what it ought to have done. If you aren't familiar with the command, or don't know for certain how the command is supposed to work, then it might actually be working right. Rather than jumping to conclusions, show the problem to someone who knows for certain.
Finally, a command's intended definition may not be best for editing with. This is a very important sort of problem, but it is also a matter of judgment. Also, it is easy to come to such a conclusion out of ignorance of some of the existing features. It is probably best not to complain about such a problem until you have checked the documentation in the usual ways, feel confident that you understand it, and know for certain that what you want is not available. If you are not sure what the command is supposed to do after a careful reading of the manual, check the index and glossary for any terms that may be unclear. If you still do not understand, this indicates a bug in the manual. The manual's job is to make everything clear. It is just as important to report documentation bugs as program bugs.
If the online documentation string of a function or variable disagrees with the manual, one of them must be wrong, so report the bug.
When you decide that there is a bug, it is important to report it and to report it in a way which is useful. What is most useful is an exact description of what commands you type, starting with the shell command to run Emacs, until the problem happens. Always include the version number of Emacs that you are using; type M-x emacs-version to print this.
The most important principle in reporting a bug is to report facts, not hypotheses or categorizations. It is always easier to report the facts, but people seem to prefer to strain to posit explanations and report them instead. If the explanations are based on guesses about how Emacs is implemented, they will be useless; we will have to try to figure out what the facts must have been to lead to such speculations. Sometimes this is impossible. But in any case, it is unnecessary work for us.
For example, suppose that you type C-x C-f /glorp/baz.ugh RET, visiting a file which (you know) happens to be rather large, and Emacs prints out `I feel pretty today'. The best way to report the bug is with a sentence like the preceding one, because it gives all the facts and nothing but the facts.
Do not assume that the problem is due to the size of the file and say, "When I visit a large file, Emacs prints out `I feel pretty today'." This is what we mean by "guessing explanations". The problem is just as likely to be due to the fact that there is a `z' in the file name. If this is so, then when we got your report, we would try out the problem with some "large file", probably with no `z' in its name, and not find anything wrong. There is no way in the world that we could guess that we should try visiting a file with a `z' in its name.
Alternatively, the problem might be due to the fact that the file starts with exactly 25 spaces. For this reason, you should make sure that you inform us of the exact contents of any file that is needed to reproduce the bug. What if the problem only occurs when you have typed the C-x a l command previously? This is why we ask you to give the exact sequence of characters you typed since starting to use Emacs.
You should not even say "visit a file" instead of C-x C-f unless you know that it makes no difference which visiting command is used. Similarly, rather than saying "if I have three characters on the line," say "after I type RET A B C RET C-p," if that is the way you entered the text.
If you are not in Fundamental mode when the problem occurs, you should say what mode you are in.
If the manifestation of the bug is an Emacs error message, it is
important to report not just the text of the error message but a backtrace
showing how the Lisp program in Emacs arrived at the error. To make the
backtrace, you must execute the Lisp expression
(setq debug-on-error t)
before the error happens (that is to
say, you must execute that expression and then make the bug happen). This
causes the Lisp debugger to run (see section The Emacs-Lisp Debugger). The debugger's
backtrace can be copied as text into the bug report. This use of the
debugger is possible only if you know how to make the bug happen again. Do
note the error message the first time the bug happens, so if you can't make
it happen again, you can report at least that.
Check whether any programs you have loaded into the Lisp world, including
your `.emacs' file, set any variables that may affect the functioning
of Emacs. Also, see whether the problem happens in a freshly started Emacs
without loading your `.emacs' file (start Emacs with the -q
switch
to prevent loading the init file). If the problem does not occur
then, it is essential that we know the contents of any programs that you
must load into the Lisp world in order to cause the problem to occur.
If the problem does depend on an init file or other Lisp programs that are not part of the standard Emacs system, then you should make sure it is not a bug in those programs by complaining to their maintainers first. After they verify that they are using Emacs in a way that is supposed to work, they should report the bug.
If you can tell us a way to cause the problem without visiting any files, please do so. This makes it much easier to debug. If you do need files, make sure you arrange for us to see their exact contents. For example, it can often matter whether there are spaces at the ends of lines, or a newline after the last line in the buffer (nothing ought to care whether the last line is terminated, but tell that to the bugs).
The easy way to record the input to Emacs precisely is to to write a dribble file; execute the Lisp expression:
(open-dribble-file "~/dribble")
using Meta-ESC or from the `*scratch*' buffer just after starting Emacs. From then on, all Emacs input will be written in the specified dribble file until the Emacs process is killed.
For possible display bugs, it is important to report the terminal type
(the value of environment variable TERM
), the complete termcap entry
for the terminal from `/etc/termcap' (since that file is not identical
on all machines), and the output that Emacs actually sent to the terminal.
The way to collect this output is to execute the Lisp expression:
(open-termscript "~/termscript")
using Meta-ESC or from the `*scratch*' buffer just after starting Emacs. From then on, all output from Emacs to the terminal will be written in the specified termscript file as well, until the Emacs process is killed. If the problem happens when Emacs starts up, put this expression into your `.emacs' file so that the termscript file will be open when Emacs displays the screen for the first time. Be warned: it is often difficult, and sometimes impossible, to fix a terminal-dependent bug without access to a terminal of the type that stimulates the bug.
The newsgroup `alt.lucid-emacs.bug' may be used for bug reports, and the newsgroup `alt.lucid-emacs.help' may be used for other discussions and requests for assistance.
If you don't have access to these newgroups, you can subscribe to the mailing list versions: the newsgroups are bidirectionally gatewayed into the mailing lists `bug-lucid-emacs@cs.uiuc.edu' and `help-lucid-emacs@cs.uiuc.edu' respectively.
To be added or removed from these mailing lists, send mail to `bug-lucid-emacs-request@cs.uiuc.edu' or `help-lucid-emacs-request@cs.uiuc.edu'. Do not send requests for addition to the mailing lists themselves.
The mailing lists and newsgroups are archived on our anonymous FTP server, `cs.uiuc.edu', and at various other archive sites around the net. You should also check the `FAQ' in `/pub/lemacs' on our anonymous FTP server. It provides some introductory information and help for initial configuration problems.
This section describes the difference between Emacs Version 18 and XEmacs.
(setq term-setup-hook 'evi)See section Using evi Mode for a brief discussion.
define-key
or
map-keymap
and set-keymap-parent
(the new keymap
functions). See section Customizing Key Bindings for more information.
load-path
is computed when Emacs starts up,
instead of being hardcoded in when Emacs is compiled. As a result, you
can now move the Emacs executable and Lisp library to a
different location in the file system without having to recompile.
load-path
by default.
sound-alist
variable. See section Changing the Bell Sound for more information.
zmacs-regions
is set to highlight the region
between point and the mark. This unifies X selection and Emacs selection
behavior.
byte-compile-and-load-file
and byte-compile-buffer
byte-compile the contents of a file or buffer.
The new conx
function lets you generate random sentences for your
amusement.
compile-defun
compiles and evaluates the current top-level
form.
find-this-file
and find-this-file-other-window
can be used
interactively with a prefix argument to switch to the filename at point
in the buffer. find-this-file-other-window
displays the file in
another window.
invert-face
,
make-face-bold
, make-face-bold-italic
,
make-face-italic
, make-face-unbold
,
make-face-unitalic
, set-face-background
,
set-face-background-pixmap
, set-face-font
,
set-face-foreground
, and set-face-underline-p
.
load-default-sounds
and load-sound-file
allow you to
customize the audible bell sound. load-default-sounds
loads and
installs sound files. load-sound-file
reads in audio files and
adds them to the sound alist. play-sound
plays the specified
sound type.
locate-library
finds the file that the function
load-library
loads, and it displays the file's full pathname.
make-directory
creates a directory, while remove-directory
removes a directory.
mark-beginning-of-buffer
and mark-end-of-buffer
push the
mark to the beginning or end of a buffer, respectively.
Several functions have been added that allow you to perform various
editing, region, and window operations using the mouse:
mouse-del-char
, mouse-delete-window
,
mouse-keep-one-window
, mouse-kill-line
,
mouse-line-length
, mouse-scroll
, mouse-select
,
mouse-select-and-split
, mouse-set-mark
,
mouse-set-point
, mouse-track
, mouse-track-adjust
,
mouse-track-and-copy-to-cutbuffer
,
mouse-track-delete-and-insert
, mouse-track-insert
, and
mouse-window-to-region
.
compare-windows
takes an argument ignore-whitespace.
The argument means ignore changes in whitespace.
You can conditionalize your `.emacs' file as follows so that XEmacs commands are invoked only when you are in XEmacs:
(cond ((string-match "Lucid" emacs-version) ;; ;; Code for any version of XEmacs goes here ;; )) (cond ((and (string-match "Lucid" emacs-version) (>= emacs-major-version 19) (>= emacs-minor-version 6)) ;; ;; Code which requires XEmacs version 19.6 or newer goes here ;; )) (cond ((>= emacs-major-version 19) ;; ;; Code for any vintage-19 emacs goes here ;; )) (cond ((and (not (string-match "Lucid" emacs-version)) (= emacs-major-version 19)) ;; ;; Code specific to FSF Emacs 19 (not XEmacs) goes here ;; )) (cond ((< emacs-major-version 19) ;; ;; Code specific to emacs 18 goes here ;; ))
Of particular interest for use in `.emacs' files are:
add-menu
lets you add a new menu to the menubar or a submenu to a
pull-down menu. add-menu-item
, disable-menu-item
,
delete-menu-item
, enable-menu-item
, and
relabel-menu-item
allow you to customize the XEmacs
pull-down menus.
make-screen
creates a new Emacs screen (X window).
These new variables are only present in XEmacs:
minibuffer-confirm-incomplete
prompts for confirmation in
contexts where completing-read
allows answers that are not valid
completions.
x-pointer-background-color
,
x-pointer-foreground-color
, x-mode-pointer-shape
,
x-pointer-shape
, and x-nontext-pointer-shape
.
zmacs-regions
determines whether LISPM-style active regions
should be used.
XEmacs has the following new default function keybindings:
backward-char
.
forward-char
.
previous-line
.
next-line
.
defun
. See section Defuns.
next-line
.
See section Keymaps.
GNU, which stands for GNU's Not Unix, is the name for the complete Unix-compatible software system which I am writing so that I can give it away free to everyone who can use it. Several other volunteers are helping me. Contributions of time, money, programs, and equipment are greatly needed.
So far we have an Emacs text editor with Lisp for writing editor commands, a source level debugger, a yacc-compatible parser generator, a linker, and around 35 utilities. A shell (command interpreter) is nearly completed. A new portable optimizing C compiler has compiled itself and may be released this year. An initial kernel exists, but many more features are needed to emulate Unix. When the kernel and compiler are finished, it will be possible to distribute a GNU system suitable for program development. We will use TeX as our text formatter, but an nroff is being worked on. We will use the free, portable X window system as well. After this we will add a portable Common Lisp, an Empire game, a spreadsheet, and hundreds of other things, plus online documentation. We hope to supply, eventually, everything useful that normally comes with a Unix system, and more.
GNU will be able to run Unix programs, but will not be identical to Unix. We will make all improvements that are convenient, based on our experience with other operating systems. In particular, we plan to have longer filenames, file version numbers, a crashproof file system, filename completion perhaps, terminal-independent display support, and perhaps eventually a Lisp-based window system through which several Lisp programs and ordinary Unix programs can share a screen. Both C and Lisp will be available as system programming languages. We will try to support UUCP, MIT Chaosnet, and Internet protocols for communication.
GNU is aimed initially at machines in the 68000/16000 class with virtual memory, because they are the easiest machines to make it run on. The extra effort to make it run on smaller machines will be left to someone who wants to use it on them.
To avoid horrible confusion, please pronounce the `G' in the word `GNU' when it is the name of this project.
I consider that the golden rule requires that if I like a program I must share it with other people who like it. Software sellers want to divide the users and conquer them, making each user agree not to share with others. I refuse to break solidarity with other users in this way. I cannot in good conscience sign a nondisclosure agreement or a software license agreement. For years I worked within the Artificial Intelligence Lab to resist such tendencies and other inhospitalities, but eventually they had gone too far: I could not remain in an institution where such things are done for me against my will.
So that I can continue to use computers without dishonor, I have decided to put together a sufficient body of free software so that I will be able to get along without any software that is not free. I have resigned from the AI lab to deny MIT any legal excuse to prevent me from giving GNU away.
Unix is not my ideal system, but it is not too bad. The essential features of Unix seem to be good ones, and I think I can fill in what Unix lacks without spoiling them. And a system compatible with Unix would be convenient for many other people to adopt.
GNU is not in the public domain. Everyone will be permitted to modify and redistribute GNU, but no distributor will be allowed to restrict its further redistribution. That is to say, proprietary modifications will not be allowed. I want to make sure that all versions of GNU remain free.
I have found many other programmers who are excited about GNU and want to help.
Many programmers are unhappy about the commercialization of system software. It may enable them to make more money, but it requires them to feel in conflict with other programmers in general rather than feel as comrades. The fundamental act of friendship among programmers is the sharing of programs; marketing arrangements now typically used essentially forbid programmers to treat others as friends. The purchaser of software must choose between friendship and obeying the law. Naturally, many decide that friendship is more important. But those who believe in law often do not feel at ease with either choice. They become cynical and think that programming is just a way of making money.
By working on and using GNU rather than proprietary programs, we can be hospitable to everyone and obey the law. In addition, GNU serves as an example to inspire and a banner to rally others to join us in sharing. This can give us a feeling of harmony which is impossible if we use software that is not free. For about half the programmers I talk to, this is an important happiness that money cannot replace.
I am asking computer manufacturers for donations of machines and money. I'm asking individuals for donations of programs and work.
One consequence you can expect if you donate machines is that GNU will run on them at an early date. The machines should be complete, ready-to-use systems, approved for use in a residential area, and not in need of sophisticated cooling or power.
I have found very many programmers eager to contribute part-time work for GNU. For most projects, such part-time distributed work would be very hard to coordinate; the independently-written parts would not work together. But for the particular task of replacing Unix, this problem is absent. A complete Unix system contains hundreds of utility programs, each of which is documented separately. Most interface specifications are fixed by Unix compatibility. If each contributor can write a compatible replacement for a single Unix utility, and make it work properly in place of the original on a Unix system, then these utilities will work right when put together. Even allowing for Murphy to create a few unexpected problems, assembling these components will be a feasible task. (The kernel will require closer communication and will be worked on by a small, tight group.)
If I get donations of money, I may be able to hire a few people full or part time. The salary won't be high by programmers' standards, but I'm looking for people for whom building community spirit is as important as making money. I view this as a way of enabling dedicated people to devote their full energies to working on GNU by sparing them the need to make a living in another way.
Once GNU is written, everyone will be able to obtain good system software free, just like air.
This means much more than just saving everyone the price of a Unix license. It means that much wasteful duplication of system programming effort will be avoided. This effort can go instead into advancing the state of the art.
Complete system sources will be available to everyone. As a result, a user who needs changes in the system will always be free to make them himself, or hire any available programmer or company to make them for him. Users will no longer be at the mercy of one programmer or company which owns the sources and is in sole position to make changes.
Schools will be able to provide a much more educational environment by encouraging all students to study and improve the system code. Harvard's computer lab used to have the policy that no program could be installed on the system if its sources were not on public display, and upheld it by actually refusing to install certain programs. I was very much inspired by this.
Finally, the overhead of considering who owns the system software and what one is or is not entitled to do with it will be lifted.
Arrangements to make people pay for using a program, including licensing of copies, always incur a tremendous cost to society through the cumbersome mechanisms necessary to figure out how much (that is, which programs) a person must pay for. And only a police state can force everyone to obey them. Consider a space station where air must be manufactured at great cost: charging each breather per liter of air may be fair, but wearing the metered gas mask all day and all night is intolerable even if everyone can afford to pay the air bill. And the TV cameras everywhere to see if you ever take the mask off are outrageous. It's better to support the air plant with a head tax and chuck the masks.
Copying all or parts of a program is as natural to a programmer as breathing, and as productive. It ought to be as free.
"Nobody will use it if it is free, because that means they can't rely on any support.""You have to charge for the program to pay for providing the support."
If people would rather pay for GNU plus service than get GNU free without service, a company to provide just service to people who have obtained GNU free ought to be profitable.
We must distinguish between support in the form of real programming work and mere handholding. The former is something one cannot rely on from a software vendor. If your problem is not shared by enough people, the vendor will tell you to get lost.
If your business needs to be able to rely on support, the only way is to have all the necessary sources and tools. Then you can hire any available person to fix your problem; you are not at the mercy of any individual. With Unix, the price of sources puts this out of consideration for most businesses. With GNU this will be easy. It is still possible for there to be no available competent person, but this problem cannot be blamed on distibution arrangements. GNU does not eliminate all the world's problems, only some of them.
Meanwhile, the users who know nothing about computers need handholding: doing things for them which they could easily do themselves but don't know how.
Such services could be provided by companies that sell just hand-holding and repair service. If it is true that users would rather spend money and get a product with service, they will also be willing to buy the service having got the product free. The service companies will compete in quality and price; users will not be tied to any particular one. Meanwhile, those of us who don't need the service should be able to use the program without paying for the service.
"You cannot reach many people without advertising, and you must charge for the program to support that.""It's no use advertising a program people can get free."
There are various forms of free or very cheap publicity that can be used to inform numbers of computer users about something like GNU. But it may be true that one can reach more microcomputer users with advertising. If this is really so, a business which advertises the service of copying and mailing GNU for a fee ought to be successful enough to pay for its advertising and more. This way, only the users who benefit from the advertising pay for it.
On the other hand, if many people get GNU from their friends, and such companies don't succeed, this will show that advertising was not really necessary to spread GNU. Why is it that free market advocates don't want to let the free market decide this?
"My company needs a proprietary operating system to get a competitive edge."
GNU will remove operating system software from the realm of competition. You will not be able to get an edge in this area, but neither will your competitors be able to get an edge over you. You and they will compete in other areas, while benefitting mutually in this one. If your business is selling an operating system, you will not like GNU, but that's tough on you. If your business is something else, GNU can save you from being pushed into the expensive business of selling operating systems.
I would like to see GNU development supported by gifts from many manufacturers and users, reducing the cost to each.
"Don't programmers deserve a reward for their creativity?"
If anything deserves a reward, it is social contribution. Creativity can be a social contribution, but only in so far as society is free to use the results. If programmers deserve to be rewarded for creating innovative programs, by the same token they deserve to be punished if they restrict the use of these programs.
"Shouldn't a programmer be able to ask for a reward for his creativity?"
There is nothing wrong with wanting pay for work, or seeking to maximize one's income, as long as one does not use means that are destructive. But the means customary in the field of software today are based on destruction.
Extracting money from users of a program by restricting their use of it is destructive because the restrictions reduce the amount and the ways that the program can be used. This reduces the amount of wealth that humanity derives from the program. When there is a deliberate choice to restrict, the harmful consequences are deliberate destruction.
The reason a good citizen does not use such destructive means to become wealthier is that, if everyone did so, we would all become poorer from the mutual destructiveness. This is Kantian ethics; or, the Golden Rule. Since I do not like the consequences that result if everyone hoards information, I am required to consider it wrong for one to do so. Specifically, the desire to be rewarded for one's creativity does not justify depriving the world in general of all or part of that creativity.
"Won't programmers starve?"
I could answer that nobody is forced to be a programmer. Most of us cannot manage to get any money for standing on the street and making faces. But we are not, as a result, condemned to spend our lives standing on the street making faces, and starving. We do something else.
But that is the wrong answer because it accepts the questioner's implicit assumption: that without ownership of software, programmers cannot possibly be paid a cent. Supposedly it is all or nothing.
The real reason programmers will not starve is that it will still be possible for them to get paid for programming; just not paid as much as now.
Restricting copying is not the only basis for business in software. It is the most common basis because it brings in the most money. If it were prohibited, or rejected by the customer, software business would move to other bases of organization which are now used less often. There are always numerous ways to organize any kind of business.
Probably programming will not be as lucrative on the new basis as it is now. But that is not an argument against the change. It is not considered an injustice that sales clerks make the salaries that they now do. If programmers made the same, that would not be an injustice either. (In practice they would still make considerably more than that.)
"Don't people have a right to control how their creativity is used?"
"Control over the use of one's ideas" really constitutes control over other people's lives; and it is usually used to make their lives more difficult.
People who have studied the issue of intellectual property rights carefully (such as lawyers) say that there is no intrinsic right to intellectual property. The kinds of supposed intellectual property rights that the government recognizes were created by specific acts of legislation for specific purposes.
For example, the patent system was established to encourage inventors to disclose the details of their inventions. Its purpose was to help society rather than to help inventors. At the time, the life span of 17 years for a patent was short compared with the rate of advance of the state of the art. Since patents are an issue only among manufacturers, for whom the cost and effort of a license agreement are small compared with setting up production, the patents often do not do much harm. They do not obstruct most individuals who use patented products.
The idea of copyright did not exist in ancient times, when authors frequently copied other authors at length in works of non-fiction. This practice was useful, and is the only way many authors' works have survived even in part. The copyright system was created expressly for the purpose of encouraging authorship. In the domain for which it was invented--books, which could be copied economically only on a printing press--it did little harm, and did not obstruct most of the individuals who read the books.
All intellectual property rights are just licenses granted by society because it was thought, rightly or wrongly, that society as a whole would benefit by granting them. But in any particular situation, we have to ask: are we really better off granting such license? What kind of act are we licensing a person to do?
The case of programs today is very different from that of books a hundred years ago. The fact that the easiest way to copy a program is from one neighbor to another, the fact that a program has both source code and object code which are distinct, and the fact that a program is used rather than read and enjoyed, combine to create a situation in which a person who enforces a copyright is harming society as a whole both materially and spiritually; in which a person should not do so regardless of whether the law enables him to.
"Competition makes things get done better."
The paradigm of competition is a race: by rewarding the winner, we encourage everyone to run faster. When capitalism really works this way, it does a good job; but its defenders are wrong in assuming it always works this way. If the runners forget why the reward is offered and become intent on winning, no matter how, they may find other strategies--such as, attacking other runners. If the runners get into a fist fight, they will all finish late.
Proprietary and secret software is the moral equivalent of runners in a fist fight. Sad to say, the only referee we've got does not seem to object to fights; he just regulates them ("For every ten yards you run, you can fire one shot"). He really ought to break them up, and penalize runners for even trying to fight.
"Won't everyone stop programming without a monetary incentive?"
Actually, many people will program with absolutely no monetary incentive. Programming has an irresistible fascination for some people, usually the people who are best at it. There is no shortage of professional musicians who keep at it even though they have no hope of making a living that way.
But really this question, though commonly asked, is not appropriate to the situation. Pay for programmers will not disappear, only become less. So the right question is, will anyone program with a reduced monetary incentive? My experience shows that they will.
For more than ten years, many of the world's best programmers worked at the Artificial Intelligence Lab for far less money than they could have had anywhere else. They got many kinds of non-monetary rewards: fame and appreciation, for example. And creativity is also fun, a reward in itself. Then most of them left when offered a chance to do the same interesting work for a lot of money.
What the facts show is that people will program for reasons other than riches; but if given a chance to make a lot of money as well, they will come to expect and demand it. Low-paying organizations do poorly in competition with high-paying ones, but they do not have to do badly if the high-paying ones are banned.
"We need the programmers desperately. If they demand that we stop helping our neighbors, we have to obey."
You're never so desperate that you have to obey this sort of demand. Remember: millions for defense, but not a cent for tribute!
"Programmers need to make a living somehow."
In the short run, this is true. However, there are plenty of ways that programmers could make a living without selling the right to use a program. This way is customary now because it brings programmers and businessmen the most money, not because it is the only way to make a living. It is easy to find other ways if you want to find them. Here are a number of examples.
A manufacturer introducing a new computer will pay for the porting of operating systems onto the new hardware.
The sale of teaching, hand-holding, and maintenance services could also employ programmers.
People with new ideas could distribute programs as freeware and ask for donations from satisfied users or sell hand-holding services. I have met people who are already working this way successfully.
Users with related needs can form users' groups and pay dues. A group would contract with programming companies to write programs that the group's members would like to use.
All sorts of development can be funded with a Software Tax:
Suppose everyone who buys a computer has to pay a certain percent of the price as a software tax. The government gives this to an agency like the NSF to spend on software development.But if the computer buyer makes a donation to software development himself, he can take a credit against the tax. He can donate to the project of his own choosing--often, chosen because he hopes to use the results when it is done. He can take a credit for any amount of donation up to the total tax he had to pay.
The total tax rate could be decided by a vote of the payers of the tax, weighted according to the amount they will be taxed on.
The consequences:
- The computer-using community supports software development.
- This community decides what level of support is needed.
- Users who care which projects their share is spent on can choose this for themselves.
In the long run, making programs free is a step toward the post-scarcity world, where nobody will have to work very hard just to make a living. People will be free to devote themselves to activities that are fun, such as programming, after spending the necessary ten hours a week on required tasks such as legislation, family counseling, robot repair, and asteroid prospecting. There will be no need to be able to make a living from programming.
We have already greatly reduced the amount of work that the whole society must do for its actual productivity, but only a little of this has translated itself into leisure for workers because much nonproductive activity is required to accompany productive activity. The main causes of this are bureaucracy and isometric struggles against competition. Free software will greatly reduce these drains in the area of software production. We must do this, in order for technical gains in productivity to translate into less work for us.