Section 19.1. Editing Files Using vi

This section covers the use of the vi (pronounced "vee-eye") text editor. vi was the first real screen-based editor for Unix systems. It is also simple, small, and sleek. If you're a system administrator, learning vi can be invaluable; in many cases, larger editors, such as (X)Emacs, won't be available in emergency situations (for instance, when booting Linux from a maintenance disk).

vi is based on the same principles as many other Unix applications: that each program provides a small, specific function and is able to interact with other programs. For example, vi doesn't include its own spell checker or paragraph filler, but those features are provided by other programs that are easy to fire off from within vi. Therefore, vi itself is a bit limited, but is able to interact with other applications to provide virtually any functionality you might want.

At first, vi may appear to be somewhat complex and unwieldy. However, its single-letter commands are fast and powerful once you've learned them. The next section describes Emacs, a more flexible editor (really an integrated work environment) with an easier learning curve. Do keep in mind that knowing vi may be essential to you if you are in a situation where (X)Emacs is not available, so we encourage you to learn the basics, as odd as they may seem. It should also be added that a number of vi clones are now available that are much more comfortable to use than the original vi, the most popular of which is vim (vi improved). Chances are that your distribution has things set up so that when starting vi, you actually start one of those. We stick to the basics here, though, so that you can use the information presented here no matter which version of vi you use. You can find coverage of the newer versions in the book Learning the vi Editor by Linda Lamb and Arnold Robbins (O'Reilly).

19.1.1. Starting vi

Let's fire up vi and edit a file. The syntax for vi is:

vi filename

For example:

eggplant$ vi test

will edit the file test. Your screen should look like Figure 19-1.

Figure 19-1. vi when opening a new file

The column of ~ characters indicates that you are at the end of the file.

19.1.2. Inserting Text and Moving Around

While using vi, at any one time you are in one of two (or three, depending on how you look at it) modes of operation. These modes are known as command mode, edit mode, and ex mode.

After starting vi, you are in command mode. This mode allows you to use a number of (usually single-letter) commands to modify text , as we'll see soon. Text is actually inserted and modified within edit mode. To begin inserting text, press i (which will place you into edit mode) and begin typing. See Figure 19-2.

Figure 19-2. Entering text into vi buffer

While inserting text, you may type as many lines as you wish (pressing the Enter key after each, of course), and you may correct mistakes using the Backspace key. To end edit mode and return to command mode, press the Esc key.

While in command mode, you can use the arrow keys to move around the file. Alternatively, or when the arrow keys don't work, you may use h, j, k, and l, which move the cursor left, down, up, and right, respectively.

There are several ways to insert text other than using the i command. The a command (for "append") inserts text after the current cursor position. For example, use the left arrow key to move the cursor between the words good and men (Figure 19-3).

Figure 19-3. Positioning cursor in vi

Press a, type wo, and then press Esc to return to command mode (Figure 19-4).

Figure 19-4. vi after insertion

To open a line below the current one and begin inserting text, use the o command. Press o and type another line or two (Figure 19-5).

Figure 19-5. vi after appending text

Remember that at any time you're either in command mode (where commands such as i, a, or o are valid) or in edit mode (where you're inserting text, followed by Esc to return to command mode). If you're not sure which mode you're in, press Esc. This takes you out of edit mode, if you are in it, and does nothing except beep if you're already in command mode.

19.1.3. Deleting Text and Undoing Changes

From command mode, the x command deletes the character under the cursor. If you press x five times in our example, you end up with the screen shown in Figure 19-6.

Figure 19-6. vi after removing text

Now press a and insert some text, followed by Esc (Figure 19-7).

Figure 19-7. vi with new text

You can delete entire lines using the command dd (that is, press d twice in a row). If your cursor is on the second line in our example, dd will produce the screen shown in Figure 19-8.

Figure 19-8. vi after deleting lines

Text that is deleted may be reinserted using the p command (for "put"). Pressing p now will return the deleted line to the buffer after the current line. Using P (uppercase) instead will insert the text before the current line. By default, p and P insert text from the "undo buffer"; you can also yank and replace text from other buffers, as we'll see later.

The u command undoes the latest change (in this case, pressing u after dd is equivalent to p). If you inserted a large amount of text using the i command, pressing u immediately after returning to command mode would undo it.

To delete the word beneath the cursor, use the dw command. Place the cursor on the word Diet and type dw (see Figure 19-9).

Figure 19-9. vi after deleting a word

19.1.4. Changing Text

You can replace text using the R command, which overwrites the text beginning at the cursor. Place the cursor on the first letter in pizza, press R, and type (Figure 19-10).

The r command replaces the single character under the cursor. r does not place you in insert mode per se, so there is no reason to use Esc to return to command mode.

The ~ command changes the case of the letter under the cursor from upper- to lowercase, and vice versa. If you place the cursor on the o in Now in the previous example, and repeatedly press ~, you end up with the screen shown in Figure 19-11.

Another useful command for changing words is the cw command, which lets you simply type in the new word andafter pressing Escremoves anything that might

Figure 19-10. vi after replacing text

Figure 19-11. Changing case in vi

be left over from the original word. If the new text is longer than the one being changed, the space is automatically expanded as needed.

19.1.5. Moving Around the File

You already know how to use the arrow keys to move around the document. In addition, the w command moves the cursor to the beginning of the next word, and b moves it to the beginning of the current word. The 0 (that's a zero) command moves the cursor to the beginning of the current line, and the $ command moves it to the end of the line.

When editing large files, you'll want to move forward or backward through the file one screen at a time. Pressing Ctrl-F moves the cursor one screen forward, and Ctrl-B moves it one screen backward.

To move the cursor to the end of the file, type G. You can also move to an arbitrary line: the command 10G would move the cursor to line 10 in the file. To move to the beginning of the file, use 1G.

Typing / followed by a pattern and the Enter key causes you to jump to the first occurrence of that pattern in the text following the cursor. For example, placing the cursor on the first line of text in our example and typing /burg moves the cursor to the beginning of the word "burgers." Using ? instead of / searches backward through the file.

The pattern following a / or ? command is actually a regular expression. Regular expressions are a powerful way to specify patterns for search and replace operations and are used by many Unix utilities. You can find more information about regular expressions in the section "Regular Expressions," later in this chapter. Using regular expressions, you could, for example, search for the next uppercase letter, using the command

/[A-Z]

Therefore, if the pattern you're searching for is not a static string, regular expressions can be used to specify just what you want.

You can couple navigation commands with other commands, such as deletion. For example, the command d$ will delete everything from the cursor to the end of the line; dG will delete everything from the cursor to the end of the file.

19.1.6. Saving Files and Quitting vi

Most of the commands dealing with files within vi are invoked from ex mode. You enter ex mode when you press the : key from command mode. This places the cursor on the last line of the display, allowing you to enter various extended commands.

For example, to write the file being edited, use the command :w. Typing : causes you to enter ex mode, and typing w followed by the Enter key completes the command. The command :wq writes the file and exits vi. (The command ZZ--from command mode, without the ":"--is similar to :wq, but checks first whether the file has been changed, and writes it only in this case.)

To quit vi without saving changes to the file, use the command :q!. Using :q alone will quit vi, but only if modifications to the file have been saved. The ! in :q! means to quit vi--and that you really mean it.

19.1.7. Editing Another File

To edit another file, use the :e command. For example, to stop editing test, and edit the file foo instead, use the command shown at the bottom of Figure 19-12.

Figure 19-12. Editing antoher file with vi

If you use :e without writing the file first, you'll get the following error message:

No write since last change (:edit! overrides)

At this point, you can use :w to save the original file, and then use :e, or you can use the command :e! foo, which tells vi to edit the new file without saving changes to the original. This can be useful if you edit a file and realize that you have screwed up. You can then use the :e! command; if you don't specify a filename, vi discards the changes and re-edits the current file.

19.1.8. Including Other Files

If you use the :r command, you can include the contents of another file in the vi buffer. For example, the command

:r foo.txt

inserts the contents of the file foo.txt after the current line.

19.1.9. Running Shell Commands

The :! command allows you to enter the name of a shell command, which is executed within vi. For example, the command

:!ls -F

executes the ls command and displays the results on your screen.

The :r! command is similar to :!, but includes the standard output of the command in the buffer. The command:

:r!ls -F

produces the screen shown in Figure 19-13.

Figure 19-13. Inserting results of a command in vi

If you need to execute a series of shell commands , it's often easier to use the suspend key (usually Ctrl-Z), provided you're using a shell that supports job control, such as zsh or bash.

19.1.10. Global Searching and Replacing

There are many more features of vi than are documented here; most of these features are implemented through combinations of the simple features we've seen. Here are one or two other tidbits most vi users find useful.

The command

:[x,y]s/pattern/replacement/flags

searches for pattern between lines x and y in the buffer, and replaces instances of pattern with the replacement text . pattern is a regular expression; replacement is literal text but can contain several special characters to refer to elements in the original pattern. The following command replaces the first occurrence of weeble with wobble on lines 1 through 10, inclusive:

:1,10s/weeble/wobble

Instead of giving line-number specification, you can use the % symbol to refer to the entire file. Other special symbols can be used in place of x and y. $ refers to the last line of the file. Leave x or y blank to refer to the current line.

Among the flags you can use are g to replace all instances of pattern on each line, and c to ask for confirmation for each replacement. In most instances, you will want to use the g flag, unless you want to replace only the first occurrence of pattern on each line.

You can also use marks to refer to lines. Marks are just single-letter names that are given to cursor locations within the document. Moving the cursor to a location in the file and typing ma will set the mark a at that point. (Marks may be named any of the letters a-z or A-Z.) You can move the cursor directly to the mark a with the command `a (with a backquote). Using a regular single quote (as in 'a) will move the cursor to the beginning of the line that the mark a is on.

Marks allow you to "remember" cursor locations that denote a region of text. For example, if you want to search and replace a block of text, you can move the cursor to the beginning of the text, set a mark, move the cursor to the end of the text, and use the command:

:'a,.s/weeble/wobble/

where 'a refers to the line containing mark a, and . refers to the current line.

19.1.11. Moving Text and Using Registers

One way to copy and move text is to delete it (using the d or dd commands) and then replace it with the P command, as described earlier. For example, if you want to delete 10 lines, starting with the line that contains your cursor, and paste them somewhere else, just use the command 10dd (to delete 10 lines), move the cursor to the new location for the text, and type p. You can copy text in this way as well: typing 10dd followed by P (at the same cursor location) deletes the text and immediately replaces it. You can then paste the text elsewhere by moving the cursor and using p multiple times.

Similar to dd is the yy command, which "yanks" text without deleting it. You use p to paste the yanked text as with dd. But note that each yank operation will delete the previously yanked text from the clipboard.

The deletion and yank commands can be used on more general regions than lines. Recall that the d command deletes text through a move command; for example, d$ deletes text from the cursor to the end of the line. Similarly, y$ yanks text from the cursor to the end of the line.

Let's say you want to yank (or delete) a region of text. This can be done with marks as well. Move the cursor to the beginning of the text to be yanked and set a mark, such as ma. Move the cursor to the end of the text to be yanked and use the command y`a. This yanks text from the cursor position to the mark a. (Remember that the command `a moves the cursor to the mark a.) Using d instead of y deletes the text from the cursor to the mark.

The most convenient way to cut, copy, and paste portions of text within vi is to use registers. A register is just a named temporary storage space for text you wish to copy between locations, cut and paste within the document, and so forth.

Registers are given single-letter names; any of the characters a to z or A to Z are valid. The " command (a quotation mark) specifies a register; it is followed by the name of the register, as in "a for register a. The lowercase letters and their uppercase counterparts refer to the same registers: using the lowercase letter overwrites the previous contents of the register, and using the uppercase letter appends to it.

For instance, if we move the cursor to the first line, as in Figure 19-14, and use the command "ayy, the current line is yanked into the register a. We can then move the cursor to the second line, and use the command "ap to paste the text from register a after the current line (see Figure 19-15).

Figure 19-14. vi buffer before a yank

Figure 19-15. vi buffer after a yank

Similarly, the command "ay`a yanks text from the cursor to mark a into register a. Note that there is no correspondence between mark and register names!

Using registers allows you to copy text between files. Just copy the text to a register, use the :e command to edit a new file, and paste the text from the register.

19.1.12. Extending vi

vi is extensible in many ways. Most of the commands we've introduced can be generalized to arbitrary regions of text. As we've already seen, commands such as d and y operate on the text from the cursor to a move operation, such as $ or G. (dG deletes text from the cursor to the end of the file.) Many other commands operate on text through a move command in the same way. Using marks, you can operate on any region of text.

As we mentioned before, vi is just a text editor; it doesn't have facilities for spell checking text, compiling programs, and other such features. However, vi executes other programs that you can use to extend the editor. The command:

:x,y!command

executes the named command with the text on lines x through y as standard input, and replaces the lines with the standard output of the command. As with the s (search and replace) command, other specifications, such as % and $, can be used for the line numbers.

For example, let's say you want to prepend a quote character to all the lines in a region of text. One way to do this is to write a short shell or Perl script (see "Programming Languages and Utilities" in Chapter 1) that reads lines of input and outputs those same lines with the quote character prepended. (Or use a sed command there are many alternatives.) You can then send lines of text through this filter, which replaces them with the quoted text within vi. If the script is called quote, just use a command such as:

:`a,.!quote

which quotes the region of text between the cursor location and the mark a.

Be familiar with the various ex-mode commands that are available. The :set command allows you to set various options; for example, :set ai turns on auto indentation of text. (:set noai turns it off.)

You can specify ex-mode commands (such as :set) to execute when starting up vi in the file .exrc in your home directory. (The name of this file can be changed with the EXINIT environment variable.) For example, your .exrc file might contain:

set ai

to turn on auto indentation. You don't need the : before ex commands in this file.

A number of good tutorials and references for vi are availableboth online as well as in print. Learning the vi Editor is a good place to look for more information. One popular web site for vi information is The vi Lovers Home Page, http://thomer.com/vi/vi.html. The home of vim on the Web is http://www.vim.org.