GNU Emacs Manual

Seventh Edition, Emacs Version 18.58

for Unix Users

February 1988, revised September 1992

(General Public License upgraded, January 1991)

Richard M. Stallman

Copyright © 1985, 1986, 1988, 1992 Richard M. Stallman.

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.

Cover art by Etienne Suvasa.

Indexes, nodes containing large menus

Important General Concepts

Important Text-Changing Commands

Larger Units of Text

Advanced Features

Recovery from Problems.

Here are some other nodes which are really inferiors of the ones
already listed, mentioned here so you can get to them in one step:

Subnodes of Screen

Subnodes of Basic

Subnodes of Minibuffer

Subnodes of Mark

Subnodes of Yanking

Subnodes of Registers

Subnodes of Display

Subnodes of Search

Subnodes of Fixit

Subnodes of Files

Subnodes of Buffers

Subnodes of Windows

Subnodes of Indentation

Subnodes of Text

Subnodes of Programs

Subnodes of Compiling/Testing

Subnodes of Abbrevs

Subnodes of Picture

Subnodes of Sending Mail

Subnodes of Rmail

Subnodes of Shell

Subnodes of Customization

Subnodes of Lossage (and recovery)

Distribution

GNU Emacs is free; this means that everyone is free to use it and free to redistribute it on a free basis. GNU Emacs 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 distribution version of GNU Emacs from host ‘prep.ai.mit.edu’ using anonymous login. See the file ‘/u2/emacs/GETTING.GNU.SOFTWARE’ 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, our distribution service is not. An order form is included at the end of manuals printed by the Foundation. It is also 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.

GNU GENERAL PUBLIC LICENSE

Version 1, February 1989

Copyright © 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.

Preamble

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.

1. This License Agreement applies to any program or other work which contains a notice placed by the copyright holder saying it may be distributed under the terms of this General Public License. The “Program”, below, refers to any such program or work, and a “work based on the Program” means either the Program or any work containing the Program or a portion of it, either verbatim or with modifications. Each licensee is addressed as “you”.
2. You may copy and distribute verbatim copies of the Program’s source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice and disclaimer of warranty; keep intact all the notices that refer to this General Public License and to the absence of any warranty; and give any other recipients of the Program a copy of this General Public License along with the Program. You may charge a fee for the physical act of transferring a copy.
3. You may modify your copy or copies of the Program or any portion of it, and copy and distribute such modifications under the terms of Paragraph 1 above, provided that you also do the following:
   • cause the modified files to carry prominent notices stating that you changed the files and the date of any change; and
   • cause the whole of any work that you distribute or publish, that in whole or in part contains the Program or any part thereof, either with or without modifications, to be licensed at no charge to all third parties under the terms of this General Public License (except that you may choose to grant warranty protection to some or all third parties, at your option).
   • If the modified program normally reads commands interactively when run, you must cause it, when started running for such interactive use in the simplest and most usual way, to print or display an announcement including an appropriate copyright notice and a notice that there is no warranty (or else, saying that you provide a warranty) and that users may redistribute the program under these conditions, and telling the user how to view a copy of this General Public License.
   • You may charge a fee for the physical act of transferring a copy, and you may at your option offer warranty protection in exchange for a fee.

   Mere aggregation of another independent work with the Program (or its derivative) on a volume of a storage or distribution medium does not bring the other work under the scope of these terms.

4. You may copy and distribute the Program (or a portion or derivative of it, under Paragraph 2) in object code or executable form under the terms of Paragraphs 1 and 2 above provided that you also do one of the following:
   • accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Paragraphs 1 and 2 above; or,
   • accompany it with a written offer, valid for at least three years, to give any third party free (except for a nominal charge for the cost of distribution) a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Paragraphs 1 and 2 above; or,
   • accompany it with the information you received as to where the corresponding source code may be obtained. (This alternative is allowed only for noncommercial distribution and only if you received the program in object code or executable form alone.)

   Source code for a work means the preferred form of the work for making modifications to it. For an executable file, complete source code means all the source code for all modules it contains; but, as a special exception, it need not include source code for modules which are standard libraries that accompany the operating system on which the executable file runs, or for standard header files or definitions files that accompany that operating system.

5. You may not copy, modify, sublicense, distribute or transfer the Program except as expressly provided under this General Public License. Any attempt otherwise to copy, modify, sublicense, distribute or transfer the Program is void, and will automatically terminate your rights to use the Program under this License. However, parties who have received copies, or rights to use copies, from you under this General Public License will not have their licenses terminated so long as such parties remain in full compliance.
6. By copying, distributing or modifying the Program (or any work based on the Program) you indicate your acceptance of this license to do so, and all its terms and conditions.
7. Each time you redistribute the Program (or any work based on the Program), the recipient automatically receives a license from the original licensor to copy, distribute or modify the Program subject to these terms and conditions. You may not impose any further restrictions on the recipients’ exercise of the rights granted herein.
8. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns.

   Each version is given a distinguishing version number. If the Program specifies a version number of the license which applies to it and “any later version”, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of the license, you may choose any version ever published by the Free Software Foundation.

9. If you wish to incorporate parts of the Program into other free programs whose distribution conditions are different, write to the author to ask for permission. For software which is copyrighted by the Free Software Foundation, write to the Free Software Foundation; we sometimes make exceptions for this. Our decision will be guided by the two goals of preserving the free status of all derivatives of our free software and of promoting the sharing and reuse of software generally.
10. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
11. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

How to Apply These Terms to Your New Programs

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 is 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!

Introduction

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 your commands. See section Display.

We call it 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.

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 it to find out what any command does, or to find all the commands that pertain to a topic. See section Help.

Customizable means that you can change the definitions of Emacs commands in little ways. 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, if you prefer the four basic cursor motion commands (up, down, left and right) on keys in a diamond pattern on the keyboard, you can have it. See section Customization.

Extensible means that 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, which means that it is divided into many functions that call each other, any of which can be redefined in the middle of an editing session. Any part of Emacs can be replaced without making a separate copy of all of Emacs. Most of the editing commands of Emacs are written in Lisp already; the few exceptions could have been written in Lisp but are written in C for efficiency. Although only a programmer can write an extension, anybody can use it afterward.

1 The Organization of the Screen

Emacs divides the screen into several areas, each of which contains its own sorts of information. The biggest area, of course, is the one in which you usually see the text you are editing.

When you are using Emacs, the screen is divided into a number of windows. Initially there is one text window occupying all but the last line, plus the special echo area or minibuffer window in the last line. The text window can be subdivided horizontally or vertically into multiple text windows, each of which can be used for a different file (see section Multiple Windows). The window that the cursor is in is the selected window, in which editing takes place. The other windows are just for reference unless you select one of them.

Each text window’s last line is a mode line which describes what is going on in that window. It is in inverse video if the terminal supports that, and contains text that starts like ‘-----Emacs: something’. Its purpose is to indicate what buffer is being displayed above it in the window; what major and minor modes are in use; and whether the buffer’s text has been changed.

1.1 Point

When Emacs is running, the terminal’s cursor shows the location at which editing commands will take effect. This location is called point. Other commands move point through the text, so that you can edit at different places in it.

While the cursor appears to point at a character, point should be thought of as between two characters; it points before the character that the cursor appears on top of. Sometimes people speak of “the cursor” when they mean “point”, or speak of commands that move point as “cursor motion” commands.

Terminals have only one cursor, and when output is in progress it 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 so that it can be seen when you look at that 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. This also is how you can tell 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’.

1.2 The Echo Area

The line at the bottom of the screen (below the mode line) is the echo area. It is used to display small amounts of text for several purposes.

Echoing means printing out the characters that you type. Emacs never echoes single-character commands, and multi-character commands are echoed only if you pause while typing them. As soon as you pause for more than a second in the middle of a command, all the characters of the command so far are echoed. This is intended to prompt you for the rest of the command. Once echoing has started, the rest of the command is echoed immediately when you type it. This behavior is designed to give confident users fast response, while giving hesitant users maximum feedback. You can change this behavior by setting a variable (see section Variables Controlling Display).

If a command cannot be executed, it may print an error message in the echo area. Error messages are accompanied by a beep or by flashing the screen. Also, any input you have typed ahead is thrown away when an error happens.

Some commands print informative messages in the echo area. These messages look much like error messages, but they are not announced with a beep and do not throw away input. Sometimes the message tells you what the command has done, when this is not obvious from looking at the text being edited. Sometimes the sole purpose of a command is to print a message giving you specific information. For example, the command C-x = is used to print a message describing the character position of point in the text and its current column in the window. Commands that take a long time often display messages ending in ‘...’ while they are working, and add ‘done’ at the end when they are finished.

The echo area is also used to display the minibuffer, a window that is used for reading arguments to commands, such as the name of a file to be edited. When the minibuffer is in use, the echo area begins with a prompt string that usually ends with a colon; also, the cursor appears in that line because it is the selected window. You can always get out of the minibuffer by typing C-g. See section The Minibuffer.

1.3 The Mode Line

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’, then 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 ‘--’ 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 one that editing takes place in. When we speak of what some command does to “the buffer”, we are talking about 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 all on the screen, pos is ‘All’. Otherwise, it 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 of the possible major modes. The major modes available include Fundamental mode (the least specialized), Text mode, Lisp mode, and C mode. See section Major Modes, for details of how the modes differ and how to select one.

minor is a list of some of the minor modes that are turned on at the moment in the window’s chosen buffer. ‘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.

In addition, 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 this information pertains 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 user 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.

1.4 The Emacs Character Set

GNU Emacs uses the ASCII character set, which defines 128 different character codes. Some of these codes are assigned graphic symbols such as ‘a’ and ‘=’; the rest are control characters, such as Control-a (also called C-a for short). C-a gets its name from the fact that you type it by holding down the <CTRL> key and then pressing a. There is no distinction between C-a and C-A; they are the same character.

Some control characters have special names, and special keys you can type them with: <RET>, <TAB>, <LFD>, <DEL> and <ESC>. The space character is usually referred to below as <SPC>, even though strictly speaking it is a graphic character whose graphic happens to be blank.

Emacs extends the 7-bit ASCII code to an 8-bit code by adding an extra bit to each character. This makes 256 possible command characters. The additional bit is called Meta. Any ASCII character can be made Meta; examples of Meta characters include Meta-a (M-a, for short), M-A (not the same character as M-a, but those two characters normally have the same meaning in Emacs), M-<RET>, and M-C-a. For traditional reasons, M-C-a is usually called C-M-a; logically speaking, the order in which the modifier keys <CTRL> and <META> are mentioned does not matter.

Some terminals have a <META> key, and allow you to type Meta characters by holding this key down. Thus, Meta-a is typed by holding down <META> and pressing a. The <META> key works much like the <SHIFT> key. Such a key is not always labeled <META>, however, as this function is often a special option for a key with some other primary purpose.

If there is no <META> key, you can still type Meta characters using two-character sequences starting with <ESC>. Thus, 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.

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.

Emacs buffers also use an 8-bit character set, because bytes have 8 bits, but only the ASCII characters are considered meaningful. 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 columns). 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’.

1.5 Keys

A complete key—where ‘key’ is short for key sequence—is a sequence of keystrokes that are understood by Emacs as a unit, as a single command (possibly undefined). Most single characters constitute complete keys in the standard Emacs command set; there are also some multi-character keys. Examples of complete keys are C-a, X, <RET>, C-x C-f and C-x 4 C-f.

A prefix key is a sequence of keystrokes that are the beginning of a complete key, but not a whole one. Prefix keys and complete keys are collectively called keys.

A prefix key is the beginning of a series of longer sequences that are valid keys; adding any single character to the end of the prefix gives a valid key, which could be defined as an Emacs command, or could be a prefix itself. For example, C-x is standardly defined as a prefix, so C-x and the next input character combine to make a two-character key. There are 256 different two-character keys starting with C-x, one for each possible second character. Many of these two-character keys starting with C-x are standardly defined as Emacs commands. Notable examples include C-x C-f and C-x s (see section File Handling).

Adding one character to a prefix key does not have to form a complete key. It could make another, longer prefix. For example, C-x 4 is itself a prefix that leads to 256 different three-character keys, including C-x 4 f, C-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 C-f C-k is not a key, because the C-f is a complete key in itself. It’s impossible to give C-f C-k an independent meaning as a command as long as C-f retains its meaning. C-f C-k is two commands.

All told, the prefix keys in Emacs are C-c, C-x, C-h, C-x 4, and <ESC>. But this is not built in; it is just a matter of Emacs’s standard key bindings. In customizing Emacs, you could make new prefix keys, or eliminate these. See section Customizing Key Bindings.

Whether a sequence is a key can be changed by customization. For example, if you redefine C-f as a prefix, C-f C-k automatically becomes a key (complete, unless you define it too as a prefix). Conversely, if you remove the prefix definition of C-x 4, then C-x 4 f (or C-x 4 anything) is no longer a key.

1.6 Keys and Commands

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 which 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; this is what makes the function do what it does. 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. See Defining Commands in The GNU Emacs Lisp Manual, for more information.

The bindings between keys and functions are recorded in various tables called keymaps. See section Keymaps.

When we say that “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. It is the function next-line that 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.

In the rest of this manual, we usually ignore this subtlety to keep things simple. To give the customizer the information he needs, we state the name of the command which 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 examines the variable and behaves differently accordingly. 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.

2 Entering and Exiting Emacs

The usual way to invoke Emacs is just to type emacs <RET> at the shell. Emacs clears the screen and then displays an initial advisor 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 your 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. But Emacs must always have a current buffer for editing. In an attempt to do something useful, Emacs presents a buffer named ‘*scratch*’ which 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 also possible to specify files to be visited, Lisp files to be loaded, and functions to be called, by giving Emacs arguments in the shell command line. See section Command Line Switches and Arguments.

2.1 Exiting Emacs

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.

C-z

Suspend Emacs (suspend-emacs).

C-x C-c

Kill Emacs (save-buffers-kill-emacs).

To suspend Emacs, type C-z (suspend-emacs). This takes you back to the shell from which you invoked Emacs. You can resume Emacs with the command %emacs if you are using the C shell or the Bourne-Again shell.

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. The only way on these systems to get back to the shell from which Emacs was run (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. Unless a numeric argument is used, this command first offers to save any modified buffers. If you do not save them all, it asks for reconfirmation with yes before killing Emacs, since any changes not saved before that will be lost forever. Also, if any subprocesses are still running, C-x C-c asks for confirmation about them, since killing Emacs will kill the subprocesses immediately.

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).

2.2 Command Line Switches and Arguments

GNU Emacs supports command line arguments to request various actions when invoking Emacs. These are for compatibility with other editors and for sophisticated activities. They are not needed for ordinary editing with Emacs, so new users can skip this section.

You may be used to using command line arguments with other editors to specify which file to edit. That’s because many other editors are designed to be started afresh each time you want to edit. You edit one file and then exit the editor. The next time you want to edit either another file or the same one, you must run the editor again. With these editors, 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 different file, you visit it with the existing Emacs, which eventually comes to have many files in it ready for editing. Usually you do not kill the Emacs until you are about to log out.

In the usual style of Emacs use, files are nearly always read by typing commands to an editor that is already running. So command line arguments for specifying a file when the editor is started are seldom used.

Emacs accepts command-line arguments that specify files to visit, functions to call, and other activities and operating modes.

The command arguments are processed in the order they appear in the command argument list; however, certain arguments (the ones in the second table) must be at the front of the list if they are used.

Here are the arguments allowed:

‘file’

Visit file using find-file. See section Visiting Files.

‘+linenum file’

Visit file using find-file, then go to line number linenum in it.

‘-l file’

‘-load file’

Load a file file of Lisp code with the function load. See section Libraries of Lisp Code for Emacs.

‘-f function’

‘-funcall function’

Call Lisp function function with no arguments.

‘-i file’

‘-insert file’

Insert the contents of file into the current buffer. This is like what M-x insert-buffer does; see Miscellaneous File Operations.

‘-kill’

Exit from Emacs without asking for confirmation.

The following switches are recognized only at the beginning of the command line. If more than one of them appears, they must appear in the order that they appear in this table.

‘-t device’

Use device as the device for terminal input and output.

‘-d display’

When running with the X window system, use the display named display to make Emacs’s X window.

‘-nw’

Don’t use a window system; display text only, using an ordinary terminal device. Thus, if you run an X-capable Emacs in an Xterm with ‘emacs -nw’, it displays in the Xterm’s own window instead of making its own.

‘-batch’

Run Emacs in batch mode, which means that the text being edited is not displayed and the standard Unix interrupt characters such as C-z and C-c continue to have their normal effect. Emacs in batch mode outputs to stdout 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 exit after all command switches have been processed. In addition, auto-saving is not done except in buffers for which it has been explicitly requested.

‘-q’

‘-no-init-file’

Do not load your Emacs init file ‘~/.emacs’.

‘-u user’

‘-user user’

Load user’s Emacs init file ‘~user/.emacs’ instead of your own.

With X Windows, you can use these additional options to specify how to display the window. Each option has a corresponding resource name (used with ‘emacs’ unless you specify another name with ‘-rn name’), listed with the option, which lets you specify the same parameter using the usual X Windows defaulting mechanism. The corresponding generic resource name (used with ‘Emacs’) is usually made by capitalizing the first letter of the individual resource name, but in some cases it is a completely different string (which is listed below).

-rn name

Use name instead of ‘emacs’ when looking for X resources.

-font fontname

-fn fontname

Use font fontname.
(Resource ‘font’.)

-wn name

Name the window name.
(Resource ‘title’.)

-i

Use a bitmap icon (showing the kitchen sink) rather than a textual icon.
(Resource ‘bitmapIcon’.)

-in name

Name the icon name. (Resource ‘iconName’; ‘Title’).

-geometry coords

-w coords

Specify the shape and optionally the position of the Emacs window in the usual X way.
(Resource ‘geometry’.)

-b width

Specify that the window border is width pixels thick.
(Resource ‘borderWidth’.)

-ib width

Leave width blank pixels between the border and the window contents.
(Resource ‘internalBorder’; ‘BorderWidth’.)

-r

Use reverse video.
(Resource ‘reverseVideo’.)

-fg color

Use color color for text in the window.
(Resource ‘foreground’.)

-bg color

Use the color color for the background of the window.
(Resource ‘background’.)

-bd color

Use color color for the window border.
(Resource ‘borderColor’.)

-cr color

Specify the color, color, to use for the cursor.
(Resource ‘cursorColor’; ‘Foreground’.)

-ms color

Use color color for the mouse cursor.
(Resource ‘pointerColor’; ‘Foreground’.)

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 arguments 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 current buffer, expected to be a C program.

emacs -batch foo.c -l hack-c-program -f save-buffer -kill > log

This says 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 that this command was done with. ‘-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.

3 Basic Editing Commands

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).

3.1 Inserting Text

To insert printing characters into the text you are editing, just type them. This inserts the character 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’, then if you type XX, you get ‘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 will split lines automatically when they become too long, if you turn on a special mode called Auto Fill mode. See section Filling Text, for how to use Auto Fill mode.

Customization information: <DEL> in most modes runs the command named 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:

• Control-q followed by any non-graphic character (even C-g) inserts that character.
• Control-q followed by three octal digits inserts the character with the specified character code.

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 rather than shove it to the right, you can enable Overwrite mode, a minor mode. See section Minor Modes.

3.2 Changing the Location of Point

To do more than insert characters, you have to know how to move point (see section Point). Here are a few of the commands for doing that.

C-a

Move to the beginning of the line (beginning-of-line).

C-e

Move to the end of the line (end-of-line).

C-f

Move forward one character (forward-char).

C-b

Move backward one character (backward-char).

M-f

Move forward one word (forward-word).

M-b

Move backward one word (backward-word).

C-n

Move down one line, vertically (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.

C-p

Move up one line, vertically (previous-line).

C-l

Clear the screen and reprint everything (recenter). Text moves on the screen to bring point to the center of the window.

M-r

Move point to left margin on the line halfway down the screen or 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).

C-t

Transpose two characters, the ones before and after the cursor (transpose-chars).

M-<

Move to the top of the buffer (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.

M->

Move to the end of the buffer (end-of-buffer).

M-x goto-char

Read a number n and move cursor to character number n. Position 1 is the beginning of the buffer.

M-x goto-line

Read a number n and move cursor to line number n. Line 1 is the beginning of the buffer.

C-x C-n

Use the current column of point as the semipermanent goal column for C-n and C-p (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.

C-u C-x C-n

Cancel the goal column. Henceforth, C-n and C-p once again try to avoid changing the horizontal position, as usual.

If you set the variable track-eol to a non-nil value, then C-n and C-p when at the end of the starting line move to the end of the line. Normally, track-eol is nil.

3.3 Erasing Text

<DEL>

Delete the character before the cursor (delete-backward-char).

C-d

Delete the character after the cursor (delete-char).

C-k

Kill to the end of the line (kill-line).

M-d

Kill forward to the end of the next word (kill-word).

M-<DEL>

Kill back to the beginning of the previous word (backward-kill-word).

You already know about the <DEL> key which deletes the character before the cursor. Another key, 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 together.

To erase a larger amount of text, use the Control-k key, which kills a line at a time. If C-k is done at the beginning or middle of a line, it kills all the text up to the end of the line. If C-k is done 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.

3.4 Files

The commands above are sufficient for creating and altering text in an Emacs buffer; the 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 for you to retrieve later 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’. In Emacs, to begin editing this file, type

C-x C-f /usr/rms/foo.c <RET>

Here the file name is given as an argument to the command C-x C-f (find-file). That command uses the minibuffer to read the argument, and you type <RET> to terminate the argument (see section The Minibuffer).

Emacs obeys the command by visiting the file: creating a buffer, copying the contents of the file into the buffer, and then displaying the buffer for you to edit. You can make changes in it, and then save the file by typing C-x C-s (save-buffer). 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, and the file ‘foo.c’ is not changed.

To create a file, just visit the file with C-x C-f as if it already existed. Emacs will make an empty buffer in which you can insert the text you want to put in the file. When you save your text with C-x C-s, the file will be created.

Of course, there is a lot more to learn about using files. See section File Handling.

3.5 Help

If you forget what a key does, you can find out with the Help character, which is C-h. Type C-h k followed by the key you want to know about; for example, C-h k C-n tells you all about 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.

3.6 Blank Lines

Here are special commands and techniques for putting in and taking out blank lines.

C-o

Insert one or more blank lines after the cursor (open-line).

C-x C-o

Delete all but one of many consecutive blank lines (delete-blank-lines).

When you want to insert a new line of text before an existing line, you can do it by typing the new line of text, followed by <RET>. However, it may be easier to see what you are doing if you first make a blank line and then insert the desired text into it. This is easy to do using the key C-o (open-line), which inserts a newline after point but leaves point in front of the newline. After C-o, type the text for the new line. C-o F O O has the same effect as F O O <RET>, except for the final location of point.

You can make several blank lines by typing C-o several times, or by giving it an argument to tell it how many blank lines to make. See section Numeric Arguments, for how.

If you have many blank lines in a row and want to get rid of them, use C-x C-o (delete-blank-lines). When 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 consecutive blank lines, leaving exactly one. With point on a blank line with no other blank line adjacent to it, the sole blank line is deleted, leaving none. When point is on a nonblank line, C-x C-o deletes any blank lines following that nonblank line.

3.7 Continuation Lines

If you add too many characters to one line, without breaking it with a <RET>, the line will grow to occupy two (or more) lines on the screen, with a ‘\’ at the extreme right margin of all but the last of them. The ‘\’ says 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. Sometimes it is nice to have Emacs insert newlines automatically when a line gets too long; for this, use Auto Fill mode (see section Filling Text).

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. ‘$’ is used in the last column instead of ‘\’ to inform you that truncation is in effect.

Continuation can be turned off for a particular buffer by setting 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.

3.8 Cursor Position Information

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. This is because the text is stored in a way that makes it difficult to compute this information. Displaying them all the time would be intolerably slow. They are not needed very often in Emacs anyway, but there are commands to compute them and print them.

M-x what-page

Print page number of point, and line number within page.

M-x what-line

Print line number of point in the buffer.

M-=

Print number of lines in the current region (count-lines-region).

C-x =

Print character code of character after point, character position of point, and column of point (what-cursor-position).

There are two commands for printing line numbers. M-x what-line counts lines from the beginning of the file and prints the line number point is on. The first line of the file is line number 1. These numbers can be used as arguments to M-x goto-line. By contrast, M-x what-page counts pages from the beginning of the file, and counts lines within the page, printing both of them. See section Pages.

While on this subject, we might as well mention M-= (count-lines-region), which prints the number of lines in the region (see section The Mark and the Region). See section Pages, for the command C-x l which counts the lines in the current page.

The command C-x = (what-cursor-position) can be used to find out the column that the cursor is in, and other miscellaneous information about point. It prints a line in the echo area that looks like this:

Char: x (0170)  point=65986 of 563027(12%)  x=44

(In fact, this is the output produced when point is before the ‘x=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.

‘x=’ 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>  x=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%)  x=0

3.9 Numeric Arguments

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 terminal 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 <META> key. For example,

M-5 C-n

would move 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 nonblank 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 for reasons of convenience of use of the individual command.

To insert multiple copies of a digit, you can type C-u count C-u digit. The second C-u ends the numeric argument, so that the following character always acts a key sequence to be executed.

4 Undoing Changes

Emacs allows all changes made in the text of a buffer to be undone, 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.

C-x u

Undo one batch of changes (usually, one command worth) (undo).

C-_

The same.

The command C-x u or C-_ is how you undo. 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, to redo changes you have undone, type C-f or any other command that will have no important effect, and then give 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. At this time, all the modifications you made have been cancelled. If you do not remember whether you changed the buffer deliberately, type C-_ once, and when you see the last change you made undone, you will remember why you made it. If it 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, it means that the buffer contents are the same as they were 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. Also, minibuffers, help buffers and documentation buffers don’t record undo information. Use the command buffer-enable-undo to enable recording undo information in the current buffer.

As editing continues, undo lists get longer and longer. To prevent them from using up all available memory space, garbage collection trims back their sizes to thresholds you can set. (For this purpose, the “size” of an undo list measures the cons cells that make up the list, plus the strings of deleted text.)

Two variables control the range of acceptable sizes: undo-limit and undo-strong-limit. Normally, the most recent changes are kept until their size exceeds undo-limit; all older changes are discarded. But if a change pushes the size above undo-strong-limit, it is discarded as well as all older changes. One exception: the most recent set of changes is sacred; garbage collection never discards that. (In Emacs versions 18.57 and 18.58, these variables are called undo-threshold and undo-high-threshold.)

The reason the undo command has two keys, C-x u and C-_, set up to run it is that it is worthy of a single-character key, but the way to type C-_ on some keyboards is not obvious. C-x u is an alternative you can type in the same fashion on any terminal.

5 The Minibuffer

The minibuffer is the facility used by Emacs commands 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. The usual Emacs editing commands can be used in the minibuffer to edit the argument.

When the minibuffer is in use, it appears in the echo area, and the terminal’s cursor moves there. The beginning of the minibuffer line displays a prompt which says what kind of input you should supply and how it will be used. Often this prompt is derived from the name of the command that 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 will be used as the argument value if you enter an empty argument (e.g., just type <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> which exits the minibuffer. You can get out of the minibuffer, canceling the command that it was for, by typing 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:

• If a command gets an error while you are in the minibuffer, this does not cancel the minibuffer. However, the echo area is needed for the error message and therefore the minibuffer itself is hidden for a while. It comes back after a few seconds, or as soon as you type anything.
• If in the minibuffer you use a command whose purpose is to print a message in the echo area, such as C-x =, the message is printed normally, and the minibuffer is hidden for a while. It comes back after a few seconds, or as soon as you type anything.
• Echoing of keystrokes does not take place while the minibuffer is in use.

5.1 Minibuffers for File Names

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 is to inform you which directory the file will be found in if you do not specify a directory. 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 kill with M-<DEL> the 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.

5.2 Editing in the Minibuffer

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, inserting a newline into the minibuffer must be done with C-o or with C-q <LFD>. (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. When the minibuffer is in use, its 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 some restrictions on the use of the minibuffer window, however. You cannot switch buffers in it—the minibuffer and its window are permanently attached. Also, you cannot split or kill the minibuffer window. But you can make it taller in the normal fashion with C-x ^ (see section Deleting and Rearranging Windows).

If while in the minibuffer you issue a command that displays help text of any sort in another window, then that window is identified as the one to scroll if you type C-M-v while in the minibuffer. This lasts until you exit the minibuffer. This feature comes into play if a completing minibuffer gives you a list of possible completions.

Recursive use of the minibuffer is supported by Emacs. 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.

5.3 Completion

When appropriate, the minibuffer provides a completion facility. This means that you type enough of the argument to determine the rest, based on Emacs’s knowledge of which arguments make sense, and Emacs visibly fills in the rest, or as much as can be determined from the part you have 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 the program said to do so.

5.3.1 Completion Example

A concrete example may help here. If you type Meta-x au <TAB>, the <TAB> looks for alternatives (in this case, command names) that start with ‘au’. There are only two: auto-fill-mode and auto-save-mode. These 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. 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.

5.3.2 Completion Commands

Here is a list of all the completion commands, defined in the minibuffer when completion is available.

<TAB>

Complete the text in the minibuffer as much as possible
(minibuffer-complete).

<SPC>

Complete the text in the minibuffer but don’t add or fill out more than one word (minibuffer-complete-word).

<RET>

Submit the text in the minibuffer as the argument, possibly completing first as described below (minibuffer-complete-and-exit).

?

Print a list of all possible completions of the text in the minibuffer (minibuffer-list-completions).

<SPC> completes much like <TAB>, but 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-’. This gives ‘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.

• Strict completion is used when it is meaningless to give any argument except one of the known alternatives. For example, when C-x k reads the name of a buffer to kill, it is meaningless to give anything but the name of an existing buffer. In strict completion, <RET> refuses to exit if the text in the minibuffer does not complete to an exact match.
• Cautious completion is similar to strict completion, except that <RET> exits only if the text was an exact match already, not needing completion. If the text is not an exact match, <RET> does not exit, but it does complete the text. If it completes to an exact match, a second <RET> will exit.

  Cautious completion is used for reading file names for files that must already exist.

• Permissive completion is used when any string whatever is meaningful, and the list of completion alternatives is just a guide. For example, when C-x C-f reads the name of a file to visit, any file name is allowed, in case you want to create a file. In permissive completion, <RET> takes the text in the minibuffer exactly as given, without completing it.

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. Also, 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’ can complete to ‘foo.c’ even though ‘foo.o’ exists as well. If the only possible completions are files that end in “ignored” strings, then they are not ignored.

Normally, a completion command that finds the next character is undetermined 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.

5.4 Repeating Minibuffer Commands

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.

C-x <ESC>

Re-execute a recent minibuffer command repeat-complex-command).

M-p

Within C-x <ESC>, move to the previous recorded command
(previous-complex-command).

M-n

Within C-x <ESC>, move to the next (more recent) recorded command (next-complex-command).

M-x list-command-history

Display the entire command history, showing all the commands C-x <ESC> can repeat, most recent first.

C-x <ESC> is used to re-execute a recent minibuffer-using command. With no argument, it repeats the last such command. A numeric argument specifies which command to repeat; 1 means the last one, and larger numbers specify earlier ones.

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. Whatever expression you finally submit is what 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 it will repeat exactly as before.

Once inside the minibuffer for C-x <ESC>, if 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 them with the next later command. After finding the desired previous command, you can edit its expression as usual and then resubmit it by typing <RET> as usual. Any editing you have done on the command to be repeated is lost if you use M-n or M-p.

M-p is more useful than M-n, since more often you will initially request to repeat the most recent command and then decide to repeat an older one instead. These keys are specially defined within C-x <ESC> to run the commands previous-complex-command and next-complex-command.

The list of previous minibuffer-using commands is stored as a Lisp list in the variable command-history. Each element 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.

6 Running Commands by Name

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 do not need to be brief are not bound to keys; to run them, you must refer to them by name.

A command name is, by convention, made up 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 is more characters to type. Any command can be run by name, even if it is also runnable by keys.

The way to run a command by name is to start with M-x, type the command name, and finish it 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 the command name. For example, the command forward-char can be invoked by name by typing

M-x forward-char <RET>

or

M-x fo <TAB> c <RET>

Note that forward-char is the same command that you invoke with the key C-f. Any command (interactively callable function) defined in Emacs can be called by its name using M-x whether or not any keys are bound to it.

If you type C-g while the command name is being read, you cancel the M-x command and get out of the minibuffer, ending up at top level.

To pass a numeric argument to the command you are invoking with M-x, specify the numeric argument before the M-x. M-x passes the argument along to the function which it calls. The argument value appears in the prompt while the command name is being read.

Normally, when describing a command that is run by name, we omit the <RET> that is needed to terminate the name. Thus we might speak of M-x auto-fill-mode rather than M-x auto-fill-mode <RET>. We mention the <RET> only when there is a need to emphasize its presence, such as 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.

7 Help

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 that you can type after C-h are called help options. One help option is C-h; that is how you ask for help about using C-h.

C-h C-h prints a list of the possible help options, and then asks you to go ahead and type the option. It prompts with a string

A B C F I K L M N S T V W C-c C-d C-n C-w.  Type C-h again for more help:

and you should type one of those characters.

Typing a third C-h displays a description of what the options mean; it still waits for you to type an option. To cancel, type C-g.

Here is a summary of the defined help commands.

C-h a string <RET>

Display a list of commands whose names contain string
(command-apropos).

C-h b

Display a table of all key bindings in effect now; local bindings of the current major mode first, followed by all global bindings (describe-bindings).

C-h c key

Print the name of the command that key runs (describe-key-briefly). c is for ‘character’. For more extensive information on key, use C-h k.

C-h f function <RET>

Display documentation on the Lisp function named function (describe-function). Note that commands are Lisp functions, so a command name may be used.

C-h i

Run Info, the program for browsing documentation files (info). The complete Emacs manual is available on-line in Info.

C-h k key

Display name and documentation of the command key runs (describe-key).

C-h l

Display a description of the last 100 characters you typed (view-lossage).

C-h m

Display documentation of the current major mode (describe-mode).

C-h n

Display documentation of Emacs changes, most recent first (view-emacs-news).

C-h s

Display current contents of the syntax table, plus an explanation of what they mean (describe-syntax).

C-h t

Display the Emacs tutorial (help-with-tutorial).

C-h v var <RET>

Display the documentation of the Lisp variable var (describe-variable).

C-h w command <RET>

Print which keys run the command named command (where-is).

7.1 Documentation for a Key

The most basic C-h options are C-h c (describe-key-briefly) and C-h k (describe-key). C-h c key prints in the echo area the name of the command that key is bound to. For example, C-h c C-f prints ‘forward-char’. Since command names are chosen to describe what the command does, this is a good way to get a very brief description of what key does.

C-h k key is similar but gives more information. It displays the documentation string of the command key is bound to as well as its name. This is too big for the echo area, so a window is used for the display.

7.2 Help by Command or Variable Name

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 this to get the documentation of a command that is known by name. For example,

C-h f auto-fill-mode <RET>

displays the documentation of auto-fill-mode. This is the only way to see the documentation of a command that is not bound to any key (one which you would normally call using M-x).

C-h f is also useful for Lisp functions that you are planning to use in a Lisp program. For example, if you have just written the code (make-vector len) and want to be sure that 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.

The function name for C-h f to describe has a default which is used if you type <RET> leaving the minibuffer empty. The default is the function called by the innermost Lisp expression in the buffer around point, provided 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 that starts 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 tells you what you want to know, just type C-g to cancel the C-h f command and go on editing.

C-h w command <RET> tells you what keys are bound to command. It prints a list of the keys in the echo area. Alternatively, it says that the command is not on any keys, which implies that you must use M-x to call it.

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, but only if that is the name of a known Lisp variable. See section Variables.

7.3 Apropos

A more sophisticated sort of question to ask is, “What are the commands for working with files?” For 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 appears a brief description of how to use the command, and what keys you can currently invoke it with. For example, it would say 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 which 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, 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 the standard Emacs commands. By giving you a feel for the naming conventions, this set should also serve to aid you in developing 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.

7.4 Other Help Commands

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 in. If you see commands that you don’t know, you can use C-h c to find out what they do.

Emacs has several major modes, each of which 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 other information about the current Emacs mode. C-h b displays a list of all the key bindings now in effect; 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 displays the full 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.

8 The Mark and the Region

There are many Emacs commands which operate on an arbitrary contiguous part of the current buffer. To specify the text for such a command to operate on, you set the mark at one end of it, and move point to the other end. The text between point and the mark is called the region. You can move point or the mark to adjust the boundaries of the region. It doesn’t matter which one is set first chronologically, or which one comes earlier in the text.

Once the mark has been set, it remains until it is set again at another place. The mark remains fixed with respect to the preceding character if text is inserted or deleted in the buffer. Each Emacs buffer has its own mark, so that when you return to a buffer that had been selected previously, it has the same mark it had before.

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 also useful for remembering 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.

8.1 Setting the Mark

Here are some commands for setting the mark:

C-<SPC>

Set the mark where point is (set-mark-command).

C-@

The same.

C-x C-x

Interchange mark and point (exchange-point-and-mark).

For example, if you wish 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. You can first go to the beginning of the text to be capitalized, type C-<SPC> to put the mark there, move to the end, and then type C-x C-u. Or, you can 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 sets the mark where point is. Then you can 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>, what you get on most terminals is the character C-@. This is the key 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 might as well think of this character as C-<SPC>.

Since terminals have only one cursor, there is no way for Emacs to show you where the mark is located. You have to remember. The usual solution to this problem is to set the mark and then use it soon, before you 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.

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.

8.2 Operating on the Region

Once you have created an active region, you can do many things to the text in it:

• Kill it with C-w (see section Deletion and Killing).
• Save it in a register with C-x x (see section Registers).
• Save it in a buffer or a file (see section Accumulating Text).
• Convert case with C-x C-l or C-x C-u
  (see section Case Conversion Commands).
• Evaluate it as Lisp code with M-x eval-region (see section Evaluating Emacs-Lisp Expressions).
• Fill it as text with M-g (see section Filling Text).
• Print hardcopy with M-x print-region (see section Hardcopy Output).
• Indent it with C-x <TAB> or C-M-\
  (see section Indentation).

8.3 Commands to Mark Textual Objects

There are commands for placing point and the mark around a textual object such as a word, list, paragraph or page.

M-@

Set mark after end of next word (mark-word). This command and the following one do not move point.

C-M-@

Set mark after end of next Lisp expression (mark-sexp).

M-h

Put region around current paragraph (mark-paragraph).

C-M-h

Put region around current Lisp defun (mark-defun).

C-x h

Put region around entire buffer (mark-whole-buffer).

C-x C-p

Put region around current page (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 allow you to save a little typing or redisplay, sometimes.

Other commands set both point and mark, to 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). M-h does all that’s necessary if you wish to indent, case-convert, or kill a whole paragraph. C-M-h (mark-defun) similarly 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, according to 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.

8.4 The Mark Ring

Aside from delimiting the region, the mark is also useful for remembering 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. It moves point to where the mark was, and restores the mark from the ring of former marks. So repeated use of this command moves point to all of the old marks on the ring, one by one. The marks you see 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 is to make it easier for you to move back later. Searches do this except when they do not actually move point. You can tell when a command sets the mark because ‘Mark Set’ is printed in the echo area.

Another way of remembering positions so you can go back to them is with registers (see section Saving Positions in Registers).

The variable mark-ring-max is the maximum number of entries to keep in the mark ring. If that many entries exist and another one is pushed, the last one in the list is discarded. Repeating C-u C-<SPC> circulates through the limited number of 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.

8.5 Deletion and Killing

Most commands which erase text from the buffer save it so that you can get it back if you change your mind, or move or copy it to other parts of the buffer. These commands are known as kill commands. The rest of the commands that erase text do not save it; they are known as delete commands. (This distinction is made only for erasure of text in the buffer.)

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 generally kill. The commands’ names and individual descriptions use the words ‘kill’ and ‘delete’ to say which they do. If you do a kill or delete command by mistake, you can use the C-x u (undo) command to undo it (see section Undoing Changes).

8.5.1 Deletion

C-d

Delete next character (delete-char).

<DEL>

Delete previous character (delete-backward-char).

M-\

Delete spaces and tabs around point (delete-horizontal-space).

M-<SPC>

Delete spaces and tabs around point, leaving one space (just-one-space).

C-x C-o

Delete blank lines around the current line (delete-blank-lines).

M-^

Join two lines by deleting the intervening newline, and any indentation following it (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. Newlines can be deleted like any other characters in the buffer; deleting a newline joins two lines. Actually, C-d and <DEL> aren’t always delete commands; if given an argument, they kill instead, since they can erase more than one character this way.

The other delete commands are those which delete only formatting characters: spaces, tabs and newlines. M-\ (delete-horizontal-space) deletes all the spaces and tab characters before and after point. M-<SPC> (just-one-space) does likewise 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, and if the current line is blank deletes all blank lines preceding the current line as well (leaving one blank line, the current line). M-^ (delete-indentation) joins the current line and the previous line, or the current line and the next line if given an argument, by deleting a newline and all surrounding spaces, possibly leaving a single space. See section M-^.

8.5.2 Killing by Lines

C-k

Kill rest of line or one or more lines (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 it blank. If given on a blank line, the blank line disappears. As a consequence, if you go to the front of a non-blank line and type C-k twice, the line disappears completely.

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. Invisible spaces and tabs at the end of the line are ignored when deciding which case applies, so if point appears to be at the end of the line, you can be sure the newline will be killed.

If C-k is given a positive argument, it kills that many lines and the newlines that follow them (however, text on the current line before point is spared). With a negative argument, it 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.

8.5.3 Other Kill Commands

C-w

Kill region (from point to the mark) (kill-region). See section Words.

M-d

Kill word (kill-word).

M-<DEL>

Kill word backwards (backward-kill-word).

C-x <DEL>

Kill back to beginning of sentence (backward-kill-sentence). See section Sentences.

M-k

Kill to end of sentence (kill-sentence).

C-M-k

Kill sexp (kill-sexp). See section Lists and Sexps.

M-z char

Kill up to next occurrence of char (zap-to-char).

A kill command which is very general is C-w (kill-region), which kills everything between point and the mark. With this command, you can kill any contiguous sequence of characters, if you first set the mark at one end of them and go to the other end.

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).

8.6 Yanking

Yanking is getting back text which was killed. This is what some systems call “pasting”. The usual way to move or copy text is to kill it and then yank it one or more times.

C-y

Yank last killed text (yank).

M-y

Replace re-inserted killed text with the previously killed text (yank-pop).

M-w

Save region as last killed text without actually killing it (copy-region-as-kill).

C-M-w

Append next kill to last batch of killed text (append-next-kill).

8.6.1 The Kill Ring

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.)

The command C-y (yank) reinserts the text of the most recent kill. It leaves the cursor at the end of the text. It sets the mark at the beginning of the text. See section The Mark and the Region.

C-u C-y leaves the cursor in front of the text, and sets the mark after it. This is only if the argument is specified with just a C-u, precisely. Any other sort of argument, including C-u and digits, has an effect described below (under “Yanking Earlier Kills”).

If you wish to copy a block of text, you might want to use M-w (copy-region-as-kill), which copies the region into the kill ring without removing it from the buffer. This is approximately equivalent to C-w followed by C-y, except that M-w does not mark the buffer as “modified” and does not temporarily change the screen.

8.6.2 Appending Kills

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 gets it all back as it was before it was killed. This means that you don’t have to kill all the text in one command; you can keep killing line after line, or word after word, until you have killed it all, and you can still get it all back at once. (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 first<RET>line of sample text<RET>’—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. But you can force it to append by first typing the command C-M-w (append-next-kill) in front of it. The 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.

8.6.3 Yanking Earlier Kills

To recover killed text that is no longer the most recent kill, you need the Meta-y (yank-pop) command. M-y can be used 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. So, to recover the text of the next-to-the-last kill, you first use C-y to recover the last kill, and then use 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 newly made 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 it 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.

M-y can take a numeric argument, which tells it how many items to advance the “last yank” pointer by. 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 it will stay there. It’s just a copy of the kill ring item, so editing it in the buffer does not change what’s in the ring. As long as no new killing is done, the “last yank” pointer remains at the same place in the kill ring, so 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 length of the kill ring is controlled by the variable kill-ring-max; no more than that many blocks of killed text are saved.

8.7 Accumulating Text

Usually we 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.

You can accumulate blocks of text from scattered locations either into a buffer or into a file if you like. These commands are described here. You can also use Emacs registers for storing and accumulating text. See section Registers.

C-x a

Append region to contents of specified buffer (append-to-buffer).

M-x prepend-to-buffer

Prepend region to contents of specified buffer.

M-x copy-to-buffer

Copy region into specified buffer, deleting that buffer’s old contents.

M-x insert-buffer

Insert contents of specified buffer into current buffer at point.

M-x append-to-file

Append region to contents of specified file, at the end.

To accumulate text into a buffer, use the command C-x a buffername (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 that name, one is created. If you append text into a buffer which has been used for editing, the copied text goes into the middle of the text of the buffer, wherever point happens to be in it.

Point in that buffer is left at the end of the copied text, so successive uses of C-x a accumulate the text in the specified buffer in the same order as they were copied. Strictly speaking, C-x a does not always append to the text already in the buffer; but if C-x a 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 just like C-x a except that 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; this too 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 C-x a. 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 text inside Emacs for that file will not change while the file itself changes.

8.8 Rectangles

The rectangle commands affect rectangular areas of the text: all the 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, such as perhaps code with comments at the right, or for changing text into or out of such formats.

When you must specify a rectangle for a command to work on, you do it by putting the mark at one corner and point at the opposite corner. The rectangle thus specified is called the region-rectangle because it is controlled about the same way the region is controlled. But 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.

M-x delete-rectangle

Delete the text of the region-rectangle, moving any following text on each line leftward to the left edge of the region-rectangle.

M-x kill-rectangle

Similar, but also save the contents of the region-rectangle as the “last killed rectangle”.

M-x yank-rectangle

Yank the last killed rectangle with its upper left corner at point.

M-x open-rectangle

Insert blank space to fill the space of the region-rectangle. The previous contents of the region-rectangle are pushed rightward.

M-x clear-rectangle

Clear the region-rectangle by replacing its contents with spaces.

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 can only record the most recent rectangle killed. This is because yanking a rectangle is so different from yanking linear text that different yank commands have to be used and yank-popping is hard to make sense of.

Inserting a rectangle is the opposite of deleting one. All you need to specify is where to put the upper left corner; that is done by putting point there. The rectangle’s first line is inserted there, 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 to blank out existing text, and M-x open-rectangle to insert a blank rectangle. Clearing a rectangle is equivalent to deleting it and then inserting as blank rectangle of the same size.

Rectangles can also be copied into and out of registers. See section Rectangle Registers.

9 Registers

Emacs registers are places 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 or 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 <RET> r

Display a description of what register r contains.

M-x view-register reads a register name as an argument and then displays the contents of the specified register.

9.1 Saving Positions in Registers

Saving a position records a spot in a buffer so that you can move back there later. Moving to a saved position reselects the buffer and moves point to the spot.

C-x / r

Save location of point in register r (point-to-register).

C-x j r

Jump to the location saved in register r (register-to-point).

To save the current location of point in a register, choose a name r and type C-x / r. The register r retains the location thus saved until you store something else in that register.

The command C-x 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 any number of times.

9.2 Saving Text in Registers

When you want to insert a copy of the same piece of text frequently, it may 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 what argument is needed to retrieve the same text with C-y. An alternative is to store the text in a register with C-x x (copy-to-register) and then retrieve it with C-x g (insert-register).

C-x x r

Copy region into register r (copy-to-register).

C-x g r

Insert text contents of register r (insert-register).

C-x x r stores a copy of the text of the region into the register named r. Given a numeric argument, C-x x deletes the text from the buffer as well.

C-x g r inserts in the buffer the text from register r. Normally it leaves point before the text and places the mark after, but with a numeric argument it puts point after the text and the mark before.

9.3 Saving Rectangles in Registers

A register can contain a rectangle instead of linear text. The rectangle is represented as a list of strings. See section Rectangles, for basic information on rectangles and how rectangles in the buffer are specified.

C-x r r

Copy the region-rectangle into register r (copy-region-to-rectangle). With numeric argument, delete it as well.

C-x g r

Insert the rectangle stored in register r (if it contains a rectangle) (insert-register).

The C-x g command inserts linear text if the register contains that, or inserts a rectangle if the register contains one.

10 Controlling the Display

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.

C-l

Clear screen and redisplay, scrolling the selected window to center point vertically within it (recenter).

C-v

Scroll forward (a windowful or a specified number of lines) (scroll-up).

M-v

Scroll backward (scroll-down).

arg C-l

Scroll so point is on line arg (recenter).

C-x <

Scroll text in current window to the left (scroll-left).

C-x >

Scroll to the right (scroll-right).

C-x $

Make deeply indented lines invisible (set-selective-display).

10.1 Scrolling

If a buffer contains text that is too large to fit entirely within a 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, the selected window is scrolled 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. It 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. If point was in the text scrolled off the top, it moves to the new top of the window. M-v with no argument moves backward with overlap similarly. 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 do scrolling is with 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, it 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 the scrolling is done so as to put point vertically centered within the window. However, if the variable scroll-step has a nonzero 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 is done.

10.2 Horizontal Scrolling

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.

10.3 Selective Display

Emacs has the ability to hide lines indented more than a certain number of columns (you specify how many columns). You can use this 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 how to give the argument.) Then lines with at least n columns of indentation disappear from the screen. The only indication of their presence is that three dots (‘…’) 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.

10.4 Variables Controlling Display

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.

If the variable visible-bell is non-nil, Emacs attempts to make the whole screen blink when it would normally make an audible bell sound. This variable has no effect if your terminal does not have a way to make the screen blink.

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. Then you might want to set the variable no-redraw-on-reenter 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. Altering the value of ctl-arrow makes it local to the current buffer; until that time, the default value is in effect. The default is initially t. 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 do not appear at the end of a line that precedes invisible lines. Then there is no visible indication of the invisible lines. This variable too becomes local automatically when set.

11 Searching and Replacement

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 nonincremental 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.

11.1 Incremental Search

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) would be found. When you have typed enough characters to identify the place you want, you can stop. Depending on what you will do next, you may or may not need to terminate the search explicitly with an <ESC> first.

C-s

Incremental search forward (isearch-forward).

C-r

Incremental search backward (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 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 <ESC>, which stops searching, leaving the cursor where the search brought it. Also, any command not specially meaningful in searches stops the searching and is then executed. Thus, typing C-a would exit the search and then move to the beginning of the line. <ESC> 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 not the one you expected to find. There was a second ‘FOO’ that you forgot about, before the one you were looking for. In this event, type another C-s to move to the next occurrence of the search string. This can be done any number of times. 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 your string is not found at all, the echo area says ‘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’, you might see the cursor after the ‘FOO’ in ‘FOOL’. At this point there are several things you can do. If your string was mistyped, you can rub some of it out and correct it. If you like the place you have found, you can type <ESC> 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 reverse 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. With them gone, the search is now successful and waiting for more input, so a second C-g will cancel 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. You can also specify a quoted character in octal while searching, just as you can for insertion. See section Basic Editing Commands.

You can change to searching backwards with C-r. If a search fails because the place you started was too late in the file, you should do this. Repeated C-r keeps looking for more occurrences backwards. A C-s starts going forwards again. C-r in a search can be cancelled with <DEL>.

If you know initially that you want to search backwards, you can use C-r instead of C-s to start the search, because C-r is also a key running a command (isearch-backward) to search backward.

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 all the rest of the current line into the search string.

All the characters special in incremental search can be changed by setting the following variables:

search-delete-char

Character to delete from incremental search string (normally <DEL>).

search-exit-char

Character to exit incremental search (normally <ESC>).

search-quote-char

Character to quote special characters for incremental search (normally C-q).

search-repeat-char

Character to repeat incremental search forwards (normally C-s).

search-reverse-char

Character to repeat incremental search backwards (normally C-r).

search-yank-line-char

Character to pull rest of line from buffer into search string (normally C-y).

search-yank-word-char

Character to pull next word from buffer into search string (normally C-w).

11.1.1 Slow Terminal Incremental Search

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 that the search has found. The single-line window comes into play 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 is the window in which the search was done 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. 1 is its normal value.

11.2 Nonincremental Search

Emacs also has conventional nonincremental search commands, which require you to type the entire search string before searching begins.

C-s <ESC> string <RET>

Search for string.

C-r <ESC> string <RET>

Search backward for string.

To do a nonincremental search, first type C-s <ESC>. This enters the minibuffer to read the search string; terminate the string with <RET>, and then the search is done. If the string is not found the search command gets an error.

The way C-s <ESC> works is that the C-s invokes incremental search, which is specially programmed to invoke nonincremental search if the argument you give it is empty. (Such an empty argument would otherwise be useless.) C-r <ESC> also works this way.

Forward and backward nonincremental searches are implemented by the commands search-forward and search-backward. These commands may be bound to keys in the usual manner. The reason that incremental search is programmed to invoke them as well is that C-s <ESC> is the traditional sequence of characters used in Emacs to invoke nonincremental search.

However, nonincremental searches performed using C-s <ESC> do not call search-forward right away. The first thing done is to see if the next character is C-w, which requests a word search.

11.3 Word Search

Word search searches 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 can be 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. With word search, you can search without having to know them.

C-s <ESC> C-w words <RET>

Search for words, ignoring differences in punctuation.

C-r <ESC> C-w words <RET>

Search backward for words, ignoring differences in punctuation.

Word search is a special case of nonincremental search and is invoked with C-s <ESC> C-w. This is followed by the search string, which must always be terminated with <RET>. Being nonincremental, 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.

A backward word search can be done by C-r <ESC> C-w.

Forward and backward word searches are implemented by the commands word-search-forward and word-search-backward. These commands may be bound to keys in the usual manner. The reason that incremental search is programmed to invoke them as well is that C-s <ESC> C-w is the traditional Emacs sequence of keys for word search.

11.4 Regular Expression 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 very 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 C-M-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 function in incremental regexp search. Typing C-s or C-r immediately after starting the search retrieves the last incremental search regexp used; that is to say, incremental regexp and non-regexp searches have independent defaults.

Note that adding characters to the regexp in an incremental regexp search does not make the cursor move back and start again. Perhaps it ought to; I am not sure. As it stands, if you have searched for ‘foo’ and you add ‘\|bar’, the search will not check for a ‘bar’ in the buffer before the ‘foo’.

Nonincremental search for a regexp is done by the functions re-search-forward and re-search-backward. You can invoke these with M-x, or bind them to keys. Also, you can call re-search-forward by way of incremental regexp search with C-M-s <ESC>.

11.5 Syntax of Regular Expressions

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, we can concatenate the regular expressions ‘f’ and ‘o’ to get the regular expression ‘fo’, which matches only the string ‘fo’. Still trivial. To do something nontrivial, you need to use one of the special characters. Here is a list of them.

. (Period)

is a special character that matches any single character except a newline. Using concatenation, we can make regular expressions like ‘a.b’ which matches any three-character string which begins with ‘a’ and ends with ‘b’.

*

is not a construct by itself; it is a suffix, which means the preceding regular expression is to be repeated as many times as possible. In ‘fo*’, the ‘*’ applies to the ‘o’, so ‘fo*’ matches one ‘f’ followed by any number of ‘o’s. The case of zero ‘o’s is allowed: ‘fo*’ does match ‘f’.

‘*’ always applies to the smallest possible preceding expression. Thus, ‘fo*’ has a repeating ‘o’, not a repeating ‘fo’.

The matcher processes a ‘*’ construct by matching, immediately, as many repetitions as can be found. Then it continues with the rest of the pattern. If that fails, backtracking occurs, discarding some of the matches of the ‘*’-modified construct in case that makes it possible to match the rest of the pattern. For example, matching ‘ca*ar’ against the string ‘caaar’, the ‘a*’ first tries to match all three ‘a’s; but the rest of the pattern is ‘ar’ and there is only ‘r’ left to match, so this try fails. The next alternative is for ‘a*’ to match only two ‘a’s. With this choice, the rest of the regexp matches successfully.

+

Is a suffix character similar to ‘*’ except that it requires that the preceding expression be matched at least once. So, for example, ‘ca+r’ will match the strings ‘car’ and ‘caaaar’ but not the string ‘cr’, whereas ‘ca*r’ would match all three strings.

?

Is a suffix character similar to ‘*’ except that it can match the preceding expression either once or not at all. For example, ‘ca?r’ will match ‘car’ or ‘cr’; nothing else.

[ … ]

‘[’ begins a character set, which is terminated by a ‘]’. In the simplest case, the characters between the two form the set. Thus, ‘[ad]’ matches either one ‘a’ or one ‘d’, and ‘[ad]*’ matches any string composed of just ‘a’s and ‘d’s (including the empty string), from which it follows that ‘c[ad]*r’ matches ‘cr’, ‘car’, ‘cdr’, ‘caddaar’, etc.

Character ranges can also be included in a character set, by writing two characters with a ‘-’ between them. Thus, ‘[a-z]’ matches any lower-case letter. Ranges may be intermixed freely with individual characters, as in ‘[a-z$%.]’, which matches any lower case letter or ‘$’, ‘%’ or period.

Note that the usual special characters are not special any more inside a character set. A completely different set of special characters exists inside character sets: ‘]’, ‘-’ and ‘^’.

To include a ‘]’ in a character set, you must make it the first character. For example, ‘[]a]’ matches ‘]’ or ‘a’. To include a ‘-’, write ‘---’, which is a range containing only ‘-’. To include ‘^’, make it other than the first character in the set.

[^ … ]

‘[^’ begins a complement character set, which matches any character except the ones specified. Thus, ‘[^a-z0-9A-Z]’ matches all characters except letters and digits.

‘^’ is not special in a character set unless it is the first character. The character following the ‘^’ is treated as if it were first (‘-’ and ‘]’ are not special there).

Note that a complement character set can match a newline, unless newline is mentioned as one of the characters not to match.

^

is a special character that matches the empty string, but only if at the beginning of a line in the text being matched. Otherwise it fails to match anything. Thus, ‘^foo’ matches a ‘foo’ which occurs at the beginning of a line.

$

is similar to ‘^’ but matches only at the end of a line. Thus, ‘xx*$’ matches a string of one ‘x’ or more at the end of a line.

\

has two functions: it quotes the special characters (including ‘\’), and it introduces additional special constructs.

Because ‘\’ quotes special characters, ‘\$’ is a regular expression which matches only ‘$’, and ‘\[’ is a regular expression which matches only ‘[’, and so on.

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.

For the most part, ‘\’ 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.

\|

specifies an alternative. Two regular expressions a and b with ‘\|’ in between form an expression that matches anything that either a or b will match.

Thus, ‘foo\|bar’ matches either ‘foo’ or ‘bar’ but no other string.

‘\|’ applies to the largest possible surrounding expressions. Only a surrounding ‘\( … \)’ grouping can limit the grouping power of ‘\|’.

Full backtracking capability exists to handle multiple uses of ‘\|’.

\( … \)

is a grouping construct that serves three purposes:

1. To enclose a set of ‘\|’ alternatives for other operations. Thus, ‘\(foo\|bar\)x’ matches either ‘foox’ or ‘barx’.
2. To enclose a complicated expression for the postfix ‘*’ to operate on. Thus, ‘ba\(na\)*’ matches ‘bananana’, etc., with any (zero or more) number of ‘na’ strings.
3. To mark a matched substring for future reference.

This last application is not a consequence of the idea of a parenthetical grouping; it is a separate feature which happens to be assigned as a second meaning to the same ‘\( … \)’ construct because there is no conflict in practice between the two meanings. Here is an explanation of this feature:

\digit

after the end of a ‘\( … \)’ construct, the matcher remembers the beginning and end of the text matched by that construct. Then, later on in the regular expression, you can use ‘\’ followed by digit to mean “match the same text matched the digit’th time by the ‘\( … \)’ construct.”

The strings matching the first nine ‘\( … \)’ constructs appearing in a regular expression are assigned numbers 1 through 9 in order that the open-parentheses appear in the regular expression. ‘\1’ through ‘\9’ may be used to refer to the text matched by the corresponding ‘\( … \)’ construct.

For example, ‘\(.*\)\1’ matches any newline-free string that is composed of two identical halves. The ‘\(.*\)’ matches the first half, which may be anything, but the ‘\1’ that follows must match the same exact text.

\`

matches the empty string, provided it is at the beginning of the buffer.

\'

matches the empty string, provided it is at the end of the buffer.

\b

matches the empty string, provided it is at the beginning or end of a word. Thus, ‘\bfoo\b’ matches any occurrence of ‘foo’ as a separate word. ‘\bballs?\b’ matches ‘ball’ or ‘balls’ as a separate word.

\B

matches the empty string, provided it is not at the beginning or end of a word.

\<

matches the empty string, provided it is at the beginning of a word.

\>

matches the empty string, provided it is at the end of a word.

\w

matches any word-constituent character. The editor syntax table determines which characters these are.

\W

matches any character that is not a word-constituent.

\scode

matches any character whose syntax is code. code is a character which represents a syntax code: thus, ‘w’ for word constituent, ‘-’ for whitespace, ‘(’ for open-parenthesis, etc. See section The Syntax Table.

\Scode

matches any character whose syntax is not code.

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 contains four parts in succession: 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.

Note that the above example shows how to write this regexp when entering it as part of an Emacs Lisp program. To enter the same regexp in an interactive command such as re-search-forward you must spell it differently:

[.?!][]"')]*\($\|^Q^I\| \)[ ^Q^I^Q^J]*

11.6 Searching and Case

All sorts of searches in Emacs normally ignore the case of the text they are searching through; if you specify searching for ‘FOO’, then ‘Foo’ and ‘foo’ are also considered a match. Regexps, and in particular character sets, are included: ‘[aB]’ would match ‘a’ or ‘A’ or ‘b’ or ‘B’.

If you do not want this feature, set the variable case-fold-search to nil. Then all letters must match exactly, including case. 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.

11.7 Replacement Commands

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 the 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.

11.7.1 Unconditional Replacement

M-x replace-string <RET> string <RET> newstring <RET>

Replace every occurrence of string with newstring.

M-x replace-regexp <RET> regexp <RET> newstring <RET>

Replace every match for regexp with newstring.

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, so if you want to cover the whole buffer you must go to the beginning first. 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.

11.7.2 Regexp Replacement

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 ‘\\’.

11.7.3 Replace Commands and Case

If the arguments to a replace command are in lower case, it preserves case when it makes a replacement. Thus, the 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.

11.7.4 Query Replace

M-% string <RET> newstring <RET>

M-x query-replace <RET> string <RET> newstring <RET>

Replace some occurrences of string with newstring.

M-x query-replace-regexp <RET> regexp <RET> newstring <RET>

Replace some matches for regexp with newstring.

If you want to change only some of the occurrences of ‘foo’ to ‘bar’, not all of them, then you cannot use an ordinary replace-string. Instead, use M-% (query-replace). 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:

<SPC>

to replace the occurrence with newstring. This preserves case, just like replace-string, provided case-replace is non-nil, as it normally is.

<DEL>

to skip to the next occurrence without replacing this one.

, (Comma)

to replace this occurrence and display the result. You are then asked for another input character, except that since the replacement has already been made, <DEL> and <SPC> are equivalent. You could type C-r at this point (see below) to alter the replaced text. You could also type C-x u to undo the replacement; this exits the query-replace, so if you want to do further replacement you must use C-x ESC to restart (see section Repeating Minibuffer Commands).

<ESC>

to exit without doing any more replacements.

. (Period)

to replace this occurrence and then exit.

!

to replace all remaining occurrences without asking again.

^

to go back to the location of the previous occurrence (or what used to be an occurrence), in case you changed it by mistake. This works by popping the mark ring. Only one ^ in a row is allowed, because only one previous replacement location is kept during query-replace.

C-r

to enter a recursive editing level, in case the occurrence needs to be edited rather than just replaced with newstring. When you are done, exit the recursive editing level with C-M-c and the next occurrence will be displayed. See section Recursive Editing Levels.

C-w

to delete the occurrence, and then enter a recursive editing level as in C-r. Use the recursive edit to insert text to replace the deleted occurrence of string. When done, exit the recursive editing level with C-M-c and the next occurrence will be displayed.

C-l

to redisplay the screen and then give another answer.

C-h

to display a message summarizing these options, then give another answer.

If you type any other character, the query-replace is exited, and the character executed 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 C-x ESC.

To replace every occurrence, you can start query-replace at the beginning of the buffer and type !, or you can use the replace-string command at the beginning of the buffer. To replace every occurrence in a part of the buffer, narrow to that part and then run replace-string or query-replace at the beginning of it. See section Narrowing.

11.8 Other Search-and-Loop Commands

Here are some other commands that find matches for a regular expression. They all operate from point to the end of the buffer.

M-x occur

Print each line that follows point and contains a match for the specified regexp. A numeric argument specifies the number of context lines to print before and after each matching line; the default is none.

The buffer ‘*Occur*’ containing the output serves as a menu for finding the occurrences in their original context. Find an occurrence as listed in ‘*Occur*’, position point there and type C-c C-c; this switches to the buffer that was searched and moves point to the original of the same occurrence.

M-x list-matching-lines

Synonym for M-x occur.

M-x count-matches

Print the number of matches following point for the specified regexp.

M-x delete-non-matching-lines

Delete each line that follows point and does not contain a match for the specified regexp.

M-x delete-matching-lines

Delete each line that follows point and contains a match for the specified regexp.

12 Commands for Fixing Typos

In this chapter we describe the commands that are especially useful for the times when you catch a mistake in your text just after you have made it, or change your mind while composing text on line.

12.1 Killing Your Mistakes

<DEL>

Delete last character (delete-backward-char).

M-<DEL>

Kill last word (backward-kill-word).

C-x <DEL>

Kill to beginning of sentence (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 again.

12.2 Transposing Text

C-t

Transpose two characters (transpose-chars).

M-t

Transpose two words (transpose-words).

C-M-t

Transpose two balanced expressions (transpose-sexps).

C-x C-t

Transpose two lines (transpose-lines).

The common error of transposing two characters can be fixed, when they are adjacent, 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, rather than transposing the last character of the line with the newline, which would be useless, C-t transposes the last two characters on the line. So, if you catch your transposition error right away, you can fix it with just a C-t. If you don’t catch it so fast, you must 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. They work like M-t except in determining the division of the text into syntactic units.

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 is assigned a special meaning (because otherwise a command with a repeat count of zero would do nothing): to transpose the character (word, sexp, line) ending after point with the one ending after the mark.

12.3 Case Conversion

M-- M-l

Convert last word to lower case. Note Meta-- is Meta-minus.

M-- M-u

Convert last word to all upper case.

M-- M-c

Convert last word to lower case with capital initial.

A very 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 have a special feature when used with a negative argument: they do not move the cursor. As soon as you see you have mistyped the last word, you can simply case-convert it and go on typing. See section Case Conversion Commands.

12.4 Checking and Correcting Spelling

M-$

Check and correct spelling of word (spell-word).

M-x spell-buffer

Check and correct spelling of each word in the buffer.

M-x spell-region

Check and correct spelling of each word in the region.

M-x spell-string

Check spelling of specified word.

To check the spelling of the word before point, and optionally correct it as well, 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 does 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. Then 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 if appropriate for every incorrect word.

M-x spell-region is similar 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 in the echo area a message giving the answer.

13 File Handling

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.

13.1 File Names

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 using 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 will be used if you type just <RET>, entering an empty argument. 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, if you do not specify a directory, the default directory is used. 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 is kept in the variable default-directory, which has a separate value in every buffer.

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 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 made randomly with C-x b, its default directory is copied from that of the buffer that was current at the time.

The default directory 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. This insertion of the default directory is inhibited if the variable insert-default-directory is set 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 as part of the text. The final minibuffer contents may look invalid, but that is not so. See section Minibuffers for File Names.

‘$’ in a file name is used to substitute environment variables. For example, if you have used the C shell command ‘setenv FOO rms/hacks’ to set up an environment variable named ‘FOO’, then you can use ‘/u/$FOO/test.c’ or ‘/u/${FOO}/test.c’ as an abbreviation for ‘/u/rms/hacks/test.c’. (In the Bourne-Again shell, write export FOO=rms/hacks to define FOO.) The environment variable name consists of all the alphanumeric characters after the ‘$’; 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 ‘$’ at the same time as variable substitution is performed for single ‘$’. 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.

13.2 Visiting Files

C-x C-f

Visit a file (find-file).

C-x C-v

Visit a different file instead of the one visited last (find-alternate-file).

C-x 4 C-f

Visit a file, in another window (find-file-other-window). Don’t change this window.

Visiting a file means copying its contents into Emacs where you can edit them. Emacs makes a new buffer for each file that 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, keeping just the name proper. For example, a file named ‘/usr/rms/emacs.tex’ would get a buffer named ‘emacs.tex’. If there is already a buffer with that name, a unique name is constructed by appending ‘<2>’, ‘<3>’, or 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 visited, or any place permanent, 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>.

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.

Your confirmation that C-x C-f has completed successfully is the appearance of new text on the screen and a new buffer name in the mode line. If the specified file does not exist and could not be created, or cannot be read, then an error results. The error message is printed in the echo area, and includes the file name which 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, a warning message is printed. See section Simultaneous Editing.

What 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). This can be inhibited by setting 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 this command is used when only one window is being displayed, that window is split in two, with one window showing the same before as before, and the other one showing the newly requested file. See section Multiple Windows.

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, and the functions 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 a 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.

You can put a local variable specification at the end of a file which specifies values for Emacs local variables whenever you visit the file. See section Local Variables in Files.

13.3 Saving Files

Saving a buffer in Emacs means writing its contents back into the file that was visited in the buffer.

C-x C-s

Save the current buffer in its visited file (save-buffer).

C-x s

Save any or all buffers in their visited files (save-some-buffers).

M-~

Forget that the current buffer has been changed
(not-modified).

C-x C-w

Save the current buffer in a specified file, and record that file as the one visited in the buffer (write-file).

M-x set-visited-file-name

Change file the name under which the current buffer will be saved.

When you wish to save the 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), saving is not really done, because it would have no effect. Instead, C-x C-s prints a message in the echo area saying

(No changes need to be written)

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. These 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, each time you use save-some-buffers you are liable to save it by mistake. One thing you can do is type M-~ (not-modified), which clears out 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, one which is not in use for anything important. Alternatively, you can 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. You could also undo all the changes by repeating the undo command C-x u until you have undone all the changes; 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 reads the new file name using the minibuffer. It can be used on a buffer that is not visiting a file, too. The buffer’s name is changed to correspond to the file it is now visiting in the usual fashion (unless the new name is in use already for some other 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, if you save the buffer, it will be saved in that file. It also marks the buffer as “modified” so that C-x C-s will save.

If you wish to mark the 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 with 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 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.

You can implement other ways to write files, and other things to be done before writing them, using the hook variable write-file-hooks. 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.

13.3.1 Backup Files

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 (provided 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 that you edit.

Emacs makes a backup for a file only the first time the 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 will be made by the next save.

13.3.1.1 Single or Numbered Backups

If you choose to have a single backup file (this is the default), the backup file’s name is constructed by appending ‘~’ to the file name being edited; thus, the backup file for ‘eval.c’ would be ‘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 such backup is available.

The choice of single backup or numbered backups is controlled by the variable version-control. Its possible values are

t

Make numbered backups.

nil

Make numbered backups for files that have numbered backups already. Otherwise, make single backups.

never

Do not in any case make numbered backups; always make single backups.

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.

13.3.1.2 Automatic Deletion of Backups

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. Recall that these 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, the excess middle versions are deleted without a murmur. If it is nil, the default, then you are asked whether the excess middle versions should really be deleted.

Dired’s . (Period) command can also be used to delete old versions. See section Dired, the Directory Editor.

13.3.1.3 Copying vs. Renaming

Backup files can be made by copying the old file or by renaming it. This makes a difference when the old file has multiple names. If the old file is renamed into the backup file, then the alternate names become names for the backup file. If the old file is copied instead, then the alternate names remain names for the file that you are editing, and the contents accessed by those names will be the new contents.

The method of making a backup file may also affect the file’s owner and group. If copying is used, these 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 always shows who last edited the file. Also, the owners of the backups show who produced those versions. Occasionally there is a file whose owner should not change; it is a good idea for such files to contain local variable lists to set backup-by-copying-when-mismatch for them alone (see section Local Variables in Files).

The choice of renaming or copying is controlled by three variables. 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, then copying is done for files that have multiple names, but renaming may still done when the file being edited has only one name. If the variable backup-by-copying-when-mismatch is non-nil, then copying is done if renaming would cause the file’s owner or group to change.

13.3.2 Protection against Simultaneous Editing

Simultaneous editing occurs when two users visit the same file, both make changes, and then both save them. If nobody were informed that this was happening, whichever user saved first would later find that his changes were lost. On some systems, Emacs notices immediately when the second user starts to change the file, and issues an immediate warning. When this is not possible, or if the second user has gone on to change the file despite the warning, Emacs checks later 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 the editing user takes the proper corrective action at this point, he can 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 for the sake of customization. The standard definition of this function asks you a question and accepts three possible answers:

s

Steal the lock. Whoever was already changing the file loses the lock, and you gain the lock.

p

Proceed. Go ahead and edit the file despite its being locked by someone else.

q

Quit. This causes an error (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 user from editing it simultaneously under different names. However, basing locking on names means that Emacs can interlock the editing of new files that will 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 still can 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 Directory Listing). This will show the file’s current author. You should attempt to contact that person to warn him or her 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 with C-x C-f a file that is already visited and when you start to modify a file. This is not strictly necessary, but it can cause you to find out about the problem earlier, when perhaps correction takes less work.

13.4 Reverting a Buffer

If you have made extensive changes to a file and then change your mind about them, you can get rid of them by reading in the previous version of the file. To do this, use M-x revert-buffer, which operates on the current buffer. Since this is a very dangerous thing to do, you must confirm it with yes.

If the current buffer has been auto-saved more recently than it has been saved for real, revert-buffer offers to read the auto save file instead of the visited file (see section Auto-Saving: Protection Against Disasters). This question comes before the usual request for confirmation, and demands y or n as an answer. If you have started to type yes for confirmation without realizing that the other question was going to be asked, the y will answer that question, but the es will not be valid confirmation. So you will have a chance to cancel the operation with C-g and try it again with the answers that 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 drastic changes, the same value of point in the old file may address a totally different piece of text.

A buffer reverted from its visited file is marked “not modified” until another change is made.

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 base. Buffers created randomly with C-x b cannot be reverted; revert-buffer reports an error when asked to do so.

13.5 Auto-Saving: Protection Against Disasters

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 that 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 been 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.

13.5.1 Auto-Save Files

Auto-saving does not normally save in the files that you visited, because it can be very undesirable to save a program that is in an inconsistent state when you have made 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 request saving explicitly (such as with C-x C-s).

Normally, the auto-save file name is made by appending ‘#’ to the front and rear 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 made by appending ‘#%’ to the front and ‘#’ to the rear of buffer name. For example, the ‘*mail*’ buffer in which you compose messages to be sent is auto-saved in a file named ‘#%*mail*#’. Auto-save file names are made this way unless you reprogram parts of Emacs to do something different (the functions make-auto-save-file-name and auto-save-file-name-p). The file name to be used for auto-saving in a buffer is calculated when 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.

A buffer’s auto-save file is deleted when you save the buffer in its visited file. To inhibit this, 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 go with the new visited name.

13.5.2 Controlling Auto-Saving

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 auto-saving is the usual practice for file-visiting buffers. Auto-saving can be turned on or off for any existing buffer with the command M-x auto-save-mode. 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 does auto-saving periodically based on counting how many characters you have typed since the last time auto-saving was done. The variable auto-save-interval specifies how many characters there are 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.

13.5.3 Recovering Data from Auto-Saves

The way to use the contents of an auto-save file to recover from a loss of data is with the command M-x recover-file <RET> file <RET>. This visits file and then (after your confirmation) restores the contents from its auto-save file ‘#file#’. You can then save 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, you should turn auto-saving back on with M-x auto-save-mode.

13.6 Listing a File Directory

Files are classified 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).

C-x C-d dir-or-pattern

Print a brief directory listing (list-directory).

C-u C-x C-d dir-or-pattern

Print a verbose directory listing.

The command to print a directory listing is C-x C-d (list-directory). It reads using the minibuffer a file name which is either a directory to be listed or a wildcard-containing pattern for the files to be listed. For example,

C-x C-d /u2/emacs/etc <RET>