CONTENTS |
In this chapter, we'll cover the following topics:
Regular expressions
vi
Many UNIX users have a Graphical User Interface (GUI) through which they access their UNIX system. The Common Desktop Environment (CDE) is the most commonly used GUI on UNIX systems. It is based on the X Windows System and Motif, which together provide an advanced windowing environment. A chapter in this book is devoted to CDE. Most UNIX GUIs provide a graphical editor. Despite the fact that these graphical editors are a standard part of most GUIs, the visual editor, vi, still remains the most popular UNIX editor. With many fine graphics-based editors as a standard part of most UNIX GUIs and a plethora of editors available as part of personal computer windowing environments, why am I covering vi? The answer is two-fold. First, not everyone using a UNIX system has access to a graphics display and may therefore need to know and use vi. Because vi comes with most UNIX-based systems and is a powerful editor, many new UNIX users end up using and liking it. Second, vi has traditionally been thought of as the UNIX editor. Few UNIX users have not used vi. This fact does not mean that it is everyone's primary editor; however, virtually all UNIX users have had some experience with vi.
Also, a line editor called ed comes with many UNIX systems. It is now seldom used because vi is a screen editor. Also available is an enhanced version of ed called ex. vi is much more widely used than either of the line editors, so I'll cover only vi in this chapter.
I'll cover the basics of using vi in this chapter. You can experiment with what is covered here, and if you really like it, you can investigate some of the more advanced features of vi. A quick reference card summarizing all the vi commands covered in this chapter is included with this book.
Table 8-1 is a list of tables in this chapter that summarize some of the more commonly used vi commands by function:
Table Number | vi Function |
---|---|
Expr | Regular Expressions |
Introduction | Modes and Notations in vi |
1 | Starting a vi Session |
2 | Cursor Control Commands in vi |
3 | AddingTextinvi |
4 | Deleting Text in vi |
5 | Changing Text in vi |
6 | Search and Replace in vi |
7 | Copying in vi |
8 | Undoinvi |
9 | Saving Text and Exiting vi |
10 | Options in vi |
11 | Status in vi |
12 | Positioning and Marking in vi |
13 | Joining Lines in vi |
14 | Cursor Placement and Adjusting Screen in vi |
15 | Shell Escape Commands in vi |
16 | Macros and Abbreviations in vi |
17 | Indenting Text in vi |
18 | Shell Filters in vi |
19 | Pattern Matching in vi |
Regular expressions describe patterns for which you are searching. The regular expression usually defines the pattern for which you are searching using wildcards. Since a regular expression defines a pattern you are searching for, the terms "regular expression" and "pattern matching" are often used interchangably.
Let's get down to a couple of words-of-caution immediately:
Regular expressions are different from file-matching patterns used by the shell. Regular expressions are used by both the shell and many programs, including those covered in this chapter. The file matching done by the shell and programs such as find are different from the regular expressions covered in this chapter.
Use single quotes around regular expressions. The meta-characters used in this chapter must be quoted in order to be passed to the shell as an argument. You will, therefore, see most regular expressions in this chapter quoted.
When using the programs in this book, such as grep and vi, you provide a regular expression that the program evaluates. The command will search for the pattern you supply. The pattern could be as simple as a string or it could include wildcards. The wildcards used by many programs are called meta-characters.
Table 8-2 shows a list of meta-characters and the program(s) to which they apply. Only the programs covered in this book (awk, grep, sed, and vi) are shown in Table 8-2. These meta-characters may be used with other programs, such as ed and egrep, as well, which are not covered in the book. Table 8-2 describes the meta-characters and their use.
Meta Character | awk | grep | sed | vi | Use |
---|---|---|---|---|---|
. | Yes | Yes | Yes | Yes | Match any single character. |
* | Yes | Yes | Yes | Yes | Match any number of the single character that precedes *. |
[...] | Yes | Yes | Yes | Yes | Match any one of the characters in the set [...] |
$ | Yes | Yes | Yes | Yes | Matches the end of the line. |
^ | Yes | Yes | Yes | Yes | Matches the beginning of the line. |
\ | Yes | Yes | Yes | Yes | Escape the special character that follows \. |
\{n,m\} | Yes | Yes | No | No | Match a range of occurrences of a single character between n and m. |
+ | Yes | No | No | No | Match one or more occurrences of the preceding regular expression. |
? | Yes | No | No | No | Match zero or one occurrence of the preceding regular expression. |
| | Yes | No | No | No | The preceding or following regular expression can be matched. |
() | Yes | No | No | No | Groups regular expressions in a typical parenthesis fashion. |
\{ \} | No | No | No | Yes | Match a word's beginning or end. |
You may want to refer to this table when regular expressions are used for one of the commands in the table.
We're first going to cover some of the fundamentals of the operation of vi called modes, and then go over some of the notations used in the tables in this chapter.
A feature of vi that often confuses new users is that it has modes. When you are in command mode, everything you type is interpreted as a command. In command mode, you can specify such actions as the location to which you want the cursor to move. When you are in input mode, everything you type is information to be added to the file. Command mode is the default when you start vi. You can move into command mode from input mode atany time by pressing the escape key. You move into insert mode from command mode by typing one of the input mode commands covered shortly.
vi commands don't really have a standard form. For this reason, I cover common notations. Table 8-3 summarizes modes and commands in vi:
Mode or Notation | Description |
---|---|
Command Mode | You are issuing commands such as moving the cursor or deleting text, rather than inserting or changing text when in command mode. You can switch to insert mode by issuing an insert mode command such as i for insert or a for add text. |
Insert Mode | You are in insert mode when changing or inserting more than one character of text. You canswitch to commandmodeby pressing the escape key. |
: (colon commands) | Commands that start with a: are completed by pressing the return key. |
control (^) commands | When a command uses the control (^) key, you pressand hold down thecontrol key and then press the next key that is part of the command. For instance, ^g means press and hold control and then g to get the status on the file you are editing. |
file for the name of a file | Many commands require you to specify the name of a file. For instance, in the command vi file, you would substitute the name of the file you wish to edit for file. |
char for the name of a character | Many commands require you to specify a single character. For instance, in the command fchar, you would substitute the character you wish to search for in place of char. |
cursor_command for a cursor movement command | Many commands require you to specify a cursor command to execute. For instance, in the command dcursor_ command, you would substitute for cursor_command the commandyou wish to execute. |
string for a character string | Many commands require you to specify a character string. For instance, in the command /string, you would substitute for string the character string for which you wish to search. |
Let's jump right in and edit a file. For most of the examples in this chapter, I perform various vi commands and capture the results in an X Window. The best way to learn any topic is by example. I not only provide many examples, but I also capture each example in an X Window so that you can see the results of each command. From the command line, we type vi and the name of the file we wish to edit, in this case, wisdom:
$ vi wisdom
We are then editing the file wisdom as shown in Figure 8-1. wisdom contains a passage from Tao Te Ching or "Book of the Way." We use this file throughout this chapter.
The bottom line in Figure 8-1 is the message line in vi. After invoking vi, the message line indicates the name of the file, the number of lines, and the number of characters in the file. Different messages appear on the message line, depending on the command you issue, as we see in upcoming examples. If a tilde appears on any lines in the file, as it does in the two lines above the message line in wisdom, it means that not enough lines exist to fill up the screen. The cursor is the dark box that appears at line 1 in Figure 8-1.
We can specify several file names, and after saving the first file we move on to the second file by entering:n, and continue going through the list of files in this way. Or we can specify a file and position the cursor on the last line in the file. The default is for the cursor to appear over the first character in the file, as shown in Figure 8-1.
Table 8-4 shows some of the ways we can start a vi session:
Command | Description |
---|---|
vi file | Edit file. |
vi -r file | Edit last saved version of file after a crash. |
vi -R file | Edit file in read-only mode. |
vi + n file | Edit file and place cursor at line n. |
vi + file | Edit file and place cursor on last line. |
vi file1 file2 file3 ... | Edit file1 through file3, and after saving changes in file1, you can move to file2 by entering:n. |
vi +/ string file | Edit file and place cursor at the beginning of the line containing string. |
Figure 8-2 shows editing wisdom and placing the cursor at line 5 with the command vi +5 wisdom:
Figure 8-3 shows editing wisdom and placing the cursor at the last line of the file with the command vi + wisdom.
Figure 8-4 shows editing wisdom and placing the cursor at the line containing task with vi +/task wisdom.
A key skill to develop in vi is getting the cursor to the desired position. You do this in command mode. You have a variety of ways to move the cursor around the screen. Table 8-5 summarizes some of the more commonly used cursor movements:
Command | Cursor Movement |
---|---|
h or ^h | Move left one character. |
j or ^j or ^n | Move down one line. |
k or ^p | Move up one line. |
l or space | Move right one character. |
G | Go to the last line of the file. |
nG | Go to line number n. |
G$ | Go to the last character in the file. |
1G | Go to the first line in the file. |
w | Go to the beginning of the next word. |
W | Go to the beginning of next word, ignore punctuation. |
b | Go to the beginning of the previous word. |
B | Go to the start of previous word, ignore punctuation. |
L | Go to the last line of the screen. |
M | Go to the middle line of the screen. |
H | Go to the first line of the screen. |
e | Move to the end of the next word. |
E | Move to the end of the next word, ignore punctuation. |
( | Go to the beginning of the sentence. |
) | Go to the end of the sentence. |
{ | Go to the beginning of the paragraph. |
} | Go to the beginning of the next paragraph. |
0 or | Go to the first column in the current line. |
n | Go to column n in the current line. |
^ (caret) | Go to the first non-blank character in the current line. |
$ | Go to the last character in the current line. |
+ | or return Go to the first character in the next line. |
- | Go to the first non-blank character in the previous line. |
I know that the fact that you have to remember these commands in order to get the cursor to the desired position may seem a little strange at first, but this is the way vi works. Let's use wisdom to show how some of these cursor movements work. Figures 8-5 and 8-6 show some cursor movements. Like all of the upcoming figures, Figures 8-5 and 8-6 show wisdom before a command is entered on the left and the result after the command is entered on the right. The command issued appears in the middle. Some of the commands in upcoming figures use the enter and escape keys.
Now that we know how to move around the cursor, let's do something with it. You need to first learn about cursor movement, because the commands for adding text take place relative to the position of the cursor. Table 8-6 summarizes some commands for adding text:
Command | Insertion Action |
---|---|
a | Append new text after the cursor. |
A | Append new text after the end of the current line. |
i | Insert new text before the cursor. |
I | Insert new text before the beginning of the current line. |
o | Open a line below the current line and insert. |
O | Opena line abovethe current line and insert. |
:r file | Read file and insert after the current line. |
:nr file | Read file and insert after line n. |
escape | Get back to command mode. |
^v char | Ignore special meaning of char when inserting. This is for inserting special characters. |
Let's now look at some examples of adding text into wisdom in Figure 8-7:
We also need to learn about cursor movement before learning how to delete text, because the commands for deleting text take place relative to the position of the cursor. Table 8-7 summarizes some commands for deleting text:
Command | Deletion Action |
---|---|
x | Delete the character at the cursor. You can also put a number in front of x to specify the number of characters to delete. |
nx | Delete n characters beginning with the current. |
X | Delete the previous character. You can also put a number in front of X to specify the number of previous characters to delete. |
nX | Delete previous n characters. |
dw | Delete to the beginning of the next word. |
ndw | Delete the next n words beginning with the current. |
dG | Delete lines to the end of the file. |
dd | Delete the entire line. |
ndd | Delete n lines beginning with the current. |
db | Delete the previous word. |
ndb | Delete the previous n words beginning with the current. |
:n,md | Deletes lines n through m. |
D or d$ | Delete from the cursor to the end of the line. |
dcursor_command | Delete text to the cursor_command. dG would delete from the current line to the end of the file. |
^h or backspace | While inserting, delete the previous character. |
^w | While inserting, delete the previous word. |
Let's now look at some examples of deleting text from wisdom in Figures 8-8 and 8-9:
Okay, you've added text and deleted text, and now you want to change text. vi isn't so bad so far, is it? Table 8-8 summarizes some commands for changing text:
Command (Preceding these commands with a number repeats the commands any number of times.) | Replacement Action |
---|---|
rchar | Replace the current character with char. |
Rtext escape | Replace the current characters with text until escape is entered. |
stext escape | Substitute text for the current character. |
S or cctext escape | Substitute text for the entire line. |
cwtext escape | Change the current word to text. |
Ctext escape | Change the rest of the current line to text. |
cG escape | Change to the end of the file. |
ccursor_cmd text escape | Change to text from the current position to cursor_cmd. |
Let's now look at some examples of replacing text from wisdom in Figures 8-10 and 8-11:
You have a lot of search and replace functionality in vi. Table 8-9 summarizes some of the more common search-and-replace functionality in vi:
Command | Search and Replace Action |
---|---|
/text | Search for text going forward into the file. |
?text | Search for text going backward into the file. |
n | Repeat the search in the same direction as the original search. |
N | Repeat the search in the opposite direction as the original search. |
ftext | Search for text going forward in the current line. |
Ftext | Search for text going backward in the current line. |
ttext | Search for text going forward in the current line and stop at the character before text. |
Ttext | Search for text going backward in the current line to character after text. |
:set ic | Ignore case when searching. |
:set noic | Make searching case-sensitive. |
:s/oldtext/newtext/ | Substitute newtext for oldtext. |
:m,ns/oldtext/newtext/ | Substitute newtext for oldtext in lines m through n. |
& | Repeat the last:s command. |
:g/text1/s/text2/text3 | Find line containing text1, replace text2 with text3. |
:g/text/command | Run command on all lines that contain text. |
:v/text/command | Run command on all lines that do not contain text. |
Let's now look at some examples of searching and replacing text in wisdom in Figure 8-12:
You can perform advanced searches with:g and:v. You can find and display all the lines in a file containing while with the following command:
:g/while/p
The /p in this command line is the print command used with the ex editor. You can find all the lines in the file that contain while and delete those lines with the following command:
:g/while/d
You can also specify the specific line numbers for which you want the search to take place. The following command finds all the lines between 10 and 20 that contain while and prints the line number on which they appear:
:10,20g/while/nu
:g runs a command on the lines that contain the text for which we are searching, and:v runs a command on the lines that do not contain the specified text. The following three commands act on the lines that do not contain while, in the same way that the previous three act on the lines that do contain while:
:v/while/p :v/while/d :10,20v/while/nu
The first command prints lines that do not contain while. The second command deletes the lines on which while does not appear. The third command prints the line number between 10 and 20 on which while does not appear.
You can copy text in vi. Some commands for copying are shown in Table 8-10:
Command | Copy Action |
---|---|
yy | Yank the current line. |
nyy | Yank n lines. |
p (lowercase) | Put the yanked text after the cursor. |
p (uppercase) | Put the yanked text before the cursor. |
"(a-z)nyy | Copy n lines into the buffer named in parentheses. Omit n for the current line. |
"(a-z)ndd | Delete n lines into the buffer named in parentheses. Omit n for the current line. |
"(a-z)p | Put lines named in the buffer in parentheses after current line. |
"(a-z)P | Put lines named in the buffer in parentheses before the current line. |
Let's now look at some examples of copying text in wisdom in Figure 8-13:
You can easily undo and repeat changes in vi with the commands shown in Table 8-11:
Command | Undo Action |
---|---|
u | Undo the last change. |
U | Undo all changes to the current line. |
. (period) | Repeat the last change. |
, (comma) | Repeat, in reverse direction, last f, F, t, or T search command. |
; (semi-colon) | Repeat last f, F, t, or T search command. |
"np | Retrieve the last nth delete (a limited number of deletes are in the buffer, usually nine). |
n | Repeat last / or ? search command. |
N | Repeat, in reverse direction, last / or ? search command. |
You have a number of different ways to save files and exit vi, some of which are summarized in Table 8-12:
Command | Save and/or Quit Action |
---|---|
:w | Save the file but don't exit vi. |
:w file | Save changes in file but don't quit vi. |
:wq or ZZ or:x | Save the file and quit vi. |
:q! | Quit vi without saving the file. |
:e! | Re-edit the file, discarding changes since the last write. |
There are many options you can set and unset in vi. To set an option, you type:set option. To unset an option, you type:set nooption. Table 8-13 summarizes some of the more commonly used options:
Option | Action |
---|---|
:set all | Print all options. |
:set nooption | Turn off option. |
:set nu | Prefix lines with line number. |
:set showmode | Show whether input or replace mode. |
:set noic | Ignore case when searching. |
:set list | Show tabs (^l) and end of line ($). |
:set ts=8 | Set tab stops for text input. |
:set window=n | Set number of lines in a text window to n. |
Let's now prefix lines with line numbers and show input or replace mode in Figure 8-14:
Many additional options are available beyond those in Table 8-13. The following is a list of options produced on a UNIX system from the:set all command. You should issue this command when in vi to see the options available to you:
:set all noautoindent autoprint noautowrite nobeautify directory=/var/tmp nodoubleescape noedcompatible noerrorbells noexrc flash hardtabs=8 noignorecase keyboardedit nokeyboardedit! nolisp nolist magic mesg nomodelines nonumber nonovice nooptimize paragraphs=IPLPPPQPP LIpplpipnpbp prompt noreadonly redraw remap report=5 scroll=11 sections=NHSHH HUuhsh+c shell=/sbin/sh shiftwidth=8 noshowmatch noshowmode noslowopen tabstop=8 taglength=0 tags=tags /usr/lib/tags tagstack term=hp noterse timeout timeoutlen=500 ttytype=hp warn window=23 wrapscan wrapmargin=0 nowriteany noshowmatch noshowmode
Many of the options are preceded by a "no," indicating that the option is not set. You may want to list your options with:set all and then experiment with the options of interest to you to see the effect they will have on your vi session.
You can obtain a lot of useful status information with some simple commands in vi. You can display the current line number, number of lines in the file, file name, and other status information with the commands shown in Table 8-14:
Option | Action |
---|---|
:.= | Print the current line number. |
:= | Print the number of lines in the file. |
^g | Show the file name, current line number, total lines in the file, and percent of file location. |
:l | Use the letter "l" to display various special characters such as tab and newline. |
You can define sections of text to which you can move as well as mark text with characters and move to those marks. Table 8-15 summarizes positioning and marking in vi:
Option | Action |
---|---|
{ | Insert { in first column to define section. |
[[ | Go back to beginning of section. |
]] | Forward to beginning of next section. |
m(a-z) | Mark current position with a letter such as mz for mark z. |
'(a-z) | Move cursor to specified mark such as'z for move to z. |
You can join one or more lines in vi using the commands shown in Table 8-16:
Option | Action |
---|---|
J | Join the next line to the end of the current line. |
nJ | Join the next n lines. |
You can place the cursor anywhere in your file and adjust the screen in a variety of ways using the commands shown in Table 8-17:
Option | Action |
---|---|
H | Move cursor to top line of the screen. |
nH | Move cursor to n line from the top of the screen. |
M | Move cursor to the middle of the screen. |
L | Move cursor to the bottom line of the screen. |
nL | Move cursor to line n from the bottom of the screen. |
^e (control-e) | Move screen up one line. |
^y | Move screen down one line. |
^u | Move screen up one-half page. |
^d | Move screen down one-half page. |
^b | Move screen up one page. |
^f | Move screen down one page. |
^l (letter l) | Redraw screen. |
z - return | Make current line the top of screen. |
nz - return | Make n line the top of screen. |
z. | Make current line the middle line. |
nz. | Make line n the middle line on screen. |
z- | Make current line the bottom line. |
nz- | Make line n the bottom line on screen. |
You can run a UNIX command without exiting vi by using shell escape commands. You could do something as simple as start a subshell with the:sh command. You could also run a command outside the file you are editing without exiting vi. Table 8-18 describes shell escape commands:
Option | Action |
---|---|
:! command | Execute shell command command such as:! ls. |
:!! | Execute last shell command. |
:r! command | Read and insert output from command, such as:r! ls to run ls and read contents. |
:w !command | Send currently edited file to command as standard input and execute command, such as:w ! grep all. |
:cd directory | Change the current working directory to directory. |
:sh | Start a sub-shell and use ^d (control-d) to return to vi. |
:so file | Read and execute commands in the shell program file. |
An example of using:w would be to send the file wisdom as standard input to grep, looking for all lines that contain all, asinthe following example:
:w ! grep all Think of the small as by a series of small acts.
You can issue the:so command to read and execute the commands in a file. Issuing the following command when in vi would run the commands in the file file_with_commands:
:so file_with_commands
This file contains the following two commands:
:set nu :g/all/p
When we issue the earlier:so command, line numbers are shown with the:set nu command and the following lines containing all are printed:
Think of the small as by a series of small acts.
You are not limited to issuing individual vi commands. You can define strings of vi commands and define a key corresponding to this string that you can recall. When defining the keys for your macros, you can't use the following: K V g q v *= and function keys. There are also control keys you can't use, so stay away from control keys in general. Table 8-19 shows macros and abbreviations.
Option | Action |
---|---|
:map key command_seq | Define key to run command_seq, such as:map e ea to append text whenever you use e to move to the end of a word. |
:map | Display all defined macros on the status line. |
:umap key | Remove the macro for key. |
:ab string1 string2 | Define an abbreviation such that when string1 is inserted, replace it with string2. When inserting text type string1, press escape key and string2 will be inserted. |
:ab | Display all abbreviations. |
:cd directory | Change the current working directory to directory. |
:una string | Unabbreviate string. |
Avoid control keys, symbols, and don't use characters: K V g q v *= and function keys. |
An example of using the map command would be to automatically add text when you move to the end, as shown with the following map command:
:map e ea
This command maps e to ea. When you go to the end of the next word with e, you are also placed in insert mode with a so that you can append new text immediately after the end of the word.
You can also abbreviate long sequences with ab. For instance, you could abbreviate system administration with sa with the following command:
:ab sa system administration
Now, whenever you insert text, type sa and then presstheescape key to complete the insert; the string system administration appears. sa is an abbreviation for system administration.
You can indent text in a variety of different ways. Table 8-20 shows some of the more commonly used indenting commands:
Option | Action |
---|---|
^i (control i) or tab | While inserting text, insert on shift width. Shift width can be defined. |
:set ai | Turn on auto-indentation. |
:set sw=n | Set shift width to n characters. |
n<< | Shift n lines left by one shift width. |
n>> | Shift n lines right by one shift width. For example, 3>> shifts the next three lines right by oneshift width. |
Before you adjust the shift width, you may want to issue:set all in order to see the current number of characters to which the shift width is set. It is usually eight characters by default. To set shift width to 16 characters, issue the following command:
:set sw=16
You can then shift over the next three lines to the right by 16 characters each, with the following command:
3>>
The next three lines are then shifted right by 16 characters.
You can send information from the file you are editing to a command and then replace the original text with the output of the command. Table 8-21 shows a shell filter:
Option | Action |
---|---|
!cursor_command command | Send text from the current position to that described by cursor_command to the shell command. For example, use !} grep admin to take text from the current position to the end of the paragraph, run this text through grep looking for the word admin, and replace existing text with the output of grep. |
Pattern matching allows you to find patterns within the file you are editing. You can then perform functions such as changing what you have found in some way. Table 8-22 shows some of the most common pattern-matching commands:
Option | Action |
---|---|
^ (caret) | Match the beginning of the line. To search for Think at only the beginning of the line, you would use: /^Think You can use this in combination with $, which matches to the end of the line, to delete all blank lines with: :g/^$/d. |
$ | Match end of line. To match last. only when it is followed by a newline character, you would use: /last.$ |
. | Match any single character. |
\< | Match beginning of word. |
\> | Match end of word. |
[string] | Match any single character in string. To find mp, mP, Mp, or MP, use: /[mM][pP] Change all occurrences of input or Input to INPUT with: :%s/[Ii]nput/INPUT/g |
[^string] | Match any character not in string. |
[a-p] | Match any character between a and p. |
* | Match zero or more occurrences of previous character in expression. |
\ | Escape meaning of next character. To search for [, use the following: /\[ |
\\ | Escape the \ character. |
You may find pattern matching a little confusing when you first start to use it, so I'll go through several simple examples to get you started. Keep in mind that many of the pattern-matching techniques described here also work outside vi in your shell.
We'll begin with the square bracket operator. To match any of the single characters m, f, or p, you would use the following:
/[mfp]
A common pattern to match would be a word with the first letter in the word, either uppercase or lowercase. To match input or Input, you would use the following:
/[Ii]nput
After you match either Input or input, you could then change it to INPUT with the following command:
:%s/[Ii]nput/INPUT/g
You can use sequences of expressions to search for more than one character, as shown in the following example:
/[mM][pP]
This sequence will match mp, mP, Mp, or MP. You are, in effect, searching for any of the four two-character strings.
You can also use the square bracket operator to match single characters within a range. To find an occurrence of any digit in a file, you could use either of the two following square bracket searches:
/[0123456789]
or
/[0-9]
The hyphen denotes a range within the square bracket. To find any character, either uppercase or lowercase, you could use the following:
/[a-zA-Z]
To search for characters that normally have a special meaning, such as [, you can ignore, or escape, the special meaning by preceding the special character with a \ (backslash). To search for [in vi, for instance, you would use the following sequence:
/\[
This search finds the first occurrence of [.
You can specify that you wish your pattern match to take place at only the beginning or end of a line. To specify a beginning of the line pattern match, use the ^ (caret) preceding your desired pattern, as shown in the following example:
/^Think
This matches Think only when it appears at the beginning of a line.
To specify an end-of-the-line pattern match, use a $ (dollar sign) following your desired pattern, as shown in the following example:
/last.$
This matches last. only when it is followed by a newline.
The following section contains copies of the manual pages for vi.
vi - Run visual editor.
vi(1) vi(1) NAME vi, view, vedit - screen-oriented (visual) text editor SYNOPSIS vi [-] [-l] [-r] [-R] [-t tag] [-v] [-V] [-wsize] [-x] [+command] [file ...] XPG4 Synopsis vi [-rR] [-c command] [-t tag] [-w size] [file ...] Obsolescent Options vi [-rR] [+command] [-t tag] [-w size] [file ...] view [-] [-l] [-r] [-R] [-t tag] [-v] [-V] [-wsize] [-x] [+command] [file ...] vedit [-] [-r] [-R] [-l] [-t tag] [-v] [-V] [-wsize] [-x] [+command] [file ...] Remarks The program names ex, edit, vi, view, and vedit are separate personalities of the same program. This manual entry describes the behavior of the vi/view/vedit personality. DESCRIPTION The vi (visual) program is a display-oriented text editor that is based on the underlying ex line editor (see ex(1)). It is possible to switch back and forth between the two and to execute ex commands from within vi. The line-editor commands and the editor options are described in ex(1). Only the visual mode commands are described here. The view program is identical to vi except that the readonly editor option is set (see ex(1)). The vedit program is somewhat friendlier for beginners and casual users. The report editor option is set to 1, and the nomagic, novice, and showmode editor options are set. In vi, the terminal screen acts as a window into a memory copy of the file being edited. Changes made to the file copy are reflected in the screen display. The position of the cursor on the screen indicates the position within the file copy. The environment variable TERM must specify a terminal type that is defined in the terminfo database (see terminfo(4)). Otherwise, a message is displayed and the line-editor is invoked. As with ex, editor initialization scripts can be placed in the environment variable EXINIT, or in the file .exrc in the current or home directory. Options and Arguments vi recognizes the following command-line options and arguments: - Suppress all interactive-user feedback. This is useful when editor commands are taken from scripts. -l Set the lisp editor option (see ex(1)). Provides indents appropriate for lisp code. The (, ), {, }, [[, and ]] commands in vi are modified to function with lisp source code. -r Recover the specified files after an editor or system crash. If no file is specified, a list of all saved files is printed. You must be the owner of the saved file in order to recover it (superuser cannot recover files owned by other users). -R Set the readonly editor option to prevent overwriting a file inadvertently (see ex(1)). -t tag Execute the tag tag command to load and position a predefined file. See the tag command and the tags editor option in ex(1). -v Invoke visual mode (vi). Useful with ex, it has no effect on vi. -V Set verbose mode. Editor commands are displayed as they are executed when input from a .exrc file or a source file (see the source command in ex(1)). -wsize Set the value of the window editor option to size. If size is omitted, it defaults to 3. -x Set encryption mode. You are prompted for a key to allow for the creation or editing of an encrypted file (see the crypt command in ex(1)). -c command (XPG4 only.) +command (Obsolescent) Begin editing by executing the specified ex command-mode commands. As with the normal ex command-line entries, the command option-argument can consist of multiple ex commands separated by vertical- line commands (|). The use of commands that enter input mode in this manner produces undefined results. file Specify the file or files to be edited. If more than one file is specified, they are processed in the order given. If the -r option is also specified, the files are read from the recovery area. (XPG4 only.) If both the -t tag and -c command (or the obsolescent +command) options are given, the -t tag will be processed first, that is, the file containing the tag is selected by -t and then the command is executed. When invoked, vi is in command mode. input mode is initiated by several commands used to insert or change text. In input mode, ESC (escape) is used to leave input mode; however, two consecutive ESC characters are required to leave input mode if the doubleescape editor option is set (see ex(1)). In command mode, ESC is used to cancel a partial command; the terminal bell sounds if the editor is not in input mode and there is no partially entered command. WARNING: ESC completes a "bottom line" command (see below). The last (bottom) line of the screen is used to echo the input for search commands (/ and ?), ex commands (:), and system commands (!). It is also used to report errors or print other messages. The receipt of SIGINT during text input or during the input of a command on the bottom line terminates the input (or cancels the command) and returns the editor to command mode. During command mode, SIGINT causes the bell to be sounded. In general the bell indicates an error (such as an unrecognized key). Lines displayed on the screen containing only a Zindicatethatthe last li~e above them is the last line of the file (the Z lines are past the end o~ the file). Terminals with limited local intelligence might display lines on the screen marked with an @. These indicate space on the screen not corresponding to lines in the file. (These lines can be removed by entering a ^R, forcing the editor to retype the screen without these holes.) If the system crashes or vi aborts due to an internal error or unexpected signal, vi attempts to preserve the buffer if any unwritten changes were made. Use the -r command line option to retrieve the saved changes. The vi text editor supports the SIGWINCH signal, and redraws the screen in response to window-size changes. Command Summary Most commands accept a preceding number as an argument, either to give a size or position (for display or movement commands), or as a repeat count (for commands that change text). For simplicity, this optional argument is referred to as count when its effect is described. The following operators can be followed by a movement command to specify an extent of text to be affected: c, d, y, <, >, !, and =. The region specified begins at the current cursor position and ends just prior to the cursor position indicated by the move. If the command operates on lines only, all the lines that fall partly or wholly within this region are affected. Otherwise the exact marked region is affected. In the following description, control characters are indicated in the form ^X, which represents Ctrl-X. Whitespace is defined to be the characters space, tab, and alternative space. Alternative space is the first character of the ALT_PUNCT item described in langinfo(5) for the language specified by the LANG environment variable (see environ(5)). Unless otherwise specified, the commands are interpreted in command mode and have no special effect in input mode. ^B Scroll backward to display the previous window of text. A preceding count specifies the number of windows to go back. Two lines of overlap are kept if possible. ^D Scroll forward a half-window of text. A preceding count gives the number of (logical) lines to scroll, and is remembered for future ^D and ^U commands. ^D (input mode) Backs up over the indentation provided by autoindent or ^T to the next multiple of shiftwidth spaces. Whitespace inserted by ^T at other than the beginning of a line cannot be backed over using ^D. A preceding ^ removes all indentation for the current and subsequent input lines of the current input mode until new indentation is established by inserting leading whitespace, either by direct input or by using ^T. ^E Scroll forward one line, leaving the cursor where it is if possible. ^F Scroll forward to display the window of text following the current one. A preceding count specifies the number of windows to advance. Two lines of overlap are kept if possible. (XPG4 only.) The current line is displayed and the cursor is moved to the first nonblank character of the current line or the first character if the line is a blank line. ^G Print the current file name and other information, including the number of lines and the current position (equivalent to the ex command f). ^H Move one space to the left (stops at the left margin). A preceding count specifies the number of spaces to back up. (Same as h). ^H (input mode) Move the cursor left to the previous input character without erasing it from the screen. The character is deleted from the saved text. ^J Move the cursor down one line in the same column, if possible. A preceding count specifies the number of lines to move down. (Same as ^N and j). ^L Clear and redraw the screen. Use when the screen is scrambled for any reason. ^M Move to the first nonwhitespace character in the next line. A preceding count specifies the number of lines to advance. ^N Same as ^J and j. ^P Move the cursor up one line in the same column. A preceding count specifies the number of lines to move up (same as k). ^R Redraw the current screen, eliminating the false lines marked with @ (which do not correspond to actual lines in the file). ^T Pop the tag stack. See the pop command in ex(1). ^T (input mode) Insert shiftwidth whitespace. If at the beginning of the line, this inserted space can only be backed over using ^D. ^U Scroll up a half-window of text. A preceding count gives the number of (logical) lines to scroll, and is remembered for future ^D and ^U commands. ^V In input mode, ^V quotes the next character to permit the insertion of special characters (including ESC) into the file. ^W In input mode, ^W backs up one word; the deleted characters remain on the display. ^Y Scroll backward one line, leaving the cursor where it is, if possible. ^[ Cancel a partially formed command; ^[ sounds the bell if there is no partially formed command. In input mode, ^[ terminates input mode. However, two consecutive ESC characters are required to terminate input mode if the doubleescape editor option is set (see ex(1)). When entering a command on the bottom line of the screen (ex command line or search pattern with \ or ?), terminate input and execute command. On many terminals, ^[ can be entered by pressing the ESC or ESCAPE key. ^\ Exit vi and enter ex command mode. If in input mode, terminate the input first. ^] Take the word at or after the cursor as a tag and execute the tagMbobC editor command (see ex(1)). ^^ Return to the previous file (equivalent to :ex #). space Move one space to the right (stops at the end of the line). A preceding count specifies the number of spaces to go forward (same as l). erase Erase, where erase is the user-designated erase character (see stty(1)). Same as ^H. kill Kill, where kill is the user-designated kill character (see stty(1)). In input mode, kill backs up to the beginning of the current input line without erasing the line from the screen display. susp Suspend the editor session and return to the calling shell, where susp is the user-designated process- control suspend character (see stty(1)). See ex(1) for more information on the suspend editor command. ! An operator that passes specified lines from the buffer as standard input to the specified system command, and replaces those lines with the standard output from the command. The ! is followed by a movement command specifying the lines to be passed (lines from the current position to the end of the movement) and then the command (terminated as usual by a return). A preceding count is passed on to the movement command after !. Doubling ! and preceding it by count causes that many lines, starting with the current line, to be passed. " Use to precede a named buffer specification. There are named buffers 1 through 9 in which the editor places deleted text. The named buffers a through z are available to the user for saving deleted or yanked text; see also y, below. $ Move to the end of the current line. A preceding count specifies the number of lines to advance (for example, 2$ causes the cursor to advance to the end of the next line). % Move to the parenthesis or brace that matches the parenthesis or brace at the current cursor position. & Same as the ex command & (that is, & repeats the previous substitute command). ' When followed by a ', vi returns to the previous context, placing the cursor at the beginning of the line. (The previous context is set whenever a nonrelative move is made.) When followed by a letter a-z, returns to the line marked with that letter (see the m command), at the first nonwhitespace character in the line. When used with an operator such as d to specify an extent of text, the operation takes place over complete lines (see also `). ` When followed by a `, vi returns to the previous context, placing the cursor at the character position marked (the previous context is set whenever a nonrelative move is made). When followed by a letter a z, returns to the line marked with that letter (see the m command), at the character position marked. When used with an operator such as d to specify an extent of text, the operation takes place from the exact marked place to the current position within the line (see also '). [[ Back up to the previous section boundary. A section is defined by the value of the sections option. Lines that start with a form feed (^L) or { also stop [[. If the option lisp is set, the cursor stops at each ( at the beginning of a line. ]] Move forward to a section boundary (see [[). ^ Move to the first nonwhitespace position on the current line. ( Move backward to the beginning of a sentence. A sentence ends at a ., !, or ? followed by either the end of a line or by two spaces. Any number of closing ), ], ", and ' characters can appear between the ., !, or ? and the spaces or end of line. If a count is specified, the cursor moves back the specified number of sentences. If the lisp option is set, the cursor moves to the beginning of a lisp s-expression. Sentences also begin at paragraph and section boundaries (see { and [[). ) Move forward to the beginning of a sentence. If a count is specified, the cursor advances the specified number of sentences (see (). { Move back to the beginning of the preceding paragraph. A paragraph is defined by the value of the paragraphs option. A completely empty line and a section boundary (see [[ above) are also interpreted as the beginning of a paragraph. If a count is specified, the cursor moves backward the specified number of paragraphs. } Move forward to the beginning of the next paragraph. If a count is specified, the cursor advances the specified number of paragraphs (see {). | Requires a preceding count; the cursor moves to the specified column of the current line (if possible). + Move to the first nonwhitespace character in the next line. If a count is specified, the cursor advances the specified number of lines (same as ^M). , The comma (,) performs the reverse action of the last f, F, t, or T command issued, by searching in the opposite direction on the current line. If a count is specified, the cursor repeats the search the specified number of times. - The hyphen character (-) moves the cursor to the first nonwhitespace character in the previous line. If a count is specified, the cursor moves back the specified number of times. _ The underscore character (_) moves the cursor to the first nonwhitespace character in the current line. If a count is specified, the cursor advances the specified number of lines, with the current line being counted as the first line; no count or a count of 1 specifies the current line. . Repeat the last command that changed the buffer. If a count is specified, the command is repeated the specified number of times. / Read a string from the last line on the screen, interpret it as a regular expression, and scan forward for the next occurrence of a matching string. The search begins when the user types a carriage return to terminate the pattern; the search can be terminated by sending SIGINT (or the user-designated interrupt character). When used with an operator to specify an extent of text, the defined region begins with the current cursor position and ends at the beginning of the matched string. Entire lines can be specified by giving an offset from the matched line (by using a closing / followed by a +n or -n). 0 Move to the first character on the current line (the 0 is not interpreted as a command when preceded by a nonzero digit). : The colon character (:) begins an ex command. The : and the entered command are echoed on the bottom line; the ex command is executed when the user types a carriage return. ; Repeat the last single character find using f, F, t, or T. If a count is specified, the search is repeated the specified number of times. < An operator that shifts lines to the left by one shiftwidth. The < can be followed by a move to specify lines. A preceding count is passed through to the move command. When repeated (<<), shifts the current line (or count lines starting at the current one). > An operator that shifts lines right one shiftwidth (see <). = If the lisp option is set, = reindents the specified lines, as if they were typed in with lisp and autoindent set. = can be preceded by a count to indicate how many lines to process, or followed by a move command for the same purpose. ? Scan backwards, the reverse of / (see /). @buffer Execute the commands stored in the named buffer. Be careful not to include a <return> character at the end of the buffer contents unless the <return> is part of the command stream. Commands to be executed in ex mode should be preceded by a colon (:). ~ The tilde (~) switches the case of the character under the cursor (if it is a letter), then moves one character to the right, stopping at the end of the line). A preceding count specifies how many characters in the current line are switched. A Append at the end of line (same as $a). B Back up one word, where a word is any nonblank sequence, placing the cursor at the beginning of the word. If a count is specified, the cursor moves back the specified number of words. C Change the rest of the text on the current line (same as c$). D Delete the rest of the text on the current line (same as d$). E Move forward to the end of a word, where a word is any nonblank sequence. If a count is specified, the cursor advances the specified number of words. F Must be followed by a single character; scans backwards in the current line, searching for that character and moving the cursor to it, if found. If a count is specified, the search is repeated the specified number of times. G Go to the line number given as preceding argument, or the end of the file if no preceding count is given. H Move the cursor to the top line on the screen. If a count is given, the cursor moves to count number of lines from the top of the screen. The cursor is placed on the first nonwhitespace character on the line. If used as the target of an operator, entire lines are affected. I Insert at the beginning of a line (same as ^ followed by i). J Join the current line with the next one, supplying appropriate whitespace: one space between words, two spaces after a period, and no spaces at all if the first character of the next line is a closing parenthesis ()). A preceding count causes the specified number of lines to be joined, instead of just two. L Move the cursor to the first nonwhitespace character of the last line on the screen. If a count is given, the cursor moves to count number of lines from the bottom of the screen. When used with an operator, entire lines are affected. M Move the cursor to the middle line on the screen, at the first nonwhitespace position on the line. N Scan for the next match of the last pattern given to / or ?, but in the opposite direction; this is the reverse of n. O Open a new line above the current line and enter input mode. P Put back (replace) the last deleted or yanked text before/above the cursor. Entire lines of text are returned above the cursor if entire lines were deleted or yanked. Otherwise, the text is inserted just before the cursor. (XPG4 only.) In this case, the cursor is moved to last column position of the inserted characters. If P is preceded by a named buffer specification (x), the contents of that buffer are retrieved instead. Q Exit vi and enter ex command mode. R Replace characters on the screen with characters entered, until the input is terminated with ESC. S Change entire lines (same as cc). A preceding count changes the specified number of lines. T Must be followed by a single character; scan backwards in the current line for that character, and, if found, place the cursor just after that character. A count is equivalent to repeating the search the specified number of times. U Restore the current line to its state before the cursor was last moved to it. (XPG4 only.) The cursor position is set to the column position 1 or to the position indicated by the previous line if the autoindent is set. W Move forward to the beginning of a word in the current line, where a word is a sequence of nonblank characters. If the current position is at the beginning of a word, the current position is within a bigword or the character at that position cannot be a part of a bigword, the current position shall move to the first character of the next bigword. If no subsequent bigword exists on the current line, the current position shall move to the first character of the first bigword on the first following line that contains the bigword. For this command, an empty or blank line is considered to contain exactly one bigword. The current line is set to the line containing the bigword selected and the current position is set to the first character of the bigword selected. A preceding count specifies the number of words to advance. X Delete the character before the cursor. A preceding count repeats the effect, but only characters on the current line are deleted. Y Place (yank) a copy of the current line into the unnamed buffer (same as yy). If a count is specified, count lines are copied to the buffer. If the Y is preceded by a buffer name, the lines are copied to the named buffer. ZZ Exit the editor, writing out the buffer if it was changed since the last write (same as the ex command x). Note that if the last write was to a different file and no changes have occurred since, the editor exits without writing out the buffer. a Enter input mode, appending the entered text after the current cursor position. A preceding count causes the inserted text to be replicated the specified number of times, but only if the inserted text is all on one line. b Back up to the previous beginning of a word in the current line. A word is a sequence of alphanumerics or a sequence of special characters. A preceding count repeats the effect. c Must be followed by a movement command. Delete the specified region of text, and enter input mode to replace deleted text with new text. If more than part of a single line is affected, the deleted text is saved in the numeric buffers. If only part of the current line is affected, the last character deleted is marked with a $. A preceding count passes that value through to the move command. If the command is cc, the entire current line is changed. d Must be followed by a movement command. Delete the specified region of text. If more than part of a line is affected, the text is saved in the numeric buffers. A preceding count passes that value through to the move command. If the command is dd, the entire current line is deleted. e Move forward to the end of the next word, defined as for b. A preceding count repeats the effect. f Must be followed by a single character; scan the rest of the current line for that character, and moves the cursor to it if found. A preceding count repeats the action that many times. h Move the cursor one character to the left (same as ^H). A preceding count repeats the effect. i Enter input mode, inserting the entered text before the cursor (see a). j Move the cursor one line down in the same column (same as ^J and ^N). k Move the cursor one line up (same as ^P). l Move the cursor one character to the right (same as <space>). mx Mark the current position of the cursor. x is a lowercase letter, a-z, that is used with the ` and ' commands to refer to the marked line or line position. n Repeat the last / or ? scanning commands. o Open a line below the current line and enter input mode; otherwise like O. p Put text after/below the cursor; otherwise like P. r Must be followed by a single character; the character under the cursor is replaced by the specified one. (The new character can be a new-line.) If r is preceded by a count, count characters are replaced by the specified character. s Delete the single character under the cursor and enter input mode; the entered text replaces the deleted character. A preceding count specifies how many characters on the current line are changed. The last character being changed is marked with a $, as for c. t Must be followed by a single character; scan the remainder of the line for that character. The cursor moves to the column prior to the character if the character is found. A preceding count is equivalent to repeating the search count times. u Reverse the last change made to the current buffer. If repeated, u alternates between these two states; thus is its own inverse. When used after an insertion of text on more than one line, the lines are saved in the numerically named buffers. w Move forward to the beginning of the next word (where word is defined as in b). A preceding count specifies how many words the cursor advances. x Delete the single character under the cursor. When x is preceded by a count, x deletes the specified number of characters forward from the cursor position, but only on the current line. y Must be followed by a movement command; the specified text is copied (yanked) into the unnamed temporary buffer. If preceded by a named buffer specification, "x, the text is placed in that buffer also. If the command is yy, the entire current line is yanked. z Redraw the screen with the current line placed as specified by the following options: z<return> specifies the top of the screen, z. the center of the screen, and z- the bottom of the screen. The commands z^ and z+ are similar to ^B and ^F, respectively. However, z^ and z+ do not attempt to maintain two lines of overlap. A count after the z and before the following character to specifies the number of lines displayed in the redrawn screen. A count before the z gives the number of the line to use as the reference line instead of the default current line. Keyboard Editing Keys At initialization, the editor automatically maps some terminal keyboard editing keys to equivalent visual mode commands. These mappings are only established for keys that are listed in the following table and defined in the terminfo(4) database as valid for the current terminal (as specified by the TERM environment variable). Both command and input mode mappings are created (see the map command in ex(1)). With the exception of the insertchar keys, which simply toggle input mode on and off, the input mode mappings exit input mode, perform the same action as the command mode mapping, and then reenter input mode. On certain terminals, the character sequence sent by a keyboard editing key, which is then mapped to a visual mode command, can be the same character sequence a user might enter to perform another command or set of commands. This is most likely to happen with the input mode mappings; therefore, on these terminals, the input mode mappings are disabled by default. Users can override the disabling and enabling of both the command and input mode keyboard editing key mappings by setting the keyboardedit and keyboardedit! editor options as appropriate (see ex(1)). The timeout, timeoutlen, and doubleescape editor options are alternative methods of addressing this problem. terminfo command input map entry mode map mode map name description __________________________________________________________________ key_ic i ^[ inschar insert char key_eic i ^[ inschar end insert char key_up k ^[ka up arrow up key_down j ^[ja down arrow down key_left h ^[ha left arrow left key_right l ^[la right arrow right key_home H ^[Ha home arrow home key_il o^[ ^[o^[a insline insert line key_dl dd ^[dda delline delete line key_clear ^L ^[^La clear clear screen key_eol d$ ^[d$a clreol clear line key_sf ^E ^[^Ea scrollf scroll down key_dc x ^[xa delchar delete char key_npage ^F ^[^Fa npage next page key_ppage ^B ^[^Ba ppage previous page key_sr ^Y ^[^Ya sr scroll up key_eos dG ^[dGa clreos clear to end of screen __________________________________________________________________ EXTERNAL INFLUENCES Support for international codes and environment variables are as follows: Environment Variables UNIX95 specifies using the XPG4 behaviour for this command. COLUMNS overrides the system-selected horizontal screen size. LINES overrides the system-selected vertical screen size, used as the number of lines in a screenful and the vertical screen size in visual mode. SHELL is a variable that shall be interpreted as the preferred command-line interpreter for use in !, shell, read, and other commands with an operand of the form !string. For the shell command the program shall be invoked with the two arguments -c and string. If this variable is null or not set, the sh utility shall be used. TERM is a variable that shall be interpreted as the name of the terminal type. If this variable is unset or null, an unspecified default terminal type shall be used. PATH determines the search path for the shell command specified in the editor commands, shell, read, and write. EXINIT determines a list of ex commands that will be executed on editor startup, before reading the first file. The list can contain multiple commands by separating them using a vertical line (|) character. HOME determines a pathname of a directory that will be searched for an editor startup file named .exrc. LC_ALL This variable shall determine the locale to be used to override any values for locale categories specified by the setting of LANG or any environment variables beginning with LC_. LC_MESSAGES determines the locale that should be used to affect the format and contents of diagnostic messages written to standard error and informative messages written to standard output. LC_COLLATE determines the collating sequence used in evaluating regular expressions and in processing the tags file. LC_CTYPE determines the interpretation of text as single and/or multi-byte characters, the classification of characters as uppercase or lowercase letters, the shifting of letters between uppercase and lowercase, and the characters matched by character class expressions in regular expressions. LANG determines the language in which messages are displayed. LANGOPTS specifies options determining how text for right-to-left languages is stored in input and output files. See environ(5). If LC_COLLATE or LC_CTYPE is not specified in the environment or is set to the empty string, the value of LANG is used as a default for each unspecified or empty variable. If LANG is not specified or is set to the empty string, a default of "C" (see lang(5)) is used instead of LANG. If any internationalization variable contains an invalid setting, the editor behaves as if all internationalization variables are set to "C". See environ(5). International Code Set Support Single- and multi-byte character code sets are supported. WARNINGS See also the WARNINGS section in ex(1). Program Limits vi places the following limits on files being edited: Maximum Line Length LINE_MAX characters (defined in <limits.h>), including 2-3 bytes for overhead. Thus, if the value specified for LINE_MAX is 2048, a line length up to 2044 characters should cause no problem. If you load a file that contain lines longer than the specified limit, the lines are truncated to the stated maximum length. Saving the file will write the truncated version over the original file, thus overwriting the original lines completely. Attempting to create lines longer than the allowable maximum for the editor produces a line too long error message. Maximum File Size The maximum file length of 234,239 lines is silently enforced. Other limits: - 256 characters per global command list. - 128 characters in a file name in vi or ex open mode. On short-file-name HP-UX systems, the maximum file name length is 14 characters. - 128 characters in a previous insert/delete buffer. - 100 characters in a shell-escape command. - 63 characters in a string-valued option (:set command). - 30 characters in a program tag name. - 32 or fewer macros defined by map command. - 512 or fewer characters total in combined map macros. AUTHOR vi was developed by the University of California, Berkeley. The 16- bit extensions to vi are based in part on software of the Toshiba Corporation. SEE ALSO ctags(1), ed(1), ex(1), stty(1), write(1), terminfo(4), environ(5), lang(5), regexp(5). The Ultimate Guide to the vi and ex Text Editors, Benjamin/Cummings Publishing Company, Inc., ISBN 0-8053-4460-8, HP part number 97005-90015. STANDARDS CONFORMANCE vi: SVID2, SVID3, XPG2, XPG3, XPG4
CONTENTS |