Utilities

The REAL/IX Operating System includes the standard UNIX System V utilities. This chapter gives a brief description of the most commonly used utilities; full information on these and other utilities are given in the manual pages.

File Access Utilities

The following user commands are used for accessing and manipulating files and directories.

Editors

The REAL/IX Operating System supports a variety of editors:

When using ed and vi, you are always working on a copy of the file in a temporary buffer; the disk copy of the file is unaffected until you explicitly issue a w (write) command. Consequently, if you make a major mistake during an editing session, you can regain the unaltered file by issuing a q (quit) command without writing the file.

ed

The ed(1) editor can be used on either a VDT or a hard-copy terminal. Most new users prefer vi(1) for normal editing tasks, but ed is still useful in shell scripts.

ed is a line editor; during editing sessions, it always points to a single line in the file called the current line. Unless you specify a line or range of lines, ed applies each command you issue to the current line. When you access an existing file, ed makes the last line the current line so you can append text easily.

To begin an ed editing session, enter the command ed filename. If this is a new file, ed will respond "?new-file." If you are editing an existing file, ed responds with the number of characters in the file. Once in ed, you work in one of two modes: append mode or edit mode. In append mode, you add to the contents of a file; in edit mode you can display, delete, or replace text and move about the file. In edit mode, you can use shell regular expressions for search and substitute operations.

The basic ed commands are:

<CR>; -move down one line in the buffer

Lines can be specified to ed in the following ways:

ed also includes a substitution facility, with the syntax:

begin_line/end_lines/old_text/new_text/command

where command can be g, l, n, p, or gp.

vi

The vi(1) editor allows you to view and edit a screen-full of text. Text is inserted at the location of the cursor. vi depends on the block-mode definition of your terminal established in the terminfo files and set with the $TERM environmental variable; be sure this is set up correctly before attempting to run vi.

To begin a vi editing session, enter a "vi filename" command at the shell. For a new file, vi will return an empty screen with tilde characters (~) in the left column of each line; for an existing file, vi will display the first screen of the file. Like ed(1), vi has an append mode and an edit mode.

There are more than 100 commands within vi; the following lists identify a basic set. vi uses "control characters" for a number of functions; these are denoted using the ^ character. For instance, "control D" (^d) is typed by holding the CTRL key down and pressing the D key.

ESC -(ESCAPE key) leave append mode and return to edit mode

The search and replace functionality of ed can be accessed through the ex(1) editor, executed by entering ":" from edit mode.

sed

sed(1) is a stream-oriented version of ed(1). It applies a list of editor commands to each line in one or more files, writing the resulting modified lines to standard output (or redirected to a file). The lines to be edited can be selected by regular expressions.

sed is a convenient tool for removing unwanted fields, transforming all occurrences of one string to another string, or adding a new field to each line. As a very simple example, the following shell script replaces every occurrence of the string "cat" with "dog":

sed 's/cat/dog/g' filename

You can create a file that includes the sed specifications, then identify that file on the sed command line. For instance, if you create a file called addblank that contains the following:

a

^D

then issue the following command:
sed -f addblank filename

a blank line is inserted after every line in filename.

sed is most commonly used in scripts to edit existing files on the system. For instance, if an add-on package requires that lines be added to the /etc/profile file or /dev/README, the script can use sed to add lines at the end. Whatever is in the file will be left as is, and the required information is added at the end. See the sed manual page or a general UNIX system text for more information and instructions on using sed.

Filters

A number of REAL/IX Operating System utilities are classified as filters. A filter reads its standard input, transforms the data in some way, and writes the result to its standard output. Filters are commonly used together, with the output from one piped to the next:

filter1 < input-file | filter2 | filter3 > output-file

Bytes flow from the input file to the output file through the pipe. Each filter performs a rather small, specialized function: the first filter might translate certain characters, the second filter might rearrange the lines (typically by sorting), and the third might eliminate duplicate lines. By using several filters together, they perform a more complex function.

The following sections give an overview of some of the more common filters.

grep

grep(1) selects lines from a file, based on matching a pattern (or regular expression, as discussed in Metacharacters, Chapter 7). The pattern is commonly just a simple string. For example, to search the current directory for all references to fcntl(2), the command is:

grep fcntl *.c

Useful options to grep can select lines that do not match the pattern, can print the number of each matching line, and can make a lowercase pattern match an uppercase pattern as well. By using grep with regular expressions, you can specify complex matching criteria, such as "lines that begin or end with the specified pattern," "lines containing pattern1 followed by pattern2," or "a line that contains the nth occurrence of the specified pattern."

fgrep(1) and egrep(1) are similar to grep. fgrep can search for multiple patterns in parallel, but can only search for fixed strings, not regular expressions. egrep is faster than grep under some circumstances and generalizes the notion of regular expression. egrep can, for example, extract lines containing any of several alternate regular expressions; in other words, the logical OR of several regular expressions.

awk

awk(1) is a pattern matching programming language whose syntax is modeled after C; it is often used as a report generator. It is named for its Bell Laboratories authors, Aho, Weinberger, and Kernighan. Like grep(1), awk selects lines from its standard input according to specified criteria that are applied in a single pass over the input stream. awk can then perform actions on the data it selects.

The action statements operate on fields (by default, delimited with spaces in tabs) in the selected line and on variables you declare. Actions include arithmetic, conditional tests, and string processing. Fields, modified fields, constants, and variables can be written to the standard output. Special actions, such as writing totals, can be taken at the beginning and end of the output stream.

sort

sort(1) sorts the lines in a file. The lines can contain any ASCII characters. Options to sort permit sorting on multiple fields (delimited by blanks, semicolons, slashes, or other characters), on specific bytes within fields, into ascending or descending order, ignoring case or "white space" (blanks and tabs), and so on. sort can also merge sorted files, optionally eliminating duplicate lines.

Communication

The REAL/IX Operating System supports the UNIX System V utilities for sending and receiving mail and files.

Sending Mail

Two commands are available for sending mail: mail(1) is the simplest, mailx(1) is a more sophisticated electronic mail facility. The two are compatible: mail sent with mailx can be received with mail and vice versa.

To send mail with the mail command, do the following:

mail person1 [person2 person3 ...] <CR>
insert lines of message <CR>
and more lines of message <CR>
.<CR>

The message is terminated with a . or, optionally, a ^D character. mail can also take a file as input using the shell redirect character:

mail < filename

mailx is a more sophisticated electronic mail facility. The system prompts you for a Subject which is displayed when the recipient uses mailx to read mail. In addition, you can define an alias of several users to streamline sending mail to groups and escape to an editor while sending mail

To send mail with the mailx command, do the following:

mailx person1 [or alias or list of people] <CR>

The system will prompt you for a Subject:; after you enter that, you can type your message as for mail; terminate the input with a ^D character. To escape to an editor while entering the message, type ~V (vi(1)) or ~e (ed(1)) at the beginning of a line.

To define aliases and do other mailx customization, set up a .mailrc file in your home directory, for instance:

set askcc append metoo crt=22
set record=$HOME/mail/outbox MBOX=$HOME/mail/mbox
set sign=Joe Smith, Manager of Software Systems
group myfolks bob mary jim sue

The "set record" line establishes the path name for incoming and outgoing mail (mailx automatically archives a copy of all outgoing mail. The "set sign" line creates a standard signature that can be called with ~a (autograph). The "group" line establishes an alias called "myfolks"; the mailx myfolks command is then equivalent to mailx bob mary jim sue. The mailx(1) manual page gives a full list of options that can be set in the .mailrc file.

Receiving Mail

The mail(1) and mailx(1) commands without names are also used to receive mail. mail will display all messages in your mailbox (/usr/mail/name); with the -r option, messages are displayed in the order they were posted; otherwise the most recently posted is displayed first.

After each message, the system gives you a ? prompt. Your response determines what happens to the message, such as

Entering mailx with no names displays a listing of all messages in your mailbox including any Subject headers posted with the message. Each message has a number used to identify the message; this number is used with a command. The following list of commands all address message #2 in the mailbox:

Console Management

Proper console management can allow you to interact with more than one process per terminal. conman(1) is the REAL/IX utility to assist you with console management.

conman

conman(1) allows a user to interact with more than one shell or terminal session from a single terminal. The user controls these shells or sessions, known as layers, using various commands. The current layer is the layer which can receive input from or send output to the terminal. Other layers attempting to read from and write to the terminal are blocked. When the output of a non-current layer is important and should be viewed immediately, the alarm option can be set; if the current layer is not so enabled, the alarm layer will automatically become the current layer when it has pending output for the screen. Access between layers is controlled by entering ^z (the CTRL-z character); in the current implementation, only this character can be used as the switch character. A layer is bound to a pseudo tty device and is manipulated like a real tty device using stty(1) and ioctl(2). For more detailed information about conman and its options and commands, refer to the conman manual pages.

Go to Chapter 9 TOC