Advanced Bash-Scripting Guide

An in-depth exploration of the art of shell scripting

Mendel Cooper


10 Mar 2014

Revision History
Revision 6.505 Apr 2012Revised by: mc
Revision 6.627 Nov 2012Revised by: mc
Revision 1010 Mar 2014Revised by: mc

This tutorial assumes no previous knowledge of scripting or programming, yet progresses rapidly toward an intermediate/advanced level of instruction . . . all the while sneaking in little nuggets of UNIX® wisdom and lore. It serves as a textbook, a manual for self-study, and as a reference and source of knowledge on shell scripting techniques. The exercises and heavily-commented examples invite active reader participation, under the premise that the only way to really learn scripting is to write scripts.

This book is suitable for classroom use as a general introduction to programming concepts.

This document is herewith granted to the Public Domain. No copyright!


For Anita, the source of all the magic

Table of Contents
Part 1. Introduction
1. Shell Programming!
2. Starting Off With a Sha-Bang
Part 2. Basics
3. Special Characters
4. Introduction to Variables and Parameters
5. Quoting
6. Exit and Exit Status
7. Tests
8. Operations and Related Topics
Part 3. Beyond the Basics
9. Another Look at Variables
10. Manipulating Variables
11. Loops and Branches
12. Command Substitution
13. Arithmetic Expansion
14. Recess Time
Part 4. Commands
15. Internal Commands and Builtins
16. External Filters, Programs and Commands
17. System and Administrative Commands
Part 5. Advanced Topics
18. Regular Expressions
19. Here Documents
20. I/O Redirection
21. Subshells
22. Restricted Shells
23. Process Substitution
24. Functions
25. Aliases
26. List Constructs
27. Arrays
28. Indirect References
29. /dev and /proc
30. Network Programming
31. Of Zeros and Nulls
32. Debugging
33. Options
34. Gotchas
35. Scripting With Style
36. Miscellany
37. Bash, versions 2, 3, and 4
38. Endnotes
38.1. Author's Note
38.2. About the Author
38.3. Where to Go For Help
38.4. Tools Used to Produce This Book
38.5. Credits
38.6. Disclaimer
A. Contributed Scripts
B. Reference Cards
C. A Sed and Awk Micro-Primer
C.1. Sed
C.2. Awk
D. Parsing and Managing Pathnames
E. Exit Codes With Special Meanings
F. A Detailed Introduction to I/O and I/O Redirection
G. Command-Line Options
G.1. Standard Command-Line Options
G.2. Bash Command-Line Options
H. Important Files
I. Important System Directories
J. An Introduction to Programmable Completion
K. Localization
L. History Commands
M. Sample .bashrc and .bash_profile Files
N. Converting DOS Batch Files to Shell Scripts
O. Exercises
O.1. Analyzing Scripts
O.2. Writing Scripts
P. Revision History
Q. Download and Mirror Sites
R. To Do List
S. Copyright
T. ASCII Table
List of Tables
8-1. Operator Precedence
15-1. Job identifiers
33-1. Bash options
36-1. Numbers representing colors in Escape Sequences
B-1. Special Shell Variables
B-2. TEST Operators: Binary Comparison
B-3. TEST Operators: Files
B-4. Parameter Substitution and Expansion
B-5. String Operations
B-6. Miscellaneous Constructs
C-1. Basic sed operators
C-2. Examples of sed operators
E-1. Reserved Exit Codes
N-1. Batch file keywords / variables / operators, and their shell equivalents
N-2. DOS commands and their UNIX equivalents
P-1. Revision History
List of Examples
2-1. cleanup: A script to clean up log files in /var/log
2-2. cleanup: An improved clean-up script
2-3. cleanup: An enhanced and generalized version of above scripts.
3-1. Code blocks and I/O redirection
3-2. Saving the output of a code block to a file
3-3. Running a loop in the background
3-4. Backup of all files changed in last day
4-1. Variable assignment and substitution
4-2. Plain Variable Assignment
4-3. Variable Assignment, plain and fancy
4-4. Integer or string?
4-5. Positional Parameters
4-6. wh, whois domain name lookup
4-7. Using shift
5-1. Echoing Weird Variables
5-2. Escaped Characters
5-3. Detecting key-presses
6-1. exit / exit status
6-2. Negating a condition using !
7-1. What is truth?
7-2. Equivalence of test, /usr/bin/test, [ ], and /usr/bin/[
7-3. Arithmetic Tests using (( ))
7-4. Testing for broken links
7-5. Arithmetic and string comparisons
7-6. Testing whether a string is null
7-7. zmore
8-1. Greatest common divisor
8-2. Using Arithmetic Operations
8-3. Compound Condition Tests Using && and ||
8-4. Representation of numerical constants
8-5. C-style manipulation of variables
9-1. $IFS and whitespace
9-2. Timed Input
9-3. Once more, timed input
9-4. Timed read
9-5. Am I root?
9-6. arglist: Listing arguments with $* and $@
9-7. Inconsistent $* and $@ behavior
9-8. $* and $@ when $IFS is empty
9-9. Underscore variable
9-10. Using declare to type variables
9-11. Generating random numbers
9-12. Picking a random card from a deck
9-13. Brownian Motion Simulation
9-14. Random between values
9-15. Rolling a single die with RANDOM
9-16. Reseeding RANDOM
9-17. Pseudorandom numbers, using awk
10-1. Inserting a blank line between paragraphs in a text file
10-2. Generating an 8-character "random" string
10-3. Converting graphic file formats, with filename change
10-4. Converting streaming audio files to ogg
10-5. Emulating getopt
10-6. Alternate ways of extracting and locating substrings
10-7. Using parameter substitution and error messages
10-8. Parameter substitution and "usage" messages
10-9. Length of a variable
10-10. Pattern matching in parameter substitution
10-11. Renaming file extensions:
10-12. Using pattern matching to parse arbitrary strings
10-13. Matching patterns at prefix or suffix of string
11-1. Simple for loops
11-2. for loop with two parameters in each [list] element
11-3. Fileinfo: operating on a file list contained in a variable
11-4. Operating on a parameterized file list
11-5. Operating on files with a for loop
11-6. Missing in [list] in a for loop
11-7. Generating the [list] in a for loop with command substitution
11-8. A grep replacement for binary files
11-9. Listing all users on the system
11-10. Checking all the binaries in a directory for authorship
11-11. Listing the symbolic links in a directory
11-12. Symbolic links in a directory, saved to a file
11-13. A C-style for loop
11-14. Using efax in batch mode
11-15. Simple while loop
11-16. Another while loop
11-17. while loop with multiple conditions
11-18. C-style syntax in a while loop
11-19. until loop
11-20. Nested Loop
11-21. Effects of break and continue in a loop
11-22. Breaking out of multiple loop levels
11-23. Continuing at a higher loop level
11-24. Using continue N in an actual task
11-25. Using case
11-26. Creating menus using case
11-27. Using command substitution to generate the case variable
11-28. Simple string matching
11-29. Checking for alphabetic input
11-30. Creating menus using select
11-31. Creating menus using select in a function
12-1. Stupid script tricks
12-2. Generating a variable from a loop
12-3. Finding anagrams
15-1. A script that spawns multiple instances of itself
15-2. printf in action
15-3. Variable assignment, using read
15-4. What happens when read has no variable
15-5. Multi-line input to read
15-6. Detecting the arrow keys
15-7. Using read with file redirection
15-8. Problems reading from a pipe
15-9. Changing the current working directory
15-10. Letting let do arithmetic.
15-11. Showing the effect of eval
15-12. Using eval to select among variables
15-13. Echoing the command-line parameters
15-14. Forcing a log-off
15-15. A version of rot13
15-16. Using set with positional parameters
15-17. Reversing the positional parameters
15-18. Reassigning the positional parameters
15-19. "Unsetting" a variable
15-20. Using export to pass a variable to an embedded awk script
15-21. Using getopts to read the options/arguments passed to a script
15-22. "Including" a data file
15-23. A (useless) script that sources itself
15-24. Effects of exec
15-25. A script that exec's itself
15-26. Waiting for a process to finish before proceeding
15-27. A script that kills itself
16-1. Using ls to create a table of contents for burning a CDR disk
16-2. Hello or Good-bye
16-3. Badname, eliminate file names in current directory containing bad characters and whitespace.
16-4. Deleting a file by its inode number
16-5. Logfile: Using xargs to monitor system log
16-6. Copying files in current directory to another
16-7. Killing processes by name
16-8. Word frequency analysis using xargs
16-9. Using expr
16-10. Using date
16-11. Date calculations
16-12. Word Frequency Analysis
16-13. Which files are scripts?
16-14. Generating 10-digit random numbers
16-15. Using tail to monitor the system log
16-16. Printing out the From lines in stored e-mail messages
16-17. Emulating grep in a script
16-18. Crossword puzzle solver
16-19. Looking up definitions in Webster's 1913 Dictionary
16-20. Checking words in a list for validity
16-21. toupper: Transforms a file to all uppercase.
16-22. lowercase: Changes all filenames in working directory to lowercase.
16-23. du: DOS to UNIX text file conversion.
16-24. rot13: ultra-weak encryption.
16-25. Generating "Crypto-Quote" Puzzles
16-26. Formatted file listing.
16-27. Using column to format a directory listing
16-28. nl: A self-numbering script.
16-29. manview: Viewing formatted manpages
16-30. Using cpio to move a directory tree
16-31. Unpacking an rpm archive
16-32. Stripping comments from C program files
16-33. Exploring /usr/X11R6/bin
16-34. An "improved" strings command
16-35. Using cmp to compare two files within a script.
16-36. basename and dirname
16-37. A script that copies itself in sections
16-38. Checking file integrity
16-39. Uudecoding encoded files
16-40. Finding out where to report a spammer
16-41. Analyzing a spam domain
16-42. Getting a stock quote
16-43. Updating FC4
16-44. Using ssh
16-45. A script that mails itself
16-46. Generating prime numbers
16-47. Monthly Payment on a Mortgage
16-48. Base Conversion
16-49. Invoking bc using a here document
16-50. Calculating PI
16-51. Converting a decimal number to hexadecimal
16-52. Factoring
16-53. Calculating the hypotenuse of a triangle
16-54. Using seq to generate loop arguments
16-55. Letter Count"
16-56. Using getopt to parse command-line options
16-57. A script that copies itself
16-58. Exercising dd
16-59. Capturing Keystrokes
16-60. Preparing a bootable SD card for the Raspberry Pi
16-61. Securely deleting a file
16-62. Filename generator
16-63. Converting meters to miles
16-64. Using m4
17-1. Setting a new password
17-2. Setting an erase character
17-3. secret password: Turning off terminal echoing
17-4. Keypress detection
17-5. Checking a remote server for identd
17-6. pidof helps kill a process
17-7. Checking a CD image
17-8. Creating a filesystem in a file
17-9. Adding a new hard drive
17-10. Using umask to hide an output file from prying eyes
17-11. Backlight: changes the brightness of the (laptop) screen backlight
17-12. killall, from /etc/rc.d/init.d
19-1. broadcast: Sends message to everyone logged in
19-2. dummyfile: Creates a 2-line dummy file
19-3. Multi-line message using cat
19-4. Multi-line message, with tabs suppressed
19-5. Here document with replaceable parameters
19-6. Upload a file pair to Sunsite incoming directory
19-7. Parameter substitution turned off
19-8. A script that generates another script
19-9. Here documents and functions
19-10. "Anonymous" Here Document
19-11. Commenting out a block of code
19-12. A self-documenting script
19-13. Prepending a line to a file
19-14. Parsing a mailbox
20-1. Redirecting stdin using exec
20-2. Redirecting stdout using exec
20-3. Redirecting both stdin and stdout in the same script with exec
20-4. Avoiding a subshell
20-5. Redirected while loop
20-6. Alternate form of redirected while loop
20-7. Redirected until loop
20-8. Redirected for loop
20-9. Redirected for loop (both stdin and stdout redirected)
20-10. Redirected if/then test
20-11. Data file for above examples
20-12. Logging events
21-1. Variable scope in a subshell
21-2. List User Profiles
21-3. Running parallel processes in subshells
22-1. Running a script in restricted mode
23-1. Code block redirection without forking
23-2. Redirecting the output of process substitution into a loop.
24-1. Simple functions
24-2. Function Taking Parameters
24-3. Functions and command-line args passed to the script
24-4. Passing an indirect reference to a function
24-5. Dereferencing a parameter passed to a function
24-6. Again, dereferencing a parameter passed to a function
24-7. Maximum of two numbers
24-8. Converting numbers to Roman numerals
24-9. Testing large return values in a function
24-10. Comparing two large integers
24-11. Real name from username
24-12. Local variable visibility
24-13. Demonstration of a simple recursive function
24-14. Another simple demonstration
24-15. Recursion, using a local variable
24-16. The Fibonacci Sequence
24-17. The Towers of Hanoi
25-1. Aliases within a script
25-2. unalias: Setting and unsetting an alias
26-1. Using an and list to test for command-line arguments
26-2. Another command-line arg test using an and list
26-3. Using or lists in combination with an and list
27-1. Simple array usage
27-2. Formatting a poem
27-3. Various array operations
27-4. String operations on arrays
27-5. Loading the contents of a script into an array
27-6. Some special properties of arrays
27-7. Of empty arrays and empty elements
27-8. Initializing arrays
27-9. Copying and concatenating arrays
27-10. More on concatenating arrays
27-11. The Bubble Sort
27-12. Embedded arrays and indirect references
27-13. The Sieve of Eratosthenes
27-14. The Sieve of Eratosthenes, Optimized
27-15. Emulating a push-down stack
27-16. Complex array application: Exploring a weird mathematical series
27-17. Simulating a two-dimensional array, then tilting it
28-1. Indirect Variable References
28-2. Passing an indirect reference to awk
29-1. Using /dev/tcp for troubleshooting
29-2. Playing music
29-3. Finding the process associated with a PID
29-4. On-line connect status
30-1. Print the server environment
30-2. IP addresses
31-1. Hiding the cookie jar
31-2. Setting up a swapfile using /dev/zero
31-3. Creating a ramdisk
32-1. A buggy script
32-2. Missing keyword
32-3. test24: another buggy script
32-4. Testing a condition with an assert
32-5. Trapping at exit
32-6. Cleaning up after Control-C
32-7. A Simple Implementation of a Progress Bar
32-8. Tracing a variable
32-9. Running multiple processes (on an SMP box)
34-1. Numerical and string comparison are not equivalent
34-2. Subshell Pitfalls
34-3. Piping the output of echo to a read
36-1. shell wrapper
36-2. A slightly more complex shell wrapper
36-3. A generic shell wrapper that writes to a logfile
36-4. A shell wrapper around an awk script
36-5. A shell wrapper around another awk script
36-6. Perl embedded in a Bash script
36-7. Bash and Perl scripts combined
36-8. Python embedded in a Bash script
36-9. A script that speaks
36-10. A (useless) script that recursively calls itself
36-11. A (useful) script that recursively calls itself
36-12. Another (useful) script that recursively calls itself
36-13. A "colorized" address database
36-14. Drawing a box
36-15. Echoing colored text
36-16. A "horserace" game
36-17. A Progress Bar
36-18. Return value trickery
36-19. Even more return value trickery
36-20. Passing and returning arrays
36-21. Fun with anagrams
36-22. Widgets invoked from a shell script
36-23. Test Suite
37-1. String expansion
37-2. Indirect variable references - the new way
37-3. Simple database application, using indirect variable referencing
37-4. Using arrays and other miscellaneous trickery to deal four random hands from a deck of cards
37-5. A simple address database
37-6. A somewhat more elaborate address database
37-7. Testing characters
37-8. Reading N characters
37-9. Using a here document to set a variable
37-10. Piping input to a read
37-11. Negative array indices
37-12. Negative parameter in string-extraction construct
A-1. mailformat: Formatting an e-mail message
A-2. rn: A simple-minded file renaming utility
A-3. blank-rename: Renames filenames containing blanks
A-4. encryptedpw: Uploading to an ftp site, using a locally encrypted password
A-5. copy-cd: Copying a data CD
A-6. Collatz series
A-7. days-between: Days between two dates
A-8. Making a dictionary
A-9. Soundex conversion
A-10. Game of Life
A-11. Data file for Game of Life
A-12. behead: Removing mail and news message headers
A-13. password: Generating random 8-character passwords
A-14. fifo: Making daily backups, using named pipes
A-15. Generating prime numbers using the modulo operator
A-16. tree: Displaying a directory tree
A-17. tree2: Alternate directory tree script
A-18. string functions: C-style string functions
A-19. Directory information
A-20. Library of hash functions
A-21. Colorizing text using hash functions
A-22. More on hash functions
A-23. Mounting USB keychain storage devices
A-24. Converting to HTML
A-25. Preserving weblogs
A-26. Protecting literal strings
A-27. Unprotecting literal strings
A-28. Spammer Identification
A-29. Spammer Hunt
A-30. Making wget easier to use
A-31. A podcasting script
A-32. Nightly backup to a firewire HD
A-33. An expanded cd command
A-34. A soundcard setup script
A-35. Locating split paragraphs in a text file
A-36. Insertion sort
A-37. Standard Deviation
A-38. A pad file generator for shareware authors
A-39. A man page editor
A-40. Petals Around the Rose
A-41. Quacky: a Perquackey-type word game
A-42. Nim
A-43. A command-line stopwatch
A-44. An all-purpose shell scripting homework assignment solution
A-45. The Knight's Tour
A-46. Magic Squares
A-47. Fifteen Puzzle
A-48. The Towers of Hanoi, graphic version
A-49. The Towers of Hanoi, alternate graphic version
A-50. An alternate version of the script
A-51. The version of the example used in the Tab Expansion appendix
A-52. Cycling through all the possible color backgrounds
A-53. Morse Code Practice
A-54. Base64 encoding/decoding
A-55. Inserting text in a file using sed
A-56. The Gronsfeld Cipher
A-57. Bingo Number Generator
A-58. Basics Reviewed
A-59. Testing execution times of various commands
A-60. Associative arrays vs. conventional arrays (execution times)
C-1. Counting Letter Occurrences
J-1. Completion script for
M-1. Sample .bashrc file
M-2. .bash_profile file
N-2. Shell Script Conversion of VIEWDATA.BAT
T-1. A script that generates an ASCII table
T-2. Another ASCII table script
T-3. A third ASCII table script, using awk

Part 1. Introduction


Script: A writing; a written document. [Obs.]

--Webster's Dictionary, 1913 ed.

The shell is a command interpreter. More than just the insulating layer between the operating system kernel and the user, it's also a fairly powerful programming language. A shell program, called a script, is an easy-to-use tool for building applications by "gluing together" system calls, tools, utilities, and compiled binaries. Virtually the entire repertoire of UNIX commands, utilities, and tools is available for invocation by a shell script. If that were not enough, internal shell commands, such as testing and loop constructs, lend additional power and flexibility to scripts. Shell scripts are especially well suited for administrative system tasks and other routine repetitive tasks not requiring the bells and whistles of a full-blown tightly structured programming language.

Chapter 1. Shell Programming!


No programming language is perfect. There is not even a single best language; there are only languages well suited or perhaps poorly suited for particular purposes.

--Herbert Mayer

A working knowledge of shell scripting is essential to anyone wishing to become reasonably proficient at system administration, even if they do not anticipate ever having to actually write a script. Consider that as a Linux machine boots up, it executes the shell scripts in /etc/rc.d to restore the system configuration and set up services. A detailed understanding of these startup scripts is important for analyzing the behavior of a system, and possibly modifying it.

The craft of scripting is not hard to master, since scripts can be built in bite-sized sections and there is only a fairly small set of shell-specific operators and options [1] to learn. The syntax is simple -- even austere -- similar to that of invoking and chaining together utilities at the command line, and there are only a few "rules" governing their use. Most short scripts work right the first time, and debugging even the longer ones is straightforward.

    In the early days of personal computing, the BASIC language enabled
    anyone reasonably computer proficient to write programs on an early
    generation of microcomputers. Decades later, the Bash scripting
    language enables anyone with a rudimentary knowledge of Linux or
    UNIX to do the same on modern machines.

    We now have miniaturized single-board computers with amazing
    capabilities, such as the Raspberry Pi.
    Bash scripting provides a way to explore the capabilities of these
    fascinating devices.

A shell script is a quick-and-dirty method of prototyping a complex application. Getting even a limited subset of the functionality to work in a script is often a useful first stage in project development. In this way, the structure of the application can be tested and tinkered with, and the major pitfalls found before proceeding to the final coding in C, C++, Java, Perl, or Python.

Shell scripting hearkens back to the classic UNIX philosophy of breaking complex projects into simpler subtasks, of chaining together components and utilities. Many consider this a better, or at least more esthetically pleasing approach to problem solving than using one of the new generation of high-powered all-in-one languages, such as Perl, which attempt to be all things to all people, but at the cost of forcing you to alter your thinking processes to fit the tool.

According to Herbert Mayer, "a useful language needs arrays, pointers, and a generic mechanism for building data structures." By these criteria, shell scripting falls somewhat short of being "useful." Or, perhaps not. . . .

We will be using Bash, an acronym [3] for "Bourne-Again shell" and a pun on Stephen Bourne's now classic Bourne shell. Bash has become a de facto standard for shell scripting on most flavors of UNIX. Most of the principles this book covers apply equally well to scripting with other shells, such as the Korn Shell, from which Bash derives some of its features, [4] and the C Shell and its variants. (Note that C Shell programming is not recommended due to certain inherent problems, as pointed out in an October, 1993 Usenet post by Tom Christiansen.)

What follows is a tutorial on shell scripting. It relies heavily on examples to illustrate various features of the shell. The example scripts work -- they've been tested, insofar as possible -- and some of them are even useful in real life. The reader can play with the actual working code of the examples in the source archive ( or scriptname.bash), [5] give them execute permission (chmod u+rx scriptname), then run them to see what happens. Should the source archive not be available, then cut-and-paste from the HTML or pdf rendered versions. Be aware that some of the scripts presented here introduce features before they are explained, and this may require the reader to temporarily skip ahead for enlightenment.

Unless otherwise noted, the author of this book wrote the example scripts that follow.


His countenance was bold and bashed not.

--Edmund Spenser

Chapter 2. Starting Off With a Sha-Bang


Shell programming is a 1950s juke box . . .

--Larry Wall

In the simplest case, a script is nothing more than a list of system commands stored in a file. At the very least, this saves the effort of retyping that particular sequence of commands each time it is invoked.

Example 2-1. cleanup: A script to clean up log files in /var/log

# Cleanup
# Run as root, of course.

cd /var/log
cat /dev/null > messages
cat /dev/null > wtmp
echo "Log files cleaned up."

There is nothing unusual here, only a set of commands that could just as easily have been invoked one by one from the command-line on the console or in a terminal window. The advantages of placing the commands in a script go far beyond not having to retype them time and again. The script becomes a program -- a tool -- and it can easily be modified or customized for a particular application.

Example 2-2. cleanup: An improved clean-up script

# Proper header for a Bash script.

# Cleanup, version 2

# Run as root, of course.
# Insert code here to print error message and exit if not root.

# Variables are better than hard-coded values.

cat /dev/null > messages
cat /dev/null > wtmp

echo "Logs cleaned up."

exit #  The right and proper method of "exiting" from a script.
     #  A bare "exit" (no parameter) returns the exit status
     #+ of the preceding command. 

Now that's beginning to look like a real script. But we can go even farther . . .

Example 2-3. cleanup: An enhanced and generalized version of above scripts.

# Cleanup, version 3

#  Warning:
#  -------
#  This script uses quite a number of features that will be explained
#+ later on.
#  By the time you've finished the first half of the book,
#+ there should be nothing mysterious about it.

ROOT_UID=0     # Only users with $UID 0 have root privileges.
LINES=50       # Default number of lines saved.
E_XCD=86       # Can't change directory?
E_NOTROOT=87   # Non-root exit error.

# Run as root, of course.
if [ "$UID" -ne "$ROOT_UID" ]
  echo "Must be root to run this script."
  exit $E_NOTROOT

if [ -n "$1" ]
# Test whether command-line argument is present (non-empty).
  lines=$LINES # Default, if not specified on command-line.

#  Stephane Chazelas suggests the following,
#+ as a better way of checking command-line arguments,
#+ but this is still a bit advanced for this stage of the tutorial.
#    E_WRONGARGS=85  # Non-numerical argument (bad argument format).
#    case "$1" in
#    ""      ) lines=50;;
#    *[!0-9]*) echo "Usage: `basename $0` lines-to-cleanup";
#     exit $E_WRONGARGS;;
#    *       ) lines=$1;;
#    esac
#* Skip ahead to "Loops" chapter to decipher all this.


if [ `pwd` != "$LOG_DIR" ]  # or   if [ "$PWD" != "$LOG_DIR" ]
                            # Not in /var/log?
  echo "Can't change to $LOG_DIR."
  exit $E_XCD
fi  # Doublecheck if in right directory before messing with log file.

# Far more efficient is:
# cd /var/log || {
#   echo "Cannot change to necessary directory." >&2
#   exit $E_XCD;
# }

tail -n $lines messages > mesg.temp # Save last section of message log file.
mv mesg.temp messages               # Rename it as system log file.

#  cat /dev/null > messages
#* No longer needed, as the above method is safer.

cat /dev/null > wtmp  #  ': > wtmp' and '> wtmp'  have the same effect.
echo "Log files cleaned up."
#  Note that there are other log files in /var/log not affected
#+ by this script.

exit 0
#  A zero return value from the script upon exit indicates success
#+ to the shell.

Since you may not wish to wipe out the entire system log, this version of the script keeps the last section of the message log intact. You will constantly discover ways of fine-tuning previously written scripts for increased effectiveness.

* * *

The sha-bang ( #!) [6] at the head of a script tells your system that this file is a set of commands to be fed to the command interpreter indicated. The #! is actually a two-byte [7] magic number, a special marker that designates a file type, or in this case an executable shell script (type man magic for more details on this fascinating topic). Immediately following the sha-bang is a path name. This is the path to the program that interprets the commands in the script, whether it be a shell, a programming language, or a utility. This command interpreter then executes the commands in the script, starting at the top (the line following the sha-bang line), and ignoring comments. [8]

#!/bin/sed -f
#!/bin/awk -f

Each of the above script header lines calls a different command interpreter, be it /bin/sh, the default shell (bash in a Linux system) or otherwise. [9] Using #!/bin/sh, the default Bourne shell in most commercial variants of UNIX, makes the script portable to non-Linux machines, though you sacrifice Bash-specific features. The script will, however, conform to the POSIX [10] sh standard.

Note that the path given at the "sha-bang" must be correct, otherwise an error message -- usually "Command not found." -- will be the only result of running the script. [11]

#! can be omitted if the script consists only of a set of generic system commands, using no internal shell directives. The second example, above, requires the initial #!, since the variable assignment line, lines=50, uses a shell-specific construct. [12] Note again that #!/bin/sh invokes the default shell interpreter, which defaults to /bin/bash on a Linux machine.


This tutorial encourages a modular approach to constructing a script. Make note of and collect "boilerplate" code snippets that might be useful in future scripts. Eventually you will build quite an extensive library of nifty routines. As an example, the following script prolog tests whether the script has been invoked with the correct number of parameters.

script_parameters="-a -h -m -z"
#                  -a = all, -h = help, etc.

if [ $# -ne $Number_of_expected_args ]
  echo "Usage: `basename $0` $script_parameters"
  # `basename $0` is the script's filename.
  exit $E_WRONG_ARGS

Many times, you will write a script that carries out one particular task. The first script in this chapter is an example. Later, it might occur to you to generalize the script to do other, similar tasks. Replacing the literal ("hard-wired") constants by variables is a step in that direction, as is replacing repetitive code blocks by functions.

2.1. Invoking the script

Having written the script, you can invoke it by sh scriptname, [13] or alternatively bash scriptname. (Not recommended is using sh <scriptname, since this effectively disables reading from stdin within the script.) Much more convenient is to make the script itself directly executable with a chmod.


chmod 555 scriptname (gives everyone read/execute permission) [14]


chmod +rx scriptname (gives everyone read/execute permission)

chmod u+rx scriptname (gives only the script owner read/execute permission)

Having made the script executable, you may now test it by ./scriptname. [15] If it begins with a "sha-bang" line, invoking the script calls the correct command interpreter to run it.

As a final step, after testing and debugging, you would likely want to move it to /usr/local/bin (as root, of course), to make the script available to yourself and all other users as a systemwide executable. The script could then be invoked by simply typing scriptname [ENTER] from the command-line.

2.2. Preliminary Exercises

  1. System administrators often write scripts to automate common tasks. Give several instances where such scripts would be useful.

  2. Write a script that upon invocation shows the time and date, lists all logged-in users, and gives the system uptime. The script then saves this information to a logfile.

Chapter 3. Special Characters

What makes a character special? If it has a meaning beyond its literal meaning, a meta-meaning, then we refer to it as a special character. Along with commands and keywords, special characters are building blocks of Bash scripts.

Special Characters Found In Scripts and Elsewhere


Comments. Lines beginning with a # (with the exception of #!) are comments and will not be executed.

# This line is a comment.

Comments may also occur following the end of a command.

echo "A comment will follow." # Comment here.
#                            ^ Note whitespace before #

Comments may also follow whitespace at the beginning of a line.

     # A tab precedes this comment.

Comments may even be embedded within a pipe.

initial=( `cat "$startfile" | sed -e '/#/d' | tr -d '\n' |\
# Delete lines containing '#' comment character.
           sed -e 's/\./\. /g' -e 's/_/_ /g'` )
# Excerpted from script


A command may not follow a comment on the same line. There is no method of terminating the comment, in order for "live code" to begin on the same line. Use a new line for the next command.


Of course, a quoted or an escaped # in an echo statement does not begin a comment. Likewise, a # appears in certain parameter-substitution constructs and in numerical constant expressions.
echo "The # here does not begin a comment."
echo 'The # here does not begin a comment.'
echo The \# here does not begin a comment.
echo The # here begins a comment.

echo ${PATH#*:}       # Parameter substitution, not a comment.
echo $(( 2#101011 ))  # Base conversion, not a comment.

# Thanks, S.C.
The standard quoting and escape characters (" ' \) escape the #.

Certain pattern matching operations also use the #.


Command separator [semicolon]. Permits putting two or more commands on the same line.

echo hello; echo there

if [ -x "$filename" ]; then    #  Note the space after the semicolon.
#+                   ^^
  echo "File $filename exists."; cp $filename $filename.bak
else   #                       ^^
  echo "File $filename not found."; touch $filename
fi; echo "File test complete."

Note that the ";" sometimes needs to be escaped.


Terminator in a case option [double semicolon].

case "$variable" in
  abc)  echo "\$variable = abc" ;;
  xyz)  echo "\$variable = xyz" ;;

;;&, ;&

Terminators in a case option (version 4+ of Bash).


"dot" command [period]. Equivalent to source (see Example 15-22). This is a bash builtin.


"dot", as a component of a filename. When working with filenames, a leading dot is the prefix of a "hidden" file, a file that an ls will not normally show.
bash$ touch .hidden-file
bash$ ls -l	      
total 10
 -rw-r--r--    1 bozo      4034 Jul 18 22:04 data1.addressbook
 -rw-r--r--    1 bozo      4602 May 25 13:58 data1.addressbook.bak
 -rw-r--r--    1 bozo       877 Dec 17  2000 employment.addressbook

bash$ ls -al	      
total 14
 drwxrwxr-x    2 bozo  bozo      1024 Aug 29 20:54 ./
 drwx------   52 bozo  bozo      3072 Aug 29 20:51 ../
 -rw-r--r--    1 bozo  bozo      4034 Jul 18 22:04 data1.addressbook
 -rw-r--r--    1 bozo  bozo      4602 May 25 13:58 data1.addressbook.bak
 -rw-r--r--    1 bozo  bozo       877 Dec 17  2000 employment.addressbook
 -rw-rw-r--    1 bozo  bozo         0 Aug 29 20:54 .hidden-file

When considering directory names, a single dot represents the current working directory, and two dots denote the parent directory.

bash$ pwd

bash$ cd .
bash$ pwd

bash$ cd ..
bash$ pwd

The dot often appears as the destination (directory) of a file movement command, in this context meaning current directory.

bash$ cp /home/bozo/current_work/junk/* .
Copy all the "junk" files to $PWD.


"dot" character match. When matching characters, as part of a regular expression, a "dot" matches a single character.


partial quoting [double quote]. "STRING" preserves (from interpretation) most of the special characters within STRING. See Chapter 5.


full quoting [single quote]. 'STRING' preserves all special characters within STRING. This is a stronger form of quoting than "STRING". See Chapter 5.


comma operator. The comma operator [16] links together a series of arithmetic operations. All are evaluated, but only the last one is returned.
let "t2 = ((a = 9, 15 / 3))"
# Set "a = 9" and "t2 = 15 / 3"

The comma operator can also concatenate strings.
for file in /{,usr/}bin/*calc
#             ^    Find all executable files ending in "calc"
#+                 in /bin and /usr/bin directories.
        if [ -x "$file" ]
          echo $file

# /bin/ipcalc
# /usr/bin/kcalc
# /usr/bin/oidcalc
# /usr/bin/oocalc

# Thank you, Rory Winston, for pointing this out.

,, ,

Lowercase conversion in parameter substitution (added in version 4 of Bash).


escape [backslash]. A quoting mechanism for single characters.

\X escapes the character X. This has the effect of "quoting" X, equivalent to 'X'. The \ may be used to quote " and ', so they are expressed literally.

See Chapter 5 for an in-depth explanation of escaped characters.


Filename path separator [forward slash]. Separates the components of a filename (as in /home/bozo/projects/Makefile).

This is also the division arithmetic operator.


command substitution. The `command` construct makes available the output of command for assignment to a variable. This is also known as backquotes or backticks.


null command [colon]. This is the shell equivalent of a "NOP" (no op, a do-nothing operation). It may be considered a synonym for the shell builtin true. The ":" command is itself a Bash builtin, and its exit status is true (0).

echo $?   # 0

Endless loop:

while :

# Same as:
#    while true
#    do
#      ...
#    done

Placeholder in if/then test:

if condition
then :   # Do nothing and branch ahead
else     # Or else ...

Provide a placeholder where a binary operation is expected, see Example 8-2 and default parameters.

: ${username=`whoami`}
# ${username=`whoami`}   Gives an error without the leading :
#                        unless "username" is a command or builtin...

: ${1?"Usage: $0 ARGUMENT"}     # From " example script.

Provide a placeholder where a command is expected in a here document. See Example 19-10.

Evaluate string of variables using parameter substitution (as in Example 10-7).
: ${HOSTNAME?} ${USER?} ${MAIL?}
#  Prints error message
#+ if one or more of essential environmental variables not set.

Variable expansion / substring replacement.

In combination with the > redirection operator, truncates a file to zero length, without changing its permissions. If the file did not previously exist, creates it.
: >   # File "" now empty.	      

# Same effect as   cat /dev/null >
# However, this does not fork a new process, since ":" is a builtin.
See also Example 16-15.

In combination with the >> redirection operator, has no effect on a pre-existing target file (: >> target_file). If the file did not previously exist, creates it.


This applies to regular files, not pipes, symlinks, and certain special files.

May be used to begin a comment line, although this is not recommended. Using # for a comment turns off error checking for the remainder of that line, so almost anything may appear in a comment. However, this is not the case with :.
: This is a comment that generates an error, ( if [ $x -eq 3] ).

The ":" serves as a field separator, in /etc/passwd, and in the $PATH variable.
bash$ echo $PATH

A colon is acceptable as a function name.
  echo "The name of this function is "$FUNCNAME" "
  # Why use a colon as a function name?
  # It's a way of obfuscating your code.


# The name of this function is :
This is not portable behavior, and therefore not a recommended practice. In fact, more recent releases of Bash do not permit this usage. An underscore _ works, though.

A colon can serve as a placeholder in an otherwise empty function.

not_empty ()
} # Contains a : (null command), and so is not empty.


reverse (or negate) the sense of a test or exit status [bang]. The ! operator inverts the exit status of the command to which it is applied (see Example 6-2). It also inverts the meaning of a test operator. This can, for example, change the sense of equal ( = ) to not-equal ( != ). The ! operator is a Bash keyword.

In a different context, the ! also appears in indirect variable references.

In yet another context, from the command line, the ! invokes the Bash history mechanism (see Appendix L). Note that within a script, the history mechanism is disabled.


wild card [asterisk]. The * character serves as a "wild card" for filename expansion in globbing. By itself, it matches every filename in a given directory.

bash$ echo *

The * also represents any number (or zero) characters in a regular expression.


arithmetic operator. In the context of arithmetic operations, the * denotes multiplication.

** A double asterisk can represent the exponentiation operator or extended file-match globbing.


test operator. Within certain expressions, the ? indicates a test for a condition.

In a double-parentheses construct, the ? can serve as an element of a C-style trinary operator. [17]


(( var0 = var1<98?9:21 ))
#                ^ ^

# if [ "$var1" -lt 98 ]
# then
#   var0=9
# else
#   var0=21
# fi

In a parameter substitution expression, the ? tests whether a variable has been set.


wild card. The ? character serves as a single-character "wild card" for filename expansion in globbing, as well as representing one character in an extended regular expression.


Variable substitution (contents of a variable).

echo $var1     # 5
echo $var2     # 23skidoo

A $ prefixing a variable name indicates the value the variable holds.


end-of-line. In a regular expression, a "$" addresses the end of a line of text.

$' ... '

Quoted string expansion. This construct expands single or multiple escaped octal or hex values into ASCII [18] or Unicode characters.

$*, $@

exit status variable. The $? variable holds the exit status of a command, a function, or of the script itself.


process ID variable. The $$ variable holds the process ID [19] of the script in which it appears.


command group.
(a=hello; echo $a)


A listing of commands within parentheses starts a subshell.

Variables inside parentheses, within the subshell, are not visible to the rest of the script. The parent process, the script, cannot read variables created in the child process, the subshell.
( a=321; )	      

echo "a = $a"   # a = 123
# "a" within parentheses acts like a local variable.

array initialization.
Array=(element1 element2 element3)


Brace expansion.
echo \"{These,words,are,quoted}\"   # " prefix and suffix
# "These" "words" "are" "quoted"

cat {file1,file2,file3} > combined_file
# Concatenates the files file1, file2, and file3 into combined_file.

cp file22.{txt,backup}
# Copies "file22.txt" to "file22.backup"

A command may act upon a comma-separated list of file specs within braces. [20] Filename expansion (globbing) applies to the file specs between the braces.


No spaces allowed within the braces unless the spaces are quoted or escaped.

echo {file1,file2}\ :{\ A," B",' C'}

file1 : A file1 : B file1 : C file2 : A file2 : B file2 : C


Extended Brace expansion.
echo {a..z} # a b c d e f g h i j k l m n o p q r s t u v w x y z
# Echoes characters between a and z.

echo {0..3} # 0 1 2 3
# Echoes characters between 0 and 3.

base64_charset=( {A..Z} {a..z} {0..9} + / = )
# Initializing an array, using extended brace expansion.
# From vladz's "" example script.

The {a..z} extended brace expansion construction is a feature introduced in version 3 of Bash.


Block of code [curly brackets]. Also referred to as an inline group, this construct, in effect, creates an anonymous function (a function without a name). However, unlike in a "standard" function, the variables inside a code block remain visible to the remainder of the script.

bash$ { local a;
	      a=123; }
bash: local: can only be used in a

{ a=321; }
echo "a = $a"   # a = 321   (value inside code block)

# Thanks, S.C.

The code block enclosed in braces may have I/O redirected to and from it.

Example 3-1. Code blocks and I/O redirection

# Reading lines in /etc/fstab.


read line1
read line2
} < $File

echo "First line in $File is:"
echo "$line1"
echo "Second line in $File is:"
echo "$line2"

exit 0

# Now, how do you parse the separate fields of each line?
# Hint: use awk, or . . .
# . . . Hans-Joerg Diers suggests using the "set" Bash builtin.

Example 3-2. Saving the output of a code block to a file


#  Queries an rpm file for description, listing,
#+ and whether it can be installed.
#  Saves output to a file.
#  This script illustrates using a code block.


if [ -z "$1" ]
  echo "Usage: `basename $0` rpm-file"
  exit $E_NOARGS

{ # Begin code block.
  echo "Archive Description:"
  rpm -qpi $1       # Query description.
  echo "Archive Listing:"
  rpm -qpl $1       # Query listing.
  rpm -i --test $1  # Query whether rpm file can be installed.
  if [ "$?" -eq $SUCCESS ]
    echo "$1 can be installed."
    echo "$1 cannot be installed."
  echo              # End code block.
} > "$1.test"       # Redirects output of everything in block to file.

echo "Results of rpm test in file $1.test"

# See rpm man page for explanation of options.

exit 0


Unlike a command group within (parentheses), as above, a code block enclosed by {braces} will not normally launch a subshell. [21]

It is possible to iterate a code block using a non-standard for-loop.


placeholder for text. Used after xargs -i (replace strings option). The {} double curly brackets are a placeholder for output text.

ls . | xargs -i -t cp ./{} $1
#            ^^         ^^

# From "" ( example.

{} \;

pathname. Mostly used in find constructs. This is not a shell builtin.


The ";" ends the -exec option of a find command sequence. It needs to be escaped to protect it from interpretation by the shell.

[ ]


Test expression between [ ]. Note that [ is part of the shell builtin test (and a synonym for it), not a link to the external command /usr/bin/test.

[[ ]]


Test expression between [[ ]]. More flexible than the single-bracket [ ] test, this is a shell keyword.

See the discussion on the [[ ... ]] construct.

[ ]

array element.

In the context of an array, brackets set off the numbering of each element of that array.
echo ${Array[1]}

[ ]

range of characters.

As part of a regular expression, brackets delineate a range of characters to match.

$[ ... ]

integer expansion.

Evaluate integer expression between $[ ].

echo $[$a+$b]   # 10
echo $[$a*$b]   # 21

Note that this usage is deprecated, and has been replaced by the (( ... )) construct.

(( ))

integer expansion.

Expand and evaluate integer expression between (( )).

See the discussion on the (( ... )) construct.

> &> >& >> < <>

scriptname >filename redirects the output of scriptname to file filename. Overwrite filename if it already exists.

command &>filename redirects both the stdout and the stderr of command to filename.


This is useful for suppressing output when testing for a condition. For example, let us test whether a certain command exists.

bash$ type bogus_command &>/dev/null

bash$ echo $?

Or in a script:

command_test () { type "$1" &>/dev/null; }
#                                      ^

cmd=rmdir            # Legitimate command.
command_test $cmd; echo $?   # 0

cmd=bogus_command    # Illegitimate command
command_test $cmd; echo $?   # 1

command >&2 redirects stdout of command to stderr.

scriptname >>filename appends the output of scriptname to file filename. If filename does not already exist, it is created.

[i]<>filename opens file filename for reading and writing, and assigns file descriptor i to it. If filename does not exist, it is created.



In a different context, the "<" and ">" characters act as string comparison operators.

In yet another context, the "<" and ">" characters act as integer comparison operators. See also Example 16-9.


redirection used in a here document.


redirection used in a here string.

<, >

ASCII comparison.

if [[ "$veg1" < "$veg2" ]]
  echo "Although $veg1 precede $veg2 in the dictionary,"
  echo -n "this does not necessarily imply anything "
  echo "about my culinary preferences."
  echo "What kind of dictionary are you using, anyhow?"

\<, \>

bash$ grep '\<the\>' textfile


pipe. Passes the output (stdout) of a previous command to the input (stdin) of the next one, or to the shell. This is a method of chaining commands together.

echo ls -l | sh
#  Passes the output of "echo ls -l" to the shell,
#+ with the same result as a simple "ls -l".

cat *.lst | sort | uniq
# Merges and sorts all ".lst" files, then deletes duplicate lines.

The output of a command or commands may be piped to a script.
# : Changes input to uppercase.

tr 'a-z' 'A-Z'
#  Letter ranges must be quoted
#+ to prevent filename generation from single-letter filenames.

exit 0
Now, let us pipe the output of ls -l to this script.
bash$ ls -l | ./
-RW-RW-R--    1 BOZO  BOZO       109 APR  7 19:49 1.TXT
 -RW-RW-R--    1 BOZO  BOZO       109 APR 14 16:48 2.TXT
 -RW-R--R--    1 BOZO  BOZO       725 APR 20 20:56 DATA-FILE


The stdout of each process in a pipe must be read as the stdin of the next. If this is not the case, the data stream will block, and the pipe will not behave as expected.
cat file1 file2 | ls -l | sort
# The output from "cat file1 file2" disappears.

A pipe runs as a child process, and therefore cannot alter script variables.
echo "new_value" | read variable
echo "variable = $variable"     # variable = initial_value

If one of the commands in the pipe aborts, this prematurely terminates execution of the pipe. Called a broken pipe, this condition sends a SIGPIPE signal.


force redirection (even if the noclobber option is set). This will forcibly overwrite an existing file.


OR logical operator. In a test construct, the || operator causes a return of 0 (success) if either of the linked test conditions is true.


Run job in background. A command followed by an & will run in the background.

bash$ sleep 10 &
[1] 850
[1]+  Done                    sleep 10

Within a script, commands and even loops may run in the background.

Example 3-3. Running a loop in the background


for i in 1 2 3 4 5 6 7 8 9 10            # First loop.
  echo -n "$i "
done & # Run this loop in background.
       # Will sometimes execute after second loop.

echo   # This 'echo' sometimes will not display.

for i in 11 12 13 14 15 16 17 18 19 20   # Second loop.
  echo -n "$i "

echo   # This 'echo' sometimes will not display.

# ======================================================

# The expected output from the script:
# 1 2 3 4 5 6 7 8 9 10 
# 11 12 13 14 15 16 17 18 19 20 

# Sometimes, though, you get:
# 11 12 13 14 15 16 17 18 19 20 
# 1 2 3 4 5 6 7 8 9 10 bozo $
# (The second 'echo' doesn't execute. Why?)

# Occasionally also:
# 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
# (The first 'echo' doesn't execute. Why?)

# Very rarely something like:
# 11 12 13 1 2 3 4 5 6 7 8 9 10 14 15 16 17 18 19 20 
# The foreground loop preempts the background one.

exit 0

#  Nasimuddin Ansari suggests adding    sleep 1
#+ after the   echo -n "$i"   in lines 6 and 14,
#+ for some real fun.


A command run in the background within a script may cause the script to hang, waiting for a keystroke. Fortunately, there is a remedy for this.


AND logical operator. In a test construct, the && operator causes a return of 0 (success) only if both the linked test conditions are true.


option, prefix. Option flag for a command or filter. Prefix for an operator. Prefix for a default parameter in parameter substitution.

COMMAND -[Option1][Option2][...]

ls -al

sort -dfu $filename

if [ $file1 -ot $file2 ]
then #      ^
  echo "File $file1 is older than $file2."

if [ "$a" -eq "$b" ]
then #    ^
  echo "$a is equal to $b."

if [ "$c" -eq 24 -a "$d" -eq 47 ]
then #    ^              ^
  echo "$c equals 24 and $d equals 47."

#               ^


The double-dash -- prefixes long (verbatim) options to commands.

sort --ignore-leading-blanks

Used with a Bash builtin, it means the end of options to that particular command.


This provides a handy means of removing files whose names begin with a dash.
bash$ ls -l
-rw-r--r-- 1 bozo bozo 0 Nov 25 12:29 -badname

bash$ rm -- -badname

bash$ ls -l
total 0

The double-dash is also used in conjunction with set.

set -- $variable (as in Example 15-18)


redirection from/to stdin or stdout [dash].

bash$ cat -



As expected, cat - echoes stdin, in this case keyboarded user input, to stdout. But, does I/O redirection using - have real-world applications?

(cd /source/directory && tar cf - . ) | (cd /dest/directory && tar xpvf -)
# Move entire file tree from one directory to another
# [courtesy Alan Cox <[email protected]>, with a minor change]

# 1) cd /source/directory
#    Source directory, where the files to be moved are.
# 2) &&
#   "And-list": if the 'cd' operation successful,
#    then execute the next command.
# 3) tar cf - .
#    The 'c' option 'tar' archiving command creates a new archive,
#    the 'f' (file) option, followed by '-' designates the target file
#    as stdout, and do it in current directory tree ('.').
# 4) |
#    Piped to ...
# 5) ( ... )
#    a subshell
# 6) cd /dest/directory
#    Change to the destination directory.
# 7) &&
#   "And-list", as above
# 8) tar xpvf -
#    Unarchive ('x'), preserve ownership and file permissions ('p'),
#    and send verbose messages to stdout ('v'),
#    reading data from stdin ('f' followed by '-').
#    Note that 'x' is a command, and 'p', 'v', 'f' are options.
# Whew!

# More elegant than, but equivalent to:
#   cd source/directory
#   tar cf - . | (cd ../dest/directory; tar xpvf -)
#     Also having same effect:
# cp -a /source/directory/* /dest/directory
#     Or:
# cp -a /source/directory/* /source/directory/.[^.]* /dest/directory
#     If there are hidden files in /source/directory.

bunzip2 -c linux-2.6.16.tar.bz2 | tar xvf -
#  --uncompress tar file--      | --then pass it to "tar"--
#  If "tar" has not been patched to handle "bunzip2",
#+ this needs to be done in two discrete steps, using a pipe.
#  The purpose of the exercise is to unarchive "bzipped" kernel source.

Note that in this context the "-" is not itself a Bash operator, but rather an option recognized by certain UNIX utilities that write to stdout, such as tar, cat, etc.

bash$ echo "whatever" | cat -

Where a filename is expected, - redirects output to stdout (sometimes seen with tar cf), or accepts input from stdin, rather than from a file. This is a method of using a file-oriented utility as a filter in a pipe.

bash$ file
Usage: file [-bciknvzL] [-f namefile] [-m magicfiles] file...
By itself on the command-line, file fails with an error message.

Add a "-" for a more useful result. This causes the shell to await user input.
bash$ file -
standard input:              ASCII text

bash$ file -
standard input:              Bourne-Again shell script text executable
Now the command accepts input from stdin and analyzes it.

The "-" can be used to pipe stdout to other commands. This permits such stunts as prepending lines to a file.

Using diff to compare a file with a section of another:

grep Linux file1 | diff file2 -

Finally, a real-world example using - with tar.

Example 3-4. Backup of all files changed in last day


#  Backs up all files in current directory modified within last 24 hours
#+ in a "tarball" (tarred and gzipped file).

BACKUPFILE=backup-$(date +%m-%d-%Y)
#                 Embeds date in backup filename.
#                 Thanks, Joshua Tschida, for the idea.
#  If no backup-archive filename specified on command-line,
#+ it will default to "backup-MM-DD-YYYY.tar.gz."

tar cvf - `find . -mtime -1 -type f -print` > $archive.tar
gzip $archive.tar
echo "Directory $PWD backed up in archive file \"$archive.tar.gz\"."

#  Stephane Chazelas points out that the above code will fail
#+ if there are too many files found
#+ or if any filenames contain blank characters.

# He suggests the following alternatives:
# -------------------------------------------------------------------
#   find . -mtime -1 -type f -print0 | xargs -0 tar rvf "$archive.tar"
#      using the GNU version of "find".

#   find . -mtime -1 -type f -exec tar rvf "$archive.tar" '{}' \;
#         portable to other UNIX flavors, but much slower.
# -------------------------------------------------------------------

exit 0


Filenames beginning with "-" may cause problems when coupled with the "-" redirection operator. A script should check for this and add an appropriate prefix to such filenames, for example ./-FILENAME, $PWD/-FILENAME, or $PATHNAME/-FILENAME.

If the value of a variable begins with a -, this may likewise create problems.
echo $var		
# Has the effect of "echo -n", and outputs nothing.


previous working directory. A cd - command changes to the previous working directory. This uses the $OLDPWD environmental variable.


Do not confuse the "-" used in this sense with the "-" redirection operator just discussed. The interpretation of the "-" depends on the context in which it appears.


Minus. Minus sign in an arithmetic operation.


Equals. Assignment operator
echo $a   # 28

In a different context, the "=" is a string comparison operator.


Plus. Addition arithmetic operator.

In a different context, the + is a Regular Expression operator.


Option. Option flag for a command or filter.

Certain commands and builtins use the + to enable certain options and the - to disable them. In parameter substitution, the + prefixes an alternate value that a variable expands to.


modulo. Modulo (remainder of a division) arithmetic operation.

let "z = 5 % 3"
echo $z  # 2

In a different context, the % is a pattern matching operator.


home directory [tilde]. This corresponds to the $HOME internal variable. ~bozo is bozo's home directory, and ls ~bozo lists the contents of it. ~/ is the current user's home directory, and ls ~/ lists the contents of it.
bash$ echo ~bozo

bash$ echo ~

bash$ echo ~/

bash$ echo ~:

bash$ echo ~nonexistent-user


current working directory. This corresponds to the $PWD internal variable.


previous working directory. This corresponds to the $OLDPWD internal variable.


regular expression match. This operator was introduced with version 3 of Bash.


beginning-of-line. In a regular expression, a "^" addresses the beginning of a line of text.

^, ^^

Uppercase conversion in parameter substitution (added in version 4 of Bash).

Control Characters

change the behavior of the terminal or text display. A control character is a CONTROL + key combination (pressed simultaneously). A control character may also be written in octal or hexadecimal notation, following an escape.

Control characters are not normally useful inside a script.

  • Ctl-A

    Moves cursor to beginning of line of text (on the command-line).

  • Ctl-B

    Backspace (nondestructive).

  • Ctl-C

    Break. Terminate a foreground job.

  • Ctl-D

    Log out from a shell (similar to exit).

    EOF (end-of-file). This also terminates input from stdin.

    When typing text on the console or in an xterm window, Ctl-D erases the character under the cursor. When there are no characters present, Ctl-D logs out of the session, as expected. In an xterm window, this has the effect of closing the window.

  • Ctl-E

    Moves cursor to end of line of text (on the command-line).

  • Ctl-F

    Moves cursor forward one character position (on the command-line).

  • Ctl-G

    BEL. On some old-time teletype terminals, this would actually ring a bell. In an xterm it might beep.

  • Ctl-H

    Rubout (destructive backspace). Erases characters the cursor backs over while backspacing.

    # Embedding Ctl-H in a string.
    a="^H^H"                  # Two Ctl-H's -- backspaces
                              # ctl-V ctl-H, using vi/vim
    echo "abcdef"             # abcdef
    echo -n "abcdef$a "       # abcd f
    #  Space at end  ^              ^  Backspaces twice.
    echo -n "abcdef$a"        # abcdef
    #  No space at end               ^ Doesn't backspace (why?).
                              # Results may not be quite as expected.
    echo; echo
    # Constantin Hagemeier suggests trying:
    # a=$'\010\010'
    # a=$'\b\b'
    # a=$'\x08\x08'
    # But, this does not change the results.
    # Now, try this.
    rubout="^H^H^H^H^H"       # 5 x Ctl-H.
    echo -n "12345678"
    sleep 2
    echo -n "$rubout"
    sleep 2

  • Ctl-I

    Horizontal tab.

  • Ctl-J

    Newline (line feed). In a script, may also be expressed in octal notation -- '\012' or in hexadecimal -- '\x0a'.

  • Ctl-K

    Vertical tab.

    When typing text on the console or in an xterm window, Ctl-K erases from the character under the cursor to end of line. Within a script, Ctl-K may behave differently, as in Lee Lee Maschmeyer's example, below.

  • Ctl-L

    Formfeed (clear the terminal screen). In a terminal, this has the same effect as the clear command. When sent to a printer, a Ctl-L causes an advance to end of the paper sheet.

  • Ctl-M

    Carriage return.

    # Thank you, Lee Maschmeyer, for this example.
    read -n 1 -s -p \
    $'Control-M leaves cursor at beginning of this line. Press Enter. \x0d'
               # Of course, '0d' is the hex equivalent of Control-M.
    echo >&2   #  The '-s' makes anything typed silent,
               #+ so it is necessary to go to new line explicitly.
    read -n 1 -s -p $'Control-J leaves cursor on next line. \x0a'
               #  '0a' is the hex equivalent of Control-J, linefeed.
    echo >&2
    read -n 1 -s -p $'And Control-K\x0bgoes straight down.'
    echo >&2   #  Control-K is vertical tab.
    # A better example of the effect of a vertical tab is:
    var=$'\x0aThis is the bottom line\x0bThis is the top line\x0a'
    echo "$var"
    #  This works the same way as the above example. However:
    echo "$var" | col
    #  This causes the right end of the line to be higher than the left end.
    #  It also explains why we started and ended with a line feed --
    #+ to avoid a garbled screen.
    # As Lee Maschmeyer explains:
    # --------------------------
    #  In the [first vertical tab example] . . . the vertical tab
    #+ makes the printing go straight down without a carriage return.
    #  This is true only on devices, such as the Linux console,
    #+ that can't go "backward."
    #  The real purpose of VT is to go straight UP, not down.
    #  It can be used to print superscripts on a printer.
    #  The col utility can be used to emulate the proper behavior of VT.
    exit 0

  • Ctl-N

    Erases a line of text recalled from history buffer [23] (on the command-line).

  • Ctl-O

    Issues a newline (on the command-line).

  • Ctl-P

    Recalls last command from history buffer (on the command-line).

  • Ctl-Q

    Resume (XON).

    This resumes stdin in a terminal.

  • Ctl-R

    Backwards search for text in history buffer (on the command-line).

  • Ctl-S

    Suspend (XOFF).

    This freezes stdin in a terminal. (Use Ctl-Q to restore input.)

  • Ctl-T

    Reverses the position of the character the cursor is on with the previous character (on the command-line).

  • Ctl-U

    Erase a line of input, from the cursor backward to beginning of line. In some settings, Ctl-U erases the entire line of input, regardless of cursor position.

  • Ctl-V

    When inputting text, Ctl-V permits inserting control characters. For example, the following two are equivalent:
    echo -e '\x0a'
    echo <Ctl-V><Ctl-J>

    Ctl-V is primarily useful from within a text editor.

  • Ctl-W

    When typing text on the console or in an xterm window, Ctl-W erases from the character under the cursor backwards to the first instance of whitespace. In some settings, Ctl-W erases backwards to first non-alphanumeric character.

  • Ctl-X

    In certain word processing programs, Cuts highlighted text and copies to clipboard.

  • Ctl-Y

    Pastes back text previously erased (with Ctl-U or Ctl-W).

  • Ctl-Z

    Pauses a foreground job.

    Substitute operation in certain word processing applications.

    EOF (end-of-file) character in the MSDOS filesystem.


functions as a separator between commands and/or variables. Whitespace consists of either spaces, tabs, blank lines, or any combination thereof. [24] In some contexts, such as variable assignment, whitespace is not permitted, and results in a syntax error.

Blank lines have no effect on the action of a script, and are therefore useful for visually separating functional sections.

$IFS, the special variable separating fields of input to certain commands. It defaults to whitespace.

To preserve whitespace within a string or in a variable, use quoting.

UNIX filters can target and operate on whitespace using the POSIX character class [:space:].

Chapter 4. Introduction to Variables and Parameters

Variables are how programming and scripting languages represent data. A variable is nothing more than a label, a name assigned to a location or set of locations in computer memory holding an item of data.

Variables appear in arithmetic operations and manipulation of quantities, and in string parsing.

4.1. Variable Substitution

The name of a variable is a placeholder for its value, the data it holds. Referencing (retrieving) its value is called variable substitution.


Let us carefully distinguish between the name of a variable and its value. If variable1 is the name of a variable, then $variable1 is a reference to its value, the data item it contains. [25]

bash$ variable1=23

bash$ echo variable1

bash$ echo $variable1

The only times a variable appears "naked" -- without the $ prefix -- is when declared or assigned, when unset, when exported, in an arithmetic expression within double parentheses (( ... )), or in the special case of a variable representing a signal (see Example 32-5). Assignment may be with an = (as in var1=27), in a read statement, and at the head of a loop (for var2 in 1 2 3).

Enclosing a referenced value in double quotes (" ... ") does not interfere with variable substitution. This is called partial quoting, sometimes referred to as "weak quoting." Using single quotes (' ... ') causes the variable name to be used literally, and no substitution will take place. This is full quoting, sometimes referred to as 'strong quoting.' See Chapter 5 for a detailed discussion.

Note that $variable is actually a simplified form of ${variable}. In contexts where the $variable syntax causes an error, the longer form may work (see Section 10.2, below).

Example 4-1. Variable assignment and substitution


# Variables: assignment and substitution

#   ^ ^

# No space permitted on either side of = sign when initializing variables.
# What happens if there is a space?

#  "VARIABLE =value"
#           ^
#% Script tries to run "VARIABLE" command with one argument, "=value".

#  "VARIABLE= value"
#            ^
#% Script tries to run "value" command with
#+ the environmental variable "VARIABLE" set to "".

echo hello    # hello
# Not a variable reference, just the string "hello" ...

echo $hello   # 375
#    ^          This *is* a variable reference.
echo ${hello} # 375
#               Likewise a variable reference, as above.

# Quoting . . .
echo "$hello"    # 375
echo "${hello}"  # 375


hello="A B  C   D"
echo $hello   # A B C D
echo "$hello" # A B  C   D
# As we see, echo $hello   and   echo "$hello"   give different results.
# =======================================
# Quoting a variable preserves whitespace.
# =======================================


echo '$hello'  # $hello
#    ^      ^
#  Variable referencing disabled (escaped) by single quotes,
#+ which causes the "$" to be interpreted literally.

# Notice the effect of different types of quoting.

hello=    # Setting it to a null value.
echo "\$hello (null value) = $hello"      # $hello (null value) =
#  Note that setting a variable to a null value is not the same as
#+ unsetting it, although the end result is the same (see below).

# --------------------------------------------------------------

#  It is permissible to set multiple variables on the same line,
#+ if separated by white space.
#  Caution, this may reduce legibility, and may not be portable.

var1=21  var2=22  var3=$V3
echo "var1=$var1   var2=$var2   var3=$var3"

# May cause problems with legacy versions of "sh" . . .

# --------------------------------------------------------------

echo; echo

numbers="one two three"
#           ^   ^
other_numbers="1 2 3"
#               ^ ^
#  If there is whitespace embedded within a variable,
#+ then quotes are necessary.
#  other_numbers=1 2 3                  # Gives an error message.
echo "numbers = $numbers"
echo "other_numbers = $other_numbers"   # other_numbers = 1 2 3
#  Escaping the whitespace also works.
mixed_bag=2\ ---\ Whatever
#           ^    ^ Space after escape (\).

echo "$mixed_bag"         # 2 --- Whatever

echo; echo

echo "uninitialized_variable = $uninitialized_variable"
# Uninitialized variable has null value (no value at all!).
uninitialized_variable=   #  Declaring, but not initializing it --
                          #+ same as setting it to a null value, as above.
echo "uninitialized_variable = $uninitialized_variable"
                          # It still has a null value.

uninitialized_variable=23       # Set it.
unset uninitialized_variable    # Unset it.
echo "uninitialized_variable = $uninitialized_variable"
                                # uninitialized_variable =
                                # It still has a null value.

exit 0


An uninitialized variable has a "null" value -- no assigned value at all (not zero!).
if [ -z "$unassigned" ]
  echo "\$unassigned is NULL."
fi     # $unassigned is NULL.

Using a variable before assigning a value to it may cause problems. It is nevertheless possible to perform arithmetic operations on an uninitialized variable.
echo "$uninitialized"                                # (blank line)
let "uninitialized += 5"                             # Add 5 to it.
echo "$uninitialized"                                # 5

#  Conclusion:
#  An uninitialized variable has no value,
#+ however it evaluates as 0 in an arithmetic operation.
See also Example 15-23.

4.2. Variable Assignment


the assignment operator (no space before and after)


Do not confuse this with = and -eq, which test, rather than assign!

Note that = can be either an assignment or a test operator, depending on context.

Example 4-2. Plain Variable Assignment

# Naked variables


# When is a variable "naked", i.e., lacking the '$' in front?
# When it is being assigned, rather than referenced.

# Assignment
echo "The value of \"a\" is $a."

# Assignment using 'let'
let a=16+5
echo "The value of \"a\" is now $a."


# In a 'for' loop (really, a type of disguised assignment):
echo -n "Values of \"a\" in the loop are: "
for a in 7 8 9 11
  echo -n "$a "


# In a 'read' statement (also a type of assignment):
echo -n "Enter \"a\" "
read a
echo "The value of \"a\" is now $a."


exit 0

Example 4-3. Variable Assignment, plain and fancy


a=23              # Simple case
echo $a
echo $b

# Now, getting a little bit fancier (command substitution).

a=`echo Hello!`   # Assigns result of 'echo' command to 'a' ...
echo $a
#  Note that including an exclamation mark (!) within a
#+ command substitution construct will not work from the command-line,
#+ since this triggers the Bash "history mechanism."
#  Inside a script, however, the history functions are disabled by default.

a=`ls -l`         # Assigns result of 'ls -l' command to 'a'
echo $a           # Unquoted, however, it removes tabs and newlines.
echo "$a"         # The quoted variable preserves whitespace.
                  # (See the chapter on "Quoting.")

exit 0

Variable assignment using the $(...) mechanism (a newer method than backquotes). This is likewise a form of command substitution.

# From /etc/rc.d/rc.local
R=$(cat /etc/redhat-release)
arch=$(uname -m)

4.3. Bash Variables Are Untyped

Unlike many other programming languages, Bash does not segregate its variables by "type." Essentially, Bash variables are character strings, but, depending on context, Bash permits arithmetic operations and comparisons on variables. The determining factor is whether the value of a variable contains only digits.

Example 4-4. Integer or string?


a=2334                   # Integer.
let "a += 1"
echo "a = $a "           # a = 2335
echo                     # Integer, still.

b=${a/23/BB}             # Substitute "BB" for "23".
                         # This transforms $b into a string.
echo "b = $b"            # b = BB35
declare -i b             # Declaring it an integer doesn't help.
echo "b = $b"            # b = BB35

let "b += 1"             # BB35 + 1
echo "b = $b"            # b = 1
echo                     # Bash sets the "integer value" of a string to 0.

echo "c = $c"            # c = BB34
d=${c/BB/23}             # Substitute "23" for "BB".
                         # This makes $d an integer.
echo "d = $d"            # d = 2334
let "d += 1"             # 2334 + 1
echo "d = $d"            # d = 2335

# What about null variables?
e=''                     # ... Or e="" ... Or e=
echo "e = $e"            # e =
let "e += 1"             # Arithmetic operations allowed on a null variable?
echo "e = $e"            # e = 1
echo                     # Null variable transformed into an integer.

# What about undeclared variables?
echo "f = $f"            # f =
let "f += 1"             # Arithmetic operations allowed?
echo "f = $f"            # f = 1
echo                     # Undeclared variable transformed into an integer.
# However ...
let "f /= $undecl_var"   # Divide by zero?
#   let: f /= : syntax error: operand expected (error token is " ")
# Syntax error! Variable $undecl_var is not set to zero here!
# But still ...
let "f /= 0"
#   let: f /= 0: division by 0 (error token is "0")
# Expected behavior.

#  Bash (usually) sets the "integer value" of null to zero
#+ when performing an arithmetic operation.
#  But, don't try this at home, folks!
#  It's undocumented and probably non-portable behavior.

# Conclusion: Variables in Bash are untyped,
#+ with all attendant consequences.

exit $?

Untyped variables are both a blessing and a curse. They permit more flexibility in scripting and make it easier to grind out lines of code (and give you enough rope to hang yourself!). However, they likewise permit subtle errors to creep in and encourage sloppy programming habits.

To lighten the burden of keeping track of variable types in a script, Bash does permit declaring variables.

4.4. Special Variable Types

Local variables

Variables visible only within a code block or function (see also local variables in functions)

Environmental variables

Variables that affect the behavior of the shell and user interface


In a more general context, each process has an "environment", that is, a group of variables that the process may reference. In this sense, the shell behaves like any other process.

Every time a shell starts, it creates shell variables that correspond to its own environmental variables. Updating or adding new environmental variables causes the shell to update its environment, and all the shell's child processes (the commands it executes) inherit this environment.


The space allotted to the environment is limited. Creating too many environmental variables or ones that use up excessive space may cause problems.

bash$ eval "`seq 10000 | sed -e 's/.*/export var&=ZZZZZZZZZZZZZZ/'`"

bash$ du
bash: /usr/bin/du: Argument list too long

Note: this "error" has been fixed, as of kernel version 2.6.23.

(Thank you, St�phane Chazelas for the clarification, and for providing the above example.)

If a script sets environmental variables, they need to be "exported," that is, reported to the environment local to the script. This is the function of the export command.


A script can export variables only to child processes, that is, only to commands or processes which that particular script initiates. A script invoked from the command-line cannot export variables back to the command-line environment. Child processes cannot export variables back to the parent processes that spawned them.

Definition: A child process is a subprocess launched by another process, its parent.

Positional parameters

Arguments passed to the script from the command line [26] : $0, $1, $2, $3 . . .

$0 is the name of the script itself, $1 is the first argument, $2 the second, $3 the third, and so forth. [27] After $9, the arguments must be enclosed in brackets, for example, ${10}, ${11}, ${12}.

The special variables $* and $@ denote all the positional parameters.

Example 4-5. Positional Parameters


# Call this script with at least 10 parameters, for example
# ./scriptname 1 2 3 4 5 6 7 8 9 10


echo "The name of this script is \"$0\"."
# Adds ./ for current directory
echo "The name of this script is \"`basename $0`\"."
# Strips out path name info (see 'basename')


if [ -n "$1" ]              # Tested variable is quoted.
 echo "Parameter #1 is $1"  # Need quotes to escape #

if [ -n "$2" ]
 echo "Parameter #2 is $2"

if [ -n "$3" ]
 echo "Parameter #3 is $3"

# ...

if [ -n "${10}" ]  # Parameters > $9 must be enclosed in {brackets}.
 echo "Parameter #10 is ${10}"

echo "-----------------------------------"
echo "All the command-line parameters are: "$*""

if [ $# -lt "$MINPARAMS" ]
  echo "This script needs at least $MINPARAMS command-line arguments!"


exit 0

Bracket notation for positional parameters leads to a fairly simple way of referencing the last argument passed to a script on the command-line. This also requires indirect referencing.

args=$#           # Number of args passed.
# Note: This is an *indirect reference* to $args ...

# Or:       lastarg=${!#}             (Thanks, Chris Monson.)
# This is an *indirect reference* to the $# variable.
# Note that lastarg=${!$#} doesn't work.

Some scripts can perform different operations, depending on which name they are invoked with. For this to work, the script needs to check $0, the name it was invoked by. [28] There must also exist symbolic links to all the alternate names of the script. See Example 16-2.


If a script expects a command-line parameter but is invoked without one, this may cause a null variable assignment, generally an undesirable result. One way to prevent this is to append an extra character to both sides of the assignment statement using the expected positional parameter.

variable1_=$1_  # Rather than variable1=$1
# This will prevent an error, even if positional parameter is absent.


# The extra character can be stripped off later, like so.
# Side effects only if $variable1_ begins with an underscore.
# This uses one of the parameter substitution templates discussed later.
# (Leaving out the replacement pattern results in a deletion.)

#  A more straightforward way of dealing with this is
#+ to simply test whether expected positional parameters have been passed.
if [ -z $1 ]

#  However, as Fabian Kreutz points out,
#+ the above method may have unexpected side-effects.
#  A better method is parameter substitution:
#         ${1:-$DefaultVal}
#  See the "Parameter Substition" section
#+ in the "Variables Revisited" chapter.


Example 4-6. wh, whois domain name lookup


# Does a 'whois domain-name' lookup on any of 3 alternate servers:
#          ,,

# Place this script -- renamed 'wh' -- in /usr/local/bin

# Requires symbolic links:
# ln -s /usr/local/bin/wh /usr/local/bin/wh-ripe
# ln -s /usr/local/bin/wh /usr/local/bin/wh-apnic
# ln -s /usr/local/bin/wh /usr/local/bin/wh-tucows


if [ -z "$1" ]
  echo "Usage: `basename $0` [domain-name]"
  exit $E_NOARGS

# Check script name and call proper server.
case `basename $0` in    # Or:    case ${0##*/} in
    "wh"       ) whois [email protected];;
    "wh-ripe"  ) whois [email protected];;
    "wh-apnic" ) whois [email protected];;
    "wh-cw"    ) whois [email protected];;
    *          ) echo "Usage: `basename $0` [domain-name]";;

exit $?


The shift command reassigns the positional parameters, in effect shifting them to the left one notch.

$1 <--- $2, $2 <--- $3, $3 <--- $4, etc.

The old $1 disappears, but $0 (the script name) does not change. If you use a large number of positional parameters to a script, shift lets you access those past 10, although {bracket} notation also permits this.

Example 4-7. Using shift

# Using 'shift' to step through all the positional parameters.

#  Name this script something like,
#+ and invoke it with some parameters.
#+ For example:
#             sh a b c def 83 barndoor

until [ -z "$1" ]  # Until all parameters used up . . .
  echo -n "$1 "

echo               # Extra linefeed.

# But, what happens to the "used-up" parameters?
echo "$2"
#  Nothing echoes!
#  When $2 shifts into $1 (and there is no $3 to shift into $2)
#+ then $2 remains empty.
#  So, it is not a parameter *copy*, but a *move*.


#  See also the script for a "shiftless"
#+ alternative method of stepping through the positional params.

The shift command can take a numerical parameter indicating how many positions to shift.


shift 3    # Shift 3 positions.
#  n=3; shift $n
#  Has the same effect.

echo "$1"

exit 0

# ======================== #

$ sh 1 2 3 4 5

#  However, as Eleni Fragkiadaki, points out,
#+ attempting a 'shift' past the number of
#+ positional parameters ($#) returns an exit status of 1,
#+ and the positional parameters themselves do not change.
#  This means possibly getting stuck in an endless loop. . . .
#  For example:
#      until [ -z "$1" ]
#      do
#         echo -n "$1 "
#         shift 20    #  If less than 20 pos params,
#      done           #+ then loop never ends!
# When in doubt, add a sanity check. . . .
#           shift 20 || break
#                    ^^^^^^^^


The shift command works in a similar fashion on parameters passed to a function. See Example 36-18.

Chapter 5. Quoting

Quoting means just that, bracketing a string in quotes. This has the effect of protecting special characters in the string from reinterpretation or expansion by the shell or shell script. (A character is "special" if it has an interpretation other than its literal meaning. For example, the asterisk * represents a wild card character in globbing and Regular Expressions).

bash$ ls -l [Vv]*
-rw-rw-r--    1 bozo  bozo       324 Apr  2 15:05 VIEWDATA.BAT
 -rw-rw-r--    1 bozo  bozo       507 May  4 14:25
 -rw-rw-r--    1 bozo  bozo       539 Apr 14 17:11

bash$ ls -l '[Vv]*'
ls: [Vv]*: No such file or directory

Certain programs and utilities reinterpret or expand special characters in a quoted string. An important use of quoting is protecting a command-line parameter from the shell, but still letting the calling program expand it.

bash$ grep '[Ff]irst' *.txt
file1.txt:This is the first line of file1.txt.
 file2.txt:This is the First line of file2.txt.

Note that the unquoted grep [Ff]irst *.txt works under the Bash shell. [29]

Quoting can also suppress echo's "appetite" for newlines.

bash$ echo $(ls -l)
total 8 -rw-rw-r-- 1 bo bo 13 Aug 21 12:57 -rw-rw-r-- 1 bo bo 78 Aug 21 12:57

bash$ echo "$(ls -l)"
total 8
 -rw-rw-r--  1 bo bo  13 Aug 21 12:57
 -rw-rw-r--  1 bo bo  78 Aug 21 12:57

5.1. Quoting Variables

When referencing a variable, it is generally advisable to enclose its name in double quotes. This prevents reinterpretation of all special characters within the quoted string -- except $, ` (backquote), and \ (escape). [30] Keeping $ as a special character within double quotes permits referencing a quoted variable ("$variable"), that is, replacing the variable with its value (see Example 4-1, above).

Use double quotes to prevent word splitting. [31] An argument enclosed in double quotes presents itself as a single word, even if it contains whitespace separators.

List="one two three"

for a in $List     # Splits the variable in parts at whitespace.
  echo "$a"
# one
# two
# three

echo "---"

for a in "$List"   # Preserves whitespace in a single variable.
do #     ^     ^
  echo "$a"
# one two three

A more elaborate example:

variable1="a variable containing five words"
COMMAND This is $variable1    # Executes COMMAND with 7 arguments:
# "This" "is" "a" "variable" "containing" "five" "words"

COMMAND "This is $variable1"  # Executes COMMAND with 1 argument:
# "This is a variable containing five words"

variable2=""    # Empty.

COMMAND $variable2 $variable2 $variable2
                # Executes COMMAND with no arguments. 
COMMAND "$variable2" "$variable2" "$variable2"
                # Executes COMMAND with 3 empty arguments. 
COMMAND "$variable2 $variable2 $variable2"
                # Executes COMMAND with 1 argument (2 spaces). 

# Thanks, St�phane Chazelas.


Enclosing the arguments to an echo statement in double quotes is necessary only when word splitting or preservation of whitespace is an issue.

Example 5-1. Echoing Weird Variables

# Echoing weird variables.


echo $var        # '(]\{}$"
echo "$var"      # '(]\{}$"     Doesn't make a difference.


echo $var        # '(] {}$"     \ converted to space. Why?
echo "$var"      # '(]\{}$"

# Examples above supplied by Stephane Chazelas.


echo $var2       #   "
echo "$var2"     # \\"
# But ... var2="\\\\"" is illegal. Why?
echo "$var3"     # \\\\
# Strong quoting works, though.

# ************************************************************ #
# As the first example above shows, nesting quotes is permitted.

echo "$(echo '"')"           # "
#    ^           ^

# At times this comes in useful.

var1="Two bits"
echo "\$var1 = "$var1""      # $var1 = Two bits
#    ^                ^

# Or, as Chris Hiestand points out ...

if [[ "$(du "$My_File1")" -gt "$(du "$My_File2")" ]]
#     ^     ^         ^ ^     ^     ^         ^ ^
# ************************************************************ #

Single quotes (' ') operate similarly to double quotes, but do not permit referencing variables, since the special meaning of $ is turned off. Within single quotes, every special character except ' gets interpreted literally. Consider single quotes ("full quoting") to be a stricter method of quoting than double quotes ("partial quoting").


Since even the escape character (\) gets a literal interpretation within single quotes, trying to enclose a single quote within single quotes will not yield the expected result.
echo "Why can't I write 's between single quotes"


# The roundabout method.
echo 'Why can'\''t I write '"'"'s between single quotes'
#    |-------|  |----------|   |-----------------------|
# Three single-quoted strings, with escaped and quoted single quotes between.

# This example courtesy of St�phane Chazelas.

5.2. Escaping

Escaping is a method of quoting single characters. The escape (\) preceding a character tells the shell to interpret that character literally.


With certain commands and utilities, such as echo and sed, escaping a character may have the opposite effect - it can toggle on a special meaning for that character.

Special meanings of certain escaped characters

used with echo and sed


means newline


means return


means tab


means vertical tab


means backspace


means alert (beep or flash)


translates to the octal ASCII equivalent of 0nn, where nn is a string of digits


The $' ... ' quoted string-expansion construct is a mechanism that uses escaped octal or hex values to assign ASCII characters to variables, e.g., quote=$'\042'.

Example 5-2. Escaped Characters

# escaped characters

### First, let's show some basic escaped-character usage. ###

# Escaping a newline.
# ------------------

echo ""

echo "This will print
as two lines."
# This will print
# as two lines.

echo "This will print \
as one line."
# This will print as one line.

echo; echo

echo "============="

echo "\v\v\v\v"      # Prints \v\v\v\v literally.
# Use the -e option with 'echo' to print escaped characters.
echo "============="
echo -e "\v\v\v\v"   # Prints 4 vertical tabs.
echo "=============="

echo -e "\042"       # Prints " (quote, octal ASCII character 42).
echo "=============="

# The $'\X' construct makes the -e option unnecessary.

echo; echo "NEWLINE and (maybe) BEEP"
echo $'\n'           # Newline.
echo $'\a'           # Alert (beep).
                     # May only flash, not beep, depending on terminal.

# We have seen $'\nnn" string expansion, and now . . .

# =================================================================== #
# Version 2 of Bash introduced the $'\nnn' string expansion construct.
# =================================================================== #

echo "Introducing the \$\' ... \' string-expansion construct . . . "
echo ". . . featuring more quotation marks."

echo $'\t \042 \t'   # Quote (") framed by tabs.
# Note that  '\nnn' is an octal value.

# It also works with hexadecimal values, in an $'\xhhh' construct.
echo $'\t \x22 \t'  # Quote (") framed by tabs.
# Thank you, Greg Keraunen, for pointing this out.
# Earlier Bash versions allowed '\x022'.


# Assigning ASCII characters to a variable.
# ----------------------------------------
quote=$'\042'        # " assigned to a variable.
echo "$quote Quoted string $quote and this lies outside the quotes."


# Concatenating ASCII chars in a variable.
triple_underline=$'\137\137\137'  # 137 is octal ASCII code for '_'.
echo "$triple_underline UNDERLINE $triple_underline"


ABC=$'\101\102\103\010'           # 101, 102, 103 are octal A, B, C.
echo $ABC


escape=$'\033'                    # 033 is octal for escape.
echo "\"escape\" echoes as $escape"
#                                   no visible output.


exit 0

A more elaborate example:

Example 5-3. Detecting key-presses

# Author: Sigurd Solaas, 20 Apr 2011
# Used in ABS Guide with permission.
# Requires version 4.2+ of Bash.

key="no value yet"
while true; do
  echo "Bash Extra Keys Demo. Keys to try:"
  echo "* Insert, Delete, Home, End, Page_Up and Page_Down"
  echo "* The four arrow keys"
  echo "* Tab, enter, escape, and space key"
  echo "* The letter and number keys, etc."
  echo "    d = show date/time"
  echo "    q = quit"
  echo "================================"

 # Convert the separate home-key to home-key_num_7:
 if [ "$key" = $'\x1b\x4f\x48' ]; then
  #   Quoted string-expansion construct. 

 # Convert the separate end-key to end-key_num_1.
 if [ "$key" = $'\x1b\x4f\x46' ]; then

 case "$key" in
  $'\x1b\x5b\x32\x7e')  # Insert
   echo Insert Key
  $'\x1b\x5b\x33\x7e')  # Delete
   echo Delete Key
  $'\x1b\x5b\x31\x7e')  # Home_key_num_7
   echo Home Key
  $'\x1b\x5b\x34\x7e')  # End_key_num_1
   echo End Key
  $'\x1b\x5b\x35\x7e')  # Page_Up
   echo Page_Up
  $'\x1b\x5b\x36\x7e')  # Page_Down
   echo Page_Down
  $'\x1b\x5b\x41')  # Up_arrow
   echo Up arrow
  $'\x1b\x5b\x42')  # Down_arrow
   echo Down arrow
  $'\x1b\x5b\x43')  # Right_arrow
   echo Right arrow
  $'\x1b\x5b\x44')  # Left_arrow
   echo Left arrow
  $'\x09')  # Tab
   echo Tab Key
  $'\x0a')  # Enter
   echo Enter Key
  $'\x1b')  # Escape
   echo Escape Key
  $'\x20')  # Space
   echo Space Key
  echo Time to quit...
  exit 0
   echo You pressed: \'"$key"\'

 echo "================================"

 unset K1 K2 K3
 read -s -N1 -p "Press a key: "
 read -s -N2 -t 0.001
 read -s -N1 -t 0.001


exit $?

See also Example 37-1.


gives the quote its literal meaning

echo "Hello"                     # Hello
echo "\"Hello\" ... he said."    # "Hello" ... he said.


gives the dollar sign its literal meaning (variable name following \$ will not be referenced)

echo "\$variable01"           # $variable01
echo "The book cost \$7.98."  # The book cost $7.98.


gives the backslash its literal meaning

echo "\\"  # Results in \

# Whereas . . .

echo "\"   # Invokes secondary prompt from the command-line.
           # In a script, gives an error message.

# However . . .

echo '\'   # Results in \


The behavior of \ depends on whether it is escaped, strong-quoted, weak-quoted, or appearing within command substitution or a here document.
                      #  Simple escaping and quoting
echo \z               #  z
echo \\z              # \z
echo '\z'             # \z
echo '\\z'            # \\z
echo "\z"             # \z
echo "\\z"            # \z

                      #  Command substitution
echo `echo \z`        #  z
echo `echo \\z`       #  z
echo `echo \\\z`      # \z
echo `echo \\\\z`     # \z
echo `echo \\\\\\z`   # \z
echo `echo \\\\\\\z`  # \\z
echo `echo "\z"`      # \z
echo `echo "\\z"`     # \z

                      # Here document
cat <<EOF              
EOF                   # \z

cat <<EOF              
EOF                   # \z

# These examples supplied by St�phane Chazelas.

Elements of a string assigned to a variable may be escaped, but the escape character alone may not be assigned to a variable.
echo "$variable"
# Will not work - gives an error message:
# : command not found
# A "naked" escape cannot safely be assigned to a variable.
#  What actually happens here is that the "\" escapes the newline and
#+ the effect is        variable=echo "$variable"
#+                      invalid variable assignment

echo "$variable"        #  23skidoo
                        #  This works, since the second line
                        #+ is a valid variable assignment.

#        \^    escape followed by space
echo "$variable"        # space

echo "$variable"        # \

echo "$variable"
# Will not work - gives an error message:
# \: command not found
#  First escape escapes second one, but the third one is left "naked",
#+ with same result as first instance, above.

echo "$variable"        # \\
                        # Second and fourth escapes escaped.
                        # This is o.k.

Escaping a space can prevent word splitting in a command's argument list.
file_list="/bin/cat /bin/gzip /bin/more /usr/bin/less /usr/bin/emacs-20.7"
# List of files as argument(s) to a command.

# Add two files to the list, and list all.
ls -l /usr/X11R6/bin/xsetroot /sbin/dump $file_list

echo "-------------------------------------------------------------------------"

# What happens if we escape a couple of spaces?
ls -l /usr/X11R6/bin/xsetroot\ /sbin/dump\ $file_list
# Error: the first three files concatenated into a single argument to 'ls -l'
#        because the two escaped spaces prevent argument (word) splitting.

The escape also provides a means of writing a multi-line command. Normally, each separate line constitutes a different command, but an escape at the end of a line escapes the newline character, and the command sequence continues on to the next line.

(cd /source/directory && tar cf - . ) | \
(cd /dest/directory && tar xpvf -)
# Repeating Alan Cox's directory tree copy command,
# but split into two lines for increased legibility.

# As an alternative:
tar cf - -C /source/directory . |
tar xpvf - -C /dest/directory
# See note below.
# (Thanks, St�phane Chazelas.)


If a script line ends with a |, a pipe character, then a \, an escape, is not strictly necessary. It is, however, good programming practice to always escape the end of a line of code that continues to the following line.

echo "foo


echo 'foo
bar'    # No difference yet.


echo foo\
bar     # Newline escaped.


echo "foo\
bar"     # Same here, as \ still interpreted as escape within weak quotes.


echo 'foo\
bar'     # Escape character \ taken literally because of strong quoting.

# Examples suggested by St�phane Chazelas.

Chapter 6. Exit and Exit Status


... there are dark corners in the Bourne shell, and people use all of them.

--Chet Ramey

The exit command terminates a script, just as in a C program. It can also return a value, which is available to the script's parent process.

Every command returns an exit status (sometimes referred to as a return status or exit code). A successful command returns a 0, while an unsuccessful one returns a non-zero value that usually can be interpreted as an error code. Well-behaved UNIX commands, programs, and utilities return a 0 exit code upon successful completion, though there are some exceptions.

Likewise, functions within a script and the script itself return an exit status. The last command executed in the function or script determines the exit status. Within a script, an exit nnn command may be used to deliver an nnn exit status to the shell (nnn must be an integer in the 0 - 255 range).


When a script ends with an exit that has no parameter, the exit status of the script is the exit status of the last command executed in the script (previous to the exit).



. . .


# Will exit with status of last command.


The equivalent of a bare exit is exit $? or even just omitting the exit.



. . .


# Will exit with status of last command.

exit $?



. . . 


# Will exit with status of last command.

$? reads the exit status of the last command executed. After a function returns, $? gives the exit status of the last command executed in the function. This is Bash's way of giving functions a "return value." [32]

Following the execution of a pipe, a $? gives the exit status of the last command executed.

After a script terminates, a $? from the command-line gives the exit status of the script, that is, the last command executed in the script, which is, by convention, 0 on success or an integer in the range 1 - 255 on error.

Example 6-1. exit / exit status


echo hello
echo $?    # Exit status 0 returned because command executed successfully.

lskdf      # Unrecognized command.
echo $?    # Non-zero exit status returned -- command failed to execute.


exit 113   # Will return 113 to shell.
           # To verify this, type "echo $?" after script terminates.

#  By convention, an 'exit 0' indicates success,
#+ while a non-zero exit value means an error or anomalous condition.
#  See the "Exit Codes With Special Meanings" appendix.

$? is especially useful for testing the result of a command in a script (see Example 16-35 and Example 16-20).


The !, the logical not qualifier, reverses the outcome of a test or command, and this affects its exit status.

Example 6-2. Negating a condition using !

true    # The "true" builtin.
echo "exit status of \"true\" = $?"     # 0

! true
echo "exit status of \"! true\" = $?"   # 1
# Note that the "!" needs a space between it and the command.
#    !true   leads to a "command not found" error
# The '!' operator prefixing a command invokes the Bash history mechanism.

# No error this time, but no negation either.
# It just repeats the previous command (true).

# =========================================================== #
# Preceding a _pipe_ with ! inverts the exit status returned.
ls | bogus_command     # bash: bogus_command: command not found
echo $?                # 127

! ls | bogus_command   # bash: bogus_command: command not found
echo $?                # 0
# Note that the ! does not change the execution of the pipe.
# Only the exit status changes.
# =========================================================== #

# Thanks, St�phane Chazelas and Kristopher Newsome.


Certain exit status codes have reserved meanings and should not be user-specified in a script.

Chapter 7. Tests

Every reasonably complete programming language can test for a condition, then act according to the result of the test. Bash has the test command, various bracket and parenthesis operators, and the if/then construct.

7.1. Test Constructs

  • An if/then construct tests whether the exit status of a list of commands is 0 (since 0 means "success" by UNIX convention), and if so, executes one or more commands.

  • There exists a dedicated command called [ (left bracket special character). It is a synonym for test, and a builtin for efficiency reasons. This command considers its arguments as comparison expressions or file tests and returns an exit status corresponding to the result of the comparison (0 for true, 1 for false).

  • With version 2.02, Bash introduced the [[ ... ]] extended test command, which performs comparisons in a manner more familiar to programmers from other languages. Note that [[ is a keyword, not a command.

    Bash sees [[ $a -lt $b ]] as a single element, which returns an exit status.

  • The (( ... )) and let ... constructs return an exit status, according to whether the arithmetic expressions they evaluate expand to a non-zero value. These arithmetic-expansion constructs may therefore be used to perform arithmetic comparisons.

    (( 0 && 1 ))                 # Logical AND
    echo $?     # 1     ***
    # And so ...
    let "num = (( 0 && 1 ))"
    echo $num   # 0
    # But ...
    let "num = (( 0 && 1 ))"
    echo $?     # 1     ***
    (( 200 || 11 ))              # Logical OR
    echo $?     # 0     ***
    # ...
    let "num = (( 200 || 11 ))"
    echo $num   # 1
    let "num = (( 200 || 11 ))"
    echo $?     # 0     ***
    (( 200 | 11 ))               # Bitwise OR
    echo $?                      # 0     ***
    # ...
    let "num = (( 200 | 11 ))"
    echo $num                    # 203
    let "num = (( 200 | 11 ))"
    echo $?                      # 0     ***
    # The "let" construct returns the same exit status
    #+ as the double-parentheses arithmetic expansion.


    Again, note that the exit status of an arithmetic expression is not an error value.
    var=-2 && (( var+=2 ))
    echo $?                   # 1
    var=-2 && (( var+=2 )) && echo $var
                              # Will not echo $var!

  • An if can test any command, not just conditions enclosed within brackets.

    if cmp a b &> /dev/null  # Suppress output.
    then echo "Files a and b are identical."
    else echo "Files a and b differ."
    # The very useful "if-grep" construct:
    # ----------------------------------- 
    if grep -q Bash file
      then echo "File contains at least one occurrence of Bash."
    if echo "$word" | grep -q "$letter_sequence"
    # The "-q" option to grep suppresses output.
      echo "$letter_sequence found in $word"
      echo "$letter_sequence not found in $word"
      then echo "Command succeeded."
      else echo "Command failed."

  • These last two examples courtesy of St�phane Chazelas.

Example 7-1. What is truth?


#  Tip:
#  If you're unsure how a certain condition might evaluate,
#+ test it in an if-test.


echo "Testing \"0\""
if [ 0 ]      # zero
  echo "0 is true."
else          # Or else ...
  echo "0 is false."
fi            # 0 is true.


echo "Testing \"1\""
if [ 1 ]      # one
  echo "1 is true."
  echo "1 is false."
fi            # 1 is true.


echo "Testing \"-1\""
if [ -1 ]     # minus one
  echo "-1 is true."
  echo "-1 is false."
fi            # -1 is true.


echo "Testing \"NULL\""
if [ ]        # NULL (empty condition)
  echo "NULL is true."
  echo "NULL is false."
fi            # NULL is false.


echo "Testing \"xyz\""
if [ xyz ]    # string
  echo "Random string is true."
  echo "Random string is false."
fi            # Random string is true.


echo "Testing \"\$xyz\""
if [ $xyz ]   # Tests if $xyz is null, but...
              # it's only an uninitialized variable.
  echo "Uninitialized variable is true."
  echo "Uninitialized variable is false."
fi            # Uninitialized variable is false.


echo "Testing \"-n \$xyz\""
if [ -n "$xyz" ]            # More pedantically correct.
  echo "Uninitialized variable is true."
  echo "Uninitialized variable is false."
fi            # Uninitialized variable is false.


xyz=          # Initialized, but set to null value.

echo "Testing \"-n \$xyz\""
if [ -n "$xyz" ]
  echo "Null variable is true."
  echo "Null variable is false."
fi            # Null variable is false.


# When is "false" true?

echo "Testing \"false\""
if [ "false" ]              #  It seems that "false" is just a string ...
  echo "\"false\" is true." #+ and it tests true.
  echo "\"false\" is false."
fi            # "false" is true.


echo "Testing \"\$false\""  # Again, uninitialized variable.
if [ "$false" ]
  echo "\"\$false\" is true."
  echo "\"\$false\" is false."
fi            # "$false" is false.
              # Now, we get the expected result.

#  What would happen if we tested the uninitialized variable "$true"?


exit 0

Exercise. Explain the behavior of Example 7-1, above.

if [ condition-true ]
   command 1
   command 2
else  # Or else ...
      # Adds default code block executing if original condition tests false.
   command 3
   command 4


When if and then are on same line in a condition test, a semicolon must terminate the if statement. Both if and then are keywords. Keywords (or commands) begin statements, and before a new statement on the same line begins, the old one must terminate.

if [ -x "$filename" ]; then

Else if and elif


elif is a contraction for else if. The effect is to nest an inner if/then construct within an outer one.

if [ condition1 ]
elif [ condition2 ]
# Same as else if

The if test condition-true construct is the exact equivalent of if [ condition-true ]. As it happens, the left bracket, [ , is a token [33] which invokes the test command. The closing right bracket, ] , in an if/test should not therefore be strictly necessary, however newer versions of Bash require it.


The test command is a Bash builtin which tests file types and compares strings. Therefore, in a Bash script, test does not call the external /usr/bin/test binary, which is part of the sh-utils package. Likewise, [ does not call /usr/bin/[, which is linked to /usr/bin/test.

bash$ type test
test is a shell builtin
bash$ type '['
[ is a shell builtin
bash$ type '[['
[[ is a shell keyword
bash$ type ']]'
]] is a shell keyword
bash$ type ']'
bash: type: ]: not found

If, for some reason, you wish to use /usr/bin/test in a Bash script, then specify it by full pathname.

Example 7-2. Equivalence of test, /usr/bin/test, [ ], and /usr/bin/[



if test -z "$1"
  echo "No command-line arguments."
  echo "First command-line argument is $1."


if /usr/bin/test -z "$1"      # Equivalent to "test" builtin.
#  ^^^^^^^^^^^^^              # Specifying full pathname.
  echo "No command-line arguments."
  echo "First command-line argument is $1."


if [ -z "$1" ]                # Functionally identical to above code blocks.
#   if [ -z "$1"                should work, but...
#+  Bash responds to a missing close-bracket with an error message.
  echo "No command-line arguments."
  echo "First command-line argument is $1."


if /usr/bin/[ -z "$1" ]       # Again, functionally identical to above.
# if /usr/bin/[ -z "$1"       # Works, but gives an error message.
#                             # Note:
#                               This has been fixed in Bash, version 3.x.
  echo "No command-line arguments."
  echo "First command-line argument is $1."


exit 0


Following an if, neither the test command nor the test brackets ( [ ] or [[ ]] ) are strictly necessary.

if cd "$dir" 2>/dev/null; then   # "2>/dev/null" hides error message.
  echo "Now in $dir."
  echo "Can't change to $dir."
The "if COMMAND" construct returns the exit status of COMMAND.

Similarly, a condition within test brackets may stand alone without an if, when used in combination with a list construct.
[ "$var1" -ne "$var2" ] && echo "$var1 is not equal to $var2"

[ -d "$home" ] || echo "$home directory does not exist."

The (( )) construct expands and evaluates an arithmetic expression. If the expression evaluates as zero, it returns an exit status of 1, or "false". A non-zero expression returns an exit status of 0, or "true". This is in marked contrast to using the test and [ ] constructs previously discussed.

Example 7-3. Arithmetic Tests using (( ))

# Arithmetic tests.

# The (( ... )) construct evaluates and tests numerical expressions.
# Exit status opposite from [ ... ] construct!

(( 0 ))
echo "Exit status of \"(( 0 ))\" is $?."         # 1

(( 1 ))
echo "Exit status of \"(( 1 ))\" is $?."         # 0

(( 5 > 4 ))                                      # true
echo "Exit status of \"(( 5 > 4 ))\" is $?."     # 0

(( 5 > 9 ))                                      # false
echo "Exit status of \"(( 5 > 9 ))\" is $?."     # 1

(( 5 == 5 ))                                     # true
echo "Exit status of \"(( 5 == 5 ))\" is $?."    # 0
# (( 5 = 5 ))  gives an error message.

(( 5 - 5 ))                                      # 0
echo "Exit status of \"(( 5 - 5 ))\" is $?."     # 1

(( 5 / 4 ))                                      # Division o.k.
echo "Exit status of \"(( 5 / 4 ))\" is $?."     # 0

(( 1 / 2 ))                                      # Division result < 1.
echo "Exit status of \"(( 1 / 2 ))\" is $?."     # Rounded off to 0.
                                                 # 1

(( 1 / 0 )) 2>/dev/null                          # Illegal division by 0.
#           ^^^^^^^^^^^
echo "Exit status of \"(( 1 / 0 ))\" is $?."     # 1

# What effect does the "2>/dev/null" have?
# What would happen if it were removed?
# Try removing it, then rerunning the script.

# ======================================= #

# (( ... )) also useful in an if-then test.


if (( var1 > var2 ))
then #^      ^      Note: Not $var1, $var2. Why?
  echo "$var1 is greater than $var2"
fi     # 5 is greater than 4

exit 0

7.2. File test operators

Returns true if...


file exists


file exists

This is identical in effect to -e. It has been "deprecated," [34] and its use is discouraged.


file is a regular file (not a directory or device file)


file is not zero size


file is a directory


file is a block device


file is a character device

device0="/dev/sda2"    # /   (root directory)
if [ -b "$device0" ]
  echo "$device0 is a block device."

# /dev/sda2 is a block device.

device1="/dev/ttyS1"   # PCMCIA modem card.
if [ -c "$device1" ]
  echo "$device1 is a character device."

# /dev/ttyS1 is a character device.


file is a pipe

function show_input_type()
   [ -p /dev/fd/0 ] && echo PIPE || echo STDIN

show_input_type "Input"                           # STDIN
echo "Input" | show_input_type                    # PIPE

# This example courtesy of Carl Anderson.


file is a symbolic link


file is a symbolic link


file is a socket


file (descriptor) is associated with a terminal device

This test option may be used to check whether the stdin [ -t 0 ] or stdout [ -t 1 ] in a given script is a terminal.


file has read permission (for the user running the test)


file has write permission (for the user running the test)


file has execute permission (for the user running the test)


set-group-id (sgid) flag set on file or directory

If a directory has the sgid flag set, then a file created within that directory belongs to the group that owns the directory, not necessarily to the group of the user who created the file. This may be useful for a directory shared by a workgroup.


set-user-id (suid) flag set on file

A binary owned by root with set-user-id flag set runs with root privileges, even when an ordinary user invokes it. [35] This is useful for executables (such as pppd and cdrecord) that need to access system hardware. Lacking the suid flag, these binaries could not be invoked by a non-root user.

	      -rwsr-xr-t    1 root       178236 Oct  2  2000 /usr/sbin/pppd

A file with the suid flag set shows an s in its permissions.


sticky bit set

Commonly known as the sticky bit, the save-text-mode flag is a special type of file permission. If a file has this flag set, that file will be kept in cache memory, for quicker access. [36] If set on a directory, it restricts write permission. Setting the sticky bit adds a t to the permissions on the file or directory listing. This restricts altering or deleting specific files in that directory to the owner of those files.

	      drwxrwxrwt    7 root         1024 May 19 21:26 tmp/

If a user does not own a directory that has the sticky bit set, but has write permission in that directory, she can only delete those files that she owns in it. This keeps users from inadvertently overwriting or deleting each other's files in a publicly accessible directory, such as /tmp. (The owner of the directory or root can, of course, delete or rename files there.)


you are owner of file


group-id of file same as yours


file modified since it was last read

f1 -nt f2

file f1 is newer than f2

f1 -ot f2

file f1 is older than f2

f1 -ef f2

files f1 and f2 are hard links to the same file


"not" -- reverses the sense of the tests above (returns true if condition absent).

Example 7-4. Testing for broken links

# Written by Lee bigelow <[email protected]>
# Used in ABS Guide with permission.

#  A pure shell script to find dead symlinks and output them quoted
#+ so they can be fed to xargs and dealt with :)
#+ eg. sh /somedir /someotherdir|xargs rm
#  This, however, is a better method:
#  find "somedir" -type l -print0|\
#  xargs -r0 file|\
#  grep "broken symbolic"|
#  sed -e 's/^\|: *broken symbolic.*$/"/g'
#+ but that wouldn't be pure Bash, now would it.
#  Caution: beware the /proc file system and any circular links!

#  If no args are passed to the script set directories-to-search 
#+ to current directory.  Otherwise set the directories-to-search 
#+ to the args passed.

[ $# -eq 0 ] && directorys=`pwd` || directorys=$@

#  Setup the function linkchk to check the directory it is passed 
#+ for files that are links and don't exist, then print them quoted.
#  If one of the elements in the directory is a subdirectory then 
#+ send that subdirectory to the linkcheck function.

linkchk () {
    for element in $1/*; do
      [ -h "$element" -a ! -e "$element" ] && echo \"$element\"
      [ -d "$element" ] && linkchk $element
    # Of course, '-h' tests for symbolic link, '-d' for directory.

#  Send each arg that was passed to the script to the linkchk() function
#+ if it is a valid directoy.  If not, then print the error message
#+ and usage info.
for directory in $directorys; do
    if [ -d $directory ]
	then linkchk $directory
	    echo "$directory is not a directory"
	    echo "Usage: $0 dir1 dir2 ..."

exit $?

Example 31-1, Example 11-8, Example 11-3, Example 31-3, and Example A-1 also illustrate uses of the file test operators.

7.3. Other Comparison Operators

A binary comparison operator compares two variables or quantities. Note that integer and string comparison use a different set of operators.

integer comparison


is equal to

if [ "$a" -eq "$b" ]


is not equal to

if [ "$a" -ne "$b" ]


is greater than

if [ "$a" -gt "$b" ]


is greater than or equal to

if [ "$a" -ge "$b" ]


is less than

if [ "$a" -lt "$b" ]


is less than or equal to

if [ "$a" -le "$b" ]


is less than (within double parentheses)

(("$a" < "$b"))


is less than or equal to (within double parentheses)

(("$a" <= "$b"))


is greater than (within double parentheses)

(("$a" > "$b"))


is greater than or equal to (within double parentheses)

(("$a" >= "$b"))

string comparison


is equal to

if [ "$a" = "$b" ]


Note the whitespace framing the =.

if [ "$a"="$b" ] is not equivalent to the above.


is equal to

if [ "$a" == "$b" ]

This is a synonym for =.


The == comparison operator behaves differently within a double-brackets test than within single brackets.
[[ $a == z* ]]   # True if $a starts with an "z" (pattern matching).
[[ $a == "z*" ]] # True if $a is equal to z* (literal matching).

[ $a == z* ]     # File globbing and word splitting take place.
[ "$a" == "z*" ] # True if $a is equal to z* (literal matching).

# Thanks, St�phane Chazelas


is not equal to

if [ "$a" != "$b" ]

This operator uses pattern matching within a [[ ... ]] construct.


is less than, in ASCII alphabetical order

if [[ "$a" < "$b" ]]

if [ "$a" \< "$b" ]

Note that the "<" needs to be escaped within a [ ] construct.


is greater than, in ASCII alphabetical order

if [[ "$a" > "$b" ]]

if [ "$a" \> "$b" ]

Note that the ">" needs to be escaped within a [ ] construct.

See Example 27-11 for an application of this comparison operator.


string is null, that is, has zero length

 String=''   # Zero-length ("null") string variable.

if [ -z "$String" ]
  echo "\$String is null."
  echo "\$String is NOT null."
fi     # $String is null.


string is not null.


The -n test requires that the string be quoted within the test brackets. Using an unquoted string with ! -z, or even just the unquoted string alone within test brackets (see Example 7-6) normally works, however, this is an unsafe practice. Always quote a tested string. [37]

Example 7-5. Arithmetic and string comparisons



#  Here "a" and "b" can be treated either as integers or strings.
#  There is some blurring between the arithmetic and string comparisons,
#+ since Bash variables are not strongly typed.

#  Bash permits integer operations and comparisons on variables
#+ whose value consists of all-integer characters.
#  Caution advised, however.


if [ "$a" -ne "$b" ]
  echo "$a is not equal to $b"
  echo "(arithmetic comparison)"


if [ "$a" != "$b" ]
  echo "$a is not equal to $b."
  echo "(string comparison)"
  #     "4"  != "5"
  # ASCII 52 != ASCII 53

# In this particular instance, both "-ne" and "!=" work.


exit 0

Example 7-6. Testing whether a string is null

# Testing null strings and unquoted strings,
#+ but not strings and sealing wax, not to mention cabbages and kings . . .

# Using   if [ ... ]

# If a string has not been initialized, it has no defined value.
# This state is called "null" (not the same as zero!).

if [ -n $string1 ]    # string1 has not been declared or initialized.
  echo "String \"string1\" is not null."
  echo "String \"string1\" is null."
fi                    # Wrong result.
# Shows $string1 as not null, although it was not initialized.


# Let's try it again.

if [ -n "$string1" ]  # This time, $string1 is quoted.
  echo "String \"string1\" is not null."
  echo "String \"string1\" is null."
fi                    # Quote strings within test brackets!


if [ $string1 ]       # This time, $string1 stands naked.
  echo "String \"string1\" is not null."
  echo "String \"string1\" is null."
fi                    # This works fine.
# The [ ... ] test operator alone detects whether the string is null.
# However it is good practice to quote it (if [ "$string1" ]).
# As Stephane Chazelas points out,
#    if [ $string1 ]    has one argument, "]"
#    if [ "$string1" ]  has two arguments, the empty "$string1" and "]" 



if [ $string1 ]       # Again, $string1 stands unquoted.
  echo "String \"string1\" is not null."
  echo "String \"string1\" is null."
fi                    # Again, gives correct result.
# Still, it is better to quote it ("$string1"), because . . .

string1="a = b"

if [ $string1 ]       # Again, $string1 stands unquoted.
  echo "String \"string1\" is not null."
  echo "String \"string1\" is null."
fi                    # Not quoting "$string1" now gives wrong result!

exit 0   # Thank you, also, Florian Wisser, for the "heads-up".

Example 7-7. zmore

# zmore

# View gzipped files with 'more' filter.


if [ $# -eq 0 ] # same effect as:  if [ -z "$1" ]
# $1 can exist, but be empty:  zmore "" arg2 arg3
  echo "Usage: `basename $0` filename" >&2
  # Error message to stderr.
  exit $E_NOARGS
  # Returns 85 as exit status of script (error code).


if [ ! -f "$filename" ]   # Quoting $filename allows for possible spaces.
  echo "File $filename not found!" >&2   # Error message to stderr.
  exit $E_NOTFOUND

if [ ${filename##*.} != "gz" ]
# Using bracket in variable substitution.
  echo "File $1 is not a gzipped file!"
  exit $E_NOTGZIP

zcat $1 | more

# Uses the 'more' filter.
# May substitute 'less' if desired.

exit $?   # Script returns exit status of pipe.
#  Actually "exit $?" is unnecessary, as the script will, in any case,
#+ return the exit status of the last command executed.

compound comparison


logical and

exp1 -a exp2 returns true if both exp1 and exp2 are true.


logical or

exp1 -o exp2 returns true if either exp1 or exp2 is true.

These are similar to the Bash comparison operators && and ||, used within double brackets.
[[ condition1 && condition2 ]]

The -o and -a operators work with the test command or occur within single test brackets.
if [ "$expr1" -a "$expr2" ]
  echo "Both expr1 and expr2 are true."
  echo "Either expr1 or expr2 is false."


But, as rihad points out:
[ 1 -eq 1 ] && [ -n "`echo true 1>&2`" ]   # true
[ 1 -eq 2 ] && [ -n "`echo true 1>&2`" ]   # (no output)
# ^^^^^^^ False condition. So far, everything as expected.

# However ...
[ 1 -eq 2 -a -n "`echo true 1>&2`" ]       # true
# ^^^^^^^ False condition. So, why "true" output?

# Is it because both condition clauses within brackets evaluate?
[[ 1 -eq 2 && -n "`echo true 1>&2`" ]]     # (no output)
# No, that's not it.

# Apparently && and || "short-circuit" while -a and -o do not.

Refer to Example 8-3, Example 27-17, and Example A-29 to see compound comparison operators in action.

7.4. Nested if/then Condition Tests

Condition tests using the if/then construct may be nested. The net result is equivalent to using the && compound comparison operator.


if [ "$a" -gt 0 ]
  if [ "$a" -lt 5 ]
    echo "The value of \"a\" lies somewhere between 0 and 5."

# Same result as:

if [ "$a" -gt 0 ] && [ "$a" -lt 5 ]
  echo "The value of \"a\" lies somewhere between 0 and 5."

Example 37-4 and Example 17-11 demonstrate nested if/then condition tests.

7.5. Testing Your Knowledge of Tests

The systemwide xinitrc file can be used to launch the X server. This file contains quite a number of if/then tests. The following is excerpted from an "ancient" version of xinitrc (Red Hat 7.1, or thereabouts).

if [ -f $HOME/.Xclients ]; then
  exec $HOME/.Xclients
elif [ -f /etc/X11/xinit/Xclients ]; then
  exec /etc/X11/xinit/Xclients
     # failsafe settings.  Although we should never get here
     # (we provide fallbacks in Xclients as well) it can't hurt.
     xclock -geometry 100x100-5+5 &
     xterm -geometry 80x50-50+150 &
     if [ -f /usr/bin/netscape -a -f /usr/share/doc/HTML/index.html ]; then
             netscape /usr/share/doc/HTML/index.html &

Explain the test constructs in the above snippet, then examine an updated version of the file, /etc/X11/xinit/xinitrc, and analyze the if/then test constructs there. You may need to refer ahead to the discussions of grep, sed, and regular expressions.

Chapter 8. Operations and Related Topics

8.1. Operators


variable assignment

Initializing or changing the value of a variable


All-purpose assignment operator, which works for both arithmetic and string assignments.

category=minerals  # No spaces allowed after the "=".


Do not confuse the "=" assignment operator with the = test operator.

#   =  as a test operator

if [ "$string1" = "$string2" ]

#  if [ "X$string1" = "X$string2" ] is safer,
#+ to prevent an error message should one of the variables be empty.
#  (The prepended "X" characters cancel out.)

arithmetic operators











# Bash, version 2.02, introduced the "**" exponentiation operator.

let "z=5**3"    # 5 * 5 * 5
echo "z = $z"   # z = 125


modulo, or mod (returns the remainder of an integer division operation)

bash$ expr 5 % 3
5/3 = 1, with remainder 2

This operator finds use in, among other things, generating numbers within a specific range (see Example 9-11 and Example 9-15) and formatting program output (see Example 27-16 and Example A-6). It can even be used to generate prime numbers, (see Example A-15). Modulo turns up surprisingly often in numerical recipes.

Example 8-1. Greatest common divisor

# greatest common divisor
#         Uses Euclid's algorithm

#  The "greatest common divisor" (gcd) of two integers
#+ is the largest integer that will divide both, leaving no remainder.

#  Euclid's algorithm uses successive division.
#    In each pass,
#+      dividend <---  divisor
#+      divisor  <---  remainder
#+   until remainder = 0.
#    The gcd = dividend, on the final pass.
#  For an excellent discussion of Euclid's algorithm, see
#+ Jim Loy's site,

# ------------------------------------------------------
# Argument check

if [ $# -ne "$ARGS" ]
  echo "Usage: `basename $0` first-number second-number"
  exit $E_BADARGS
# ------------------------------------------------------

gcd ()

  dividend=$1             #  Arbitrary assignment.
  divisor=$2              #! It doesn't matter which of the two is larger.
                          #  Why not?

  remainder=1             #  If an uninitialized variable is used inside
                          #+ test brackets, an error message results.

  until [ "$remainder" -eq 0 ]
  do    #  ^^^^^^^^^^  Must be previously initialized!
    let "remainder = $dividend % $divisor"
    dividend=$divisor     # Now repeat with 2 smallest numbers.
  done                    # Euclid's algorithm

}                         # Last $dividend is the gcd.

gcd $1 $2

echo; echo "GCD of $1 and $2 = $dividend"; echo

# Exercises :
# ---------
# 1) Check command-line arguments to make sure they are integers,
#+   and exit the script with an appropriate error message if not.
# 2) Rewrite the gcd () function to use local variables.

exit 0

plus-equal (increment variable by a constant) [38]

let "var += 5" results in var being incremented by 5.


minus-equal (decrement variable by a constant)


times-equal (multiply variable by a constant)

let "var *= 4" results in var being multiplied by 4.


slash-equal (divide variable by a constant)


mod-equal (remainder of dividing variable by a constant)

Arithmetic operators often occur in an expr or let expression.

Example 8-2. Using Arithmetic Operations

# Counting to 11 in 10 different ways.

n=1; echo -n "$n "

let "n = $n + 1"   # let "n = n + 1"  also works.
echo -n "$n "

: $((n = $n + 1))
#  ":" necessary because otherwise Bash attempts
#+ to interpret "$((n = $n + 1))" as a command.
echo -n "$n "

(( n = n + 1 ))
#  A simpler alternative to the method above.
#  Thanks, David Lombard, for pointing this out.
echo -n "$n "

n=$(($n + 1))
echo -n "$n "

: $[ n = $n + 1 ]
#  ":" necessary because otherwise Bash attempts
#+ to interpret "$[ n = $n + 1 ]" as a command.
#  Works even if "n" was initialized as a string.
echo -n "$n "

n=$[ $n + 1 ]
#  Works even if "n" was initialized as a string.
#* Avoid this type of construct, since it is obsolete and nonportable.
#  Thanks, Stephane Chazelas.
echo -n "$n "

# Now for C-style increment operators.
# Thanks, Frank Wang, for pointing this out.

let "n++"          # let "++n"  also works.
echo -n "$n "

(( n++ ))          # (( ++n ))  also works.
echo -n "$n "

: $(( n++ ))       # : $(( ++n )) also works.
echo -n "$n "

: $[ n++ ]         # : $[ ++n ] also works
echo -n "$n "


exit 0


Integer variables in older versions of Bash were signed long (32-bit) integers, in the range of -2147483648 to 2147483647. An operation that took a variable outside these limits gave an erroneous result.

echo $BASH_VERSION   # 1.14

echo "a = $a"        # a = 2147483646
let "a+=1"           # Increment "a".
echo "a = $a"        # a = 2147483647
let "a+=1"           # increment "a" again, past the limit.
echo "a = $a"        # a = -2147483648
                     #      ERROR: out of range,
                     # +    and the leftmost bit, the sign bit,
                     # +    has been set, making the result negative.

As of version >= 2.05b, Bash supports 64-bit integers.


Bash does not understand floating point arithmetic. It treats numbers containing a decimal point as strings.


let "b = $a + 1.3"  # Error.
# let: b = 1.5 + 1.3: syntax error in expression
#                            (error token is ".5 + 1.3")

echo "b = $b"       # b=1

Use bc in scripts that that need floating point calculations or math library functions.

bitwise operators. The bitwise operators seldom make an appearance in shell scripts. Their chief use seems to be manipulating and testing values read from ports or sockets. "Bit flipping" is more relevant to compiled languages, such as C and C++, which provide direct access to system hardware. However, see vladz's ingenious use of bitwise operators in his (Example A-54) script.

bitwise operators


bitwise left shift (multiplies by 2 for each shift position)



let "var <<= 2" results in var left-shifted 2 bits (multiplied by 4)


bitwise right shift (divides by 2 for each shift position)


right-shift-equal (inverse of <<=)


bitwise AND


bitwise AND-equal


bitwise OR


bitwise OR-equal


bitwise NOT


bitwise XOR


bitwise XOR-equal

logical (boolean) operators



if [ ! -f $FILENAME ]



if [ $condition1 ] && [ $condition2 ]
#  Same as:  if [ $condition1 -a $condition2 ]
#  Returns true if both condition1 and condition2 hold true...

if [[ $condition1 && $condition2 ]]    # Also works.
#  Note that && operator not permitted inside brackets
#+ of [ ... ] construct.


&& may also be used, depending on context, in an and list to concatenate commands.



if [ $condition1 ] || [ $condition2 ]
# Same as:  if [ $condition1 -o $condition2 ]
# Returns true if either condition1 or condition2 holds true...

if [[ $condition1 || $condition2 ]]    # Also works.
#  Note that || operator not permitted inside brackets
#+ of a [ ... ] construct.


Bash tests the exit status of each statement linked with a logical operator.

Example 8-3. Compound Condition Tests Using && and ||



if [ "$a" -eq 24 ] && [ "$b" -eq 47 ]
  echo "Test #1 succeeds."
  echo "Test #1 fails."

# ERROR:   if [ "$a" -eq 24 && "$b" -eq 47 ]
#+         attempts to execute  ' [ "$a" -eq 24 '
#+         and fails to finding matching ']'.
#  Note:  if [[ $a -eq 24 && $b -eq 24 ]]  works.
#  The double-bracket if-test is more flexible
#+ than the single-bracket version.       
#    (The "&&" has a different meaning in line 17 than in line 6.)
#    Thanks, Stephane Chazelas, for pointing this out.

if [ "$a" -eq 98 ] || [ "$b" -eq 47 ]
  echo "Test #2 succeeds."
  echo "Test #2 fails."

#  The -a and -o options provide
#+ an alternative compound condition test.
#  Thanks to Patrick Callahan for pointing this out.

if [ "$a" -eq 24 -a "$b" -eq 47 ]
  echo "Test #3 succeeds."
  echo "Test #3 fails."

if [ "$a" -eq 98 -o "$b" -eq 47 ]
  echo "Test #4 succeeds."
  echo "Test #4 fails."

if [ "$a" = rhino ] && [ "$b" = crocodile ]
  echo "Test #5 succeeds."
  echo "Test #5 fails."

exit 0

The && and || operators also find use in an arithmetic context.

bash$ echo $(( 1 && 2 )) $((3 && 0)) $((4 || 0)) $((0 || 0))
1 0 1 0

miscellaneous operators


Comma operator

The comma operator chains together two or more arithmetic operations. All the operations are evaluated (with possible side effects. [39]

let "t1 = ((5 + 3, 7 - 1, 15 - 4))"
echo "t1 = $t1"           ^^^^^^  # t1 = 11
# Here t1 is set to the result of the last operation. Why?

let "t2 = ((a = 9, 15 / 3))"      # Set "a" and calculate "t2".
echo "t2 = $t2    a = $a"         # t2 = 5    a = 9

The comma operator finds use mainly in for loops. See Example 11-13.

8.2. Numerical Constants

A shell script interprets a number as decimal (base 10), unless that number has a special prefix or notation. A number preceded by a 0 is octal (base 8). A number preceded by 0x is hexadecimal (base 16). A number with an embedded # evaluates as BASE#NUMBER (with range and notational restrictions).

Example 8-4. Representation of numerical constants

# Representation of numbers in different bases.

# Decimal: the default
let "dec = 32"
echo "decimal number = $dec"             # 32
# Nothing out of the ordinary here.

# Octal: numbers preceded by '0' (zero)
let "oct = 032"
echo "octal number = $oct"               # 26
# Expresses result in decimal.
# --------- ------ -- -------

# Hexadecimal: numbers preceded by '0x' or '0X'
let "hex = 0x32"
echo "hexadecimal number = $hex"         # 50

echo $((0x9abc))                         # 39612
#     ^^      ^^   double-parentheses arithmetic expansion/evaluation
# Expresses result in decimal.

# Other bases: BASE#NUMBER
# BASE between 2 and 64.
# NUMBER must use symbols within the BASE range, see below.

let "bin = 2#111100111001101"
echo "binary number = $bin"              # 31181

let "b32 = 32#77"
echo "base-32 number = $b32"             # 231

let "b64 = 64#@_"
echo "base-64 number = $b64"             # 4031
# This notation only works for a limited range (2 - 64) of ASCII characters.
# 10 digits + 26 lowercase characters + 26 uppercase characters + @ + _


echo $((36#zz)) $((2#10101010)) $((16#AF16)) $((53#1aA))
                                         # 1295 170 44822 3375

#  Important note:
#  --------------
#  Using a digit out of range of the specified base notation
#+ gives an error message.

let "bad_oct = 081"
# (Partial) error message output:
#  bad_oct = 081: value too great for base (error token is "081")
#              Octal numbers use only digits in the range 0 - 7.

exit $?   # Exit value = 1 (error)

# Thanks, Rich Bartell and Stephane Chazelas, for clarification.

8.3. The Double-Parentheses Construct

Similar to the let command, the (( ... )) construct permits arithmetic expansion and evaluation. In its simplest form, a=$(( 5 + 3 )) would set a to 5 + 3, or 8. However, this double-parentheses construct is also a mechanism for allowing C-style manipulation of variables in Bash, for example, (( var++ )).

Example 8-5. C-style manipulation of variables

# Manipulating a variable, C-style, using the (( ... )) construct.


(( a = 23 ))  #  Setting a value, C-style,
              #+ with spaces on both sides of the "=".
echo "a (initial value) = $a"   # 23

(( a++ ))     #  Post-increment 'a', C-style.
echo "a (after a++) = $a"       # 24

(( a-- ))     #  Post-decrement 'a', C-style.
echo "a (after a--) = $a"       # 23

(( ++a ))     #  Pre-increment 'a', C-style.
echo "a (after ++a) = $a"       # 24

(( --a ))     #  Pre-decrement 'a', C-style.
echo "a (after --a) = $a"       # 23


#  Note that, as in C, pre- and post-decrement operators
#+ have different side-effects.

n=1; let --n && echo "True" || echo "False"  # False
n=1; let n-- && echo "True" || echo "False"  # True

#  Thanks, Jeroen Domburg.


(( t = a<45?7:11 ))   # C-style trinary operator.
#       ^  ^ ^
echo "If a < 45, then t = 7, else t = 11."  # a = 23
echo "t = $t "                              # t = 7


# -----------------
# Easter Egg alert!
# -----------------
#  Chet Ramey seems to have snuck a bunch of undocumented C-style
#+ constructs into Bash (actually adapted from ksh, pretty much).
#  In the Bash docs, Ramey calls (( ... )) shell arithmetic,
#+ but it goes far beyond that.
#  Sorry, Chet, the secret is out.

# See also "for" and "while" loops using the (( ... )) construct.

# These work only with version 2.04 or later of Bash.


See also Example 11-13 and Example 8-4.

8.4. Operator Precedence

In a script, operations execute in order of precedence: the higher precedence operations execute before the lower precedence ones. [40]

Table 8-1. Operator Precedence

var++ var--post-increment, post-decrementC-style operators
++var --varpre-increment, pre-decrement 
! ~negationlogical / bitwise, inverts sense of following operator
**exponentiationarithmetic operation
* / %multiplication, division, moduloarithmetic operation
+ -addition, subtractionarithmetic operation
<< >>left, right shiftbitwise
-z -nunary comparisonstring is/is-not null
-e -f -t -x, etc.unary comparisonfile-test
< -lt > -gt <= -le >= -gecompound comparisonstring and integer
-nt -ot -efcompound comparisonfile-test
== -eq != -neequality / inequalitytest operators, string and integer
^XORexclusive OR, bitwise
&& -aANDlogical, compound comparison
|| -oORlogical, compound comparison
?:trinary operatorC-style
=assignment(do not confuse with equality test)
*= /= %= += -= <<= >>= &=combination assignmenttimes-equal, divide-equal, mod-equal, etc.
,commalinks a sequence of operations

In practice, all you really need to remember is the following:

  • The "My Dear Aunt Sally" mantra (multiply, divide, add, subtract) for the familiar arithmetic operations.

  • The compound logical operators, &&, ||, -a, and -o have low precedence.

  • The order of evaluation of equal-precedence operators is usually left-to-right.

Now, let's utilize our knowledge of operator precedence to analyze a couple of lines from the /etc/init.d/functions file, as found in the Fedora Core Linux distro.

while [ -n "$remaining" -a "$retry" -gt 0 ]; do

# This looks rather daunting at first glance.

# Separate the conditions:
while [ -n "$remaining" -a "$retry" -gt 0 ]; do
#       --condition 1-- ^^ --condition 2-

#  If variable "$remaining" is not zero length
#+      AND (-a)
#+ variable "$retry" is greater-than zero
#+ then
#+ the [ expresion-within-condition-brackets ] returns success (0)
#+ and the while-loop executes an iteration.
#  ==============================================================
#  Evaluate "condition 1" and "condition 2" ***before***
#+ ANDing them. Why? Because the AND (-a) has a lower precedence
#+ than the -n and -gt operators,
#+ and therefore gets evaluated *last*.


if [ -f /etc/sysconfig/i18n -a -z "${NOLOCALE:-}" ] ; then

# Again, separate the conditions:
if [ -f /etc/sysconfig/i18n -a -z "${NOLOCALE:-}" ] ; then
#    --condition 1--------- ^^ --condition 2-----

#  If file "/etc/sysconfig/i18n" exists
#+      AND (-a)
#+ variable $NOLOCALE is zero length
#+ then
#+ the [ test-expresion-within-condition-brackets ] returns success (0)
#+ and the commands following execute.
#  As before, the AND (-a) gets evaluated *last*
#+ because it has the lowest precedence of the operators within
#+ the test brackets.
#  ==============================================================
#  Note:
#  ${NOLOCALE:-} is a parameter expansion that seems redundant.
#  But, if $NOLOCALE has not been declared, it gets set to *null*,
#+ in effect declaring it.
#  This makes a difference in some contexts.


To avoid confusion or error in a complex sequence of test operators, break up the sequence into bracketed sections.
if [ "$v1" -gt "$v2"  -o  "$v1" -lt "$v2"  -a  -e "$filename" ]
# Unclear what's going on here...

if [[ "$v1" -gt "$v2" ]] || [[ "$v1" -lt "$v2" ]] && [[ -e "$filename" ]]
# Much better -- the condition tests are grouped in logical sections.

Chapter 9. Another Look at Variables

Used properly, variables can add power and flexibility to scripts. This requires learning their subtleties and nuances.

9.1. Internal Variables

Builtin variables:

variables affecting bash script behavior


The path to the Bash binary itself
bash$ echo $BASH


An environmental variable pointing to a Bash startup file to be read when a script is invoked


A variable indicating the subshell level. This is a new addition to Bash, version 3.

See Example 21-1 for usage.


Process ID of the current instance of Bash. This is not the same as the $$ variable, but it often gives the same result.

bash4$ echo $$

bash4$ echo $BASHPID

bash4$ ps ax | grep bash4
11015 pts/2    R      0:00 bash4

But ...


echo "\$\$ outside of subshell = $$"                              # 9602
echo "\$BASH_SUBSHELL  outside of subshell = $BASH_SUBSHELL"      # 0
echo "\$BASHPID outside of subshell = $BASHPID"                   # 9602


( echo "\$\$ inside of subshell = $$"                             # 9602
  echo "\$BASH_SUBSHELL inside of subshell = $BASH_SUBSHELL"      # 1
  echo "\$BASHPID inside of subshell = $BASHPID" )                # 9603
  # Note that $$ returns PID of parent process.


A 6-element array containing version information about the installed release of Bash. This is similar to $BASH_VERSION, below, but a bit more detailed.

# Bash version info:

for n in 0 1 2 3 4 5
  echo "BASH_VERSINFO[$n] = ${BASH_VERSINFO[$n]}"

# BASH_VERSINFO[0] = 3                      # Major version no.
# BASH_VERSINFO[1] = 00                     # Minor version no.
# BASH_VERSINFO[2] = 14                     # Patch level.
# BASH_VERSINFO[3] = 1                      # Build version.
# BASH_VERSINFO[4] = release                # Release status.
# BASH_VERSINFO[5] = i386-redhat-linux-gnu  # Architecture
                                            # (same as $MACHTYPE).


The version of Bash installed on the system

bash$ echo $BASH_VERSION

tcsh% echo $BASH_VERSION
BASH_VERSION: Undefined variable.

Checking $BASH_VERSION is a good method of determining which shell is running. $SHELL does not necessarily give the correct answer.


A colon-separated list of search paths available to the cd command, similar in function to the $PATH variable for binaries. The $CDPATH variable may be set in the local ~/.bashrc file.

bash$ cd bash-doc
bash: cd: bash-doc: No such file or directory

bash$ CDPATH=/usr/share/doc
bash$ cd bash-doc

bash$ echo $PWD


The top value in the directory stack [41] (affected by pushd and popd)

This builtin variable corresponds to the dirs command, however dirs shows the entire contents of the directory stack.


The default editor invoked by a script, usually vi or emacs.


"effective" user ID number

Identification number of whatever identity the current user has assumed, perhaps by means of su.


The $EUID is not necessarily the same as the $UID.


Name of the current function

xyz23 ()
  echo "$FUNCNAME now executing."  # xyz23 now executing.


                                   # Null value outside a function.

See also Example A-50.


A list of filename patterns to be excluded from matching in globbing.


Groups current user belongs to

This is a listing (array) of the group id numbers for current user, as recorded in /etc/passwd and /etc/group.

root# echo $GROUPS

root# echo ${GROUPS[1]}

root# echo ${GROUPS[5]}


Home directory of the user, usually /home/username (see Example 10-7)


The hostname command assigns the system host name at bootup in an init script. However, the gethostname() function sets the Bash internal variable $HOSTNAME. See also Example 10-7.


host type

Like $MACHTYPE, identifies the system hardware.

bash$ echo $HOSTTYPE

internal field separator

This variable determines how Bash recognizes fields, or word boundaries, when it interprets character strings.

$IFS defaults to whitespace (space, tab, and newline), but may be changed, for example, to parse a comma-separated data file. Note that $* uses the first character held in $IFS. See Example 5-1.

bash$ echo "$IFS"

(With $IFS set to default, a blank line displays.)

bash$ echo "$IFS" | cat -vte
(Show whitespace: here a single space, ^I [horizontal tab],
  and newline, and display "$" at end-of-line.)

bash$ bash -c 'set w x y z; IFS=":-;"; echo "$*"'
(Read commands from string and assign any arguments to pos params.)

Set $IFS to eliminate whitespace in pathnames.
IFS="$(printf '\n\t')"   # Per David Wheeler.


$IFS does not handle whitespace the same as it does other characters.

Example 9-1. $IFS and whitespace



# The plus sign will be interpreted as a separator.
echo $var1     # a b c
echo $var2     # d-e-f
echo $var3     # g,h,i


# The plus sign reverts to default interpretation.
# The minus sign will be interpreted as a separator.
echo $var1     # a+b+c
echo $var2     # d e f
echo $var3     # g,h,i


# The comma will be interpreted as a separator.
# The minus sign reverts to default interpretation.
echo $var1     # a+b+c
echo $var2     # d-e-f
echo $var3     # g h i


IFS=" "
# The space character will be interpreted as a separator.
# The comma reverts to default interpretation.
echo $var1     # a+b+c
echo $var2     # d-e-f
echo $var3     # g,h,i

# ======================================================== #

# However ...
# $IFS treats whitespace differently than other characters.

  for arg
    echo "[$arg]"
  done #  ^    ^   Embed within brackets, for your viewing pleasure.

echo; echo "IFS=\" \""
echo "-------"

IFS=" "
var=" a  b c   "
#    ^ ^^   ^^^
output_args_one_per_line $var  # output_args_one_per_line `echo " a  b c   "`
# [a]
# [b]
# [c]

echo; echo "IFS=:"
echo "-----"

var=":a::b:c:::"               # Same pattern as above,
#    ^ ^^   ^^^                #+ but substituting ":" for " "  ...
output_args_one_per_line $var
# []
# [a]
# []
# [b]
# [c]
# []
# []

# Note "empty" brackets.
# The same thing happens with the "FS" field separator in awk.



(Many thanks, St�phane Chazelas, for clarification and above examples.)

See also Example 16-41, Example 11-8, and Example 19-14 for instructive examples of using $IFS.


Ignore EOF: how many end-of-files (control-D) the shell will ignore before logging out.


Often set in the .bashrc or /etc/profile files, this variable controls collation order in filename expansion and pattern matching. If mishandled, LC_COLLATE can cause unexpected results in filename globbing.


As of version 2.05 of Bash, filename globbing no longer distinguishes between lowercase and uppercase letters in a character range between brackets. For example, ls [A-M]* would match both File1.txt and file1.txt. To revert to the customary behavior of bracket matching, set LC_COLLATE to C by an export LC_COLLATE=C in /etc/profile and/or ~/.bashrc.


This internal variable controls character interpretation in globbing and pattern matching.


This variable is the line number of the shell script in which this variable appears. It has significance only within the script in which it appears, and is chiefly useful for debugging purposes.

last_cmd_arg=$_  # Save it.

echo "At line number $LINENO, variable \"v1\" = $v1"
echo "Last command argument processed = $last_cmd_arg"


machine type

Identifies the system hardware.

bash$ echo $MACHTYPE

Old working directory ("OLD-Print-Working-Directory", previous directory you were in).


operating system type

bash$ echo $OSTYPE

Path to binaries, usually /usr/bin/, /usr/X11R6/bin/, /usr/local/bin, etc.

When given a command, the shell automatically does a hash table search on the directories listed in the path for the executable. The path is stored in the environmental variable, $PATH, a list of directories, separated by colons. Normally, the system stores the $PATH definition in /etc/profile and/or ~/.bashrc (see Appendix H).

bash$ echo $PATH

PATH=${PATH}:/opt/bin appends the /opt/bin directory to the current path. In a script, it may be expedient to temporarily add a directory to the path in this way. When the script exits, this restores the original $PATH (a child process, such as a script, may not change the environment of the parent process, the shell).


The current "working directory", ./, is usually omitted from the $PATH as a security measure.


Array variable holding exit status(es) of last executed foreground pipe.

bash$ echo $PIPESTATUS

bash$ ls -al | bogus_command
bash: bogus_command: command not found
bash$ echo ${PIPESTATUS[1]}

bash$ ls -al | bogus_command
bash: bogus_command: command not found
bash$ echo $?

The members of the $PIPESTATUS array hold the exit status of each respective command executed in a pipe. $PIPESTATUS[0] holds the exit status of the first command in the pipe, $PIPESTATUS[1] the exit status of the second command, and so on.


The $PIPESTATUS variable may contain an erroneous 0 value in a login shell (in releases prior to 3.0 of Bash).

tcsh% bash

bash$ who | grep nobody | sort
bash$ echo ${PIPESTATUS[*]}

The above lines contained in a script would produce the expected 0 1 0 output.

Thank you, Wayne Pollock for pointing this out and supplying the above example.


The $PIPESTATUS variable gives unexpected results in some contexts.

bash$ echo $BASH_VERSION

bash$ $ ls | bogus_command | wc
bash: bogus_command: command not found
 0       0       0

bash$ echo ${PIPESTATUS[@]}
141 127 0

Chet Ramey attributes the above output to the behavior of ls. If ls writes to a pipe whose output is not read, then SIGPIPE kills it, and its exit status is 141. Otherwise its exit status is 0, as expected. This likewise is the case for tr.


$PIPESTATUS is a "volatile" variable. It needs to be captured immediately after the pipe in question, before any other command intervenes.

bash$ $ ls | bogus_command | wc
bash: bogus_command: command not found
 0       0       0

bash$ echo ${PIPESTATUS[@]}
0 127 0

bash$ echo ${PIPESTATUS[@]}


The pipefail option may be useful in cases where $PIPESTATUS does not give the desired information.


The $PPID of a process is the process ID (pid) of its parent process. [42]

Compare this with the pidof command.


A variable holding a command to be executed just before the primary prompt, $PS1 is to be displayed.


This is the main prompt, seen at the command-line.


The secondary prompt, seen when additional input is expected. It displays as ">".


The tertiary prompt, displayed in a select loop (see Example 11-30).


The quartenary prompt, shown at the beginning of each line of output when invoking a script with the -x [verbose trace] option. It displays as "+".

As a debugging aid, it may be useful to embed diagnostic information in $PS4.
P4='$(read time junk < /proc/$$/schedstat; echo "@@@ $time @@@ " )'
# Per suggestion by Erik Brandsberg.
set -x
# Various commands follow ...


Working directory (directory you are in at the time)

This is the analog to the pwd builtin command.



clear # Clear the screen.


cd $TargetDirectory
echo "Deleting stale files in $TargetDirectory."

if [ "$PWD" != "$TargetDirectory" ]
then    # Keep from wiping out wrong directory by accident.
  echo "Wrong directory!"
  echo "In $PWD, rather than $TargetDirectory!"
  echo "Bailing out!"

rm -rf *
rm .[A-Za-z0-9]*    # Delete dotfiles.
# rm -f .[^.]* ..?*   to remove filenames beginning with multiple dots.
# (shopt -s dotglob; rm -f *)   will also work.
# Thanks, S.C. for pointing this out.

#  A filename (`basename`) may contain all characters in the 0 - 255 range,
#+ except "/".
#  Deleting files beginning with weird characters, such as -
#+ is left as an exercise. (Hint: rm ./-weirdname or rm -- -weirdname)
result=$?   # Result of delete operations. If successful = 0.

ls -al              # Any files left?
echo "Done."
echo "Old files deleted in $TargetDirectory."

# Various other operations here, as necessary.

exit $result


The default value when a variable is not supplied to read. Also applicable to select menus, but only supplies the item number of the variable chosen, not the value of the variable itself.


# REPLY is the default value for a 'read' command.

echo -n "What is your favorite vegetable? "

echo "Your favorite vegetable is $REPLY."
#  REPLY holds the value of last "read" if and only if
#+ no variable supplied.

echo -n "What is your favorite fruit? "
read fruit
echo "Your favorite fruit is $fruit."
echo "but..."
echo "Value of \$REPLY is still $REPLY."
#  $REPLY is still set to its previous value because
#+ the variable $fruit absorbed the new "read" value.


exit 0


The number of seconds the script has been running.



echo "Hit Control-C to exit before $TIME_LIMIT seconds."

while [ "$SECONDS" -le "$TIME_LIMIT" ]
do   #   $SECONDS is an internal shell variable.
  if [ "$SECONDS" -eq 1 ]

  echo "This script has been running $SECONDS $units."
  #  On a slow or overburdened machine, the script may skip a count
  #+ every once in a while.
  sleep $INTERVAL

echo -e "\a"  # Beep!

exit 0


The list of enabled shell options, a readonly variable.
bash$ echo $SHELLOPTS


Shell level, how deeply Bash is nested. [43] If, at the command-line, $SHLVL is 1, then in a script it will increment to 2.


This variable is not affected by subshells. Use $BASH_SUBSHELL when you need an indication of subshell nesting.


If the $TMOUT environmental variable is set to a non-zero value time, then the shell prompt will time out after $time seconds. This will cause a logout.

As of version 2.05b of Bash, it is now possible to use $TMOUT in a script in combination with read.

# Works in scripts for Bash, versions 2.05b and later.

TMOUT=3    # Prompt times out at three seconds.

echo "What is your favorite song?"
echo "Quickly now, you only have $TMOUT seconds to answer!"
read song

if [ -z "$song" ]
  song="(no answer)"
  # Default response.

echo "Your favorite song is $song."

There are other, more complex, ways of implementing timed input in a script. One alternative is to set up a timing loop to signal the script when it times out. This also requires a signal handling routine to trap (see Example 32-5) the interrupt generated by the timing loop (whew!).

Example 9-2. Timed Input


# TMOUT=3    Also works, as of newer versions of Bash.

TIMELIMIT=3  # Three seconds in this instance.
             # May be set to different value.

  if [ "$answer" = TIMEOUT ]
    echo $answer
  else       # Don't want to mix up the two instances. 
    echo "Your favorite veggie is $answer"
    kill $!  #  Kills no-longer-needed TimerOn function
             #+ running in background.
             #  $! is PID of last job running in background.


  sleep $TIMELIMIT && kill -s 14 $$ &
  # Waits 3 seconds, then sends sigalarm to script.


trap Int14Vector $TIMER_INTERRUPT
# Timer interrupt (14) subverted for our purposes.

echo "What is your favorite vegetable "
read answer

#  Admittedly, this is a kludgy implementation of timed input.
#  However, the "-t" option to "read" simplifies this task.
#  See the "" script.
#  However, what about timing not just single user input,
#+ but an entire script?

#  If you need something really elegant ...
#+ consider writing the application in C or C++,
#+ using appropriate library functions, such as 'alarm' and 'setitimer.'

exit 0

An alternative is using stty.

Example 9-3. Once more, timed input


#  Written by Stephane Chazelas,
#+ and modified by the document author.

INTERVAL=5                # timeout interval

timedout_read() {
  old_tty_settings=`stty -g`
  stty -icanon min 0 time ${timeout}0
  eval read $varname      # or just  read $varname
  stty "$old_tty_settings"
  # See man page for "stty."

echo; echo -n "What's your name? Quick! "
timedout_read $INTERVAL your_name

#  This may not work on every terminal type.
#  The maximum timeout depends on the terminal.
#+ (it is often 25.5 seconds).


if [ ! -z "$your_name" ]  # If name input before timeout ...
  echo "Your name is $your_name."
  echo "Timed out."


# The behavior of this script differs somewhat from ""
# At each keystroke, the counter resets.

exit 0

Perhaps the simplest method is using the -t option to read.

Example 9-4. Timed read

# [time-out]
# Inspired by a suggestion from "syngin seven" (thanks).

TIMELIMIT=4         # 4 seconds

read -t $TIMELIMIT variable <&1
#                           ^^^
#  In this instance, "<&1" is needed for Bash 1.x and 2.x,
#  but unnecessary for Bash 3+.


if [ -z "$variable" ]  # Is null?
  echo "Timed out, variable still unset."
  echo "variable = $variable"

exit 0

User ID number

Current user's user identification number, as recorded in /etc/passwd

This is the current user's real id, even if she has temporarily assumed another identity through su. $UID is a readonly variable, not subject to change from the command line or within a script, and is the counterpart to the id builtin.

Example 9-5. Am I root?

#   Am I root or not?

ROOT_UID=0   # Root has $UID 0.

if [ "$UID" -eq "$ROOT_UID" ]  # Will the real "root" please stand up?
  echo "You are root."
  echo "You are just an ordinary user (but mom loves you just the same)."

exit 0

# ============================================================= #
# Code below will not execute, because the script already exited.

# An alternate method of getting to the root of matters:


username=`id -nu`              # Or...   username=`whoami`
if [ "$username" = "$ROOTUSER_NAME" ]
  echo "Rooty, toot, toot. You are root."
  echo "You are just a regular fella."

See also Example 2-3.


The variables $ENV, $LOGNAME, $MAIL, $TERM, $USER, and $USERNAME are not Bash builtins. These are, however, often set as environmental variables in one of the Bash or login startup files. $SHELL, the name of the user's login shell, may be set from /etc/passwd or in an "init" script, and it is likewise not a Bash builtin.

tcsh% echo $LOGNAME
tcsh% echo $SHELL
tcsh% echo $TERM

bash$ echo $LOGNAME
bash$ echo $SHELL
bash$ echo $TERM

Positional Parameters

$0, $1, $2, etc.

Positional parameters, passed from command line to script, passed to a function, or set to a variable (see Example 4-5 and Example 15-16)


Number of command-line arguments [44] or positional parameters (see Example 36-2)


All of the positional parameters, seen as a single word


"$*" must be quoted.


Same as $*, but each parameter is a quoted string, that is, the parameters are passed on intact, without interpretation or expansion. This means, among other things, that each parameter in the argument list is seen as a separate word.


Of course, "$@" should be quoted.

Example 9-6. arglist: Listing arguments with $* and $@

# Invoke this script with several arguments, such as "one two three" ...


if [ ! -n "$1" ]
  echo "Usage: `basename $0` argument1 argument2 etc."
  exit $E_BADARGS


index=1          # Initialize count.

echo "Listing args with \"\$*\":"
for arg in "$*"  # Doesn't work properly if "$*" isn't quoted.
  echo "Arg #$index = $arg"
  let "index+=1"
done             # $* sees all arguments as single word. 
echo "Entire arg list seen as single word."


index=1          # Reset count.
                 # What happens if you forget to do this?

echo "Listing args with \"\$@\":"
for arg in "$@"
  echo "Arg #$index = $arg"
  let "index+=1"
done             # $@ sees arguments as separate words. 
echo "Arg list seen as separate words."


index=1          # Reset count.

echo "Listing args with \$* (unquoted):"
for arg in $*
  echo "Arg #$index = $arg"
  let "index+=1"
done             # Unquoted $* sees arguments as separate words. 
echo "Arg list seen as separate words."

exit 0

Following a shift, the $@ holds the remaining command-line parameters, lacking the previous $1, which was lost.
# Invoke with ./scriptname 1 2 3 4 5

echo "$@"    # 1 2 3 4 5
echo "$@"    # 2 3 4 5
echo "$@"    # 3 4 5

# Each "shift" loses parameter $1.
# "$@" then contains the remaining parameters.

The $@ special parameter finds use as a tool for filtering input into shell scripts. The cat "$@" construction accepts input to a script either from stdin or from files given as parameters to the script. See Example 16-24 and Example 16-25.


The $* and $@ parameters sometimes display inconsistent and puzzling behavior, depending on the setting of $IFS.

Example 9-7. Inconsistent $* and $@ behavior


#  Erratic behavior of the "$*" and "$@" internal Bash variables,
#+ depending on whether or not they are quoted.
#  Demonstrates inconsistent handling of word splitting and linefeeds.

set -- "First one" "second" "third:one" "" "Fifth: :one"
# Setting the script arguments, $1, $2, $3, etc.


echo 'IFS unchanged, using "$*"'
for i in "$*"               # quoted
do echo "$((c+=1)): [$i]"   # This line remains the same in every instance.
                            # Echo args.
echo ---

echo 'IFS unchanged, using $*'
for i in $*                 # unquoted
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS unchanged, using "$@"'
for i in "$@"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS unchanged, using $@'
for i in $@
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using "$*"'
for i in "$*"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using $*'
for i in $*
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using "$var" (var=$*)'
for i in "$var"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using $var (var=$*)'
for i in $var
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using $var (var="$*")'
for i in $var
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using "$var" (var="$*")'
for i in "$var"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using "$@"'
for i in "$@"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using $@'
for i in $@
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using $var (var=$@)'
for i in $var
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using "$var" (var=$@)'
for i in "$var"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using "$var" (var="$@")'
for i in "$var"
do echo "$((c+=1)): [$i]"
echo ---

echo 'IFS=":", using $var (var="$@")'
for i in $var
do echo "$((c+=1)): [$i]"


# Try this script with ksh or zsh -y.

exit 0

#  This example script written by Stephane Chazelas,
#+ and slightly modified by the document author.


The $@ and $* parameters differ only when between double quotes.

Example 9-8. $* and $@ when $IFS is empty


#  If $IFS set, but empty,
#+ then "$*" and "$@" do not echo positional params as expected.

mecho ()       # Echo positional parameters.
echo "$1,$2,$3";

IFS=""         # Set, but empty.
set a b c      # Positional parameters.

mecho "$*"     # abc,,
#                   ^^
mecho $*       # a,b,c

mecho $@       # a,b,c
mecho "$@"     # a,b,c

#  The behavior of $* and $@ when $IFS is empty depends
#+ on which Bash or sh version being run.
#  It is therefore inadvisable to depend on this "feature" in a script.

# Thanks, Stephane Chazelas.


Other Special Parameters


Flags passed to script (using set). See Example 15-16.


This was originally a ksh construct adopted into Bash, and unfortunately it does not seem to work reliably in Bash scripts. One possible use for it is to have a script self-test whether it is interactive.


PID (process ID) of last job run in background


COMMAND1="sleep 100"

echo "Logging PIDs background commands for script: $0" >> "$LOG"
# So they can be monitored, and killed as necessary.
echo >> "$LOG"

# Logging commands.

echo -n "PID of \"$COMMAND1\":  " >> "$LOG"
echo $! >> "$LOG"
# PID of "sleep 100":  1506

# Thank you, Jacques Lederer, for suggesting this.

Using $! for job control:

possibly_hanging_job & { sleep ${TIMEOUT}; eval 'kill -9 $!' &> /dev/null; }
# Forces completion of an ill-behaved program.
# Useful, for example, in init scripts.

# Thank you, Sylvain Fourmanoit, for this creative use of the "!" variable.

Or, alternately:

# This example by Matthew Sage.
# Used with permission.

TIMEOUT=30   # Timeout value in seconds

possibly_hanging_job & {
        while ((count < TIMEOUT )); do
                eval '[ ! -d "/proc/$!" ] && ((count = TIMEOUT))'
                # /proc is where information about running processes is found.
                # "-d" tests whether it exists (whether directory exists).
                # So, we're waiting for the job in question to show up.
                sleep 1
        eval '[ -d "/proc/$!" ] && kill -15 $!'
        # If the hanging job is running, kill it.

#  -------------------------------------------------------------- #

#  However, this may not not work as specified if another process
#+ begins to run after the "hanging_job" . . .
#  In such a case, the wrong job may be killed.
#  Ariel Meragelman suggests the following fix.

# Timeout value in seconds
possibly_hanging_job & {

while ((count < TIMEOUT )); do
  eval '[ ! -d "/proc/$lastjob" ] && ((count = TIMEOUT))'
  sleep 1
eval '[ -d "/proc/$lastjob" ] && kill -15 $lastjob'




Special variable set to final argument of previous command executed.

Example 9-9. Underscore variable


echo $_              #  /bin/bash
                     #  Just called /bin/bash to run the script.
                     #  Note that this will vary according to
                     #+ how the script is invoked.

du >/dev/null        #  So no output from command.
echo $_              #  du

ls -al >/dev/null    #  So no output from command.
echo $_              #  -al  (last argument)

echo $_              #  :

Exit status of a command, function, or the script itself (see Example 24-7)


Process ID (PID) of the script itself. [45] The $$ variable often finds use in scripts to construct "unique" temp file names (see Example 32-6, Example 16-31, and Example 15-27). This is usually simpler than invoking mktemp.

9.2. Typing variables: declare or typeset

The declare or typeset builtins, which are exact synonyms, permit modifying the properties of variables. This is a very weak form of the typing [46] available in certain programming languages. The declare command is specific to version 2 or later of Bash. The typeset command also works in ksh scripts.

declare/typeset options

-r readonly

(declare -r var1 works the same as readonly var1)

This is the rough equivalent of the C const type qualifier. An attempt to change the value of a readonly variable fails with an error message.

declare -r var1=1
echo "var1 = $var1"   # var1 = 1

(( var1++ ))          # line 4: var1: readonly variable

-i integer

declare -i number
# The script will treat subsequent occurrences of "number" as an integer.		

echo "Number = $number"     # Number = 3

echo "Number = $number"     # Number = 0
# Tries to evaluate the string "three" as an integer.

Certain arithmetic operations are permitted for declared integer variables without the need for expr or let.

echo "n = $n"       # n = 6/3

declare -i n
echo "n = $n"       # n = 2

-a array

declare -a indices

The variable indices will be treated as an array.

-f function(s)

declare -f

A declare -f line with no arguments in a script causes a listing of all the functions previously defined in that script.

declare -f function_name

A declare -f function_name in a script lists just the function named.

-x export

declare -x var3

This declares a variable as available for exporting outside the environment of the script itself.

-x var=$value

declare -x var3=373

The declare command permits assigning a value to a variable in the same statement as setting its properties.

Example 9-10. Using declare to type variables


func1 ()
  echo This is a function.

declare -f        # Lists the function above.


declare -i var1   # var1 is an integer.
echo "var1 declared as $var1"
var1=var1+1       # Integer declaration eliminates the need for 'let'.
echo "var1 incremented by 1 is $var1."
# Attempt to change variable declared as integer.
echo "Attempting to change var1 to floating point value, 2367.1."
var1=2367.1       # Results in error message, with no change to variable.
echo "var1 is still $var1"


declare -r var2=13.36         # 'declare' permits setting a variable property
                              #+ and simultaneously assigning it a value.
echo "var2 declared as $var2" # Attempt to change readonly variable.
var2=13.37                    # Generates error message, and exit from script.

echo "var2 is still $var2"    # This line will not execute.

exit 0                        # Script will not exit here.


Using the declare builtin restricts the scope of a variable.
foo ()

bar ()
echo $FOO

bar   # Prints bar.

However . . .
foo (){
declare FOO="bar"

bar ()
echo $FOO

bar  # Prints nothing.

# Thank you, Michael Iatrou, for pointing this out.

9.2.1. Another use for declare

The declare command can be helpful in identifying variables, environmental or otherwise. This can be especially useful with arrays.

bash$ declare | grep HOME

bash$ zzy=68
bash$ declare | grep zzy

bash$ Colors=([0]="purple" [1]="reddish-orange" [2]="light green")
bash$ echo ${Colors[@]}
purple reddish-orange light green
bash$ declare | grep Colors
Colors=([0]="purple" [1]="reddish-orange" [2]="light green")

9.3. $RANDOM: generate random integer


Anyone who attempts to generate random numbers by deterministic means is, of course, living in a state of sin.

--John von Neumann

$RANDOM is an internal Bash function (not a constant) that returns a pseudorandom [47] integer in the range 0 - 32767. It should not be used to generate an encryption key.

Example 9-11. Generating random numbers


# $RANDOM returns a different random integer at each invocation.
# Nominal range: 0 - 32767 (signed 16-bit integer).


echo "$MAXCOUNT random numbers:"
echo "-----------------"
while [ "$count" -le $MAXCOUNT ]      # Generate 10 ($MAXCOUNT) random integers.
  echo $number
  let "count += 1"  # Increment count.
echo "-----------------"

# If you need a random int within a certain range, use the 'modulo' operator.
# This returns the remainder of a division operation.



let "number %= $RANGE"
#           ^^
echo "Random number less than $RANGE  ---  $number"


#  If you need a random integer greater than a lower bound,
#+ then set up a test to discard all numbers below that.


number=0   #initialize
while [ "$number" -le $FLOOR ]
echo "Random number greater than $FLOOR ---  $number"

   # Let's examine a simple alternative to the above loop, namely
   #       let "number = $RANDOM + $FLOOR"
   # That would eliminate the while-loop and run faster.
   # But, there might be a problem with that. What is it?

# Combine above two techniques to retrieve random number between two limits.
number=0   #initialize
while [ "$number" -le $FLOOR ]
  let "number %= $RANGE"  # Scales $number down within $RANGE.
echo "Random number between $FLOOR and $RANGE ---  $number"

# Generate binary choice, that is, "true" or "false" value.

let "number %= $BINARY"
#  Note that    let "number >>= 14"    gives a better random distribution
#+ (right shifts out everything except last binary digit).
if [ "$number" -eq $T ]
  echo "TRUE"
  echo "FALSE"


# Generate a toss of the dice.
SPOTS=6   # Modulo 6 gives range 0 - 5.
          # Incrementing by 1 gives desired range of 1 - 6.
          # Thanks, Paulo Marcel Coelho Aragao, for the simplification.
# Would it be better to just set SPOTS=7 and not add 1? Why or why not?

# Tosses each die separately, and so gives correct odds.

    let "die1 = $RANDOM % $SPOTS +1" # Roll first one.
    let "die2 = $RANDOM % $SPOTS +1" # Roll second one.
    #  Which arithmetic operation, above, has greater precedence --
    #+ modulo (%) or addition (+)?

let "throw = $die1 + $die2"
echo "Throw of the dice = $throw"

exit 0

Example 9-12. Picking a random card from a deck


# This is an example of choosing random elements of an array.

# Pick a card, any card.



# Note variables spread over multiple lines.

suite=($Suites)                # Read into array variable.

num_suites=${#suite[*]}        # Count how many elements.

echo -n "${denomination[$((RANDOM%num_denominations))]} of "
echo ${suite[$((RANDOM%num_suites))]}

# $bozo sh
# Jack of Clubs

# Thank you, "jipe," for pointing out this use of $RANDOM.
exit 0

Example 9-13. Brownian Motion Simulation

# Author: Mendel Cooper
# Reldate: 10/26/07
# License: GPL3

#  ----------------------------------------------------------------
#  This script models Brownian motion:
#+ the random wanderings of tiny particles in a fluid,
#+ as they are buffeted by random currents and collisions.
#+ This is colloquially known as the "Drunkard's Walk."

#  It can also be considered as a stripped-down simulation of a
#+ Galton Board, a slanted board with a pattern of pegs,
#+ down which rolls a succession of marbles, one at a time.
#+ At the bottom is a row of slots or catch basins in which
#+ the marbles come to rest at the end of their journey.
#  Think of it as a kind of bare-bones Pachinko game.
#  As you see by running the script,
#+ most of the marbles cluster around the center slot.
#+ This is consistent with the expected binomial distribution.
#  As a Galton Board simulation, the script
#+ disregards such parameters as
#+ board tilt-angle, rolling friction of the marbles,
#+ angles of impact, and elasticity of the pegs.
#  To what extent does this affect the accuracy of the simulation?
#  ----------------------------------------------------------------

PASSES=500            #  Number of particle interactions / marbles.
ROWS=10               #  Number of "collisions" (or horiz. peg rows).
RANGE=3               #  0 - 2 output range from $RANDOM.
POS=0                 #  Left/right position.
RANDOM=$$             #  Seeds the random number generator from PID
                      #+ of script.

declare -a Slots      # Array holding cumulative results of passes.
NUMSLOTS=21           # Number of slots at bottom of board.

Initialize_Slots () { # Zero out all elements of the array.
for i in $( seq $NUMSLOTS )

echo                  # Blank line at beginning of run.

Show_Slots () {
echo; echo
echo -n " "
for i in $( seq $NUMSLOTS )   # Pretty-print array elements.
  printf "%3d" ${Slots[$i]}   # Allot three spaces per result.

echo # Row of slots:
echo " |__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|__|"
echo "                                ||"
echo #  Note that if the count within any particular slot exceeds 99,
     #+ it messes up the display.
     #  Running only(!) 500 passes usually avoids this.

Move () {              # Move one unit right / left, or stay put.
  Move=$RANDOM         # How random is $RANDOM? Well, let's see ...
  let "Move %= RANGE"  # Normalize into range of 0 - 2.
  case "$Move" in
    0 ) ;;                   # Do nothing, i.e., stay in place.
    1 ) ((POS--));;          # Left.
    2 ) ((POS++));;          # Right.
    * ) echo -n "Error ";;   # Anomaly! (Should never occur.)

Play () {                    # Single pass (inner loop).
while [ "$i" -lt "$ROWS" ]   # One event per row.

SHIFT=11                     # Why 11, and not 10?
let "POS += $SHIFT"          # Shift "zero position" to center.
(( Slots[$POS]++ ))          # DEBUG: echo $POS

# echo -n "$POS "


Run () {                     # Outer loop.
while [ "$p" -lt "$PASSES" ]
  (( p++ ))
  POS=0                      # Reset to zero. Why?

# --------------
# main ()
# --------------

exit $?

#  Exercises:
#  ---------
#  1) Show the results in a vertical bar graph, or as an alternative,
#+    a scattergram.
#  2) Alter the script to use /dev/urandom instead of $RANDOM.
#     Will this make the results more random?
#  3) Provide some sort of "animation" or graphic output
#     for each marble played.

Jipe points out a set of techniques for generating random numbers within a range.
#  Generate random number between 6 and 30.

#  Generate random number in the same 6 - 30 range,
#+ but the number must be evenly divisible by 3.

#  Note that this will not work all the time.
#  It fails if $RANDOM%30 returns 0.

#  Frank Wang suggests the following alternative:
   rnumber=$(( RANDOM%27/3*3+6 ))

Bill Gradwohl came up with an improved formula that works for positive numbers.

Here Bill presents a versatile function that returns a random number between two specified values.

Example 9-14. Random between values

# Random number between two specified values. 
# Script by Bill Gradwohl, with minor modifications by the document author.
# Corrections in lines 187 and 189 by Anthony Le Clezio.
# Used with permission.

randomBetween() {
   #  Generates a positive or negative random number
   #+ between $min and $max
   #+ and divisible by $divisibleBy.
   #  Gives a "reasonably random" distribution of return values.
   #  Bill Gradwohl - Oct 1, 2003

   syntax() {
   # Function embedded within function.
      echo    "Syntax: randomBetween [min] [max] [multiple]"
      echo -n "Expects up to 3 passed parameters, "
      echo    "but all are completely optional."
      echo    "min is the minimum value"
      echo    "max is the maximum value"
      echo -n "multiple specifies that the answer must be "
      echo     "a multiple of this value."
      echo    "    i.e. answer must be evenly divisible by this number."
      echo    "If any value is missing, defaults area supplied as: 0 32767 1"
      echo -n "Successful completion returns 0, "
      echo     "unsuccessful completion returns"
      echo    "function syntax and 1."
      echo -n "The answer is returned in the global variable "
      echo    "randomBetweenAnswer"
      echo -n "Negative values for any passed parameter are "
      echo    "handled correctly."

   local min=${1:-0}
   local max=${2:-32767}
   local divisibleBy=${3:-1}
   # Default values assigned, in case parameters not passed to function.

   local x
   local spread

   # Let's make sure the divisibleBy value is positive.
   [ ${divisibleBy} -lt 0 ] && divisibleBy=$((0-divisibleBy))

   # Sanity check.
   if [ $# -gt 3 -o ${divisibleBy} -eq 0 -o  ${min} -eq ${max} ]; then 
      return 1

   # See if the min and max are reversed.
   if [ ${min} -gt ${max} ]; then
      # Swap them.

   #  If min is itself not evenly divisible by $divisibleBy,
   #+ then fix the min to be within range.
   if [ $((min/divisibleBy*divisibleBy)) -ne ${min} ]; then 
      if [ ${min} -lt 0 ]; then

   #  If max is itself not evenly divisible by $divisibleBy,
   #+ then fix the max to be within range.
   if [ $((max/divisibleBy*divisibleBy)) -ne ${max} ]; then 
      if [ ${max} -lt 0 ]; then

   #  ---------------------------------------------------------------------
   #  Now, to do the real work.

   #  Note that to get a proper distribution for the end points,
   #+ the range of random values has to be allowed to go between
   #+ 0 and abs(max-min)+divisibleBy, not just abs(max-min)+1.

   #  The slight increase will produce the proper distribution for the
   #+ end points.

   #  Changing the formula to use abs(max-min)+1 will still produce
   #+ correct answers, but the randomness of those answers is faulty in
   #+ that the number of times the end points ($min and $max) are returned
   #+ is considerably lower than when the correct formula is used.
   #  ---------------------------------------------------------------------

   #  Omair Eshkenazi points out that this test is unnecessary,
   #+ since max and min have already been switched around.
   [ ${spread} -lt 0 ] && spread=$((0-spread))
   let spread+=divisibleBy

   return 0

   #  However, Paulo Marcel Coelho Aragao points out that
   #+ when $max and $min are not divisible by $divisibleBy,
   #+ the formula fails.
   #  He suggests instead the following formula:
   #    rnumber = $(((RANDOM%(max-min+1)+min)/divisibleBy*divisibleBy))


# Let's test the function.

#  Generate an array of expected answers and check to make sure we get
#+ at least one of each answer if we loop long enough.

declare -a answer
   if [ $((minimum/divisibleBy*divisibleBy)) -ne ${minimum} ]; then 
      if [ ${minimum} -lt 0 ]; then

   #  If max is itself not evenly divisible by $divisibleBy,
   #+ then fix the max to be within range.

   if [ $((maximum/divisibleBy*divisibleBy)) -ne ${maximum} ]; then 
      if [ ${maximum} -lt 0 ]; then

#  We need to generate only positive array subscripts,
#+ so we need a displacement that that will guarantee
#+ positive results.

for ((i=${minimum}; i<=${maximum}; i+=divisibleBy)); do

# Now loop a large number of times to see what we get.
loopIt=1000   #  The script author suggests 100000,
              #+ but that takes a good long while.

for ((i=0; i<${loopIt}; ++i)); do

   #  Note that we are specifying min and max in reversed order here to
   #+ make the function correct for this case.

   randomBetween ${max} ${min} ${divisibleBy}

   # Report an error if an answer is unexpected.
   [ ${randomBetweenAnswer} -lt ${min} -o ${randomBetweenAnswer} -gt ${max} ] \
   && echo MIN or MAX error - ${randomBetweenAnswer}!
   [ $((randomBetweenAnswer%${divisibleBy})) -ne 0 ] \
   && echo DIVISIBLE BY error - ${randomBetweenAnswer}!

   # Store the answer away statistically.

# Let's check the results

for ((i=${minimum}; i<=${maximum}; i+=divisibleBy)); do
   [ ${answer[i+disp]} -eq 0 ] \
   && echo "We never got an answer of $i." \
   || echo "${i} occurred ${answer[i+disp]} times."

exit 0

Just how random is $RANDOM? The best way to test this is to write a script that tracks the distribution of "random" numbers generated by $RANDOM. Let's roll a $RANDOM die a few times . . .

Example 9-15. Rolling a single die with RANDOM

# How random is RANDOM?

RANDOM=$$       # Reseed the random number generator using script process ID.

PIPS=6          # A die has 6 pips.
MAXTHROWS=600   # Increase this if you have nothing better to do with your time.
throw=0         # Number of times the dice have been cast.

ones=0          #  Must initialize counts to zero,
twos=0          #+ since an uninitialized variable is null, NOT zero.

print_result ()
echo "ones =   $ones"
echo "twos =   $twos"
echo "threes = $threes"
echo "fours =  $fours"
echo "fives =  $fives"
echo "sixes =  $sixes"

case "$1" in
  0) ((ones++));;   # Since a die has no "zero", this corresponds to 1.
  1) ((twos++));;   # And this to 2.
  2) ((threes++));; # And so forth.
  3) ((fours++));;
  4) ((fives++));;
  5) ((sixes++));;


while [ "$throw" -lt "$MAXTHROWS" ]
  let "die1 = RANDOM % $PIPS"
  update_count $die1
  let "throw += 1"


exit $?

#  The scores should distribute evenly, assuming RANDOM is random.
#  With $MAXTHROWS at 600, all should cluster around 100,
#+ plus-or-minus 20 or so.
#  Keep in mind that RANDOM is a ***pseudorandom*** generator,
#+ and not a spectacularly good one at that.

#  Randomness is a deep and complex subject.
#  Sufficiently long "random" sequences may exhibit
#+ chaotic and other "non-random" behavior.

# Exercise (easy):
# ---------------
# Rewrite this script to flip a coin 1000 times.
# Choices are "HEADS" and "TAILS."

As we have seen in the last example, it is best to reseed the RANDOM generator each time it is invoked. Using the same seed for RANDOM repeats the same series of numbers. [48] (This mirrors the behavior of the random() function in C.)

Example 9-16. Reseeding RANDOM

# Seeding the RANDOM variable.
# v 1.1, reldate 09 Feb 2013

MAXCOUNT=25       # How many numbers to generate.

random_numbers ()
local count=0
local number

while [ "$count" -lt "$MAXCOUNT" ]
  echo -n "$number "
  let "count++"

echo; echo

RANDOM=$SEED      # Setting RANDOM seeds the random number generator.
echo "Random seed = $SEED"

RANDOM=$SEED      # Same seed for RANDOM . . .
echo; echo "Again, with same random seed ..."
echo "Random seed = $SEED"
random_numbers    # . . . reproduces the exact same number series.
                  # When is it useful to duplicate a "random" series?

echo; echo

RANDOM=$SEED      # Trying again, but with a different seed . . .
echo "Random seed = $SEED"
random_numbers    # . . . gives a different number series.

echo; echo

# RANDOM=$$  seeds RANDOM from process id of script.
# It is also possible to seed RANDOM from 'time' or 'date' commands.

# Getting fancy...
SEED=$(head -1 /dev/urandom | od -N 1 | awk '{ print $2 }'| sed s/^0*//)
#  Pseudo-random output fetched
#+ from /dev/urandom (system pseudo-random device-file),
#+ then converted to line of printable (octal) numbers by "od",
#+ then "awk" retrieves just one number for SEED,
#+ finally "sed" removes any leading zeros.
echo "Random seed = $SEED"

echo; echo

exit 0


The /dev/urandom pseudo-device file provides a method of generating much more "random" pseudorandom numbers than the $RANDOM variable. dd if=/dev/urandom of=targetfile bs=1 count=XX creates a file of well-scattered pseudorandom numbers. However, assigning these numbers to a variable in a script requires a workaround, such as filtering through od (as in above example, Example 16-14, and Example A-36), or even piping to md5sum (see Example 36-16).

There are also other ways to generate pseudorandom numbers in a script. Awk provides a convenient means of doing this.

Example 9-17. Pseudorandom numbers, using awk

# Returns a pseudorandom number in the range 0 - 1,
#+ to 6 decimal places. For example: 0.822725
#  Uses the awk rand() function.

AWKSCRIPT=' { srand(); print rand() } '
#           Command(s)/parameters passed to awk
# Note that srand() reseeds awk's random number generator.

echo -n "Random number between 0 and 1 = "

echo | awk "$AWKSCRIPT"
# What happens if you leave out the 'echo'?

exit 0

# Exercises:
# ---------

# 1) Using a loop construct, print out 10 different random numbers.
#      (Hint: you must reseed the srand() function with a different seed
#+     in each pass through the loop. What happens if you omit this?)

# 2) Using an integer multiplier as a scaling factor, generate random numbers 
#+   in the range of 10 to 100.

# 3) Same as exercise #2, above, but generate random integers this time.

The date command also lends itself to generating pseudorandom integer sequences.

Chapter 10. Manipulating Variables

10.1. Manipulating Strings

Bash supports a surprising number of string manipulation operations. Unfortunately, these tools lack a unified focus. Some are a subset of parameter substitution, and others fall under the functionality of the UNIX expr command. This results in inconsistent command syntax and overlap of functionality, not to mention confusion.

String Length


expr length $string

These are the equivalent of strlen() in C.

expr "$string" : '.*'


echo ${#stringZ}                 # 15
echo `expr length $stringZ`      # 15
echo `expr "$stringZ" : '.*'`    # 15

Example 10-1. Inserting a blank line between paragraphs in a text file

# Ver. 2.1, Reldate 29Jul12 [fixup]

# Inserts a blank line between paragraphs of a single-spaced text file.
# Usage: $0 <FILENAME

MINLEN=60        # Change this value? It's a judgment call.
#  Assume lines shorter than $MINLEN characters ending in a period
#+ terminate a paragraph. See exercises below.

while read line  # For as many lines as the input file has ...
  echo "$line"   # Output the line itself.

  if [[ "$len" -lt "$MINLEN" && "$line" =~ [*{\.}]$ ]]
# if [[ "$len" -lt "$MINLEN" && "$line" =~ \[*\.\] ]]
# An update to Bash broke the previous version of this script. Ouch!
# Thank you, Halim Srama, for pointing this out and suggesting a fix.
    then echo    #  Add a blank line immediately
  fi             #+ after a short line terminated by a period.


# Exercises:
# ---------
#  1) The script usually inserts a blank line at the end
#+    of the target file. Fix this.
#  2) Line 17 only considers periods as sentence terminators.
#     Modify this to include other common end-of-sentence characters,
#+    such as ?, !, and ".

Length of Matching Substring at Beginning of String

expr match "$string" '$substring'

$substring is a regular expression.

expr "$string" : '$substring'

$substring is a regular expression.

#       |------|
#       12345678

echo `expr match "$stringZ" 'abc[A-Z]*.2'`   # 8
echo `expr "$stringZ" : 'abc[A-Z]*.2'`       # 8


expr index $string $substring

Numerical position in $string of first character in $substring that matches.

#       123456 ...
echo `expr index "$stringZ" C12`             # 6
                                             # C position.

echo `expr index "$stringZ" 1c`              # 3
# 'c' (in #3 position) matches before '1'.

This is the near equivalent of strchr() in C.

Substring Extraction


Extracts substring from $string at $position.

If the $string parameter is "*" or "@", then this extracts the positional parameters, [49] starting at $position.


Extracts $length characters of substring from $string at $position.

#       0123456789.....
#       0-based indexing.

echo ${stringZ:0}                            # abcABC123ABCabc
echo ${stringZ:1}                            # bcABC123ABCabc
echo ${stringZ:7}                            # 23ABCabc

echo ${stringZ:7:3}                          # 23A
                                             # Three characters of substring.

# Is it possible to index from the right end of the string?
echo ${stringZ:-4}                           # abcABC123ABCabc
# Defaults to full string, as in ${parameter:-default}.
# However . . .

echo ${stringZ:(-4)}                         # Cabc 
echo ${stringZ: -4}                          # Cabc
# Now, it works.
# Parentheses or added space "escape" the position parameter.

# Thank you, Dan Jacobson, for pointing this out.

The position and length arguments can be "parameterized," that is, represented as a variable, rather than as a numerical constant.

Example 10-2. Generating an 8-character "random" string

# Generating an 8-character "random" string.

if [ -n "$1" ]  #  If command-line argument present,
then            #+ then set start-string to it.
else            #  Else use PID of script as start-string.

POS=2  # Starting from position 2 in the string.
LEN=8  # Extract eight characters.

str1=$( echo "$str0" | md5sum | md5sum )
#  Doubly scramble     ^^^^^^   ^^^^^^
#+ by piping and repiping to md5sum.

# Can parameterize ^^^^ ^^^^

echo "$randstring"

exit $?

# bozo$ ./ my-password
# 1bdd88c4

#  No, this is is not recommended
#+ as a method of generating hack-proof passwords.

If the $string parameter is "*" or "@", then this extracts a maximum of $length positional parameters, starting at $position.

echo ${*:2}          # Echoes second and following positional parameters.
echo ${@:2}          # Same as above.

echo ${*:2:3}        # Echoes three positional parameters, starting at second.

expr substr $string $position $length

Extracts $length characters from $string starting at $position.

#       123456789......
#       1-based indexing.

echo `expr substr $stringZ 1 2`              # ab
echo `expr substr $stringZ 4 3`              # ABC

expr match "$string" '\($substring\)'

Extracts $substring at beginning of $string, where $substring is a regular expression.

expr "$string" : '\($substring\)'

Extracts $substring at beginning of $string, where $substring is a regular expression.

#       =======	    

echo `expr match "$stringZ" '\(.[b-c]*[A-Z]..[0-9]\)'`   # abcABC1
echo `expr "$stringZ" : '\(.[b-c]*[A-Z]..[0-9]\)'`       # abcABC1
echo `expr "$stringZ" : '\(.......\)'`                   # abcABC1
# All of the above forms give an identical result.

expr match "$string" '.*\($substring\)'

Extracts $substring at end of $string, where $substring is a regular expression.

expr "$string" : '.*\($substring\)'

Extracts $substring at end of $string, where $substring is a regular expression.

#                ======

echo `expr match "$stringZ" '.*\([A-C][A-C][A-C][a-c]*\)'`    # ABCabc
echo `expr "$stringZ" : '.*\(......\)'`                       # ABCabc

Substring Removal


Deletes shortest match of $substring from front of $string.


Deletes longest match of $substring from front of $string.

#       |----|          shortest
#       |----------|    longest

echo ${stringZ#a*C}      # 123ABCabc
# Strip out shortest match between 'a' and 'C'.

echo ${stringZ##a*C}     # abc
# Strip out longest match between 'a' and 'C'.

# You can parameterize the substrings.


echo ${stringZ#$X}      # 123ABCabc
echo ${stringZ##$X}     # abc
                        # As above.


Deletes shortest match of $substring from back of $string.

For example:
# Rename all filenames in $PWD with "TXT" suffix to a "txt" suffix.
# For example, "file1.TXT" becomes "file1.txt" . . .


for i in $(ls *.$SUFF)
  mv -f $i ${i%.$SUFF}.$suff
  #  Leave unchanged everything *except* the shortest pattern match
  #+ starting from the right-hand-side of the variable $i . . .
done ### This could be condensed into a "one-liner" if desired.

# Thank you, Rory Winston.


Deletes longest match of $substring from back of $string.

#                    ||     shortest
#        |------------|     longest

echo ${stringZ%b*c}      # abcABC123ABCa
# Strip out shortest match between 'b' and 'c', from back of $stringZ.

echo ${stringZ%%b*c}     # a
# Strip out longest match between 'b' and 'c', from back of $stringZ.

This operator is useful for generating filenames.

Example 10-3. Converting graphic file formats, with filename change

#  Converts all the MacPaint image files in a directory to "pbm" format.

#  Uses the "macptopbm" binary from the "netpbm" package,
#+ which is maintained by Brian Henderson ([email protected]).
#  Netpbm is a standard part of most Linux distros.

SUFFIX=pbm          # New filename suffix. 

if [ -n "$1" ]
  directory=$1      # If directory name given as a script argument...
  directory=$PWD    # Otherwise use current working directory.
#  Assumes all files in the target directory are MacPaint image files,
#+ with a ".mac" filename suffix.

for file in $directory/*    # Filename globbing.
  filename=${file%.*c}      #  Strip ".mac" suffix off filename
                            #+ ('.*c' matches everything
			    #+ between '.' and 'c', inclusive).
  $OPERATION $file > "$filename.$SUFFIX"
                            # Redirect conversion to new filename.
  rm -f $file               # Delete original files after converting.   
  echo "$filename.$SUFFIX"  # Log what is happening to stdout.

exit 0

# Exercise:
# --------
#  As it stands, this script converts *all* the files in the current
#+ working directory.
#  Modify it to work *only* on files with a ".mac" suffix.

# *** And here's another way to do it. *** #

# Batch convert into different graphic formats.
# Assumes imagemagick installed (standard in most Linux distros).

INFMT=png   # Can be tif, jpg, gif, etc.
OUTFMT=pdf  # Can be tif, jpg, gif, pdf, etc.

for pic in *"$INFMT"
  p2=$(ls "$pic" | sed -e s/\.$INFMT//)
  # echo $p2
    convert "$pic" $p2.$OUTFMT

exit $?

Example 10-4. Converting streaming audio files to ogg

# Convert streaming audio files (*.ra) to ogg.

# Uses the "mplayer" media player program:
# Uses the "ogg" library and "oggenc":
# This script may need appropriate codecs installed, such as ...
# Possibly also the compat-libstdc++ package.

OFILEPREF=${1%%ra}      # Strip off the "ra" suffix.
OFILESUFF=wav           # Suffix for wav file.

if [ -z "$1" ]          # Must specify a filename to convert.
  echo "Usage: `basename $0` [filename]"
  exit $E_NOARGS

mplayer "$1" -ao pcm:file=$OUTFILE
oggenc "$OUTFILE"  # Correct file extension automatically added by oggenc.

rm "$OUTFILE"      # Delete intermediate *.wav file.
                   # If you want to keep it, comment out above line.

exit $?

#  Note:
#  ----
#  On a Website, simply clicking on a *.ram streaming audio file
#+ usually only downloads the URL of the actual *.ra audio file.
#  You can then use "wget" or something similar
#+ to download the *.ra file itself.

#  Exercises:
#  ---------
#  As is, this script converts only *.ra filenames.
#  Add flexibility by permitting use of *.ram and other filenames.
#  If you're really ambitious, expand the script
#+ to do automatic downloads and conversions of streaming audio files.
#  Given a URL, batch download streaming audio files (using "wget")
#+ and convert them on the fly.

A simple emulation of getopt using substring-extraction constructs.

Example 10-5. Emulating getopt

# Author: Chris Morgan
# Used in the ABS Guide with permission.

    echo "getopt_simple()"
    echo "Parameters are '$*'"
    until [ -z "$1" ]
      echo "Processing parameter of: '$1'"
      if [ ${1:0:1} = '/' ]
          tmp=${1:1}               # Strip off leading '/' . . .
          parameter=${tmp%%=*}     # Extract name.
          value=${tmp##*=}         # Extract value.
          echo "Parameter: '$parameter', value: '$value'"
          eval $parameter=$value

# Pass all options to getopt_simple().
getopt_simple $*

echo "test is '$test'"
echo "test2 is '$test2'"

exit 0  # See also,, a modified version of this script.


sh /test=value1 /test2=value2

Parameters are '/test=value1 /test2=value2'
Processing parameter of: '/test=value1'
Parameter: 'test', value: 'value1'
Processing parameter of: '/test2=value2'
Parameter: 'test2', value: 'value2'
test is 'value1'
test2 is 'value2'

Substring Replacement


Replace first match of $substring with $replacement. [50]


Replace all matches of $substring with $replacement.


echo ${stringZ/abc/xyz}       # xyzABC123ABCabc
                              # Replaces first match of 'abc' with 'xyz'.

echo ${stringZ//abc/xyz}      # xyzABC123ABCxyz
                              # Replaces all matches of 'abc' with # 'xyz'.

echo  ---------------
echo "$stringZ"               # abcABC123ABCabc
echo  ---------------
                              # The string itself is not altered!

# Can the match and replacement strings be parameterized?
echo ${stringZ/$match/$repl}  # 000ABC123ABCabc
#              ^      ^         ^^^
echo ${stringZ//$match/$repl} # 000ABC123ABC000
# Yes!          ^      ^        ^^^         ^^^


# What happens if no $replacement string is supplied?
echo ${stringZ/abc}           # ABC123ABCabc
echo ${stringZ//abc}          # ABC123ABC
# A simple deletion takes place.


If $substring matches front end of $string, substitute $replacement for $substring.


If $substring matches back end of $string, substitute $replacement for $substring.


echo ${stringZ/#abc/XYZ}          # XYZABC123ABCabc
                                  # Replaces front-end match of 'abc' with 'XYZ'.

echo ${stringZ/%abc/XYZ}          # abcABC123ABCXYZ
                                  # Replaces back-end match of 'abc' with 'XYZ'.

10.1.1. Manipulating strings using awk

A Bash script may invoke the string manipulation facilities of awk as an alternative to using its built-in operations.

Example 10-6. Alternate ways of extracting and locating substrings


#      012345678    Bash
#      123456789    awk
# Note different string indexing system:
# Bash numbers first character of string as 0.
# Awk  numbers first character of string as 1.

echo ${String:2:4} # position 3 (0-1-2), 4 characters long
                                         # skid

# The awk equivalent of ${string:pos:length} is substr(string,pos,length).
echo | awk '
{ print substr("'"${String}"'",3,4)      # skid
#  Piping an empty "echo" to awk gives it dummy input,
#+ and thus makes it unnecessary to supply a filename.

echo "----"

# And likewise:

echo | awk '
{ print index("'"${String}"'", "skid")      # 3
}                                           # (skid starts at position 3)
'   # The awk equivalent of "expr index" ...

exit 0

10.1.2. Further Reference

For more on string manipulation in scripts, refer to Section 10.2 and the relevant section of the expr command listing.

Script examples:

  1. Example 16-9

  2. Example 10-9

  3. Example 10-10

  4. Example 10-11

  5. Example 10-13

  6. Example A-36

  7. Example A-41

10.2. Parameter Substitution

Manipulating and/or expanding variables


Same as $parameter, i.e., value of the variable parameter. In certain contexts, only the less ambiguous ${parameter} form works.

May be used for concatenating variables with strings.

echo "$your_id"
echo "Old \$PATH = $PATH"
PATH=${PATH}:/opt/bin  # Add /opt/bin to $PATH for duration of script.
echo "New \$PATH = $PATH"

${parameter-default}, ${parameter:-default}

If parameter not set, use default.

# var3 is unset.

echo ${var1-$var2}   # 1
echo ${var3-$var2}   # 2
#           ^          Note the $ prefix.

echo ${username-`whoami`}
# Echoes the result of `whoami`, if variable $username is still unset.


${parameter-default} and ${parameter:-default} are almost equivalent. The extra : makes a difference only when parameter has been declared, but is null.


#  Whether a variable has been declared
#+ affects triggering of the default option
#+ even if the variable is null.

echo "username0 has been declared, but is set to null."
echo "username0 = ${username0-`whoami`}"
# Will not echo.


echo username1 has not been declared.
echo "username1 = ${username1-`whoami`}"
# Will echo.

echo "username2 has been declared, but is set to null."
echo "username2 = ${username2:-`whoami`}"
#                            ^
# Will echo because of :- rather than just - in condition test.
# Compare to first instance, above.


# Once again:

# variable has been declared, but is set to null.

echo "${variable-0}"    # (no output)
echo "${variable:-1}"   # 1
#               ^

unset variable

echo "${variable-2}"    # 2
echo "${variable:-3}"   # 3

exit 0

The default parameter construct finds use in providing "missing" command-line arguments in scripts.
#  If not otherwise specified, the following command block operates
#+ on the file "".
#  Begin-Command-Block
#  ...
#  ...
#  ...
#  End-Command-Block

#  From "hanoi2.bash" example:
DISKS=${1:-E_NOPARAM}   # Must specify how many disks.
#  Set $DISKS to $1 command-line-parameter,
#+ or to $E_NOPARAM if that is unset.

See also Example 3-4, Example 31-2, and Example A-6.

Compare this method with using an and list to supply a default command-line argument.

${parameter=default}, ${parameter:=default}

If parameter not set, set it to default.

Both forms nearly equivalent. The : makes a difference only when $parameter has been declared and is null, [51] as above.

echo ${var=abc}   # abc
echo ${var=xyz}   # abc
# $var had already been set to abc, so it did not change.

${parameter+alt_value}, ${parameter:+alt_value}

If parameter set, use alt_value, else use null string.

Both forms nearly equivalent. The : makes a difference only when parameter has been declared and is null, see below.

echo "###### \${parameter+alt_value} ########"

echo "a = $a"      # a =

echo "a = $a"      # a = xyz

echo "a = $a"      # a = xyz

echo "###### \${parameter:+alt_value} ########"

echo "a = $a"      # a =

echo "a = $a"      # a =
# Different result from   a=${param5+xyz}

echo "a = $a"      # a = xyz

${parameter?err_msg}, ${parameter:?err_msg}

If parameter set, use it, else print err_msg and abort the script with an exit status of 1.

Both forms nearly equivalent. The : makes a difference only when parameter has been declared and is null, as above.

Example 10-7. Using parameter substitution and error messages


#  Check some of the system's environmental variables.
#  This is good preventative maintenance.
#  If, for example, $USER, the name of the person at the console, is not set,
#+ the machine will not recognize you.

: ${HOSTNAME?} ${USER?} ${HOME?} ${MAIL?}
  echo "Name of the machine is $HOSTNAME."
  echo "You are $USER."
  echo "Your home directory is $HOME."
  echo "Your mail INBOX is located in $MAIL."
  echo "If you are reading this message,"
  echo "critical environmental variables have been set."

# ------------------------------------------------------

#  The ${variablename?} construction can also check
#+ for variables set within the script.

#  Note, by the way, that string variables may be set
#+ to characters disallowed in their names.
: ${ThisVariable?}
echo "Value of ThisVariable is $ThisVariable".

echo; echo

: ${ZZXy23AB?"ZZXy23AB has not been set."}
#  Since ZZXy23AB has not been set,
#+ then the script terminates with an error message.

# You can specify the error message.
# : ${variablename?"ERROR MESSAGE"}

# Same result with:   dummy_variable=${ZZXy23AB?}
#                     dummy_variable=${ZZXy23AB?"ZXy23AB has not been set."}
#                     echo ${ZZXy23AB?} >/dev/null

#  Compare these methods of checking whether a variable has been set
#+ with "set -u" . . .

echo "You will not see this message, because script already terminated."

exit $HERE   # Will NOT exit here.

# In fact, this script will return an exit status (echo $?) of 1.

Example 10-8. Parameter substitution and "usage" messages


: ${1?"Usage: $0 ARGUMENT"}
#  Script exits here if command-line parameter absent,
#+ with following error message.
# 1: Usage: ARGUMENT

echo "These two lines echo only if command-line parameter given."
echo "command-line parameter = \"$1\""

exit 0  # Will exit here only if command-line parameter present.

# Check the exit status, both with and without command-line parameter.
# If command-line parameter present, then "$?" is 0.
# If not, then "$?" is 1.

Parameter substitution and/or expansion. The following expressions are the complement to the match in expr string operations (see Example 16-9). These particular ones are used mostly in parsing file path names.

Variable length / Substring removal


String length (number of characters in $var). For an array, ${#array} is the length of the first element in the array.



  • ${#*} and ${#@} give the number of positional parameters.

  • For an array, ${#array[*]} and ${#array[@]} give the number of elements in the array.

Example 10-9. Length of a variable



if [ $# -eq 0 ]  # Must have command-line args to demo script.
  echo "Please invoke this script with one or more command-line arguments."
  exit $E_NO_ARGS

echo "var01 = ${var01}"
echo "Length of var01 = ${#var01}"
# Now, let's try embedding a space.
var02="abcd EFGH28ij"
echo "var02 = ${var02}"
echo "Length of var02 = ${#var02}"

echo "Number of command-line arguments passed to script = ${#@}"
echo "Number of command-line arguments passed to script = ${#*}"

exit 0
${var#Pattern}, ${var##Pattern}

${var#Pattern} Remove from $var the shortest part of $Pattern that matches the front end of $var.

${var##Pattern} Remove from $var the longest part of $Pattern that matches the front end of $var.

A usage illustration from Example A-7:
# Function from "" example.
# Strips leading zero(s) from argument passed.

strip_leading_zero () #  Strip possible leading zero(s)
{                     #+ from argument passed.
  return=${1#0}       #  The "1" refers to "$1" -- passed arg.
}                     #  The "0" is what to remove from "$1" -- strips zeros.

Manfred Schwarb's more elaborate variation of the above:

strip_leading_zero2 () # Strip possible leading zero(s), since otherwise
{                      # Bash will interpret such numbers as octal values.
  shopt -s extglob     # Turn on extended globbing.
  local val=${1##+(0)} # Use local variable, longest matching series of 0's.
  shopt -u extglob     # Turn off extended globbing.
                       # If input was 0, return 0 instead of "".

Another usage illustration:

echo `basename $PWD`        # Basename of current working directory.
echo "${PWD##*/}"           # Basename of current working directory.
echo `basename $0`          # Name of script.
echo $0                     # Name of script.
echo "${0##*/}"             # Name of script.
echo "${filename##*.}"      # data
                            # Extension of filename.

${var%Pattern}, ${var%%Pattern}

${var%Pattern} Remove from $var the shortest part of $Pattern that matches the back end of $var.

${var%%Pattern} Remove from $var the longest part of $Pattern that matches the back end of $var.

Version 2 of Bash added additional options.

Example 10-10. Pattern matching in parameter substitution


# Pattern matching  using the # ## % %% parameter substitution operators.

pattern1=a*c  # * (wild card) matches everything between a - c.

echo "var1 = $var1"           # abcd12345abc6789
echo "var1 = ${var1}"         # abcd12345abc6789
                              # (alternate form)
echo "Number of characters in ${var1} = ${#var1}"

echo "pattern1 = $pattern1"   # a*c  (everything between 'a' and 'c')
echo "--------------"
echo '${var1#$pattern1}  =' "${var1#$pattern1}"    #         d12345abc6789
# Shortest possible match, strips out first 3 characters  abcd12345abc6789
#                                     ^^^^^               |-|
echo '${var1##$pattern1} =' "${var1##$pattern1}"   #                  6789      
# Longest possible match, strips out first 12 characters  abcd12345abc6789
#                                    ^^^^^                |----------|

echo; echo; echo

pattern2=b*9            # everything between 'b' and '9'
echo "var1 = $var1"     # Still  abcd12345abc6789
echo "pattern2 = $pattern2"
echo "--------------"
echo '${var1%pattern2}  =' "${var1%$pattern2}"     #     abcd12345a
# Shortest possible match, strips out last 6 characters  abcd12345abc6789
#                                     ^^^^                         |----|
echo '${var1%%pattern2} =' "${var1%%$pattern2}"    #     a
# Longest possible match, strips out last 12 characters  abcd12345abc6789
#                                    ^^^^                 |-------------|

# Remember, # and ## work from the left end (beginning) of string,
#           % and %% work from the right end.


exit 0

Example 10-11. Renaming file extensions:

# Renaming file extensions.
#         rfe old_extension new_extension
# Example:
# To rename all *.gif files in working directory to *.jpg,
#          rfe gif jpg


case $# in
  0|1)             # The vertical bar means "or" in this context.
  echo "Usage: `basename $0` old_file_suffix new_file_suffix"
  exit $E_BADARGS  # If 0 or 1 arg, then bail out.

for filename in *.$1
# Traverse list of files ending with 1st argument.
  mv $filename ${filename%$1}$2
  #  Strip off part of filename matching 1st argument,
  #+ then append 2nd argument.

exit 0

Variable expansion / Substring replacement

These constructs have been adopted from ksh.


Variable var expanded, starting from offset pos.


Expansion to a max of len characters of variable var, from offset pos. See Example A-13 for an example of the creative use of this operator.


First match of Pattern, within var replaced with Replacement.

If Replacement is omitted, then the first match of Pattern is replaced by nothing, that is, deleted.


Global replacement. All matches of Pattern, within var replaced with Replacement.

As above, if Replacement is omitted, then all occurrences of Pattern are replaced by nothing, that is, deleted.

Example 10-12. Using pattern matching to parse arbitrary strings


echo "var1 = $var1"

echo "var1 (with everything, up to and including first - stripped out) = $t"
#  t=${var1#*-}  works just the same,
#+ since # matches the shortest string,
#+ and * matches everything preceding, including an empty string.
# (Thanks, Stephane Chazelas, for pointing this out.)

echo "If var1 contains a \"-\", returns empty string...   var1 = $t"

echo "var1 (with everything from the last - on stripped out) = $t"


# -------------------------------------------
# -------------------------------------------
echo "path_name = $path_name"
echo "path_name, stripped of prefixes = $t"
# Same effect as   t=`basename $path_name` in this particular case.
#  t=${path_name%/}; t=${t##*/}   is a more general solution,
#+ but still fails sometimes.
#  If $path_name ends with a newline, then `basename $path_name` will not work,
#+ but the above expression will.
# (Thanks, S.C.)

# Same effect as   t=`dirname $path_name`
echo "path_name, stripped of suffixes = $t"
# These will fail in some cases, such as "../", "/foo////", # "foo/", "/".
#  Removing suffixes, especially when the basename has no suffix,
#+ but the dirname does, also complicates matters.
# (Thanks, S.C.)


echo "$path_name, with first 11 chars stripped off = $t"
echo "$path_name, with first 11 chars stripped off, length 5 = $t"


echo "$path_name with \"bozo\" replaced  by \"clown\" = $t"
echo "$path_name with \"today\" deleted = $t"
echo "$path_name with all o's capitalized = $t"
echo "$path_name with all o's deleted = $t"

exit 0

If prefix of var matches Pattern, then substitute Replacement for Pattern.


If suffix of var matches Pattern, then substitute Replacement for Pattern.

Example 10-13. Matching patterns at prefix or suffix of string

# Demo of pattern replacement at prefix / suffix of string.

v0=abc1234zip1234abc    # Original variable.
echo "v0 = $v0"         # abc1234zip1234abc

# Match at prefix (beginning) of string.
v1=${v0/#abc/ABCDEF}    # abc1234zip1234abc
                        # |-|
echo "v1 = $v1"         # ABCDEF1234zip1234abc
                        # |----|

# Match at suffix (end) of string.
v2=${v0/%abc/ABCDEF}    # abc1234zip123abc
                        #              |-|
echo "v2 = $v2"         # abc1234zip1234ABCDEF
                        #               |----|


#  ----------------------------------------------------
#  Must match at beginning / end of string,
#+ otherwise no replacement results.
#  ----------------------------------------------------
v3=${v0/#123/000}       # Matches, but not at beginning.
echo "v3 = $v3"         # abc1234zip1234abc
                        # NO REPLACEMENT.
v4=${v0/%123/000}       # Matches, but not at end.
echo "v4 = $v4"         # abc1234zip1234abc
                        # NO REPLACEMENT.

exit 0			
${!varprefix*}, ${!varprefix@}

Matches names of all previously declared variables beginning with varprefix.
# This is a variation on indirect reference, but with a * or @.
# Bash, version 2.04, adds this feature.


a=${!xyz*}         #  Expands to *names* of declared variables
# ^ ^   ^           + beginning with "xyz".
echo "a = $a"      #  a = xyz23 xyz24
a=${!xyz@}         #  Same as above.
echo "a = $a"      #  a = xyz23 xyz24

echo "---"

echo "b = $b"      #  b = abc23
c=${!b}            #  Now, the more familiar type of indirect reference.
echo $c            #  something_else

Chapter 11. Loops and Branches


What needs this iteration, woman?

--Shakespeare, Othello

Operations on code blocks are the key to structured and organized shell scripts. Looping and branching constructs provide the tools for accomplishing this.

11.1. Loops

A loop is a block of code that iterates [52] a list of commands as long as the loop control condition is true.

for loops

for arg in [list]

This is the basic looping construct. It differs significantly from its C counterpart.

for arg in [list]


During each pass through the loop, arg takes on the value of each successive variable in the list.

for arg in "$var1" "$var2" "$var3" ... "$varN"  
# In pass 1 of the loop, arg = $var1	    
# In pass 2 of the loop, arg = $var2	    
# In pass 3 of the loop, arg = $var3	    
# ...
# In pass N of the loop, arg = $varN

# Arguments in [list] quoted to prevent possible word splitting.

The argument list may contain wild cards.

If do is on same line as for, there needs to be a semicolon after list.

for arg in [list] ; do

Example 11-1. Simple for loops

# Listing the planets.

for planet in Mercury Venus Earth Mars Jupiter Saturn Uranus Neptune Pluto
  echo $planet  # Each planet on a separate line.

echo; echo

for planet in "Mercury Venus Earth Mars Jupiter Saturn Uranus Neptune Pluto"
    # All planets on same line.
    # Entire 'list' enclosed in quotes creates a single variable.
    # Why? Whitespace incorporated into the variable.
  echo $planet

echo; echo "Whoops! Pluto is no longer a planet!"

exit 0

Each [list] element may contain multiple parameters. This is useful when processing parameters in groups. In such cases, use the set command (see Example 15-16) to force parsing of each [list] element and assignment of each component to the positional parameters.

Example 11-2. for loop with two parameters in each [list] element

# Planets revisited.

# Associate the name of each planet with its distance from the sun.

for planet in "Mercury 36" "Venus 67" "Earth 93"  "Mars 142" "Jupiter 483"
  set -- $planet  #  Parses variable "planet"
                  #+ and sets positional parameters.
  #  The "--" prevents nasty surprises if $planet is null or
  #+ begins with a dash.

  #  May need to save original positional parameters,
  #+ since they get overwritten.
  #  One way of doing this is to use an array,
  #         original_params=("$@")

  echo "$1		$2,000,000 miles from the sun"
  #-------two  tabs---concatenate zeroes onto parameter $2

# (Thanks, S.C., for additional clarification.)

exit 0

A variable may supply the [list] in a for loop.

Example 11-3. Fileinfo: operating on a file list contained in a variable


/sbin/ypbind"     # List of files you are curious about.
                  # Threw in a dummy file, /usr/bin/fakefile.


for file in $FILES

  if [ ! -e "$file" ]       # Check if file exists.
    echo "$file does not exist."; echo
    continue                # On to next.

  ls -l $file | awk '{ print $8 "         file size: " $5 }'  # Print 2 fields.
  whatis `basename $file`   # File info.
  # Note that the whatis database needs to have been set up for this to work.
  # To do this, as root run /usr/bin/makewhatis.

exit 0

The [list] in a for loop may be parameterized.

Example 11-4. Operating on a parameterized file list



for file in $filename
 echo "Contents of $file"
 echo "---"
 cat "$file"

If the [list] in a for loop contains wild cards (* and ?) used in filename expansion, then globbing takes place.

Example 11-5. Operating on files with a for loop

# Generating [list] in a for-loop, using "globbing" ...
# Globbing = filename expansion.


for file in *
#           ^  Bash performs filename expansion
#+             on expressions that globbing recognizes.
  ls -l "$file"  # Lists all files in $PWD (current directory).
  #  Recall that the wild card character "*" matches every filename,
  #+ however, in "globbing," it doesn't match dot-files.

  #  If the pattern matches no file, it is expanded to itself.
  #  To prevent this, set the nullglob option
  #+   (shopt -s nullglob).
  #  Thanks, S.C.

echo; echo

for file in [jx]*
  rm -f $file    # Removes only files beginning with "j" or "x" in $PWD.
  echo "Removed file \"$file\"".


exit 0

Omitting the in [list] part of a for loop causes the loop to operate on $@ -- the positional parameters. A particularly clever illustration of this is Example A-15. See also Example 15-17.

Example 11-6. Missing in [list] in a for loop


#  Invoke this script both with and without arguments,
#+ and see what happens.

for a
 echo -n "$a "

#  The 'in list' missing, therefore the loop operates on '$@'
#+ (command-line argument list, including whitespace).


exit 0

It is possible to use command substitution to generate the [list] in a for loop. See also Example 16-54, Example 11-11 and Example 16-48.

Example 11-7. Generating the [list] in a for loop with command substitution

# for-loop with [list]
#+ generated by command substitution.

NUMBERS="9 7 3 8 37.53"

for number in `echo $NUMBERS`  # for number in 9 7 3 8 37.53
  echo -n "$number "

exit 0

Here is a somewhat more complex example of using command substitution to create the [list].

Example 11-8. A grep replacement for binary files

# Locates matching strings in a binary file.

# A "grep" replacement for binary files.
# Similar effect to "grep -a"


if [ $# -ne 2 ]
  echo "Usage: `basename $0` search_string filename"
  exit $E_BADARGS

if [ ! -f "$2" ]
  echo "File \"$2\" does not exist."
  exit $E_NOFILE

IFS=$'\012'       # Per suggestion of Anton Filippov.
                  # was:  IFS="\n"
for word in $( strings "$2" | grep "$1" )
# The "strings" command lists strings in binary files.
# Output then piped to "grep", which tests for desired string.
  echo $word

# As S.C. points out, lines 23 - 30 could be replaced with the simpler
#    strings "$2" | grep "$1" | tr -s "$IFS" '[\n*]'

#  Try something like  "./ mem /bin/ls"
#+ to exercise this script.

exit 0

More of the same.

Example 11-9. Listing all users on the system


n=1           # User number

for name in $(awk 'BEGIN{FS=":"}{print $1}' < "$PASSWORD_FILE" )
# Field separator = :    ^^^^^^
# Print first field              ^^^^^^^^
# Get input from password file  /etc/passwd  ^^^^^^^^^^^^^^^^^
  echo "USER #$n = $name"
  let "n += 1"

# USER #1 = root
# USER #2 = bin
# USER #3 = daemon
# ...
# USER #33 = bozo

exit $?

#  Discussion:
#  ----------
#  How is it that an ordinary user, or a script run by same,
#+ can read /etc/passwd? (Hint: Check the /etc/passwd file permissions.)
#  Is this a security hole? Why or why not?

Yet another example of the [list] resulting from command substitution.

Example 11-10. Checking all the binaries in a directory for authorship

# Find a particular string in the binaries in a specified directory.

fstring="Free Software Foundation"  # See which files come from the FSF.

for file in $( find $directory -type f -name '*' | sort )
  strings -f $file | grep "$fstring" | sed -e "s%$directory%%"
  #  In the "sed" expression,
  #+ it is necessary to substitute for the normal "/" delimiter
  #+ because "/" happens to be one of the characters filtered out.
  #  Failure to do so gives an error message. (Try it.)

exit $?

#  Exercise (easy):
#  ---------------
#  Convert this script to take command-line parameters
#+ for $directory and $fstring.

A final example of [list] / command substitution, but this time the "command" is a function.

generate_list ()
  echo "one two three"

for word in $(generate_list)  # Let "word" grab output of function.
  echo "$word"

# one
# two
# three

The output of a for loop may be piped to a command or commands.

Example 11-11. Listing the symbolic links in a directory

# Lists symbolic links in a directory.

#  Defaults to current working directory,
#+ if not otherwise specified.
#  Equivalent to code block below.
# ----------------------------------------------------------
# ARGS=1                 # Expect one command-line argument.
# if [ $# -ne "$ARGS" ]  # If not 1 arg...
# then
#   directory=`pwd`      # current working directory
# else
#   directory=$1
# fi
# ----------------------------------------------------------

echo "symbolic links in directory \"$directory\""

for file in "$( find $directory -type l )"   # -type l = symbolic links
  echo "$file"
done | sort                                  # Otherwise file list is unsorted.
#  Strictly speaking, a loop isn't really necessary here,
#+ since the output of the "find" command is expanded into a single word.
#  However, it's easy to understand and illustrative this way.

#  As Dominik 'Aeneas' Schnitzer points out,
#+ failing to quote  $( find $directory -type l )
#+ will choke on filenames with embedded whitespace.
#  containing whitespace. 

exit 0

# --------------------------------------------------------
# Jean Helou proposes the following alternative:

echo "symbolic links in directory \"$directory\""
# Backup of the current IFS. One can never be too cautious.

for file in $(find $directory -type l -printf "%p$IFS")
do     #                              ^^^^^^^^^^^^^^^^
       echo "$file"

# And, James "Mike" Conley suggests modifying Helou's code thusly:

IFS='' # Null IFS means no word breaks
for file in $( find $directory -type l )
  echo $file
  done | sort

#  This works in the "pathological" case of a directory name having
#+ an embedded colon.
#  "This also fixes the pathological case of the directory name having
#+  a colon (or space in earlier example) as well."

The stdout of a loop may be redirected to a file, as this slight modification to the previous example shows.

Example 11-12. Symbolic links in a directory, saved to a file

# Lists symbolic links in a directory.

OUTFILE=symlinks.list                         # save-file

#  Defaults to current working directory,
#+ if not otherwise specified.

echo "symbolic links in directory \"$directory\"" > "$OUTFILE"
echo "---------------------------" >> "$OUTFILE"

for file in "$( find $directory -type l )"    # -type l = symbolic links
  echo "$file"
done | sort >> "$OUTFILE"                     # stdout of loop
#           ^^^^^^^^^^^^^                       redirected to save file.

# echo "Output file = $OUTFILE"

exit $?

There is an alternative syntax to a for loop that will look very familiar to C programmers. This requires double parentheses.

Example 11-13. A C-style for loop

# Multiple ways to count up to 10.


# Standard syntax.
for a in 1 2 3 4 5 6 7 8 9 10
  echo -n "$a "

echo; echo

# +==========================================+

# Using "seq" ...
for a in `seq 10`
  echo -n "$a "

echo; echo

# +==========================================+

# Using brace expansion ...
# Bash, version 3+.
for a in {1..10}
  echo -n "$a "

echo; echo

# +==========================================+

# Now, let's do the same, using C-like syntax.


for ((a=1; a <= LIMIT ; a++))  # Double parentheses, and naked "LIMIT"
  echo -n "$a "
done                           # A construct borrowed from ksh93.

echo; echo

# +=========================================================================+

# Let's use the C "comma operator" to increment two variables simultaneously.

for ((a=1, b=1; a <= LIMIT ; a++, b++))
do  # The comma concatenates operations.
  echo -n "$a-$b "

echo; echo

exit 0

See also Example 27-16, Example 27-17, and Example A-6.


Now, a for loop used in a "real-life" context.

Example 11-14. Using efax in batch mode

# Faxing (must have 'efax' package installed).

MODEM_PORT="/dev/ttyS2"   # May be different on your machine.
#                ^^^^^      PCMCIA modem card default port.

if [ $# -ne $EXPECTED_ARGS ]
# Check for proper number of command-line args.
   echo "Usage: `basename $0` phone# text-file"
   exit $E_BADARGS

if [ ! -f "$2" ]
  echo "File $2 is not a text file."
  #     File is not a regular file, or does not exist.
  exit $E_BADARGS

fax make $2              #  Create fax-formatted files from text files.

for file in $(ls $2.0*)  #  Concatenate the converted files.
                         #  Uses wild card (filename "globbing")
			 #+ in variable list.
  fil="$fil $file"

efax -d "$MODEM_PORT"  -t "T$1" $fil   # Finally, do the work.
# Trying adding  -o1  if above line fails.

#  As S.C. points out, the for-loop can be eliminated with
#     efax -d /dev/ttyS2 -o1 -t "T$1" $2.0*
#+ but it's not quite as instructive [grin].

exit $?   # Also, efax sends diagnostic messages to stdout.


The keywords do and done delineate the for-loop command block. However, these may, in certain contexts, be omitted by framing the command block within curly brackets
for((n=1; n<=10; n++)) 
# No do!
  echo -n "* $n *"
# No done!

# Outputs:
# * 1 ** 2 ** 3 ** 4 ** 5 ** 6 ** 7 ** 8 ** 9 ** 10 *
# And, echo $? returns 0, so Bash does not register an error.


#  But, note that in a classic for-loop:    for n in [list] ...
#+ a terminal semicolon is required.

for n in 1 2 3
{  echo -n "$n "; }
#               ^

# Thank you, YongYe, for pointing this out.


This construct tests for a condition at the top of a loop, and keeps looping as long as that condition is true (returns a 0 exit status). In contrast to a for loop, a while loop finds use in situations where the number of loop repetitions is not known beforehand.

while [ condition ]

The bracket construct in a while loop is nothing more than our old friend, the test brackets used in an if/then test. In fact, a while loop can legally use the more versatile double-brackets construct (while [[ condition ]]).

As is the case with for loops, placing the do on the same line as the condition test requires a semicolon.

while [ condition ] ; do

Note that the test brackets are not mandatory in a while loop. See, for example, the getopts construct.

Example 11-15. Simple while loop



while [ "$var0" -lt "$LIMIT" ]
#      ^                    ^
# Spaces, because these are "test-brackets" . . .
  echo -n "$var0 "        # -n suppresses newline.
  #             ^           Space, to separate printed out numbers.

  var0=`expr $var0 + 1`   # var0=$(($var0+1))  also works.
                          # var0=$((var0 + 1)) also works.
                          # let "var0 += 1"    also works.
done                      # Various other methods also work.


exit 0

Example 11-16. Another while loop


                               # Equivalent to:
while [ "$var1" != "end" ]     # while test "$var1" != "end"
  echo "Input variable #1 (end to exit) "
  read var1                    # Not 'read $var1' (why?).
  echo "variable #1 = $var1"   # Need quotes because of "#" . . .
  # If input is 'end', echoes it here.
  # Does not test for termination condition until top of loop.

exit 0

A while loop may have multiple conditions. Only the final condition determines when the loop terminates. This necessitates a slightly different loop syntax, however.

Example 11-17. while loop with multiple conditions



while echo "previous-variable = $previous"
      [ "$var1" != end ] # Keeps track of what $var1 was previously.
      # Four conditions on *while*, but only the final one controls loop.
      # The *last* exit status is the one that counts.
echo "Input variable #1 (end to exit) "
  read var1
  echo "variable #1 = $var1"

# Try to figure out how this all works.
# It's a wee bit tricky.

exit 0

As with a for loop, a while loop may employ C-style syntax by using the double-parentheses construct (see also Example 8-5).

Example 11-18. C-style syntax in a while loop

# Count to 10 in a "while" loop.

LIMIT=10                 # 10 iterations.

while [ "$a" -le $LIMIT ]
  echo -n "$a "
  let "a+=1"
done                     # No surprises, so far.

echo; echo

# +=================================================================+

# Now, we'll repeat with C-like syntax.

((a = 1))      # a=1
# Double parentheses permit space when setting a variable, as in C.

while (( a <= LIMIT ))   #  Double parentheses,
do                       #+ and no "$" preceding variables.
  echo -n "$a "
  ((a += 1))             # let "a+=1"
  # Yes, indeed.
  # Double parentheses permit incrementing a variable with C-like syntax.


# C and Java programmers can feel right at home in Bash.

exit 0

Inside its test brackets, a while loop can call a function.

condition ()

  if [ $t -lt 5 ]
    return 0  # true
    return 1  # false

while condition
#     ^^^^^^^^^
#     Function call -- four loop iterations.
  echo "Still going: t = $t"

# Still going: t = 1
# Still going: t = 2
# Still going: t = 3
# Still going: t = 4

By coupling the power of the read command with a while loop, we get the handy while read construct, useful for reading and parsing files.

cat $filename |   # Supply input from a file.
while read line   # As long as there is another line to read ...

# =========== Snippet from "" example script ========== #

  while read value   # Read one data point at a time.
    rt=$(echo "scale=$SC; $rt + $value" | bc)
    (( ct++ ))

  am=$(echo "scale=$SC; $rt / $ct" | bc)

  echo $am; return $ct   # This function "returns" TWO values!
  #  Caution: This little trick will not work if $ct > 255!
  #  To handle a larger number of data points,
  #+ simply comment out the "return $ct" above.
} <"$datafile"   # Feed in data file.


A while loop may have its stdin redirected to a file by a < at its end.

A while loop may have its stdin supplied by a pipe.


This construct tests for a condition at the top of a loop, and keeps looping as long as that condition is false (opposite of while loop).

until [ condition-is-true ]

Note that an until loop tests for the terminating condition at the top of the loop, differing from a similar construct in some programming languages.

As is the case with for loops, placing the do on the same line as the condition test requires a semicolon.

until [ condition-is-true ] ; do

Example 11-19. until loop



until [ "$var1" = "$END_CONDITION" ]
# Tests condition here, at top of loop.
  echo "Input variable #1 "
  echo "($END_CONDITION to exit)"
  read var1
  echo "variable #1 = $var1"

#                     ---                        #

#  As with "for" and "while" loops,
#+ an "until" loop permits C-like test constructs.


until (( var > LIMIT ))
do  # ^^ ^     ^     ^^   No brackets, no $ prefixing variables.
  echo -n "$var "
  (( var++ ))
done    # 0 1 2 3 4 5 6 7 8 9 10 

exit 0

How to choose between a for loop or a while loop or until loop? In C, you would typically use a for loop when the number of loop iterations is known beforehand. With Bash, however, the situation is fuzzier. The Bash for loop is more loosely structured and more flexible than its equivalent in other languages. Therefore, feel free to use whatever type of loop gets the job done in the simplest way.

11.2. Nested Loops

A nested loop is a loop within a loop, an inner loop within the body of an outer one. How this works is that the first pass of the outer loop triggers the inner loop, which executes to completion. Then the second pass of the outer loop triggers the inner loop again. This repeats until the outer loop finishes. Of course, a break within either the inner or outer loop would interrupt this process.

Example 11-20. Nested Loop

# Nested "for" loops.

outer=1             # Set outer loop counter.

# Beginning of outer loop.
for a in 1 2 3 4 5
  echo "Pass $outer in outer loop."
  echo "---------------------"
  inner=1           # Reset inner loop counter.

  # ===============================================
  # Beginning of inner loop.
  for b in 1 2 3 4 5
    echo "Pass $inner in inner loop."
    let "inner+=1"  # Increment inner loop counter.
  # End of inner loop.
  # ===============================================

  let "outer+=1"    # Increment outer loop counter. 
  echo              # Space between output blocks in pass of outer loop.
# End of outer loop.

exit 0

See Example 27-11 for an illustration of nested while loops, and Example 27-13 to see a while loop nested inside an until loop.

11.3. Loop Control


Tournez cent tours, tournez mille tours,

Tournez souvent et tournez toujours . . .

--Verlaine, "Chevaux de bois"

Commands affecting loop behavior

break, continue

The break and continue loop control commands [53] correspond exactly to their counterparts in other programming languages. The break command terminates the loop (breaks out of it), while continue causes a jump to the next iteration of the loop, skipping all the remaining commands in that particular loop cycle.

Example 11-21. Effects of break and continue in a loop


LIMIT=19  # Upper limit

echo "Printing Numbers 1 through 20 (but not 3 and 11)."


while [ $a -le "$LIMIT" ]

 if [ "$a" -eq 3 ] || [ "$a" -eq 11 ]  # Excludes 3 and 11.
   continue      # Skip rest of this particular loop iteration.

 echo -n "$a "   # This will not execute for 3 and 11.

# Exercise:
# Why does the loop print up to 20?

echo; echo

echo Printing Numbers 1 through 20, but something happens after 2.


# Same loop, but substituting 'break' for 'continue'.


while [ "$a" -le "$LIMIT" ]

 if [ "$a" -gt 2 ]
   break  # Skip entire rest of loop.

 echo -n "$a "

echo; echo; echo

exit 0

The break command may optionally take a parameter. A plain break terminates only the innermost loop in which it is embedded, but a break N breaks out of N levels of loop.

Example 11-22. Breaking out of multiple loop levels

# Breaking out of loops.

# "break N" breaks out of N level loops.

for outerloop in 1 2 3 4 5
  echo -n "Group $outerloop:   "

  # --------------------------------------------------------
  for innerloop in 1 2 3 4 5
    echo -n "$innerloop "

    if [ "$innerloop" -eq 3 ]
      break  # Try   break 2   to see what happens.
             # ("Breaks" out of both inner and outer loops.)
  # --------------------------------------------------------



exit 0

The continue command, similar to break, optionally takes a parameter. A plain continue cuts short the current iteration within its loop and begins the next. A continue N terminates all remaining iterations at its loop level and continues with the next iteration at the loop, N levels above.

Example 11-23. Continuing at a higher loop level

# The "continue N" command, continuing at the Nth level loop.

for outer in I II III IV V           # outer loop
  echo; echo -n "Group $outer: "

  # --------------------------------------------------------------------
  for inner in 1 2 3 4 5 6 7 8 9 10  # inner loop

    if [[ "$inner" -eq 7 && "$outer" = "III" ]]
      continue 2  # Continue at loop on 2nd level, that is "outer loop".
                  # Replace above line with a simple "continue"
                  # to see normal loop behavior.

    echo -n "$inner "  # 7 8 9 10 will not echo on "Group III."
  # --------------------------------------------------------------------


echo; echo

# Exercise:
# Come up with a meaningful use for "continue N" in a script.

exit 0

Example 11-24. Using continue N in an actual task

# Albert Reiner gives an example of how to use "continue N":
# ---------------------------------------------------------

#  Suppose I have a large number of jobs that need to be run, with
#+ any data that is to be treated in files of a given name pattern
#+ in a directory. There are several machines that access
#+ this directory, and I want to distribute the work over these
#+ different boxen.
#  Then I usually nohup something like the following on every box:

while true
  for n in .iso.*
    [ "$n" = ".iso.opts" ] && continue
    [ -r .Iso.$beta ] && continue
    [ -r .lock.$beta ] && sleep 10 && continue
    lockfile -r0 .lock.$beta || continue
    echo -n "$beta: " `date`
    run-isotherm $beta
    ls -alF .Iso.$beta
    [ -r .Iso.$beta ] && rm -f .lock.$beta
    continue 2

exit 0

#  The details, in particular the sleep N, are particular to my
#+ application, but the general pattern is:

while true
  for job in {pattern}
    {job already done or running} && continue
    {mark job as running, do job, mark job as done}
    continue 2
  break        # Or something like `sleep 600' to avoid termination.

#  This way the script will stop only when there are no more jobs to do
#+ (including jobs that were added during runtime). Through the use
#+ of appropriate lockfiles it can be run on several machines
#+ concurrently without duplication of calculations [which run a couple
#+ of hours in my case, so I really want to avoid this]. Also, as search
#+ always starts again from the beginning, one can encode priorities in
#+ the file names. Of course, one could also do this without `continue 2',
#+ but then one would have to actually check whether or not some job
#+ was done (so that we should immediately look for the next job) or not
#+ (in which case we terminate or sleep for a long time before checking
#+ for a new job).


The continue N construct is difficult to understand and tricky to use in any meaningful context. It is probably best avoided.

11.4. Testing and Branching

The case and select constructs are technically not loops, since they do not iterate the execution of a code block. Like loops, however, they direct program flow according to conditions at the top or bottom of the block.

Controlling program flow in a code block

case (in) / esac

The case construct is the shell scripting analog to switch in C/C++. It permits branching to one of a number of code blocks, depending on condition tests. It serves as a kind of shorthand for multiple if/then/else statements and is an appropriate tool for creating menus.

case "$variable" in

�"$condition1" )

�"$condition2" )



  • Quoting the variables is not mandatory, since word splitting does not take place.

  • Each test line ends with a right paren ). [54]

  • Each condition block ends with a double semicolon ;;.

  • If a condition tests true, then the associated commands execute and the case block terminates.

  • The entire case block ends with an esac (case spelled backwards).

Example 11-25. Using case

# Testing ranges of characters.

echo; echo "Hit a key, then hit return."
read Keypress

case "$Keypress" in
  [[:lower:]]   ) echo "Lowercase letter";;
  [[:upper:]]   ) echo "Uppercase letter";;
  [0-9]         ) echo "Digit";;
  *             ) echo "Punctuation, whitespace, or other";;
esac      #  Allows ranges of characters in [square brackets],
          #+ or POSIX ranges in [[double square brackets.

#  In the first version of this example,
#+ the tests for lowercase and uppercase characters were
#+ [a-z] and [A-Z].
#  This no longer works in certain locales and/or Linux distros.
#  POSIX is more portable.
#  Thanks to Frank Wang for pointing this out.

#  Exercise:
#  --------
#  As the script stands, it accepts a single keystroke, then terminates.
#  Change the script so it accepts repeated input,
#+ reports on each keystroke, and terminates only when "X" is hit.
#  Hint: enclose everything in a "while" loop.

exit 0

Example 11-26. Creating menus using case


# Crude address database

clear # Clear the screen.

echo "          Contact List"
echo "          ------- ----"
echo "Choose one of the following persons:" 
echo "[E]vans, Roland"
echo "[J]ones, Mildred"
echo "[S]mith, Julie"
echo "[Z]ane, Morris"

read person

case "$person" in
# Note variable is quoted.

  "E" | "e" )
  # Accept upper or lowercase input.
  echo "Roland Evans"
  echo "4321 Flash Dr."
  echo "Hardscrabble, CO 80753"
  echo "(303) 734-9874"
  echo "(303) 734-9892 fax"
  echo "[email protected]"
  echo "Business partner & old friend"
# Note double semicolon to terminate each option.

  "J" | "j" )
  echo "Mildred Jones"
  echo "249 E. 7th St., Apt. 19"
  echo "New York, NY 10009"
  echo "(212) 533-2814"
  echo "(212) 533-9972 fax"
  echo "[email protected]"
  echo "Ex-girlfriend"
  echo "Birthday: Feb. 11"

# Add info for Smith & Zane later.

          * )
   # Default option.	  
   # Empty input (hitting RETURN) fits here, too.
   echo "Not yet in database."



#  Exercise:
#  --------
#  Change the script so it accepts multiple inputs,
#+ instead of terminating after displaying just one address.

exit 0

An exceptionally clever use of case involves testing for command-line parameters.
#! /bin/bash

case "$1" in
  "") echo "Usage: ${0##*/} <filename>"; exit $E_PARAM;;
                      # No command-line parameters,
                      # or first parameter empty.
# Note that ${0##*/} is ${var##pattern} param substitution.
                      # Net result is $0.

  -*) FILENAME=./$1;;   #  If filename passed as argument ($1)
                      #+ starts with a dash,
                      #+ replace it with ./$1
                      #+ so further commands don't interpret it
                      #+ as an option.

  * ) FILENAME=$1;;     # Otherwise, $1.

Here is a more straightforward example of command-line parameter handling:
#! /bin/bash

while [ $# -gt 0 ]; do    # Until you run out of parameters . . .
  case "$1" in
              # "-d" or "--debug" parameter?
              if [ ! -f $CONFFILE ]; then
                echo "Error: Supplied file doesn't exist!"
                exit $E_CONFFILE     # File not found error.
  shift       # Check next set of parameters.

#  From Stefano Falsetto's "Log2Rot" script,
#+ part of his "rottlog" package.
#  Used with permission.

Example 11-27. Using command substitution to generate the case variable

# Using command substitution to generate a "case" variable.

case $( arch ) in   # $( arch ) returns machine architecture.
                    # Equivalent to 'uname -m' ...
  i386 ) echo "80386-based machine";;
  i486 ) echo "80486-based machine";;
  i586 ) echo "Pentium-based machine";;
  i686 ) echo "Pentium2+-based machine";;
  *    ) echo "Other type of machine";;

exit 0

A case construct can filter strings for globbing patterns.

Example 11-28. Simple string matching

# Simple string matching
#                  using a 'case' construct.

match_string ()
{ # Exact string match.
  PARAMS=2     # Function requires 2 arguments.

  [ $# -eq $PARAMS ] || return $E_BAD_PARAMS

  case "$1" in
  "$2") return $MATCH;;
  *   ) return $E_NOMATCH;;



match_string $a     # wrong number of parameters
echo $?             # 91

match_string $a $b  # no match
echo $?             # 90

match_string $b $d  # match
echo $?             # 0

exit 0		    

Example 11-29. Checking for alphabetic input

# Using a "case" structure to filter a string.

FAILURE=1   #  Was FAILURE=-1,
            #+ but Bash no longer allows negative return value.

isalpha ()  # Tests whether *first character* of input string is alphabetic.
if [ -z "$1" ]                # No argument passed?
  return $FAILURE

case "$1" in
  [a-zA-Z]*) return $SUCCESS;;  # Begins with a letter?
  *        ) return $FAILURE;;
}             # Compare this with "isalpha ()" function in C.

isalpha2 ()   # Tests whether *entire string* is alphabetic.
  [ $# -eq 1 ] || return $FAILURE

  case $1 in
  *[!a-zA-Z]*|"") return $FAILURE;;
               *) return $SUCCESS;;

isdigit ()    # Tests whether *entire string* is numerical.
{             # In other words, tests for integer variable.
  [ $# -eq 1 ] || return $FAILURE

  case $1 in
    *[!0-9]*|"") return $FAILURE;;
              *) return $SUCCESS;;

check_var ()  # Front-end to isalpha ().
if isalpha "$@"
  echo "\"$*\" begins with an alpha character."
  if isalpha2 "$@"
  then        # No point in testing if first char is non-alpha.
    echo "\"$*\" contains only alpha characters."
    echo "\"$*\" contains at least one non-alpha character."
  echo "\"$*\" begins with a non-alpha character."
              # Also "non-alpha" if no argument passed.



digit_check ()  # Front-end to isdigit ().
if isdigit "$@"
  echo "\"$*\" contains only digits [0 - 9]."
  echo "\"$*\" has at least one non-digit character."



e=$(echo $b)   # Command substitution.

check_var $a
check_var $b
check_var $c
check_var $d
check_var $e
check_var $f
check_var     # No argument passed, so what happens?
digit_check $g
digit_check $h
digit_check $i

exit 0        # Script improved by S.C.

# Exercise:
# --------
#  Write an 'isfloat ()' function that tests for floating point numbers.
#  Hint: The function duplicates 'isdigit ()',
#+ but adds a test for a mandatory decimal point.

The select construct, adopted from the Korn Shell, is yet another tool for building menus.

select variable [in list]

This prompts the user to enter one of the choices presented in the variable list. Note that select uses the $PS3 prompt (#? ) by default, but this may be changed.

Example 11-30. Creating menus using select


PS3='Choose your favorite vegetable: ' # Sets the prompt string.
                                       # Otherwise it defaults to #? .


select vegetable in "beans" "carrots" "potatoes" "onions" "rutabagas"
  echo "Your favorite veggie is $vegetable."
  echo "Yuck!"
  break  # What happens if there is no 'break' here?


# Exercise:
# --------
#  Fix this script to accept user input not specified in
#+ the "select" statement.
#  For example, if the user inputs "peas,"
#+ the script would respond "Sorry. That is not on the menu."

If in list is omitted, then select uses the list of command line arguments ($@) passed to the script or the function containing the select construct.

Compare this to the behavior of a

for variable [in list]

construct with the in list omitted.

Example 11-31. Creating menus using select in a function


PS3='Choose your favorite vegetable: '


select vegetable
# [in list] omitted, so 'select' uses arguments passed to function.
  echo "Your favorite veggie is $vegetable."
  echo "Yuck!"

choice_of beans rice carrots radishes rutabaga spinach
#         $1    $2   $3      $4       $5       $6
#         passed to choice_of() function

exit 0

See also Example 37-3.

Chapter 12. Command Substitution

Command substitution reassigns the output of a command [55] or even multiple commands; it literally plugs the command output into another context. [56]

The classic form of command substitution uses backquotes (`...`). Commands within backquotes (backticks) generate command-line text.
script_name=`basename $0`
echo "The name of this script is $script_name."

The output of commands can be used as arguments to another command, to set a variable, and even for generating the argument list in a for loop.

rm `cat filename`   # "filename" contains a list of files to delete.
# S. C. points out that "arg list too long" error might result.
# Better is              xargs rm -- < filename 
# ( -- covers those cases where "filename" begins with a "-" )

textfile_listing=`ls *.txt`
# Variable contains names of all *.txt files in current working directory.
echo $textfile_listing

textfile_listing2=$(ls *.txt)   # The alternative form of command substitution.
echo $textfile_listing2
# Same result.

# A possible problem with putting a list of files into a single string
# is that a newline may creep in.
# A safer way to assign a list of files to a parameter is with an array.
#      shopt -s nullglob    # If no match, filename expands to nothing.
#      textfile_listing=( *.txt )
# Thanks, S.C.


Command substitution invokes a subshell.


Command substitution may result in word splitting.
COMMAND `echo a b`     # 2 args: a and b

COMMAND "`echo a b`"   # 1 arg: "a b"

COMMAND `echo`         # no arg

COMMAND "`echo`"       # one empty arg

# Thanks, S.C.

Even when there is no word splitting, command substitution can remove trailing newlines.
# cd "`pwd`"  # This should always work.
# However...

mkdir 'dir with trailing newline

cd 'dir with trailing newline

cd "`pwd`"  # Error message:
# bash: cd: /tmp/file with trailing newline: No such file or directory

cd "$PWD"   # Works fine.

old_tty_setting=$(stty -g)   # Save old terminal setting.
echo "Hit a key "
stty -icanon -echo           # Disable "canonical" mode for terminal.
                             # Also, disable *local* echo.
key=$(dd bs=1 count=1 2> /dev/null)   # Using 'dd' to get a keypress.
stty "$old_tty_setting"      # Restore old setting. 
echo "You hit ${#key} key."  # ${#variable} = number of characters in $variable
# Hit any key except RETURN, and the output is "You hit 1 key."
# Hit RETURN, and it's "You hit 0 key."
# The newline gets eaten in the command substitution.

#Code snippet by St�phane Chazelas.


Using echo to output an unquoted variable set with command substitution removes trailing newlines characters from the output of the reassigned command(s). This can cause unpleasant surprises.
dir_listing=`ls -l`
echo $dir_listing     # unquoted

# Expecting a nicely ordered directory listing.

# However, what you get is:
# total 3 -rw-rw-r-- 1 bozo bozo 30 May 13 17:15 1.txt -rw-rw-r-- 1 bozo
# bozo 51 May 15 20:57 -rwxr-xr-x 1 bozo bozo 217 Mar 5 21:13

# The newlines disappeared.

echo "$dir_listing"   # quoted
# -rw-rw-r--    1 bozo       30 May 13 17:15 1.txt
# -rw-rw-r--    1 bozo       51 May 15 20:57
# -rwxr-xr-x    1 bozo      217 Mar  5 21:13

Command substitution even permits setting a variable to the contents of a file, using either redirection or the cat command.

variable1=`<file1`      #  Set "variable1" to contents of "file1".
variable2=`cat file2`   #  Set "variable2" to contents of "file2".
                        #  This, however, forks a new process,
                        #+ so the line of code executes slower than the above version.

#  Note that the variables may contain embedded whitespace,
#+ or even (horrors), control characters.

#  It is not necessary to explicitly assign a variable.
echo "` <$0`"           # Echoes the script itself to stdout.

#  Excerpts from system file, /etc/rc.d/rc.sysinit
#+ (on a Red Hat Linux installation)

if [ -f /fsckoptions ]; then
        fsckoptions=`cat /fsckoptions`
if [ -e "/proc/ide/${disk[$device]}/media" ] ; then
             hdmedia=`cat /proc/ide/${disk[$device]}/media`
if [ ! -n "`uname -r | grep -- "-"`" ]; then
       ktag="`cat /proc/version`"
if [ $usb = "1" ]; then
    sleep 5
    mouseoutput=`cat /proc/bus/usb/devices 2>/dev/null|grep -E "^I.*Cls=03.*Prot=02"`
    kbdoutput=`cat /proc/bus/usb/devices 2>/dev/null|grep -E "^I.*Cls=03.*Prot=01"`


Do not set a variable to the contents of a long text file unless you have a very good reason for doing so. Do not set a variable to the contents of a binary file, even as a joke.

Example 12-1. Stupid script tricks

# Don't try this at home, folks.
# From "Stupid Script Tricks," Volume I.

exit 99  ### Comment out this line if you dare.

dangerous_variable=`cat /boot/vmlinuz`   # The compressed Linux kernel itself.

echo "string-length of \$dangerous_variable = ${#dangerous_variable}"
# string-length of $dangerous_variable = 794151
# (Newer kernels are bigger.)
# Does not give same count as 'wc -c /boot/vmlinuz'.

# echo "$dangerous_variable"
# Don't try this! It would hang the script.

#  The document author is aware of no useful applications for
#+ setting a variable to the contents of a binary file.

exit 0

Notice that a buffer overrun does not occur. This is one instance where an interpreted language, such as Bash, provides more protection from programmer mistakes than a compiled language.

Command substitution permits setting a variable to the output of a loop. The key to this is grabbing the output of an echo command within the loop.

Example 12-2. Generating a variable from a loop

# Setting a variable to the output of a loop.

variable1=`for i in 1 2 3 4 5
  echo -n "$i"                 #  The 'echo' command is critical
done`                          #+ to command substitution here.

echo "variable1 = $variable1"  # variable1 = 12345

variable2=`while [ "$i" -lt 10 ]
  echo -n "$i"                 # Again, the necessary 'echo'.
  let "i += 1"                 # Increment.

echo "variable2 = $variable2"  # variable2 = 0123456789

#  Demonstrates that it's possible to embed a loop
#+ inside a variable declaration.

exit 0


The $(...) form has superseded backticks for command substitution.

output=$(sed -n /"$1"/p $file)   # From ""	example.
# Setting a variable to the contents of a text file.
File_contents1=$(cat $file1)      
File_contents2=$(<$file2)        # Bash permits this also.

The $(...) form of command substitution treats a double backslash in a different way than `...`.

bash$ echo `echo \\`

bash$ echo $(echo \\)

The $(...) form of command substitution permits nesting. [57]

word_count=$( wc -w $(echo * | awk '{print $8}') )

Or, for something a bit more elaborate . . .

Example 12-3. Finding anagrams

# Example of nested command substitution.

#  Uses "anagram" utility
#+ that is part of the author's "yawl" word list package.


if [ -z "$1" ]
  echo "Usage $0 LETTERSET"
  exit $E_NOARGS         # Script needs a command-line argument.
elif [ ${#1} -lt $MINLEN ]
  echo "Argument must have at least $MINLEN letters."
  exit $E_BADARG

FILTER='.......'         # Must have at least 7 letters.
#       1234567
Anagrams=( $(echo $(anagram $1 | grep $FILTER) ) )
#          $(     $(  nested command sub.    ) )
#        (              array assignment         )

echo "${#Anagrams[*]}  7+ letter anagrams found"
echo ${Anagrams[0]}      # First anagram.
echo ${Anagrams[1]}      # Second anagram.
                         # Etc.

# echo "${Anagrams[*]}"  # To list all the anagrams in a single line . . .

#  Look ahead to the Arrays chapter for enlightenment on
#+ what's going on here.

# See also the script for an exercise in anagram finding.

exit $?

Examples of command substitution in shell scripts:

  1. Example 11-8

  2. Example 11-27

  3. Example 9-16

  4. Example 16-3

  5. Example 16-22

  6. Example 16-17

  7. Example 16-54

  8. Example 11-14

  9. Example 11-11

  10. Example 16-32

  11. Example 20-8

  12. Example A-16

  13. Example 29-3

  14. Example 16-47

  15. Example 16-48

  16. Example 16-49

Chapter 13. Arithmetic Expansion

Arithmetic expansion provides a powerful tool for performing (integer) arithmetic operations in scripts. Translating a string into a numerical expression is relatively straightforward using backticks, double parentheses, or let.


Arithmetic expansion with backticks (often used in conjunction with expr)

z=`expr $z + 3`          # The 'expr' command performs the expansion.

Arithmetic expansion with double parentheses, and using let

The use of backticks (backquotes) in arithmetic expansion has been superseded by double parentheses -- ((...)) and $((...)) -- and also by the very convenient let construction.

z=$((z+3))                                  #  Also correct.
                                            #  Within double parentheses,
                                            #+ parameter dereferencing
                                            #+ is optional.

# $((EXPRESSION)) is arithmetic expansion.  #  Not to be confused with
                                            #+ command substitution.

# You may also use operations within double parentheses without assignment.

  echo "n = $n"                             # n = 0

  (( n += 1 ))                              # Increment.
# (( $n += 1 )) is incorrect!
  echo "n = $n"                             # n = 1

let z=z+3
let "z += 3"  #  Quotes permit the use of spaces in variable assignment.
              #  The 'let' operator actually performs arithmetic evaluation,
              #+ rather than expansion.

Examples of arithmetic expansion in scripts:

  1. Example 16-9

  2. Example 11-15

  3. Example 27-1

  4. Example 27-11

  5. Example A-16

Chapter 14. Recess Time

This bizarre little intermission gives the reader a chance to relax and maybe laugh a bit.


  Fellow Linux user, greetings! You are reading something which
  will bring you luck and good fortune. Just e-mail a copy of
  this document to 10 of your friends. Before making the copies,
  send a 100-line Bash script to the first person on the list
  at the bottom of this letter. Then delete their name and add
  yours to the bottom of the list.

  Don't break the chain! Make the copies within 48 hours.
  Wilfred P. of Brooklyn failed to send out his ten copies and
  woke the next morning to find his job description changed
  to "COBOL programmer." Howard L. of Newport News sent
  out his ten copies and within a month had enough hardware
  to build a 100-node Beowulf cluster dedicated to playing
  Tuxracer. Amelia V. of Chicago laughed at this letter
  and broke the chain. Shortly thereafter, a fire broke out
  in her terminal and she now spends her days writing
  documentation for MS Windows.

  Don't break the chain!  Send out your ten copies today!

Courtesy 'NIX "fortune cookies", with some alterations and many apologies

Part 4. Commands

Mastering the commands on your Linux machine is an indispensable prelude to writing effective shell scripts.

This section covers the following commands:

Chapter 15. Internal Commands and Builtins

A builtin is a command contained within the Bash tool set, literally built in. This is either for performance reasons -- builtins execute faster than external commands, which usually require forking off [58] a separate process -- or because a particular builtin needs direct access to the shell internals.

A builtin may be a synonym to a system command of the same name, but Bash reimplements it internally. For example, the Bash echo command is not the same as /bin/echo, although their behavior is almost identical.

echo "This line uses the \"echo\" builtin."
/bin/echo "This line uses the /bin/echo system command."

A keyword is a reserved word, token or operator. Keywords have a special meaning to the shell, and indeed are the building blocks of the shell's syntax. As examples, for, while, do, and ! are keywords. Similar to a builtin, a keyword is hard-coded into Bash, but unlike a builtin, a keyword is not in itself a command, but a subunit of a command construct. [59]



prints (to stdout) an expression or variable (see Example 4-1).
echo Hello
echo $a

An echo requires the -e option to print escaped characters. See Example 5-2.

Normally, each echo command prints a terminal newline, but the -n option suppresses this.


An echo can be used to feed a sequence of commands down a pipe.

if echo "$VAR" | grep -q txt   # if [[ $VAR = *txt* ]]
  echo "$VAR contains the substring sequence \"txt\""


An echo, in combination with command substitution can set a variable.

a=`echo "HELLO" | tr A-Z a-z`

See also Example 16-22, Example 16-3, Example 16-47, and Example 16-48.

Be aware that echo `command` deletes any linefeeds that the output of command generates.

The $IFS (internal field separator) variable normally contains \n (linefeed) as one of its set of whitespace characters. Bash therefore splits the output of command at linefeeds into arguments to echo. Then echo outputs these arguments, separated by spaces.

bash$ ls -l /usr/share/apps/kjezz/sounds
-rw-r--r--    1 root     root         1407 Nov  7  2000
 -rw-r--r--    1 root     root          362 Nov  7  2000

bash$ echo `ls -l /usr/share/apps/kjezz/sounds`
total 40 -rw-r--r-- 1 root root 716 Nov 7 2000 -rw-r--r-- 1 root root ...

So, how can we embed a linefeed within an echoed character string?
# Embedding a linefeed?
echo "Why doesn't this string \n split on two lines?"
# Doesn't split.

# Let's try something else.

echo $"A line of text containing
a linefeed."
# Prints as two distinct lines (embedded linefeed).
# But, is the "$" variable prefix really necessary?


echo "This string splits
on two lines."
# No, the "$" is not needed.

echo "---------------"

echo -n $"Another line of text containing
a linefeed."
# Prints as two distinct lines (embedded linefeed).
# Even the -n option fails to suppress the linefeed here.

echo "---------------"

# However, the following doesn't work as expected.
# Why not? Hint: Assignment to a variable.
string1=$"Yet another line of text containing
a linefeed (maybe)."

echo $string1
# Yet another line of text containing a linefeed (maybe).
#                                    ^
# Linefeed becomes a space.

# Thanks, Steve Parker, for pointing this out.


This command is a shell builtin, and not the same as /bin/echo, although its behavior is similar.

bash$ type -a echo
echo is a shell builtin
 echo is /bin/echo


The printf, formatted print, command is an enhanced echo. It is a limited variant of the C language printf() library function, and its syntax is somewhat different.

printf format-string... parameter...

This is the Bash builtin version of the /bin/printf or /usr/bin/printf command. See the printf manpage (of the system command) for in-depth coverage.


Older versions of Bash may not support printf.

Example 15-2. printf in action

# printf demo

declare -r PI=3.14159265358979     # Read-only variable, i.e., a constant.
declare -r DecimalConstant=31373



printf "Pi to 2 decimal places = %1.2f" $PI
printf "Pi to 9 decimal places = %1.9f" $PI  # It even rounds off correctly.

printf "\n"                                  # Prints a line feed,
                                             # Equivalent to 'echo' . . .

printf "Constant = \t%d\n" $DecimalConstant  # Inserts tab (\t).

printf "%s %s \n" $Message1 $Message2


# ==========================================#
# Simulation of C function, sprintf().
# Loading a variable with a formatted string.


Pi12=$(printf "%1.12f" $PI)
echo "Pi to 12 decimal places = $Pi12"      # Roundoff error!

Msg=`printf "%s %s \n" $Message1 $Message2`
echo $Msg; echo $Msg

#  As it happens, the 'sprintf' function can now be accessed
#+ as a loadable module to Bash,
#+ but this is not portable.

exit 0

Formatting error messages is a useful application of printf



  printf "$@" >&2
  # Formats positional params passed, and sends them to stderr.
  exit $E_BADDIR

cd $var || error $"Can't cd to %s." "$var"

# Thanks, S.C.

See also Example 36-17.


"Reads" the value of a variable from stdin, that is, interactively fetches input from the keyboard. The -a option lets read get array variables (see Example 27-6).

Example 15-3. Variable assignment, using read

# "Reading" variables.

echo -n "Enter the value of variable 'var1': "
# The -n option to echo suppresses newline.

read var1
# Note no '$' in front of var1, since it is being set.

echo "var1 = $var1"


# A single 'read' statement can set multiple variables.
echo -n "Enter the values of variables 'var2' and 'var3' "
echo =n "(separated by a space or tab): "
read var2 var3
echo "var2 = $var2      var3 = $var3"
#  If you input only one value,
#+ the other variable(s) will remain unset (null).

exit 0

A read without an associated variable assigns its input to the dedicated variable $REPLY.

Example 15-4. What happens when read has no variable



# -------------------------- #
echo -n "Enter a value: "
read var
echo "\"var\" = "$var""
# Everything as expected here.
# -------------------------- #


# ------------------------------------------------------------------- #
echo -n "Enter another value: "
read           #  No variable supplied for 'read', therefore...
               #+ Input to 'read' assigned to default variable, $REPLY.
echo "\"var\" = "$var""
# This is equivalent to the first code block.
# ------------------------------------------------------------------- #

echo "========================="

#  This example is similar to the "" script.
#  However, this one shows that $REPLY is available
#+ even after a 'read' to a variable in the conventional way.

# ================================================================= #

#  In some instances, you might wish to discard the first value read.
#  In such cases, simply ignore the $REPLY variable.

{ # Code block.
read            # Line 1, to be discarded.
read line2      # Line 2, saved in variable.
  } <$0
echo "Line 2 of this script is:"
echo "$line2"   #   #
echo            #   #!/bin/bash  line discarded.

# See also the script.

exit 0

Normally, inputting a \ suppresses a newline during input to a read. The -r option causes an inputted \ to be interpreted literally.

Example 15-5. Multi-line input to read



echo "Enter a string terminated by a \\, then press <ENTER>."
echo "Then, enter a second string (no \\ this time), and again press <ENTER>."

read var1     # The "\" suppresses the newline, when reading $var1.
              #     first line \
              #     second line

echo "var1 = $var1"
#     var1 = first line second line

#  For each line terminated by a "\"
#+ you get a prompt on the next line to continue feeding characters into var1.

echo; echo

echo "Enter another string terminated by a \\ , then press <ENTER>."
read -r var2  # The -r option causes the "\" to be read literally.
              #     first line \

echo "var2 = $var2"
#     var2 = first line \

# Data entry terminates with the first <ENTER>.


exit 0

The read command has some interesting options that permit echoing a prompt and even reading keystrokes without hitting ENTER.

# Read a keypress without hitting ENTER.

read -s -n1 -p "Hit a key " keypress
echo; echo "Keypress was "\"$keypress\""."

# -s option means do not echo input.
# -n N option means accept only N characters of input.
# -p option means echo the following prompt before reading input.

# Using these options is tricky, since they need to be in the correct order.

The -n option to read also allows detection of the arrow keys and certain of the other unusual keys.

Example 15-6. Detecting the arrow keys

# Detects the arrow keys, and a few more.
# Thank you, Sandro Magi, for showing me how.

# --------------------------------------------
# Character codes generated by the keypresses.
# --------------------------------------------


echo -n "Press a key...  "
# May need to also press ENTER if a key not listed above pressed.
read -n3 key                      # Read 3 characters.

echo -n "$key" | grep "$arrowup"  #Check if character code detected.
if [ "$?" -eq $SUCCESS ]
  echo "Up-arrow key pressed."
  exit $SUCCESS

echo -n "$key" | grep "$arrowdown"
if [ "$?" -eq $SUCCESS ]
  echo "Down-arrow key pressed."
  exit $SUCCESS

echo -n "$key" | grep "$arrowrt"
if [ "$?" -eq $SUCCESS ]
  echo "Right-arrow key pressed."
  exit $SUCCESS

echo -n "$key" | grep "$arrowleft"
if [ "$?" -eq $SUCCESS ]
  echo "Left-arrow key pressed."
  exit $SUCCESS

echo -n "$key" | grep "$insert"
if [ "$?" -eq $SUCCESS ]
  echo "\"Insert\" key pressed."
  exit $SUCCESS

echo -n "$key" | grep "$delete"
if [ "$?" -eq $SUCCESS ]
  echo "\"Delete\" key pressed."
  exit $SUCCESS

echo " Some other key pressed."

exit $OTHER

# ========================================= #

#  Mark Alexander came up with a simplified
#+ version of the above script (Thank you!).
#  It eliminates the need for grep.



  read -s -n3 -p "Hit an arrow key: " x

  case "$x" in
     echo "You pressed up-arrow"
     echo "You pressed down-arrow"
     echo "You pressed left-arrow"
     echo "You pressed right-arrow"

exit $?

# ========================================= #

# Antonio Macchi has a simpler alternative.


while true
  read -sn1 a
  test "$a" == `echo -en "\e"` || continue
  read -sn1 a
  test "$a" == "[" || continue
  read -sn1 a
  case "$a" in
    A)  echo "up";;
    B)  echo "down";;
    C)  echo "right";;
    D)  echo "left";;

# ========================================= #

#  Exercise:
#  --------
#  1) Add detection of the "Home," "End," "PgUp," and "PgDn" keys.


The -n option to read will not detect the ENTER (newline) key.

The -t option to read permits timed input (see Example 9-4 and Example A-41).

The -u option takes the file descriptor of the target file.

The read command may also "read" its variable value from a file redirected to stdin. If the file contains more than one line, only the first line is assigned to the variable. If read has more than one parameter, then each of these variables gets assigned a successive whitespace-delineated string. Caution!

Example 15-7. Using read with file redirection


read var1 <data-file
echo "var1 = $var1"
# var1 set to the entire first line of the input file "data-file"

read var2 var3 <data-file
echo "var2 = $var2   var3 = $var3"
# Note non-intuitive behavior of "read" here.
# 1) Rewinds back to the beginning of input file.
# 2) Each variable is now set to a corresponding string,
#    separated by whitespace, rather than to an entire line of text.
# 3) The final variable gets the remainder of the line.
# 4) If there are more variables to be set than whitespace-terminated strings
#    on the first line of the file, then the excess variables remain empty.

echo "------------------------------------------------"

# How to resolve the above problem with a loop:
while read line
  echo "$line"
done <data-file
# Thanks, Heiner Steven for pointing this out.

echo "------------------------------------------------"

# Use $IFS (Internal Field Separator variable) to split a line of input to
# "read", if you do not want the default to be whitespace.

echo "List of all users:"
OIFS=$IFS; IFS=:       # /etc/passwd uses ":" for field separator.
while read name passwd uid gid fullname ignore
  echo "$name ($fullname)"
done </etc/passwd   # I/O redirection.
IFS=$OIFS              # Restore original $IFS.
# This code snippet also by Heiner Steven.

#  Setting the $IFS variable within the loop itself
#+ eliminates the need for storing the original $IFS
#+ in a temporary variable.
#  Thanks, Dim Segebart, for pointing this out.
echo "------------------------------------------------"
echo "List of all users:"

while IFS=: read name passwd uid gid fullname ignore
  echo "$name ($fullname)"
done </etc/passwd   # I/O redirection.

echo "\$IFS still $IFS"

exit 0


Piping output to a read, using echo to set variables will fail.

Yet, piping the output of cat seems to work.

cat file1 file2 |
while read line
echo $line

However, as Bj�n Eriksson shows:

Example 15-8. Problems reading from a pipe

# This example contributed by Bjon Eriksson.

### shopt -s lastpipe

cat $0 |
while read line
    echo "{$line}"

echo "++++++++++++++++++++++"
printf "\nAll done, last: $last\n" #  The output of this line
                                   #+ changes if you uncomment line 5.
                                   #  (Bash, version -ge 4.2 required.)

exit 0  # End of code.
        # (Partial) output of script follows.
        # The 'echo' supplies extra brackets.



{cat $0 |}
{while read line}
{echo "{$line}"}
{printf "nAll done, last: $lastn"}

All done, last: (null)

The variable (last) is set within the loop/subshell
but its value does not persist outside the loop.

The gendiff script, usually found in /usr/bin on many Linux distros, pipes the output of find to a while read construct.
find $1 \( -name "*$2" -o -name ".*$2" \) -print |
while read f; do
. . .


It is possible to paste text into the input field of a read (but not multiple lines!). See Example A-38.



The familiar cd change directory command finds use in scripts where execution of a command requires being in a specified directory.

(cd /source/directory && tar cf - . ) | (cd /dest/directory && tar xpvf -)
[from the previously cited example by Alan Cox]

The -P (physical) option to cd causes it to ignore symbolic links.

cd - changes to $OLDPWD, the previous working directory.


The cd command does not function as expected when presented with two forward slashes.
bash$ cd //
bash$ pwd
The output should, of course, be /. This is a problem both from the command-line and in a script.


Print Working Directory. This gives the user's (or script's) current directory (see Example 15-9). The effect is identical to reading the value of the builtin variable $PWD.

pushd, popd, dirs

This command set is a mechanism for bookmarking working directories, a means of moving back and forth through directories in an orderly manner. A pushdown stack is used to keep track of directory names. Options allow various manipulations of the directory stack.

pushd dir-name pushes the path dir-name onto the directory stack (to the top of the stack) and simultaneously changes the current working directory to dir-name

popd removes (pops) the top directory path name off the directory stack and simultaneously changes the current working directory to the directory now at the top of the stack.

dirs lists the contents of the directory stack (compare this with the $DIRSTACK variable). A successful pushd or popd will automatically invoke dirs.

Scripts that require various changes to the current working directory without hard-coding the directory name changes can make good use of these commands. Note that the implicit $DIRSTACK array variable, accessible from within a script, holds the contents of the directory stack.

Example 15-9. Changing the current working directory



pushd $dir1
# Will do an automatic 'dirs' (list directory stack to stdout).
echo "Now in directory `pwd`." # Uses back-quoted 'pwd'.

# Now, do some stuff in directory 'dir1'.
pushd $dir2
echo "Now in directory `pwd`."

# Now, do some stuff in directory 'dir2'.
echo "The top entry in the DIRSTACK array is $DIRSTACK."
echo "Now back in directory `pwd`."

# Now, do some more stuff in directory 'dir1'.
echo "Now back in original working directory `pwd`."

exit 0

# What happens if you don't 'popd' -- then exit the script?
# Which directory do you end up in? Why?



The let command carries out arithmetic operations on variables. [60] In many cases, it functions as a less complex version of expr.

Example 15-10. Letting let do arithmetic.



let a=11            # Same as 'a=11'
let a=a+5           # Equivalent to  let "a = a + 5"
                    # (Double quotes and spaces make it more readable.)
echo "11 + 5 = $a"  # 16

let "a <<= 3"       # Equivalent to  let "a = a << 3"
echo "\"\$a\" (=16) left-shifted 3 places = $a"
                    # 128

let "a /= 4"        # Equivalent to  let "a = a / 4"
echo "128 / 4 = $a" # 32

let "a -= 5"        # Equivalent to  let "a = a - 5"
echo "32 - 5 = $a"  # 27

let "a *=  10"      # Equivalent to  let "a = a * 10"
echo "27 * 10 = $a" # 270

let "a %= 8"        # Equivalent to  let "a = a % 8"
echo "270 modulo 8 = $a  (270 / 8 = 33, remainder $a)"
                    # 6

# Does "let" permit C-style operators?
# Yes, just as the (( ... )) double-parentheses construct does.

let a++             # C-style (post) increment.
echo "6++ = $a"     # 6++ = 7
let a--             # C-style decrement.
echo "7-- = $a"     # 7-- = 6
# Of course, ++a, etc., also allowed . . .

# Trinary operator.

# Note that $a is 6, see above.
let "t = a<7?7:11"   # True
echo $t  # 7

let a++
let "t = a<7?7:11"   # False
echo $t  #     11



The let command can, in certain contexts, return a surprising exit status.

# Evgeniy Ivanov points out:

echo $?     # 0
            # As expected.

let var++
echo $?     # 1
            # The command was successful, so why isn't $?=0 ???
            # Anomaly!

let var++
echo $?     # 0
            # As expected.

# Likewise . . .

let var=0
echo $?     # 1
            # The command was successful, so why isn't $?=0 ???

#  However, as Jeff Gorak points out,
#+ this is part of the design spec for 'let' . . .
# "If the last ARG evaluates to 0, let returns 1;
#  let returns 0 otherwise." ['help let']


eval arg1 [arg2] ... [argN]

Combines the arguments in an expression or list of expressions and evaluates them. Any variables within the expression are expanded. The net result is to convert a string into a command.


The eval command can be used for code generation from the command-line or within a script.

bash$ command_string="ps ax"
bash$ process="ps ax"
bash$ eval "$command_string" | grep "$process"
26973 pts/3    R+     0:00 grep --color ps ax
 26974 pts/3    R+     0:00 ps ax

Each invocation of eval forces a re-evaluation of its arguments.

echo $a             # $b
                    # First level.
eval echo $a        # $c
                    # Second level.
eval eval echo $a   # d
                    # Third level.

# Thank you, E. Choroba.

Example 15-11. Showing the effect of eval

# Exercising "eval" ...

y=`eval ls -l`  #  Similar to y=`ls -l`
echo $y         #+ but linefeeds removed because "echoed" variable is unquoted.
echo "$y"       #  Linefeeds preserved when variable is quoted.

echo; echo

y=`eval df`     #  Similar to y=`df`
echo $y         #+ but linefeeds removed.

#  When LF's not preserved, it may make it easier to parse output,
#+ using utilities such as "awk".

echo "==========================================================="

eval "`seq 3 | sed -e 's/.*/echo var&=ABCDEFGHIJ/'`"

echo "==========================================================="

# Now, showing how to do something useful with "eval" . . .
# (Thank you, E. Choroba!)

version=3.4     #  Can we split the version into major and minor
                #+ part in one command?
echo "version = $version"
eval major=${version/./;minor=}     #  Replaces '.' in version by ';minor='
                                    #  The substitution yields '3; minor=4'
                                    #+ so eval does minor=4, major=3
echo Major: $major, minor: $minor   #  Major: 3, minor: 4

Example 15-12. Using eval to select among variables


#  Passing arguments to a function to select
#+ one particular variable out of a group.

arr0=( 10 11 12 13 14 15 )
arr1=( 20 21 22 23 24 25 )
arr2=( 30 31 32 33 34 35 )
#       0  1  2  3  4  5      Element number (zero-indexed)

choose_array ()
  eval array_member=\${arr${array_number}[element_number]}
  #                 ^       ^^^^^^^^^^^^
  #  Using eval to construct the name of a variable,
  #+ in this particular case, an array name.

  echo "Element $element_number of array $array_number is $array_member"
} #  Function can be rewritten to take parameters.

array_number=0    # First array.
choose_array      # 13

array_number=2    # Third array.
choose_array      # 34

array_number=3    # Null array (arr3 not allocated).
choose_array      # (null)

# Thank you, Antonio Macchi, for pointing this out.

Example 15-13. Echoing the command-line parameters


# Call this script with a few command-line parameters.
# For example:
#     sh first second third fourth fifth

params=$#              # Number of command-line parameters.
param=1                # Start at first command-line param.

while [ "$param" -le "$params" ]
  echo -n "Command-line parameter "
  echo -n \$$param     #  Gives only the *name* of variable.
#         ^^^          #  $1, $2, $3, etc.
                       #  Why?
                       #  \$ escapes the first "$"
                       #+ so it echoes literally,
                       #+ and $param dereferences "$param" . . .
                       #+ . . . as expected.
  echo -n " = "
  eval echo \$$param   #  Gives the *value* of variable.
# ^^^^      ^^^        #  The "eval" forces the *evaluation*
                       #+ of \$$
                       #+ as an indirect variable reference.

(( param ++ ))         # On to the next.

exit $?

# =================================================

$ sh first second third fourth fifth
Command-line parameter $1 = first
Command-line parameter $2 = second
Command-line parameter $3 = third
Command-line parameter $4 = fourth
Command-line parameter $5 = fifth

Example 15-14. Forcing a log-off

# Killing ppp to force a log-off.
# For dialup connection, of course.

# Script should be run as root user.

#  Depending on the hardware and even the kernel version,
#+ the modem port on your machine may be different --
#+ /dev/ttyS1 or /dev/ttyS2.

killppp="eval kill -9 `ps ax | awk '/ppp/ { print $1 }'`"
#                     -------- process ID of ppp -------  

$killppp                     # This variable is now a command.

# The following operations must be done as root user.

chmod 666 /dev/$SERPORT      # Restore r+w permissions, or else what?
#  Since doing a SIGKILL on ppp changed the permissions on the serial port,
#+ we restore permissions to previous state.

rm /var/lock/LCK..$SERPORT   # Remove the serial port lock file. Why?

exit $?

# Exercises:
# ---------
# 1) Have script check whether root user is invoking it.
# 2) Do a check on whether the process to be killed
#+   is actually running before attempting to kill it.   
# 3) Write an alternate version of this script based on 'fuser':
#+      if [ fuser -s /dev/modem ]; then . . .

Example 15-15. A version of rot13

# A version of "rot13" using 'eval'.
# Compare to "" example.

setvar_rot_13()              # "rot13" scrambling
  local varname=$1 varvalue=$2
  eval $varname='$(echo "$varvalue" | tr a-z n-za-m)'

setvar_rot_13 var "foobar"   # Run "foobar" through rot13.
echo $var                    # sbbone

setvar_rot_13 var "$var"     # Run "sbbone" through rot13.
                             # Back to original variable.
echo $var                    # foobar

# This example by Stephane Chazelas.
# Modified by document author.

exit 0

Here is another example of using eval to evaluate a complex expression, this one from an earlier version of YongYe's Tetris game script.

eval ${1}+=\"${x} ${y} \"

Example A-53 uses eval to convert array elements into a command list.

The eval command occurs in the older version of indirect referencing.
eval var=\$$var


The eval command can be used to parameterize brace expansion.


The eval command can be risky, and normally should be avoided when there exists a reasonable alternative. An eval $COMMANDS executes the contents of COMMANDS, which may contain such unpleasant surprises as rm -rf *. Running an eval on unfamiliar code written by persons unknown is living dangerously.


The set command changes the value of internal script variables/options. One use for this is to toggle option flags which help determine the behavior of the script. Another application for it is to reset the positional parameters that a script sees as the result of a command (set `command`). The script can then parse the fields of the command output.

Example 15-16. Using set with positional parameters

# Script "set-test"

# Invoke this script with three command-line parameters,
# for example, "sh one two three".

echo "Positional parameters before  set \`uname -a\` :"
echo "Command-line argument #1 = $1"
echo "Command-line argument #2 = $2"
echo "Command-line argument #3 = $3"

set `uname -a` # Sets the positional parameters to the output
               # of the command `uname -a`

echo +++++
echo $_        # +++++
# Flags set in script.
echo $-        # hB
#                Anomalous behavior?

echo "Positional parameters after  set \`uname -a\` :"
# $1, $2, $3, etc. reinitialized to result of `uname -a`
echo "Field #1 of 'uname -a' = $1"
echo "Field #2 of 'uname -a' = $2"
echo "Field #3 of 'uname -a' = $3"
echo \#\#\#
echo $_        # ###

exit 0

More fun with positional parameters.

Example 15-17. Reversing the positional parameters

# Reverse positional parameters.
# Script by Dan Jacobson, with stylistic revisions by document author.

set a\ b c d\ e;
#     ^      ^     Spaces escaped 
#       ^ ^        Spaces not escaped
#              ^   Saving old IFS and setting new one.


until [ $# -eq 0 ]
do          #      Step through positional parameters.
  echo "### k0 = "$k""     # Before
  k=$1:$k;  #      Append each pos param to loop variable.
#     ^
  echo "### k = "$k""      # After

set $k  #  Set new positional parameters.
echo -
echo $# #  Count of positional parameters.
echo -

for i   #  Omitting the "in list" sets the variable -- i --
        #+ to the positional parameters.
  echo $i  # Display new positional parameters.

IFS=$OIFS  # Restore IFS.

#  Question:
#  Is it necessary to set an new IFS, internal field separator,
#+ in order for this script to work properly?
#  What happens if you don't? Try it.
#  And, why use the new IFS -- a colon -- in line 17,
#+ to append to the loop variable?
#  What is the purpose of this?

exit 0

$ ./

### k0 = 
### k = a b

### k0 = a b
### k = c a b

### k0 = c a b
### k = d e c a b


d e
a b

Invoking set without any options or arguments simply lists all the environmental and other variables that have been initialized.

bash$ set

Using set with the -- option explicitly assigns the contents of a variable to the positional parameters. If no variable follows the -- it unsets the positional parameters.

Example 15-18. Reassigning the positional parameters


variable="one two three four five"

set -- $variable
# Sets positional parameters to the contents of "$variable".

shift; shift        # Shift past first two positional params.
# shift 2             also works.

echo "first parameter = $first_param"             # one
echo "second parameter = $second_param"           # two
echo "remaining parameters = $remaining_params"   # three four five

echo; echo

# Again.
set -- $variable
echo "first parameter = $first_param"             # one
echo "second parameter = $second_param"           # two

# ======================================================

set --
# Unsets positional parameters if no variable specified.

echo "first parameter = $first_param"             # (null value)
echo "second parameter = $second_param"           # (null value)

exit 0

See also Example 11-2 and Example 16-56.


The unset command deletes a shell variable, effectively setting it to null. Note that this command does not affect positional parameters.

bash$ unset PATH

bash$ echo $PATH


Example 15-19. "Unsetting" a variable

# Unsetting a variable.

variable=hello                       #  Initialized.
echo "variable = $variable"

unset variable                       #  Unset.
                                     #  In this particular context,
                                     #+ same effect as:   variable=
echo "(unset) variable = $variable"  #  $variable is null.

if [ -z "$variable" ]                #  Try a string-length test.
  echo "\$variable has zero length."

exit 0


In most contexts, an undeclared variable and one that has been unset are equivalent. However, the ${parameter:-default} parameter substitution construct can distinguish between the two.


The export [61] command makes available variables to all child processes of the running script or shell. One important use of the export command is in startup files, to initialize and make accessible environmental variables to subsequent user processes.


Unfortunately, there is no way to export variables back to the parent process, to the process that called or invoked the script or shell.

Example 15-20. Using export to pass a variable to an embedded awk script


#  Yet another version of the "column totaler" script (
#+ that adds up a specified column (of numbers) in the target file.
#  This uses the environment to pass a script variable to 'awk' . . .
#+ and places the awk script in a variable.


if [ $# -ne "$ARGS" ] # Check for proper number of command-line args.
   echo "Usage: `basename $0` filename column-number"
   exit $E_WRONGARGS


#===== Same as original script, up to this point =====#

export column_number
# Export column number to environment, so it's available for retrieval.

# -----------------------------------------------
awkscript='{ total += $ENVIRON["column_number"] }
END { print total }'
# Yes, a variable can hold an awk script.
# -----------------------------------------------

# Now, run the awk script.
awk "$awkscript" "$filename"

# Thanks, Stephane Chazelas.

exit 0


It is possible to initialize and export variables in the same operation, as in export var1=xxx.

However, as Greg Keraunen points out, in certain situations this may have a different effect than setting a variable, then exporting it.

bash$ export var=(a b); echo ${var[0]}
(a b)

bash$ var=(a b); export var; echo ${var[0]}


A variable to be exported may require special treatment. See Example M-2.

declare, typeset

The declare and typeset commands specify and/or restrict properties of variables.


Same as declare -r, sets a variable as read-only, or, in effect, as a constant. Attempts to change the variable fail with an error message. This is the shell analog of the C language const type qualifier.


This powerful tool parses command-line arguments passed to the script. This is the Bash analog of the getopt external command and the getopt library function familiar to C programmers. It permits passing and concatenating multiple options [62] and associated arguments to a script (for example scriptname -abc -e /usr/local).

The getopts construct uses two implicit variables. $OPTIND is the argument pointer (OPTion INDex) and $OPTARG (OPTion ARGument) the (optional) argument attached to an option. A colon following the option name in the declaration tags that option as having an associated argument.

A getopts construct usually comes packaged in a while loop, which processes the options and arguments one at a time, then increments the implicit $OPTIND variable to point to the next.


  1. The arguments passed from the command-line to the script must be preceded by a dash (-). It is the prefixed - that lets getopts recognize command-line arguments as options. In fact, getopts will not process arguments without the prefixed -, and will terminate option processing at the first argument encountered lacking them.

  2. The getopts template differs slightly from the standard while loop, in that it lacks condition brackets.

  3. The getopts construct is a highly functional replacement for the traditional getopt external command.

while getopts ":abcde:fg" Option
# Initial declaration.
# a, b, c, d, e, f, and g are the options (flags) expected.
# The : after option 'e' shows it will have an argument passed with it.
  case $Option in
    a ) # Do something with variable 'a'.
    b ) # Do something with variable 'b'.
    e)  # Do something with 'e', and also with $OPTARG,
        # which is the associated argument passed with option 'e'.
    g ) # Do something with variable 'g'.
shift $(($OPTIND - 1))
# Move argument pointer to next.

# All this is not nearly as complicated as it looks <grin>.

Example 15-21. Using getopts to read the options/arguments passed to a script

# Exercising getopts and OPTIND
#          Script modified 10/09/03 at the suggestion of Bill Gradwohl.

# Here we observe how 'getopts' processes command-line arguments to script.
# The arguments are parsed as "options" (flags) and associated arguments.

# Try invoking this script with:
#   'scriptname -mn'
#   'scriptname -oq qOption' (qOption can be some arbitrary string.)
#   'scriptname -qXXX -r'
#   'scriptname -qr'
#+      - Unexpected result, takes "r" as the argument to option "q"
#   'scriptname -q -r' 
#+      - Unexpected result, same as above
#   'scriptname -mnop -mnop'  - Unexpected result
#   (OPTIND is unreliable at stating where an option came from.)
#  If an option expects an argument ("flag:"), then it will grab
#+ whatever is next on the command-line.


if [ $# -eq "$NO_ARGS" ]    # Script invoked with no command-line args?
  echo "Usage: `basename $0` options (-mnopqrs)"
  exit $E_OPTERROR          # Exit and explain usage.
                            # Usage: scriptname -options
                            # Note: dash (-) necessary

while getopts ":mnopq:rs" Option
  case $Option in
    m     ) echo "Scenario #1: option -m-   [OPTIND=${OPTIND}]";;
    n | o ) echo "Scenario #2: option -$Option-   [OPTIND=${OPTIND}]";;
    p     ) echo "Scenario #3: option -p-   [OPTIND=${OPTIND}]";;
    q     ) echo "Scenario #4: option -q-\
                  with argument \"$OPTARG\"   [OPTIND=${OPTIND}]";;
    #  Note that option 'q' must have an associated argument,
    #+ otherwise it falls through to the default.
    r | s ) echo "Scenario #5: option -$Option-";;
    *     ) echo "Unimplemented option chosen.";;   # Default.

shift $(($OPTIND - 1))
#  Decrements the argument pointer so it points to next argument.
#  $1 now references the first non-option item supplied on the command-line
#+ if one exists.

exit $?

#   As Bill Gradwohl states,
#  "The getopts mechanism allows one to specify:  scriptname -mnop -mnop
#+  but there is no reliable way to differentiate what came
#+ from where by using OPTIND."
#  There are, however, workarounds.

Script Behavior

source, . (dot command)

This command, when invoked from the command-line, executes a script. Within a script, a source file-name loads the file file-name. Sourcing a file (dot-command) imports code into the script, appending to the script (same effect as the #include directive in a C program). The net result is the same as if the "sourced" lines of code were physically present in the body of the script. This is useful in situations when multiple scripts use a common data file or function library.

Example 15-22. "Including" a data file

#  Note that this example must be invoked with bash, i.e., bash
#+ not  sh !

. data-file    # Load a data file.
# Same effect as "source data-file", but more portable.

#  The file "data-file" must be present in current working directory,
#+ since it is referred to by its basename.

# Now, let's reference some data from that file.

echo "variable1 (from data-file) = $variable1"
echo "variable3 (from data-file) = $variable3"

let "sum = $variable2 + $variable4"
echo "Sum of variable2 + variable4 (from data-file) = $sum"
echo "message1 (from data-file) is \"$message1\""
#                                  Escaped quotes
echo "message2 (from data-file) is \"$message2\""

print_message This is the message-print function in the data-file.

exit $?

File data-file for Example 15-22, above. Must be present in same directory.

# This is a data file loaded by a script.
# Files of this type may contain variables, functions, etc.
# It loads with a 'source' or '.' command from a shell script.

# Let's initialize some variables.


message1="Greetings from *** line $LINENO *** of the data file!"
message2="Enough for now. Goodbye."

print_message ()
{   # Echoes any message passed to it.

  if [ -z "$1" ]
    return 1 # Error, if argument missing.


  until [ -z "$1" ]
  do             # Step through arguments passed to function.
    echo -n "$1" # Echo args one at a time, suppressing line feeds.
    echo -n " "  # Insert spaces between words.
    shift        # Next one.


  return 0

If the sourced file is itself an executable script, then it will run, then return control to the script that called it. A sourced executable script may use a return for this purpose.

Arguments may be (optionally) passed to the sourced file as positional parameters.
source $filename $arg1 arg2

It is even possible for a script to source itself, though this does not seem to have any practical applications.

Example 15-23. A (useless) script that sources itself

# a script sourcing itself "recursively."
# From "Stupid Script Tricks," Volume II.

MAXPASSCNT=100    # Maximum number of execution passes.

echo -n  "$pass_count  "
#  At first execution pass, this just echoes two blank spaces,
#+ since $pass_count still uninitialized.

let "pass_count += 1"
#  Assumes the uninitialized variable $pass_count
#+ can be incremented the first time around.
#  This works with Bash and pdksh, but
#+ it relies on non-portable (and possibly dangerous) behavior.
#  Better would be to initialize $pass_count to 0 before incrementing.

while [ "$pass_count" -le $MAXPASSCNT ]
  . $0   # Script "sources" itself, rather than calling itself.
         # ./$0 (which would be true recursion) doesn't work here. Why?

#  What occurs here is not actually recursion,
#+ since the script effectively "expands" itself, i.e.,
#+ generates a new section of code
#+ with each pass through the 'while' loop',
#  with each 'source' in line 20.
#  Of course, the script interprets each newly 'sourced' "#!" line
#+ as a comment, and not as the start of a new script.


exit 0   # The net effect is counting from 1 to 100.
         # Very impressive.

# Exercise:
# --------
# Write a script that uses this trick to actually do something useful.

Unconditionally terminates a script. [63] The exit command may optionally take an integer argument, which is returned to the shell as the exit status of the script. It is good practice to end all but the simplest scripts with an exit 0, indicating a successful run.


If a script terminates with an exit lacking an argument, the exit status of the script is the exit status of the last command executed in the script, not counting the exit. This is equivalent to an exit $?.


An exit command may also be used to terminate a subshell.


This shell builtin replaces the current process with a specified command. Normally, when the shell encounters a command, it forks off a child process to actually execute the command. Using the exec builtin, the shell does not fork, and the command exec'ed replaces the shell. When used in a script, therefore, it forces an exit from the script when the exec'ed command terminates. [64]

Example 15-24. Effects of exec


exec echo "Exiting \"$0\" at line $LINENO."   # Exit from script here.
# $LINENO is an internal Bash variable set to the line number it's on.

# ----------------------------------
# The following lines never execute.

echo "This echo fails to echo."

exit 99                       #  This script will not exit here.
                              #  Check exit value after script terminates
                              #+ with an 'echo $?'.
                              #  It will *not* be 99.

Example 15-25. A script that exec's itself


# Note: Set permissions on this script to 555 or 755,
#       then call it with ./ or sh ./


echo "This line appears ONCE in the script, yet it keeps echoing."
echo "The PID of this instance of the script is still $$."
#     Demonstrates that a subshell is not forked off.

echo "==================== Hit Ctl-C to exit ===================="

sleep 1

exec $0   #  Spawns another instance of this same script
          #+ that replaces the previous one.

echo "This line will never echo!"  # Why not?

exit 99                            # Will not exit here!
                                   # Exit code will not be 99!

An exec also serves to reassign file descriptors. For example, exec <zzz-file replaces stdin with the file zzz-file.


The -exec option to find is not the same as the exec shell builtin.


This command permits changing shell options on the fly (see Example 25-1 and Example 25-2). It often appears in the Bash startup files, but also has its uses in scripts. Needs version 2 or later of Bash.

shopt -s cdspell
# Allows minor misspelling of directory names with 'cd'
# Option -s sets, -u unsets.

cd /hpme  # Oops! Mistyped '/home'.
pwd       # /home
          # The shell corrected the misspelling.


Putting a caller command inside a function echoes to stdout information about the caller of that function.


function1 ()
  # Inside function1 ().
  caller 0   # Tell me about it.

function1    # Line 9 of script.

# 9 main
# ^                 Line number that the function was called from.
#   ^^^^            Invoked from "main" part of script.
#        ^^^^^^^    Name of calling script.

caller 0     # Has no effect because it's not inside a function.

A caller command can also return caller information from a script sourced within another script. Analogous to a function, this is a "subroutine call."

You may find this command useful in debugging.



A command that returns a successful (zero) exit status, but does nothing else.

bash$ true
bash$ echo $?

# Endless loop
while true   # alias for ":"
   # Need a way to break out of loop or script will hang.


A command that returns an unsuccessful exit status, but does nothing else.

bash$ false
bash$ echo $?

# Testing "false" 
if false
  echo "false evaluates \"true\""
  echo "false evaluates \"false\""
# false evaluates "false"

# Looping while "false" (null loop)
while false
   # The following code will not execute.
   # Nothing happens!

type [cmd]

Similar to the which external command, type cmd identifies "cmd." Unlike which, type is a Bash builtin. The useful -a option to type identifies keywords and builtins, and also locates system commands with identical names.

bash$ type '['
[ is a shell builtin
bash$ type -a '['
[ is a shell builtin
 [ is /usr/bin/[

bash$ type type
type is a shell builtin

The type command can be useful for testing whether a certain command exists.

hash [cmds]

Records the path name of specified commands -- in the shell hash table [65] -- so the shell or script will not need to search the $PATH on subsequent calls to those commands. When hash is called with no arguments, it simply lists the commands that have been hashed. The -r option resets the hash table.


The bind builtin displays or modifies readline [66] key bindings.


Gets a short usage summary of a shell builtin. This is the counterpart to whatis, but for builtins. The display of help information got a much-needed update in the version 4 release of Bash.

bash$ help exit
exit: exit [n]
    Exit the shell with a status of N.  If N is omitted, the exit status
    is that of the last command executed.

15.1. Job Control Commands

Certain of the following job control commands take a job identifier as an argument. See the table at end of the chapter.


Lists the jobs running in the background, giving the job number. Not as useful as ps.


It is all too easy to confuse jobs and processes. Certain builtins, such as kill, disown, and wait accept either a job number or a process number as an argument. The fg, bg and jobs commands accept only a job number.

bash$ sleep 100 &
[1] 1384

bash $ jobs
[1]+  Running                 sleep 100 &

"1" is the job number (jobs are maintained by the current shell). "1384" is the PID or process ID number (processes are maintained by the system). To kill this job/process, either a kill %1 or a kill 1384 works.

Thanks, S.C.


Remove job(s) from the shell's table of active jobs.

fg, bg

The fg command switches a job running in the background into the foreground. The bg command restarts a suspended job, and runs it in the background. If no job number is specified, then the fg or bg command acts upon the currently running job.


Suspend script execution until all jobs running in background have terminated, or until the job number or process ID specified as an option terminates. Returns the exit status of waited-for command.

You may use the wait command to prevent a script from exiting before a background job finishes executing (this would create a dreaded orphan process).

Example 15-26. Waiting for a process to finish before proceeding


ROOT_UID=0   # Only users with $UID 0 have root privileges.

if [ "$UID" -ne "$ROOT_UID" ]
  echo "Must be root to run this script."
  # "Run along kid, it's past your bedtime."
  exit $E_NOTROOT

if [ -z "$1" ]
  echo "Usage: `basename $0` find-string"
  exit $E_NOPARAMS

echo "Updating 'locate' database..."
echo "This may take a while."
updatedb /usr &     # Must be run as root.

# Don't run the rest of the script until 'updatedb' finished.
# You want the the database updated before looking up the file name.

locate $1

#  Without the 'wait' command, in the worse case scenario,
#+ the script would exit while 'updatedb' was still running,
#+ leaving it as an orphan process.

exit 0

Optionally, wait can take a job identifier as an argument, for example, wait%1 or wait $PPID. [67] See the job id table.


Within a script, running a command in the background with an ampersand (&) may cause the script to hang until ENTER is hit. This seems to occur with commands that write to stdout. It can be a major annoyance.

ls -l &
echo "Done."
bash$ ./
 [bozo@localhost test-scripts]$ total 1
 -rwxr-xr-x    1 bozo     bozo           34 Oct 11 15:09

    As Walter Brameld IV explains it:

    As far as I can tell, such scripts don't actually hang. It just
    seems that they do because the background command writes text to
    the console after the prompt. The user gets the impression that
    the prompt was never displayed. Here's the sequence of events:

    1. Script launches background command.
    2. Script exits.
    3. Shell displays the prompt.
    4. Background command continues running and writing text to the
    5. Background command finishes.
    6. User doesn't see a prompt at the bottom of the output, thinks script
       is hanging.

Placing a wait after the background command seems to remedy this.

ls -l &
echo "Done."
bash$ ./
 [bozo@localhost test-scripts]$ total 1
 -rwxr-xr-x    1 bozo     bozo           34 Oct 11 15:09
Redirecting the output of the command to a file or even to /dev/null also takes care of this problem.


This has a similar effect to Control-Z, but it suspends the shell (the shell's parent process should resume it at an appropriate time).


Exit a login shell, optionally specifying an exit status.


Gives statistics on the system time elapsed when executing commands, in the following form:
0m0.020s 0m0.020s

This capability is of relatively limited value, since it is not common to profile and benchmark shell scripts.


Forcibly terminate a process by sending it an appropriate terminate signal (see Example 17-6).

Example 15-27. A script that kills itself


kill $$  # Script kills its own process here.
         # Recall that "$$" is the script's PID.

echo "This line will not echo."
# Instead, the shell sends a "Terminated" message to stdout.

exit 0   # Normal exit? No!

#  After this script terminates prematurely,
#+ what exit status does it return?
# sh
# echo $?
# 143
# 143 = 128 + 15
#             TERM signal


kill -l lists all the signals (as does the file /usr/include/asm/signal.h). A kill -9 is a sure kill, which will usually terminate a process that stubbornly refuses to die with a plain kill. Sometimes, a kill -15 works. A zombie process, that is, a child process that has terminated, but that the parent process has not (yet) killed, cannot be killed by a logged-on user -- you can't kill something that is already dead -- but init will generally clean it up sooner or later.


The killall command kills a running process by name, rather than by process ID. If there are multiple instances of a particular command running, then doing a killall on that command will terminate them all.


This refers to the killall command in /usr/bin, not the killall script in /etc/rc.d/init.d.


The command directive disables aliases and functions for the command immediately following it.

bash$ command ls


This is one of three shell directives that effect script command processing. The others are builtin and enable.


Invoking builtin BUILTIN_COMMAND runs the command BUILTIN_COMMAND as a shell builtin, temporarily disabling both functions and external system commands with the same name.


This either enables or disables a shell builtin command. As an example, enable -n kill disables the shell builtin kill, so that when Bash subsequently encounters kill, it invokes the external command /bin/kill.

The -a option to enable lists all the shell builtins, indicating whether or not they are enabled. The -f filename option lets enable load a builtin as a shared library (DLL) module from a properly compiled object file. [68].


This is a port to Bash of the ksh autoloader. With autoload in place, a function with an autoload declaration will load from an external file at its first invocation. [69] This saves system resources.

Note that autoload is not a part of the core Bash installation. It needs to be loaded in with enable -f (see above).

Table 15-1. Job identifiers

%NJob number [N]
%SInvocation (command-line) of job begins with string S
%?SInvocation (command-line) of job contains within it string S
%%"current" job (last job stopped in foreground or started in background)
%+"current" job (last job stopped in foreground or started in background)
%-Last job
$!Last background process

Chapter 16. External Filters, Programs and Commands

Standard UNIX commands make shell scripts more versatile. The power of scripts comes from coupling system commands and shell directives with simple programming constructs.

16.1. Basic Commands

The first commands a novice learns


The basic file "list" command. It is all too easy to underestimate the power of this humble command. For example, using the -R, recursive option, ls provides a tree-like listing of a directory structure. Other useful options are -S, sort listing by file size, -t, sort by file modification time, -v, sort by (numerical) version numbers embedded in the filenames, [70] -b, show escape characters, and -i, show file inodes (see Example 16-4).

bash$ ls -l
-rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter10.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter11.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter12.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter1.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter2.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter3.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:49 Chapter_headings.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:49 Preface.txt

bash$ ls -lv
 total 0
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:49 Chapter_headings.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:49 Preface.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter1.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter2.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter3.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter10.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter11.txt
 -rw-rw-r-- 1 bozo bozo 0 Sep 14 18:44 chapter12.txt


The ls command returns a non-zero exit status when attempting to list a non-existent file.
bash$ ls abc
ls: abc: No such file or directory

bash$ echo $?

Example 16-1. Using ls to create a table of contents for burning a CDR disk

# (
# Script to automate burning a CDR.

SPEED=10         # May use higher speed if your hardware supports it.
# DEVICE=/dev/cdrom     For older versions of cdrecord
DEFAULTDIR=/opt  # This is the directory containing the data to be burned.
                 # Make sure it exists.
                 # Exercise: Add a test for this.

# Uses Joerg Schilling's "cdrecord" package:

#  If this script invoked as an ordinary user, may need to suid cdrecord
#+ chmod u+s /usr/bin/cdrecord, as root.
#  Of course, this creates a security hole, though a relatively minor one.

if [ -z "$1" ]
  # Default directory, if not specified on command-line.

# Create a "table of contents" file.
# The "l" option gives a "long" file listing.
# The "R" option makes the listing recursive.
# The "F" option marks the file types (directories get a trailing /).
echo "Creating table of contents."

# Create an image file preparatory to burning it onto the CDR.
echo "Creating ISO9660 file system image ($IMAGEFILE)."

# Burn the CDR.
echo "Burning the disk."
echo "Please be patient, this will take a while."
wodim -v -isosize dev=$DEVICE $IMAGEFILE
#  In newer Linux distros, the "wodim" utility assumes the
#+ functionality of "cdrecord."
echo "Exit code = $exitcode"

exit $exitcode
cat, tac

cat, an acronym for concatenate, lists a file to stdout. When combined with redirection (> or >>), it is commonly used to concatenate files.
# Uses of 'cat'
cat filename                          # Lists the file.

cat file.1 file.2 file.3 > file.123   # Combines three files into one.
The -n option to cat inserts consecutive numbers before all lines of the target file(s). The -b option numbers only the non-blank lines. The -v option echoes nonprintable characters, using ^ notation. The -s option squeezes multiple consecutive blank lines into a single blank line.

See also Example 16-28 and Example 16-24.


In a pipe, it may be more efficient to redirect the stdin to a file, rather than to cat the file.

cat filename | tr a-z A-Z

tr a-z A-Z < filename   #  Same effect, but starts one less process,
                        #+ and also dispenses with the pipe.

tac, is the inverse of cat, listing a file backwards from its end.


reverses each line of a file, and outputs to stdout. This does not have the same effect as tac, as it preserves the order of the lines, but flips each one around (mirror image).

bash$ cat file1.txt
This is line 1.
 This is line 2.

bash$ tac file1.txt
This is line 2.
 This is line 1.

bash$ rev file1.txt
.1 enil si sihT
 .2 enil si sihT


This is the file copy command. cp file1 file2 copies file1 to file2, overwriting file2 if it already exists (see Example 16-6).


Particularly useful are the -a archive flag (for copying an entire directory tree), the -u update flag (which prevents overwriting identically-named newer files), and the -r and -R recursive flags.

cp -u source_dir/* dest_dir
#  "Synchronize" dest_dir to source_dir
#+  by copying over all newer and not previously existing files.


This is the file move command. It is equivalent to a combination of cp and rm. It may be used to move multiple files to a directory, or even to rename a directory. For some examples of using mv in a script, see Example 10-11 and Example A-2.


When used in a non-interactive script, mv takes the -f (force) option to bypass user input.

When a directory is moved to a preexisting directory, it becomes a subdirectory of the destination directory.

bash$ mv source_directory target_directory

bash$ ls -lF target_directory
total 1
 drwxrwxr-x    2 bozo  bozo      1024 May 28 19:20 source_directory/


Delete (remove) a file or files. The -f option forces removal of even readonly files, and is useful for bypassing user input in a script.


The rm command will, by itself, fail to remove filenames beginning with a dash. Why? Because rm sees a dash-prefixed filename as an option.

bash$ rm -badname
rm: invalid option -- b
 Try `rm --help' for more information.

One clever workaround is to precede the filename with a " -- " (the end-of-options flag).
bash$ rm -- -badname

Another method to is to preface the filename to be removed with a dot-slash .
bash$ rm ./-badname


When used with the recursive flag -r, this command removes files all the way down the directory tree from the current directory. A careless rm -rf * can wipe out a big chunk of a directory structure.


Remove directory. The directory must be empty of all files -- including "invisible" dotfiles [71] -- for this command to succeed.


Make directory, creates a new directory. For example, mkdir -p project/programs/December creates the named directory. The -p option automatically creates any necessary parent directories.


Changes the attributes of an existing file or directory (see Example 15-14).

chmod +x filename
# Makes "filename" executable for all users.

chmod u+s filename
# Sets "suid" bit on "filename" permissions.
# An ordinary user may execute "filename" with same privileges as the file's owner.
# (This does not apply to shell scripts.)

chmod 644 filename
#  Makes "filename" readable/writable to owner, readable to others
#+ (octal mode).

chmod 444 filename
#  Makes "filename" read-only for all.
#  Modifying the file (for example, with a text editor)
#+ not allowed for a user who does not own the file (except for root),
#+ and even the file owner must force a file-save
#+ if she modifies the file.
#  Same restrictions apply for deleting the file.

chmod 1777 directory-name
#  Gives everyone read, write, and execute permission in directory,
#+ however also sets the "sticky bit".
#  This means that only the owner of the directory,
#+ owner of the file, and, of course, root
#+ can delete any particular file in that directory.

chmod 111 directory-name
#  Gives everyone execute-only permission in a directory.
#  This means that you can execute and READ the files in that directory
#+ (execute permission necessarily includes read permission
#+ because you can't execute a file without being able to read it).
#  But you can't list the files or search for them with the "find" command.
#  These restrictions do not apply to root.

chmod 000 directory-name
#  No permissions at all for that directory.
#  Can't read, write, or execute files in it.
#  Can't even list files in it or "cd" to it.
#  But, you can rename (mv) the directory
#+ or delete it (rmdir) if it is empty.
#  You can even symlink to files in the directory,
#+ but you can't read, write, or execute the symlinks.
#  These restrictions do not apply to root.


Change file attributes. This is analogous to chmod above, but with different options and a different invocation syntax, and it works only on ext2/ext3 filesystems.

One particularly interesting chattr option is i. A chattr +i filename marks the file as immutable. The file cannot be modified, linked to, or deleted, not even by root. This file attribute can be set or removed only by root. In a similar fashion, the a option marks the file as append only.

root# chattr +i file1.txt

root# rm file1.txt

rm: remove write-protected regular file `file1.txt'? y
 rm: cannot remove `file1.txt': Operation not permitted

If a file has the s (secure) attribute set, then when it is deleted its block is overwritten with binary zeroes. [72]

If a file has the u (undelete) attribute set, then when it is deleted, its contents can still be retrieved (undeleted).

If a file has the c (compress) attribute set, then it will automatically be compressed on writes to disk, and uncompressed on reads.


The file attributes set with chattr do not show in a file listing (ls -l).


Creates links to pre-existings files. A "link" is a reference to a file, an alternate name for it. The ln command permits referencing the linked file by more than one name and is a superior alternative to aliasing (see Example 4-6).

The ln creates only a reference, a pointer to the file only a few bytes in size.

The ln command is most often used with the -s, symbolic or "soft" link flag. Advantages of using the -s flag are that it permits linking across file systems or to directories.

The syntax of the command is a bit tricky. For example: ln -s oldfile newfile links the previously existing oldfile to the newly created link, newfile.


If a file named newfile has previously existed, an error message will result.

Links give the ability to invoke a script (or any other type of executable) with multiple names, and having that script behave according to how it was invoked.

Example 16-2. Hello or Good-bye

# Saying "hello" or "goodbye"
#+          depending on how script is invoked.

# Make a link in current working directory ($PWD) to this script:
#    ln -s goodbye
# Now, try invoking this script both ways:
# ./
# ./goodbye


if [ $0 = "./goodbye" ]
  echo "Good-bye!"
  # Some other goodbye-type commands, as appropriate.

echo "Hello!"
# Some other hello-type commands, as appropriate.
man, info

These commands access the manual and information pages on system commands and installed utilities. When available, the info pages usually contain more detailed descriptions than do the man pages.

There have been various attempts at "automating" the writing of man pages. For a script that makes a tentative first step in that direction, see Example A-39.

16.2. Complex Commands

Commands for more advanced users


-exec COMMAND \;

Carries out COMMAND on each file that find matches. The command sequence terminates with ; (the ";" is escaped to make certain the shell passes it to find literally, without interpreting it as a special character).

bash$ find ~/ -name '*.txt'

If COMMAND contains {}, then find substitutes the full path name of the selected file for "{}".

find ~/ -name 'core*' -exec rm {} \;
# Removes all core dump files from user's home directory.

find /home/bozo/projects -mtime -1
#                               ^   Note minus sign!
#  Lists all files in /home/bozo/projects directory tree
#+ that were modified within the last day (current_day - 1).
find /home/bozo/projects -mtime 1
#  Same as above, but modified *exactly* one day ago.
#  mtime = last modification time of the target file
#  ctime = last status change time (via 'chmod' or otherwise)
#  atime = last access time

find "$DIR" -type f -atime +5 -exec rm {} \;
#                          ^           ^^
#  Curly brackets are placeholder for the path name output by "find."
#  Deletes all files in "/home/bozo/junk_files"
#+ that have not been accessed in *at least* 5 days (plus sign ... +5).
#  "-type filetype", where
#  f = regular file
#  d = directory
#  l = symbolic link, etc.
#  (The 'find' manpage and info page have complete option listings.)

find /etc -exec grep '[0-9][0-9]*[.][0-9][0-9]*[.][0-9][0-9]*[.][0-9][0-9]*' {} \;

# Finds all IP addresses ( in /etc directory files.
# There a few extraneous hits. Can they be filtered out?

# Possibly by:

find /etc -type f -exec cat '{}' \; | tr -c '.[:digit:]' '\n' \
| grep '^[^.][^.]*\.[^.][^.]*\.[^.][^.]*\.[^.][^.]*$'
#  [:digit:] is one of the character classes
#+ introduced with the POSIX 1003.2 standard. 

# Thanks, St�phane Chazelas. 


The -exec option to find should not be confused with the exec shell builtin.

Example 16-3. Badname, eliminate file names in current directory containing bad characters and whitespace.

# Delete filenames in current directory containing bad characters.

for filename in *
  badname=`echo "$filename" | sed -n /[\+\{\;\"\\\=\?~\(\)\<\>\&\*\|\$]/p`
# badname=`echo "$filename" | sed -n '/[+{;"\=?~()<>&*|$]/p'`  also works.
# Deletes files containing these nasties:     + { ; " \ = ? ~ ( ) < > & * | $
  rm $badname 2>/dev/null
#             ^^^^^^^^^^^ Error messages deep-sixed.

# Now, take care of files containing all manner of whitespace.
find . -name "* *" -exec rm -f {} \;
# The path name of the file that _find_ finds replaces the "{}".
# The '\' ensures that the ';' is interpreted literally, as end of command.

exit 0

# Commands below this line will not execute because of _exit_ command.

# An alternative to the above script:
find . -name '*[+{;"\\=?~()<>&*|$ ]*' -maxdepth 0 \
-exec rm -f '{}' \;
#  The "-maxdepth 0" option ensures that _find_ will not search
#+ subdirectories below $PWD.

# (Thanks, S.C.)

Example 16-4. Deleting a file by its inode number

# Deleting a file by its inode number.

#  This is useful when a filename starts with an illegal character,
#+ such as ? or -.

ARGCOUNT=1                      # Filename arg must be passed to script.

if [ $# -ne "$ARGCOUNT" ]
  echo "Usage: `basename $0` filename"

if [ ! -e "$1" ]
  echo "File \""$1"\" does not exist."

inum=`ls -i | grep "$1" | awk '{print $1}'`
# inum = inode (index node) number of file
# -----------------------------------------------------------------------
# Every file has an inode, a record that holds its physical address info.
# -----------------------------------------------------------------------

echo; echo -n "Are you absolutely sure you want to delete \"$1\" (y/n)? "
# The '-v' option to 'rm' also asks this.
read answer
case "$answer" in
[nN]) echo "Changed your mind, huh?"
      exit $E_CHANGED_MIND
*)    echo "Deleting file \"$1\".";;

find . -inum $inum -exec rm {} \;
#                           ^^
#        Curly brackets are placeholder
#+       for text output by "find."
echo "File "\"$1"\" deleted!"

exit 0

The find command also works without the -exec option.

#  Find suid root files.
#  A strange suid file might indicate a security hole,
#+ or even a system intrusion.

# Might also try /sbin, /bin, /usr/bin, /usr/local/bin, etc.
permissions="+4000"  # suid root (dangerous!)

for file in $( find "$directory" -perm "$permissions" )
  ls -ltF --author "$file"

See Example 16-30, Example 3-4, and Example 11-10 for scripts using find. Its manpage provides more detail on this complex and powerful command.


A filter for feeding arguments to a command, and also a tool for assembling the commands themselves. It breaks a data stream into small enough chunks for filters and commands to process. Consider it as a powerful replacement for backquotes. In situations where command substitution fails with a too many arguments error, substituting xargs often works. [73] Normally, xargs reads from stdin or from a pipe, but it can also be given the output of a file.

The default command for xargs is echo. This means that input piped to xargs may have linefeeds and other whitespace characters stripped out.

bash$ ls -l
total 0
 -rw-rw-r--    1 bozo  bozo         0 Jan 29 23:58 file1
 -rw-rw-r--    1 bozo  bozo         0 Jan 29 23:58 file2

bash$ ls -l | xargs
total 0 -rw-rw-r-- 1 bozo bozo 0 Jan 29 23:58 file1 -rw-rw-r-- 1 bozo bozo 0 Jan...

bash$ find ~/mail -type f | xargs grep "Linux"
./misc:User-Agent: slrn/ (Linux)
 ./sent-mail-jul-2005: hosted by the Linux Documentation Project.
 ./sent-mail-jul-2005: (Linux Documentation Project Site, rtf version)
 ./sent-mail-jul-2005: Subject: Criticism of Bozo's Windows/Linux article
 ./sent-mail-jul-2005: while mentioning that the Linux ext2/ext3 filesystem
 . . .

ls | xargs -p -l gzip gzips every file in current directory, one at a time, prompting before each operation.


Note that xargs processes the arguments passed to it sequentially, one at a time.

bash$ find /usr/bin | xargs file
/usr/bin:          directory
 /usr/bin/foomatic-ppd-options:          perl script text executable
 . . .


An interesting xargs option is -n NN, which limits to NN the number of arguments passed.

ls | xargs -n 8 echo lists the files in the current directory in 8 columns.


Another useful option is -0, in combination with find -print0 or grep -lZ. This allows handling arguments containing whitespace or quotes.

find / -type f -print0 | xargs -0 grep -liwZ GUI | xargs -0 rm -f

grep -rliwZ GUI / | xargs -0 rm -f

Either of the above will remove any file containing "GUI". (Thanks, S.C.)

cat /proc/"$pid"/"$OPTION" | xargs -0 echo
#  Formats output:         ^^^^^^^^^^^^^^^
#  From Han Holl's fixup of ""
#+ script in "/dev and /proc" chapter.


The -P option to xargs permits running processes in parallel. This speeds up execution in a machine with a multicore CPU.


ls *gif | xargs -t -n1 -P2 gif2png
# Converts all the gif images in current directory to png.

# Options:
# =======
# -t    Print command to stderr.
# -n1   At most 1 argument per command line.
# -P2   Run up to 2 processes simultaneously.

# Thank you, Roberto Polli, for the inspiration.

Example 16-5. Logfile: Using xargs to monitor system log


# Generates a log file in current directory
# from the tail end of /var/log/messages.

# Note: /var/log/messages must be world readable
# if this script invoked by an ordinary user.
#         #root chmod 644 /var/log/messages


( date; uname -a ) >>logfile
# Time and machine name
echo ---------------------------------------------------------- >>logfile
tail -n $LINES /var/log/messages | xargs | fmt -s >>logfile
echo >>logfile
echo >>logfile

exit 0

#  Note:
#  ----
#  As Frank Wang points out,
#+ unmatched quotes (either single or double quotes) in the source file
#+ may give xargs indigestion.
#  He suggests the following substitution for line 15:
#  tail -n $LINES /var/log/messages | tr -d "\"'" | xargs | fmt -s >>logfile

#  Exercise:
#  --------
#  Modify this script to track changes in /var/log/messages at intervals
#+ of 20 minutes.
#  Hint: Use the "watch" command. 

As in find, a curly bracket pair serves as a placeholder for replacement text.

Example 16-6. Copying files in current directory to another


#  Copy (verbose) all files in current directory ($PWD)
#+ to directory specified on command-line.


if [ -z "$1" ]   # Exit if no argument given.
  echo "Usage: `basename $0` directory-to-copy-to"
  exit $E_NOARGS

ls . | xargs -i -t cp ./{} $1
#            ^^ ^^      ^^
#  -t is "verbose" (output command-line to stderr) option.
#  -i is "replace strings" option.
#  {} is a placeholder for output text.
#  This is similar to the use of a curly-bracket pair in "find."
#  List the files in current directory (ls .),
#+ pass the output of "ls" as arguments to "xargs" (-i -t options),
#+ then copy (cp) these arguments ({}) to new directory ($1).  
#  The net result is the exact equivalent of
#+   cp * $1
#+ unless any of the filenames has embedded "whitespace" characters.

exit 0

Example 16-7. Killing processes by name

# Killing processes by name.
# Compare this script with

#  For instance,
#+ try "./ xterm" --
#+ and watch all the xterms on your desktop disappear.

#  Warning:
#  -------
#  This is a fairly dangerous script.
#  Running it carelessly (especially as root)
#+ can cause data loss and other undesirable effects.


if test -z "$1"  # No command-line arg supplied?
  echo "Usage: `basename $0` Process(es)_to_kill"
  exit $E_BADARGS

ps ax | grep "$PROCESS_NAME" | awk '{print $1}' | xargs -i kill {} 2&>/dev/null
#                                                       ^^      ^^

# ---------------------------------------------------------------
# Notes:
# -i is the "replace strings" option to xargs.
# The curly brackets are the placeholder for the replacement.
# 2&>/dev/null suppresses unwanted error messages.
# Can  grep "$PROCESS_NAME" be replaced by pidof "$PROCESS_NAME"?
# ---------------------------------------------------------------

exit $?

#  The "killall" command has the same effect as this script,
#+ but using it is not quite as educational.

Example 16-8. Word frequency analysis using xargs

# Crude word frequency analysis on a text file.

# Uses 'xargs' to decompose lines of text into single words.
# Compare this example to the "" script later on.

# Check for input file on command-line.

if [ $# -ne "$ARGS" ]
# Correct number of arguments passed to script?
  echo "Usage: `basename $0` filename"
  exit $E_BADARGS

if [ ! -f "$1" ]       # Does file exist?
  echo "File \"$1\" does not exist."
  exit $E_NOFILE

cat "$1" | xargs -n1 | \
#  List the file, one word per line. 
tr A-Z a-z | \
#  Shift characters to lowercase.
sed -e 's/\.//g'  -e 's/\,//g' -e 's/ /\
/g' | \
#  Filter out periods and commas, and
#+ change space between words to linefeed,
sort | uniq -c | sort -nr
#  Finally remove duplicates, prefix occurrence count
#+ and sort numerically.

#  This does the same job as the "" example,
#+ but a bit more ponderously, and it runs more slowly (why?).

exit $?

All-purpose expression evaluator: Concatenates and evaluates the arguments according to the operation given (arguments must be separated by spaces). Operations may be arithmetic, comparison, string, or logical.

expr 3 + 5

returns 8

expr 5 % 3

returns 2

expr 1 / 0

returns the error message, expr: division by zero

Illegal arithmetic operations not allowed.

expr 5 \* 3

returns 15

The multiplication operator must be escaped when used in an arithmetic expression with expr.

y=`expr $y + 1`

Increment a variable, with the same effect as let y=y+1 and y=$(($y+1)). This is an example of arithmetic expansion.

z=`expr substr $string $position $length`

Extract substring of $length characters, starting at $position.

Example 16-9. Using expr


# Demonstrating some of the uses of 'expr'
# =======================================


# Arithmetic Operators
# ---------- ---------

echo "Arithmetic Operators"
a=`expr 5 + 3`
echo "5 + 3 = $a"

a=`expr $a + 1`
echo "a + 1 = $a"
echo "(incrementing a variable)"

a=`expr 5 % 3`
# modulo
echo "5 mod 3 = $a"


# Logical Operators
# ------- ---------

#  Returns 1 if true, 0 if false,
#+ opposite of normal Bash convention.

echo "Logical Operators"

b=`expr $x = $y`         # Test equality.
echo "b = $b"            # 0  ( $x -ne $y )

b=`expr $a \> 10`
echo 'b=`expr $a \> 10`, therefore...'
echo "If a > 10, b = 0 (false)"
echo "b = $b"            # 0  ( 3 ! -gt 10 )

b=`expr $a \< 10`
echo "If a < 10, b = 1 (true)"
echo "b = $b"            # 1  ( 3 -lt 10 )
# Note escaping of operators.

b=`expr $a \<= 3`
echo "If a <= 3, b = 1 (true)"
echo "b = $b"            # 1  ( 3 -le 3 )
# There is also a "\>=" operator (greater than or equal to).


# String Operators
# ------ ---------

echo "String Operators"

echo "The string being operated upon is \"$a\"."

# length: length of string
b=`expr length $a`
echo "Length of \"$a\" is $b."

# index: position of first character in substring
#        that matches a character in string
b=`expr index $a 23`
echo "Numerical position of first \"2\" in \"$a\" is \"$b\"."

# substr: extract substring, starting position & length specified
b=`expr substr $a 2 6`
echo "Substring of \"$a\", starting at position 2,\
and 6 chars long is \"$b\"."

#  The default behavior of the 'match' operations is to
#+ search for the specified match at the BEGINNING of the string.
#       Using Regular Expressions ...
b=`expr match "$a" '[0-9]*'`               #  Numerical count.
echo Number of digits at the beginning of \"$a\" is $b.
b=`expr match "$a" '\([0-9]*\)'`           #  Note that escaped parentheses
#                   ==      ==             #+ trigger substring match.
echo "The digits at the beginning of \"$a\" are \"$b\"."


exit 0


The : (null) operator can substitute for match. For example, b=`expr $a : [0-9]*` is the exact equivalent of b=`expr match $a [0-9]*` in the above listing.


echo "String operations using \"expr \$string : \" construct"
echo "==================================================="


echo "The string being operated upon is \"`expr "$a" : '\(.*\)'`\"."
#     Escaped parentheses grouping operator.            ==  ==

#       ***************************
#+          Escaped parentheses
#+           match a substring
#       ***************************

#  If no escaped parentheses ...
#+ then 'expr' converts the string operand to an integer.

echo "Length of \"$a\" is `expr "$a" : '.*'`."   # Length of string

echo "Number of digits at the beginning of \"$a\" is `expr "$a" : '[0-9]*'`."

# ------------------------------------------------------------------------- #


echo "The digits at the beginning of \"$a\" are `expr "$a" : '\([0-9]*\)'`."
#                                                             ==      ==
echo "The first 7 characters of \"$a\" are `expr "$a" : '\(.......\)'`."
#         =====                                          ==       ==
# Again, escaped parentheses force a substring match.
echo "The last 7 characters of \"$a\" are `expr "$a" : '.*\(.......\)'`."
#         ====                  end of string operator  ^^
#  (In fact, means skip over one or more of any characters until specified
#+  substring found.)


exit 0

The above script illustrates how expr uses the escaped parentheses -- \( ... \) -- grouping operator in tandem with regular expression parsing to match a substring. Here is a another example, this time from "real life."
# Strip the whitespace from the beginning and end.
LRFDATE=`expr "$LRFDATE" : '[[:space:]]*\(.*\)[[:space:]]*$'`

#  From Peter Knowles' "" script
#+ for converting files to Sony Librie/PRS-50X format.
#  (

Perl, sed, and awk have far superior string parsing facilities. A short sed or awk "subroutine" within a script (see Section 36.2) is an attractive alternative to expr.

See Section 10.1 for more on using expr in string operations.

16.3. Time / Date Commands

Time/date and timing


Simply invoked, date prints the date and time to stdout. Where this command gets interesting is in its formatting and parsing options.

Example 16-10. Using date

# Exercising the 'date' command

echo "The number of days since the year's beginning is `date +%j`."
# Needs a leading '+' to invoke formatting.
# %j gives day of year.

echo "The number of seconds elapsed since 01/01/1970 is `date +%s`."
#  %s yields number of seconds since "UNIX epoch" began,
#+ but how is this useful?

suffix=$(date +%s)  # The "+%s" option to 'date' is GNU-specific.
echo "Temporary filename = $filename"
#  It's great for creating "unique and random" temp filenames,
#+ even better than using $$.

# Read the 'date' man page for more formatting options.

exit 0

The -u option gives the UTC (Universal Coordinated Time).

bash$ date
Fri Mar 29 21:07:39 MST 2002

bash$ date -u
Sat Mar 30 04:07:42 UTC 2002

This option facilitates calculating the time between different dates.

Example 16-11. Date calculations

# Author: Nathan Coulter
# Used in ABS Guide with permission (thanks!).

MPHR=60    # Minutes per hour.
HPD=24     # Hours per day.

diff () {
        printf '%s' $(( $(date -u -d"$TARGET" +%s) -
                        $(date -u -d"$CURRENT" +%s)))
#                       %d = day of month.

CURRENT=$(date -u -d '2007-09-01 17:30:24' '+%F %T.%N %Z')
TARGET=$(date -u -d'2007-12-25 12:30:00' '+%F %T.%N %Z')
# %F = full date, %T = %H:%M:%S, %N = nanoseconds, %Z = time zone.

printf '\nIn 2007, %s ' \
       "$(date -d"$CURRENT +
        $(( $(diff) /$MPHR /$MPHR /$HPD / 2 )) days" '+%d %B')" 
#       %B = name of month                ^ halfway
printf 'was halfway between %s ' "$(date -d"$CURRENT" '+%d %B')"
printf 'and %s\n' "$(date -d"$TARGET" '+%d %B')"

printf '\nOn %s at %s, there were\n' \
        $(date -u -d"$CURRENT" +%F) $(date -u -d"$CURRENT" +%T)
DAYS=$(( $(diff) / $MPHR / $MPHR / $HPD ))
CURRENT=$(date -d"$CURRENT +$DAYS days" '+%F %T.%N %Z')
HOURS=$(( $(diff) / $MPHR / $MPHR ))
CURRENT=$(date -d"$CURRENT +$HOURS hours" '+%F %T.%N %Z')
MINUTES=$(( $(diff) / $MPHR ))
CURRENT=$(date -d"$CURRENT +$MINUTES minutes" '+%F %T.%N %Z')
printf '%s days, %s hours, ' "$DAYS" "$HOURS"
printf '%s minutes, and %s seconds ' "$MINUTES" "$(diff)"
printf 'until Christmas Dinner!\n\n'

#  Exercise:
#  --------
#  Rewrite the diff () function to accept passed parameters,
#+ rather than using global variables.

The date command has quite a number of output options. For example %N gives the nanosecond portion of the current time. One interesting use for this is to generate random integers.
date +%N | sed -e 's/000$//' -e 's/^0//'
#  Strip off leading and trailing zeroes, if present.
#  Length of generated integer depends on
#+ how many zeroes stripped off.

# 115281032
# 63408725
# 394504284

There are many more options (try man date).

date +%j
# Echoes day of the year (days elapsed since January 1).

date +%k%M
# Echoes hour and minute in 24-hour format, as a single digit string.

# The 'TZ' parameter permits overriding the default time zone.
date                 # Mon Mar 28 21:42:16 MST 2005
TZ=EST date          # Mon Mar 28 23:42:16 EST 2005
# Thanks, Frank Kannemann and Pete Sjoberg, for the tip.

SixDaysAgo=$(date --date='6 days ago')
OneMonthAgo=$(date --date='1 month ago')  # Four weeks back (not a month!)
OneYearAgo=$(date --date='1 year ago')

See also Example 3-4 and Example A-43.


Time zone dump: echoes the time in a specified time zone.

bash$ zdump EST
EST  Tue Sep 18 22:09:22 2001 EST


Outputs verbose timing statistics for executing a command.

time ls -l / gives something like this:

real    0m0.067s
 user    0m0.004s
 sys     0m0.005s

See also the very similar times command in the previous section.


As of version 2.0 of Bash, time became a shell reserved word, with slightly altered behavior in a pipeline.


Utility for updating access/modification times of a file to current system time or other specified time, but also useful for creating a new file. The command touch zzz will create a new file of zero length, named zzz, assuming that zzz did not previously exist. Time-stamping empty files in this way is useful for storing date information, for example in keeping track of modification times on a project.


The touch command is equivalent to : >> newfile or >> newfile (for ordinary files).


Before doing a cp -u (copy/update), use touch to update the time stamp of files you don't wish overwritten.

As an example, if the directory /home/bozo/tax_audit contains the files,, and, then doing a touch spreadsheet*.data will protect these files from being overwritten by files with the same names during a cp -u /home/bozo/financial_info/spreadsheet*data /home/bozo/tax_audit.


The at job control command executes a given set of commands at a specified time. Superficially, it resembles cron, however, at is chiefly useful for one-time execution of a command set.

at 2pm January 15 prompts for a set of commands to execute at that time. These commands should be shell-script compatible, since, for all practical purposes, the user is typing in an executable shell script a line at a time. Input terminates with a Ctl-D.

Using either the -f option or input redirection (<), at reads a command list from a file. This file is an executable shell script, though it should, of course, be non-interactive. Particularly clever is including the run-parts command in the file to execute a different set of scripts.

bash$ at 2:30 am Friday < at-jobs.list
job 2 at 2000-10-27 02:30


The batch job control command is similar to at, but it runs a command list when the system load drops below .8. Like at, it can read commands from a file with the -f option.


Prints a neatly formatted monthly calendar to stdout. Will do current year or a large range of past and future years.


This is the shell equivalent of a wait loop. It pauses for a specified number of seconds, doing nothing. It can be useful for timing or in processes running in the background, checking for a specific event every so often (polling), as in Example 32-6.
sleep 3     # Pauses 3 seconds.


The sleep command defaults to seconds, but minute, hours, or days may also be specified.
sleep 3 h   # Pauses 3 hours!


The watch command may be a better choice than sleep for running commands at timed intervals.


Microsleep (the u may be read as the Greek mu, or micro- prefix). This is the same as sleep, above, but "sleeps" in microsecond intervals. It can be used for fine-grained timing, or for polling an ongoing process at very frequent intervals.

usleep 30     # Pauses 30 microseconds.

This command is part of the Red Hat initscripts / rc-scripts package.


The usleep command does not provide particularly accurate timing, and is therefore unsuitable for critical timing loops.

hwclock, clock

The hwclock command accesses or adjusts the machine's hardware clock. Some options require root privileges. The /etc/rc.d/rc.sysinit startup file uses hwclock to set the system time from the hardware clock at bootup.

The clock command is a synonym for hwclock.

16.4. Text Processing Commands

Commands affecting text and text files


File sort utility, often used as a filter in a pipe. This command sorts a text stream or file forwards or backwards, or according to various keys or character positions. Using the -m option, it merges presorted input files. The info page lists its many capabilities and options. See Example 11-10, Example 11-11, and Example A-8.


Topological sort, reading in pairs of whitespace-separated strings and sorting according to input patterns. The original purpose of tsort was to sort a list of dependencies for an obsolete version of the ld linker in an "ancient" version of UNIX.

The results of a tsort will usually differ markedly from those of the standard sort command, above.


This filter removes duplicate lines from a sorted file. It is often seen in a pipe coupled with sort.

cat list-1 list-2 list-3 | sort | uniq > final.list
# Concatenates the list files,
# sorts them,
# removes duplicate lines,
# and finally writes the result to an output file.

The useful -c option prefixes each line of the input file with its number of occurrences.

bash$ cat testfile
This line occurs only once.
 This line occurs twice.
 This line occurs twice.
 This line occurs three times.
 This line occurs three times.
 This line occurs three times.

bash$ uniq -c testfile
      1 This line occurs only once.
       2 This line occurs twice.
       3 This line occurs three times.

bash$ sort testfile | uniq -c | sort -nr
      3 This line occurs three times.
       2 This line occurs twice.
       1 This line occurs only once.

The sort INPUTFILE | uniq -c | sort -nr command string produces a frequency of occurrence listing on the INPUTFILE file (the -nr options to sort cause a reverse numerical sort). This template finds use in analysis of log files and dictionary lists, and wherever the lexical structure of a document needs to be examined.

Example 16-12. Word Frequency Analysis

# Crude word frequency analysis on a text file.
# This is a more efficient version of the "" script.

# Check for input file on command-line.

if [ $# -ne "$ARGS" ]  # Correct number of arguments passed to script?
  echo "Usage: `basename $0` filename"
  exit $E_BADARGS

if [ ! -f "$1" ]       # Check if file exists.
  echo "File \"$1\" does not exist."
  exit $E_NOFILE

# main ()
sed -e 's/\.//g'  -e 's/\,//g' -e 's/ /\
/g' "$1" | tr 'A-Z' 'a-z' | sort | uniq -c | sort -nr
#                           =========================
#                            Frequency of occurrence

#  Filter out periods and commas, and
#+ change space between words to linefeed,
#+ then shift characters to lowercase, and
#+ finally prefix occurrence count and sort numerically.

#  Arun Giridhar suggests modifying the above to:
#  . . . | sort | uniq -c | sort +1 [-f] | sort +0 -nr
#  This adds a secondary sort key, so instances of
#+ equal occurrence are sorted alphabetically.
#  As he explains it:
#  "This is effectively a radix sort, first on the
#+ least significant column
#+ (word or string, optionally case-insensitive)
#+ and last on the most significant column (frequency)."
#  As Frank Wang explains, the above is equivalent to
#+       . . . | sort | uniq -c | sort +0 -nr
#+ and the following also works:
#+       . . . | sort | uniq -c | sort -k1nr -k

exit 0

# Exercises:
# ---------
# 1) Add 'sed' commands to filter out other punctuation,
#+   such as semicolons.
# 2) Modify the script to also filter out multiple spaces and
#+   other whitespace.

bash$ cat testfile
This line occurs only once.
 This line occurs twice.
 This line occurs twice.
 This line occurs three times.
 This line occurs three times.
 This line occurs three times.

bash$ ./ testfile
      6 this
       6 occurs
       6 line
       3 times
       3 three
       2 twice
       1 only
       1 once

expand, unexpand

The expand filter converts tabs to spaces. It is often used in a pipe.

The unexpand filter converts spaces to tabs. This reverses the effect of expand.


A tool for extracting fields from files. It is similar to the print $N command set in awk, but more limited. It may be simpler to use cut in a script than awk. Particularly important are the -d (delimiter) and -f (field specifier) options.

Using cut to obtain a listing of the mounted filesystems:
cut -d ' ' -f1,2 /etc/mtab

Using cut to list the OS and kernel version:
uname -a | cut -d" " -f1,3,11,12

Using cut to extract message headers from an e-mail folder:
bash$ grep '^Subject:' read-messages | cut -c10-80
Re: Linux suitable for mission-critical apps?
 Spam complaint
 Re: Spam complaint

Using cut to parse a file:
# List all the users in /etc/passwd.


for user in $(cut -d: -f1 $FILENAME)
  echo $user

# Thanks, Oleg Philon for suggesting this.

cut -d ' ' -f2,3 filename is equivalent to awk -F'[ ]' '{ print $2, $3 }' filename


It is even possible to specify a linefeed as a delimiter. The trick is to actually embed a linefeed (RETURN) in the command sequence.

bash$ cut -d'
 ' -f3,7,19 testfile
This is line 3 of testfile.
 This is line 7 of testfile.
 This is line 19 of testfile.

Thank you, Jaka Kranjc, for pointing this out.

See also Example 16-48.


Tool for merging together different files into a single, multi-column file. In combination with cut, useful for creating system log files.

bash$ cat items
alphabet blocks
 building blocks

bash$ cat prices
 $2.50 ea.

bash$ paste items prices
alphabet blocks $1.00/dozen
 building blocks $2.50 ea.
 cables  $3.75


Consider this a special-purpose cousin of paste. This powerful utility allows merging two files in a meaningful fashion, which essentially creates a simple version of a relational database.

The join command operates on exactly two files, but pastes together only those lines with a common tagged field (usually a numerical label), and writes the result to stdout. The files to be joined should be sorted according to the tagged field for the matchups to work properly.


100 Shoes
200 Laces
300 Socks


100 $40.00
200 $1.00
300 $2.00

bash$ join

 100 Shoes $40.00
 200 Laces $1.00
 300 Socks $2.00


The tagged field appears only once in the output.


lists the beginning of a file to stdout. The default is 10 lines, but a different number can be specified. The command has a number of interesting options.

Example 16-13. Which files are scripts?

# Detects scripts within a directory.

TESTCHARS=2    # Test first 2 characters.
SHABANG='#!'   # Scripts begin with a "sha-bang."

for file in *  # Traverse all the files in current directory.
  if [[ `head -c$TESTCHARS "$file"` = "$SHABANG" ]]
  #      head -c2                      #!
  #  The '-c' option to "head" outputs a specified
  #+ number of characters, rather than lines (the default).
    echo "File \"$file\" is a script."
    echo "File \"$file\" is *not* a script."
exit 0

#  Exercises:
#  ---------
#  1) Modify this script to take as an optional argument
#+    the directory to scan for scripts
#+    (rather than just the current working directory).
#  2) As it stands, this script gives "false positives" for
#+    Perl, awk, and other scripting language scripts.
#     Correct this.

Example 16-14. Generating 10-digit random numbers

# Outputs a 10-digit random number

# Script by Stephane Chazelas.

head -c4 /dev/urandom | od -N4 -tu4 | sed -ne '1s/.* //p'

# =================================================================== #

# Analysis
# --------

# head:
# -c4 option takes first 4 bytes.

# od:
# -N4 option limits output to 4 bytes.
# -tu4 option selects unsigned decimal format for output.

# sed: 
# -n option, in combination with "p" flag to the "s" command,
# outputs only matched lines.

# The author of this script explains the action of 'sed', as follows.

# head -c4 /dev/urandom | od -N4 -tu4 | sed -ne '1s/.* //p'
# ----------------------------------> |

# Assume output up to "sed" --------> |
# is 0000000 1198195154\n

#  sed begins reading characters: 0000000 1198195154\n.
#  Here it finds a newline character,
#+ so it is ready to process the first line (0000000 1198195154).
#  It looks at its <range><action>s. The first and only one is

#   range     action
#   1         s/.* //p

#  The line number is in the range, so it executes the action:
#+ tries to substitute the longest string ending with a space in the line
#  ("0000000 ") with nothing (//), and if it succeeds, prints the result
#  ("p" is a flag to the "s" command here, this is different
#+ from the "p" command).

#  sed is now ready to continue reading its input. (Note that before
#+ continuing, if -n option had not been passed, sed would have printed
#+ the line once again).

#  Now, sed reads the remainder of the characters, and finds the
#+ end of the file.
#  It is now ready to process its 2nd line (which is also numbered '$' as
#+ it's the last one).
#  It sees it is not matched by any <range>, so its job is done.

#  In few word this sed commmand means:
#  "On the first line only, remove any character up to the right-most space,
#+ then print it."

# A better way to do this would have been:
#           sed -e 's/.* //;q'

# Here, two <range><action>s (could have been written
#           sed -e 's/.* //' -e q):

#   range                    action
#   nothing (matches line)   s/.* //
#   nothing (matches line)   q (quit)

#  Here, sed only reads its first line of input.
#  It performs both actions, and prints the line (substituted) before
#+ quitting (because of the "q" action) since the "-n" option is not passed.

# =================================================================== #

# An even simpler altenative to the above one-line script would be:
#           head -c4 /dev/urandom| od -An -tu4

See also Example 16-39.


lists the (tail) end of a file to stdout. The default is 10 lines, but this can be changed with the -n option. Commonly used to keep track of changes to a system logfile, using the -f option, which outputs lines appended to the file.

Example 16-15. Using tail to monitor the system log



cat /dev/null > $filename; echo "Creating / cleaning out file."
#  Creates the file if it does not already exist,
#+ and truncates it to zero length if it does.
#  : > filename   and   > filename also work.

tail /var/log/messages > $filename  
# /var/log/messages must have world read permission for this to work.

echo "$filename contains tail end of system log."

exit 0


To list a specific line of a text file, pipe the output of head to tail -n 1. For example head -n 8 database.txt | tail -n 1 lists the 8th line of the file database.txt.

To set a variable to a given block of a text file:
var=$(head -n $m $filename | tail -n $n)

# filename = name of file
# m = from beginning of file, number of lines to end of block
# n = number of lines to set variable to (trim from end of block)


Newer implementations of tail deprecate the older tail -$LINES filename usage. The standard tail -n $LINES filename is correct.

See also Example 16-5, Example 16-39 and Example 32-6.


A multi-purpose file search tool that uses Regular Expressions. It was originally a command/filter in the venerable ed line editor: g/re/p -- global - regular expression - print.

grep pattern [file...]

Search the target file(s) for occurrences of pattern, where pattern may be literal text or a Regular Expression.

bash$ grep '[rst]ystem.$' osinfo.txt
The GPL governs the distribution of the Linux operating system.

If no target file(s) specified, grep works as a filter on stdout, as in a pipe.

bash$ ps ax | grep clock
765 tty1     S      0:00 xclock
 901 pts/1    S      0:00 grep clock

The -i option causes a case-insensitive search.

The -w option matches only whole words.

The -l option lists only the files in which matches were found, but not the matching lines.

The -r (recursive) option searches files in the current working directory and all subdirectories below it.

The -n option lists the matching lines, together with line numbers.

bash$ grep -n Linux osinfo.txt
2:This is a file containing information about Linux.
 6:The GPL governs the distribution of the Linux operating system.

The -v (or --invert-match) option filters out matches.
grep pattern1 *.txt | grep -v pattern2

# Matches all lines in "*.txt" files containing "pattern1",
# but ***not*** "pattern2".	      

The -c (--count) option gives a numerical count of matches, rather than actually listing the matches.
grep -c txt *.sgml   # (number of occurrences of "txt" in "*.sgml" files)

#   grep -cz .
#            ^ dot
# means count (-c) zero-separated (-z) items matching "."
# that is, non-empty ones (containing at least 1 character).
printf 'a b\nc  d\n\n\n\n\n\000\n\000e\000\000\nf' | grep -cz .     # 3
printf 'a b\nc  d\n\n\n\n\n\000\n\000e\000\000\nf' | grep -cz '$'   # 5
printf 'a b\nc  d\n\n\n\n\n\000\n\000e\000\000\nf' | grep -cz '^'   # 5
printf 'a b\nc  d\n\n\n\n\n\000\n\000e\000\000\nf' | grep -c '$'    # 9
# By default, newline chars (\n) separate items to match. 

# Note that the -z option is GNU "grep" specific.

# Thanks, S.C.

The --color (or --colour) option marks the matching string in color (on the console or in an xterm window). Since grep prints out each entire line containing the matching pattern, this lets you see exactly what is being matched. See also the -o option, which shows only the matching portion of the line(s).

Example 16-16. Printing out the From lines in stored e-mail messages


#  Emulates the useful 'from' utility in Solaris, BSD, etc.
#  Echoes the "From" header line in all messages
#+ in your e-mail directory.

MAILDIR=~/mail/*               #  No quoting of variable. Why?
# Maybe check if-exists $MAILDIR:   if [ -d $MAILDIR ] . . .
GREP_OPTS="-H -A 5 --color"    #  Show file, plus extra context lines
                               #+ and display "From" in color.
TARGETSTR="^From"              # "From" at beginning of line.

for file in $MAILDIR           #  No quoting of variable.
  grep $GREP_OPTS "$TARGETSTR" "$file"
  #    ^^^^^^^^^^              #  Again, do not quote this variable.

exit $?

#  You might wish to pipe the output of this script to 'more'
#+ or redirect it to a file . . .

When invoked with more than one target file given, grep specifies which file contains matches.

bash$ grep Linux osinfo.txt misc.txt
osinfo.txt:This is a file containing information about Linux.
 osinfo.txt:The GPL governs the distribution of the Linux operating system.
 misc.txt:The Linux operating system is steadily gaining in popularity.


To force grep to show the filename when searching only one target file, simply give /dev/null as the second file.

bash$ grep Linux osinfo.txt /dev/null
osinfo.txt:This is a file containing information about Linux.
 osinfo.txt:The GPL governs the distribution of the Linux operating system.

If there is a successful match, grep returns an exit status of 0, which makes it useful in a condition test in a script, especially in combination with the -q option to suppress output.
SUCCESS=0                      # if grep lookup succeeds

grep -q "$word" "$filename"    #  The "-q" option
                               #+ causes nothing to echo to stdout.
if [ $? -eq $SUCCESS ]
# if grep -q "$word" "$filename"   can replace lines 5 - 7.
  echo "$word found in $filename"
  echo "$word not found in $filename"

Example 32-6 demonstrates how to use grep to search for a word pattern in a system logfile.

Example 16-17. Emulating grep in a script

# Rudimentary reimplementation of grep.


if [ -z "$1" ]    # Check for argument to script.
  echo "Usage: `basename $0` pattern"
  exit $E_BADARGS


for file in *     # Traverse all files in $PWD.
  output=$(sed -n /"$1"/p $file)  # Command substitution.

  if [ ! -z "$output" ]           # What happens if "$output" is not quoted?
    echo -n "$file: "
    echo "$output"
  fi              #  sed -ne "/$1/s|^|${file}: |p"  is equivalent to above.



exit 0

# Exercises:
# ---------
# 1) Add newlines to output, if more than one match in any given file.
# 2) Add features.

How can grep search for two (or more) separate patterns? What if you want grep to display all lines in a file or files that contain both "pattern1" and "pattern2"?

One method is to pipe the result of grep pattern1 to grep pattern2.

For example, given the following file:

# Filename: tstfile

This is a sample file.
This is an ordinary text file.
This file does not contain any unusual text.
This file is not unusual.
Here is some text.

Now, let's search this file for lines containing both "file" and "text" . . .

bash$ grep file tstfile
# Filename: tstfile
 This is a sample file.
 This is an ordinary text file.
 This file does not contain any unusual text.
 This file is not unusual.

bash$ grep file tstfile | grep text
This is an ordinary text file.
 This file does not contain any unusual text.

Now, for an interesting recreational use of grep . . .

Example 16-18. Crossword puzzle solver

# This is actually a wrapper around a one-liner (line 46).

#  Crossword puzzle and anagramming word game solver.
#  You know *some* of the letters in the word you're looking for,
#+ so you need a list of all valid words
#+ with the known letters in given positions.
#  For example: w...i....n
#               1???5????10
# w in position 1, 3 unknowns, i in the 5th, 4 unknowns, n at the end.
# (See comments at end of script.)

#                    ^^^^^^^^   Looks for word list here.
#  ASCII word list, one word per line.
#  If you happen to need an appropriate list,
#+ download the author's "yawl" word list package.
#  or

if [ -z "$1" ]   #  If no word pattern specified
then             #+ as a command-line argument . . .
  echo           #+ . . . then . . .
  echo "Usage:"  #+ Usage message.
  echo ""$0" \"pattern,\""
  echo "where \"pattern\" is in the form"
  echo "xxx..x.x..."
  echo "The x's represent known letters,"
  echo "and the periods are unknown letters (blanks)."
  echo "Letters and periods can be in any position."
  echo "For example, try:   sh w...i....n"
  exit $E_NOPATT

# ===============================================
# This is where all the work gets done.
grep ^"$1"$ "$DICT"   # Yes, only one line!
#    |    |
# ^ is start-of-word regex anchor.
# $ is end-of-word regex anchor.

#  From _Stupid Grep Tricks_, vol. 1,
#+ a book the ABS Guide author may yet get around
#+ to writing . . . one of these days . . .
# ===============================================

exit $?  # Script terminates here.
#  If there are too many words generated,
#+ redirect the output to a file.

$ sh w...i....n


egrep -- extended grep -- is the same as grep -E. This uses a somewhat different, extended set of Regular Expressions, which can make the search a bit more flexible. It also allows the boolean | (or) operator.
bash $ egrep 'matches|Matches' file.txt
Line 1 matches.
 Line 3 Matches.
 Line 4 contains matches, but also Matches

fgrep -- fast grep -- is the same as grep -F. It does a literal string search (no Regular Expressions), which generally speeds things up a bit.


On some Linux distros, egrep and fgrep are symbolic links to, or aliases for grep, but invoked with the -E and -F options, respectively.

Example 16-19. Looking up definitions in Webster's 1913 Dictionary


#  This script looks up definitions in the 1913 Webster's Dictionary.
#  This Public Domain dictionary is available for download
#+ from various sites, including
#+ Project Gutenberg (
#  Convert it from DOS to UNIX format (with only LF at end of line)
#+ before using it with this script.
#  Store the file in plain, uncompressed ASCII text.
#  Set DEFAULT_DICTFILE variable below to path/filename.

MAXCONTEXTLINES=50                        # Maximum number of lines to show.
                                          # Default dictionary file pathname.
                                          # Change this as necessary.
#  Note:
#  ----
#  This particular edition of the 1913 Webster's
#+ begins each entry with an uppercase letter
#+ (lowercase for the remaining characters).
#  Only the *very first line* of an entry begins this way,
#+ and that's why the search algorithm below works.

if [[ -z $(echo "$1" | sed -n '/^[A-Z]/p') ]]
#  Must at least specify word to look up, and
#+ it must start with an uppercase letter.
  echo "Usage: `basename $0` Word-to-define [dictionary-file]"
  echo "Note: Word to look up must start with capital letter,"
  echo "with the rest of the word in lowercase."
  echo "--------------------------------------------"
  echo "Examples: Abandon, Dictionary, Marking, etc."
  exit $E_BADARGS

if [ -z "$2" ]                            #  May specify different dictionary
                                          #+ as an argument to this script.

# ---------------------------------------------------------
Definition=$(fgrep -A $MAXCONTEXTLINES "$1 \\" "$dictfile")
#                  Definitions in form "Word \..."
#  And, yes, "fgrep" is fast enough
#+ to search even a very large text file.

# Now, snip out just the definition block.

echo "$Definition" |
sed -n '1,/^[A-Z]/p' |
#  Print from first line of output
#+ to the first line of the next entry.
sed '$d' | sed '$d'
#  Delete last two lines of output
#+ (blank line and first line of next entry).
# ---------------------------------------------------------

exit $?

# Exercises:
# ---------
# 1)  Modify the script to accept any type of alphabetic input
#   + (uppercase, lowercase, mixed case), and convert it
#   + to an acceptable format for processing.
# 2)  Convert the script to a GUI application,
#   + using something like 'gdialog' or 'zenity' . . .
#     The script will then no longer take its argument(s)
#   + from the command-line.
# 3)  Modify the script to parse one of the other available
#   + Public Domain Dictionaries, such as the U.S. Census Bureau Gazetteer.


See also Example A-41 for an example of speedy fgrep lookup on a large text file.

agrep (approximate grep) extends the capabilities of grep to approximate matching. The search string may differ by a specified number of characters from the resulting matches. This utility is not part of the core Linux distribution.


To search compressed files, use zgrep, zegrep, or zfgrep. These also work on non-compressed files, though slower than plain grep, egrep, fgrep. They are handy for searching through a mixed set of files, some compressed, some not.

To search bzipped files, use bzgrep.


The command look works like grep, but does a lookup on a "dictionary," a sorted word list. By default, look searches for a match in /usr/dict/words, but a different dictionary file may be specified.

Example 16-20. Checking words in a list for validity

# lookup: Does a dictionary lookup on each word in a data file.  # Data file from which to read words to test.

echo "Testing file $file"

while [ "$word" != end ]  # Last word in data file.
do               # ^^^
  read word      # From data file, because of redirection at end of loop.
  look $word > /dev/null  # Don't want to display lines in dictionary file.
  #  Searches for words in the file /usr/share/dict/words
  #+ (usually a link to linux.words).
  lookup=$?      # Exit status of 'look' command.

  if [ "$lookup" -eq 0 ]
    echo "\"$word\" is valid."
    echo "\"$word\" is invalid."

done <"$file"    # Redirects stdin to $file, so "reads" come from there.


exit 0

# ----------------------------------------------------------------
# Code below line will not execute because of "exit" command above.

# Stephane Chazelas proposes the following, more concise alternative:

while read word && [[ $word != end ]]
do if look "$word" > /dev/null
   then echo "\"$word\" is valid."
   else echo "\"$word\" is invalid."
done <"$file"

exit 0
sed, awk

Scripting languages especially suited for parsing text files and command output. May be embedded singly or in combination in pipes and shell scripts.


Non-interactive "stream editor", permits using many ex commands in batch mode. It finds many uses in shell scripts.


Programmable file extractor and formatter, good for manipulating and/or extracting fields (columns) in structured text files. Its syntax is similar to C.


wc gives a "word count" on a file or I/O stream:
bash $ wc /usr/share/doc/sed-4.1.2/README
13  70  447 README
[13 lines  70 words  447 characters]

wc -w gives only the word count.

wc -l gives only the line count.

wc -c gives only the byte count.

wc -m gives only the character count.

wc -L gives only the length of the longest line.

Using wc to count how many .txt files are in current working directory:
$ ls *.txt | wc -l
#  Will work as long as none of the "*.txt" files
#+ have a linefeed embedded in their name.

#  Alternative ways of doing this are:
#      find . -maxdepth 1 -name \*.txt -print0 | grep -cz .
#      (shopt -s nullglob; set -- *.txt; echo $#)

#  Thanks, S.C.

Using wc to total up the size of all the files whose names begin with letters in the range d - h
bash$ wc [d-h]* | grep total | awk '{print $3}'

Using wc to count the instances of the word "Linux" in the main source file for this book.
bash$ grep Linux abs-book.sgml | wc -l

See also Example 16-39 and Example 20-8.

Certain commands include some of the functionality of wc as options.
... | grep foo | wc -l
# This frequently used construct can be more concisely rendered.

... | grep -c foo
# Just use the "-c" (or "--count") option of grep.

# Thanks, S.C.


character translation filter.


Must use quoting and/or brackets, as appropriate. Quotes prevent the shell from reinterpreting the special characters in tr command sequences. Brackets should be quoted to prevent expansion by the shell.

Either tr "A-Z" "*" <filename or tr A-Z \* <filename changes all the uppercase letters in filename to asterisks (writes to stdout). On some systems this may not work, but tr A-Z '[**]' will.

The -d option deletes a range of characters.
echo "abcdef"                 # abcdef
echo "abcdef" | tr -d b-d     # aef

tr -d 0-9 <filename
# Deletes all digits from the file "filename".

The --squeeze-repeats (or -s) option deletes all but the first instance of a string of consecutive characters. This option is useful for removing excess whitespace.
bash$ echo "XXXXX" | tr --squeeze-repeats 'X'

The -c "complement" option inverts the character set to match. With this option, tr acts only upon those characters not matching the specified set.

bash$ echo "acfdeb123" | tr -c b-d +

Note that tr recognizes POSIX character classes. [74]

bash$ echo "abcd2ef1" | tr '[:alpha:]' -

Example 16-21. toupper: Transforms a file to all uppercase.

# Changes a file to all uppercase.


if [ -z "$1" ]  # Standard check for command-line arg.
  echo "Usage: `basename $0` filename"
  exit $E_BADARGS

tr a-z A-Z <"$1"

# Same effect as above, but using POSIX character set notation:
#        tr '[:lower:]' '[:upper:]' <"$1"
# Thanks, S.C.

#     Or even . . .
#     cat "$1" | tr a-z A-Z
#     Or dozens of other ways . . .

exit 0

#  Exercise:
#  Rewrite this script to give the option of changing a file
#+ to *either* upper or lowercase.
#  Hint: Use either the "case" or "select" command.

Example 16-22. lowercase: Changes all filenames in working directory to lowercase.

#  Changes every filename in working directory to all lowercase.
#  Inspired by a script of John Dubois,
#+ which was translated into Bash by Chet Ramey,
#+ and considerably simplified by the author of the ABS Guide.

for filename in *                # Traverse all files in directory.
   fname=`basename $filename`
   n=`echo $fname | tr A-Z a-z`  # Change name to lowercase.
   if [ "$fname" != "$n" ]       # Rename only files not already lowercase.
     mv $fname $n

exit $?

# Code below this line will not execute because of "exit".
# To run it, delete script above line.

# The above script will not work on filenames containing blanks or newlines.
# Stephane Chazelas therefore suggests the following alternative:

for filename in *    # Not necessary to use basename,
                     # since "*" won't return any file containing "/".
do n=`echo "$filename/" | tr '[:upper:]' '[:lower:]'`
#                             POSIX char set notation.
#                    Slash added so that trailing newlines are not
#                    removed by command substitution.
   # Variable substitution:
   n=${n%/}          # Removes trailing slash, added above, from filename.
   [[ $filename == $n ]] || mv "$filename" "$n"
                     # Checks if filename already lowercase.

exit $?

Example 16-23. du: DOS to UNIX text file conversion.

# DOS to UNIX text file converter.


if [ -z "$1" ]
  echo "Usage: `basename $0` filename-to-convert"


CR='\015'  # Carriage return.
           # 015 is octal ASCII code for CR.
           # Lines in a DOS text file end in CR-LF.
           # Lines in a UNIX text file end in LF only.

tr -d $CR < $1 > $NEWFILENAME
# Delete CR's and write to new file.

echo "Original DOS text file is \"$1\"."
echo "Converted UNIX text file is \"$NEWFILENAME\"."

exit 0

# Exercise:
# --------
# Change the above script to convert from UNIX to DOS.

Example 16-24. rot13: ultra-weak encryption.

# Classic rot13 algorithm,
#           encryption that might fool a 3-year old
#           for about 10 minutes.

# Usage: ./ filename
# or     ./ <filename
# or     ./ and supply keyboard input (stdin)

cat "$@" | tr 'a-zA-Z' 'n-za-mN-ZA-M'   # "a" goes to "n", "b" to "o" ...
#  The   cat "$@"   construct
#+ permits input either from stdin or from files.

exit 0

Example 16-25. Generating "Crypto-Quote" Puzzles

# Encrypt quotes

#  Will encrypt famous quotes in a simple monoalphabetic substitution.
#  The result is similar to the "Crypto Quote" puzzles
#+ seen in the Op Ed pages of the Sunday paper.

# The "key" is nothing more than a scrambled alphabet.
# Changing the "key" changes the encryption.

# The 'cat "$@"' construction gets input either from stdin or from files.
# If using stdin, terminate input with a Control-D.
# Otherwise, specify filename as command-line parameter.

cat "$@" | tr "a-z" "A-Z" | tr "A-Z" "$key"
#        |  to uppercase  |     encrypt       
# Will work on lowercase, uppercase, or mixed-case quotes.
# Passes non-alphabetic characters through unchanged.

# Try this script with something like:
# "Nothing so needs reforming as other people's habits."
# --Mark Twain
# Output is:

# To reverse the encryption:
# cat "$@" | tr "$key" "A-Z"

#  This simple-minded cipher can be broken by an average 12-year old
#+ using only pencil and paper.

exit 0

#  Exercise:
#  --------
#  Modify the script so that it will either encrypt or decrypt,
#+ depending on command-line argument(s).

Of course, tr lends itself to code obfuscation.


echo $x | tr "a-z" 'oh, turtleneck Phrase Jar!'

# Based on the Wikipedia "Just another Perl hacker" article.


A filter that wraps lines of input to a specified width. This is especially useful with the -s option, which breaks lines at word spaces (see Example 16-26 and Example A-1).


Simple-minded file formatter, used as a filter in a pipe to "wrap" long lines of text output.

Example 16-26. Formatted file listing.


WIDTH=40                    # 40 columns wide.

b=`ls /usr/local/bin`       # Get a file listing...

echo $b | fmt -w $WIDTH

# Could also have been done by
#    echo $b | fold - -s -w $WIDTH
exit 0

See also Example 16-5.


A powerful alternative to fmt is Kamil Toman's par utility, available from


This deceptively named filter removes reverse line feeds from an input stream. It also attempts to replace whitespace with equivalent tabs. The chief use of col is in filtering the output from certain text processing utilities, such as groff and tbl.


Column formatter. This filter transforms list-type text output into a "pretty-printed" table by inserting tabs at appropriate places.

Example 16-27. Using column to format a directory listing

# A minor modification of the example file in the "column" man page.

; ls -l | sed 1d) | column -t
#         ^^^^^^           ^^

#  The "sed 1d" in the pipe deletes the first line of output,
#+ which would be "total        N",
#+ where "N" is the total number of files found by "ls -l".

# The -t option to "column" pretty-prints a table.

exit 0

Column removal filter. This removes columns (characters) from a file and writes the file, lacking the range of specified columns, back to stdout. colrm 2 4 <filename removes the second through fourth characters from each line of the text file filename.


If the file contains tabs or nonprintable characters, this may cause unpredictable behavior. In such cases, consider using expand and unexpand in a pipe preceding colrm.


Line numbering filter: nl filename lists filename to stdout, but inserts consecutive numbers at the beginning of each non-blank line. If filename omitted, operates on stdin.

The output of nl is very similar to cat -b, since, by default nl does not list blank lines.

Example 16-28. nl: A self-numbering script.


# This script echoes itself twice to stdout with its lines numbered.

echo "     line number = $LINENO" # 'nl' sees this as line 4
#                                   (nl does not number blank lines).
#                                   'cat -n' sees it correctly as line #6.

nl `basename $0`

echo; echo  # Now, let's try it with 'cat -n'

cat -n `basename $0`
# The difference is that 'cat -n' numbers the blank lines.
# Note that 'nl -ba' will also do so.

exit 0
# -----------------------------------------------------------------

Print formatting filter. This will paginate files (or stdout) into sections suitable for hard copy printing or viewing on screen. Various options permit row and column manipulation, joining lines, setting margins, numbering lines, adding page headers, and merging files, among other things. The pr command combines much of the functionality of nl, paste, fold, column, and expand.

pr -o 5 --width=65 fileZZZ | more gives a nice paginated listing to screen of fileZZZ with margins set at 5 and 65.

A particularly useful option is -d, forcing double-spacing (same effect as sed -G).


The GNU gettext package is a set of utilities for localizing and translating the text output of programs into foreign languages. While originally intended for C programs, it now supports quite a number of programming and scripting languages.

The gettext program works on shell scripts. See the info page.


A program for generating binary message catalogs. It is used for localization.


A utility for converting file(s) to a different encoding (character set). Its chief use is for localization.

# Convert a string from UTF-8 to UTF-16 and print to the BookList
function write_utf8_string {
    echo -n "$STRING" | iconv -f UTF8 -t UTF16 | \
    cut -b 3- | tr -d \\n >> "$BOOKLIST"

#  From Peter Knowles' "" script
#+ for converting files to Sony Librie/PRS-50X format.
#  (


Consider this a fancier version of iconv, above. This very versatile utility for converting a file to a different encoding scheme. Note that recode is not part of the standard Linux installation.

TeX, gs

TeX and Postscript are text markup languages used for preparing copy for printing or formatted video display.

TeX is Donald Knuth's elaborate typsetting system. It is often convenient to write a shell script encapsulating all the options and arguments passed to one of these markup languages.

Ghostscript (gs) is a GPL-ed Postscript interpreter.


Utility for processing TeX and pdf files. Found in /usr/bin on many Linux distros, it is actually a shell wrapper that calls Perl to invoke Tex.

texexec --pdfarrange --result=Concatenated.pdf *pdf

#  Concatenates all the pdf files in the current working directory
#+ into the merged file, Concatenated.pdf . . .
#  (The --pdfarrange option repaginates a pdf file. See also --pdfcombine.)
#  The above command-line could be parameterized and put into a shell script.


Utility for converting plain text file to PostScript

For example, enscript filename.txt -p produces the PostScript output file

groff, tbl, eqn

Yet another text markup and display formatting language is groff. This is the enhanced GNU version of the venerable UNIX roff/troff display and typesetting package. Manpages use groff.

The tbl table processing utility is considered part of groff, as its function is to convert table markup into groff commands.

The eqn equation processing utility is likewise part of groff, and its function is to convert equation markup into groff commands.

Example 16-29. manview: Viewing formatted manpages

# Formats the source of a man page for viewing.

#  This script is useful when writing man page source.
#  It lets you look at the intermediate results on the fly
#+ while working on it.


if [ -z "$1" ]
  echo "Usage: `basename $0` filename"

# ---------------------------
groff -Tascii -man $1 | less
# From the man page for groff.
# ---------------------------

#  If the man page includes tables and/or equations,
#+ then the above code will barf.
#  The following line can handle such cases.
#   gtbl < "$1" | geqn -Tlatin1 | groff -Tlatin1 -mtty-char -man
#   Thanks, S.C.

exit $?   # See also the "" script.

See also Example A-39.

lex, yacc

The lex lexical analyzer produces programs for pattern matching. This has been replaced by the nonproprietary flex on Linux systems.

The yacc utility creates a parser based on a set of specifications. This has been replaced by the nonproprietary bison on Linux systems.

16.5. File and Archiving Commands



The standard UNIX archiving utility. [75] Originally a Tape ARchiving program, it has developed into a general purpose package that can handle all manner of archiving with all types of destination devices, ranging from tape drives to regular files to even stdout (see Example 3-4). GNU tar has been patched to accept various compression filters, for example: tar czvf archive_name.tar.gz *, which recursively archives and gzips all files in a directory tree except dotfiles in the current working directory ($PWD). [76]

Some useful tar options:

  1. -c create (a new archive)

  2. -x extract (files from existing archive)

  3. --delete delete (files from existing archive)


    This option will not work on magnetic tape devices.

  4. -r append (files to existing archive)

  5. -A append (tar files to existing archive)

  6. -t list (contents of existing archive)

  7. -u update archive

  8. -d compare archive with specified filesystem

  9. --after-date only process files with a date stamp after specified date

  10. -z gzip the archive

    (compress or uncompress, depending on whether combined with the -c or -x) option

  11. -j bzip2 the archive


It may be difficult to recover data from a corrupted gzipped tar archive. When archiving important files, make multiple backups.


Shell archiving utility. The text and/or binary files in a shell archive are concatenated without compression, and the resultant archive is essentially a shell script, complete with #!/bin/sh header, containing all the necessary unarchiving commands, as well as the files themselves. Unprintable binary characters in the target file(s) are converted to printable ASCII characters in the output shar file. Shar archives still show up in Usenet newsgroups, but otherwise shar has been replaced by tar/gzip. The unshar command unpacks shar archives.

The mailshar command is a Bash script that uses shar to concatenate multiple files into a single one for e-mailing. This script supports compression and uuencoding.


Creation and manipulation utility for archives, mainly used for binary object file libraries.


The Red Hat Package Manager, or rpm utility provides a wrapper for source or binary archives. It includes commands for installing and checking the integrity of packages, among other things.

A simple rpm -i package_name.rpm usually suffices to install a package, though there are many more options available.


rpm -qf identifies which package a file originates from.

bash$ rpm -qf /bin/ls


rpm -qa gives a complete list of all installed rpm packages on a given system. An rpm -qa package_name lists only the package(s) corresponding to package_name.

bash$ rpm -qa

bash$ rpm -qa docbook-utils

bash$ rpm -qa docbook | grep docbook


This specialized archiving copy command (copy input and output) is rarely seen any more, having been supplanted by tar/gzip. It still has its uses, such as moving a directory tree. With an appropriate block size (for copying) specified, it can be appreciably faster than tar.

Example 16-30. Using cpio to move a directory tree


# Copying a directory tree using cpio.

# Advantages of using 'cpio':
#   Speed of copying. It's faster than 'tar' with pipes.
#   Well suited for copying special files (named pipes, etc.)
#+  that 'cp' may choke on.


if [ $# -ne "$ARGS" ]
  echo "Usage: `basename $0` source destination"
  exit $E_BADARGS


find "$source" -depth | cpio -admvp "$destination"
#               ^^^^^         ^^^^^
#  Read the 'find' and 'cpio' info pages to decipher these options.
#  The above works only relative to $PWD (current directory) . . .
#+ full pathnames are specified.

# Exercise:
# --------

#  Add code to check the exit status ($?) of the 'find | cpio' pipe
#+ and output appropriate error messages if anything went wrong.

exit $?

This command extracts a cpio archive from an rpm one.

Example 16-31. Unpacking an rpm archive

# Unpack an 'rpm' archive

: ${1?"Usage: `basename $0` target-file"}
# Must specify 'rpm' archive name as an argument.

TEMPFILE=$$.cpio                         #  Tempfile with "unique" name.
                                         #  $$ is process ID of script.

rpm2cpio < $1 > $TEMPFILE                #  Converts rpm archive into
                                         #+ cpio archive.
cpio --make-directories -F $TEMPFILE -i  #  Unpacks cpio archive.
rm -f $TEMPFILE                          #  Deletes cpio archive.

exit 0

#  Exercise:
#  Add check for whether 1) "target-file" exists and
#+                       2) it is an rpm archive.
#  Hint:                    Parse output of 'file' command.

The pax portable archive exchange toolkit facilitates periodic file backups and is designed to be cross-compatible between various flavors of UNIX. It was designed to replace tar and cpio.

pax -wf daily_backup.pax ~/linux-server/files 
#  Creates a tar archive of all files in the target directory.
#  Note that the options to pax must be in the correct order --
#+ pax -fw     has an entirely different effect.

pax -f daily_backup.pax
#  Lists the files in the archive.

pax -rf daily_backup.pax ~/bsd-server/files
#  Restores the backed-up files from the Linux machine
#+ onto a BSD one.

Note that pax handles many of the standard archiving and compression commands.



The standard GNU/UNIX compression utility, replacing the inferior and proprietary compress. The corresponding decompression command is gunzip, which is the equivalent of gzip -d.


The -c option sends the output of gzip to stdout. This is useful when piping to other commands.

The zcat filter decompresses a gzipped file to stdout, as possible input to a pipe or redirection. This is, in effect, a cat command that works on compressed files (including files processed with the older compress utility). The zcat command is equivalent to gzip -dc.


On some commercial UNIX systems, zcat is a synonym for uncompress -c, and will not work on gzipped files.

See also Example 7-7.


An alternate compression utility, usually more efficient (but slower) than gzip, especially on large files. The corresponding decompression command is bunzip2.

Similar to the zcat command, bzcat decompresses a bzipped2-ed file to stdout.


Newer versions of tar have been patched with bzip2 support.

compress, uncompress

This is an older, proprietary compression utility found in commercial UNIX distributions. The more efficient gzip has largely replaced it. Linux distributions generally include a compress workalike for compatibility, although gunzip can unarchive files treated with compress.


The znew command transforms compressed files into gzipped ones.


Yet another compression (squeeze) utility, a filter that works only on sorted ASCII word lists. It uses the standard invocation syntax for a filter, sq < input-file > output-file. Fast, but not nearly as efficient as gzip. The corresponding uncompression filter is unsq, invoked like sq.


The output of sq may be piped to gzip for further compression.

zip, unzip

Cross-platform file archiving and compression utility compatible with DOS pkzip.exe. "Zipped" archives seem to be a more common medium of file exchange on the Internet than "tarballs."

unarc, unarj, unrar

These Linux utilities permit unpacking archives compressed with the DOS arc.exe, arj.exe, and rar.exe programs.

lzma, unlzma, lzcat

Highly efficient Lempel-Ziv-Markov compression. The syntax of lzma is similar to that of gzip. The 7-zip Website has more information.

xz, unxz, xzcat

A new high-efficiency compression tool, backward compatible with lzma, and with an invocation syntax similar to gzip. For more information, see the Wikipedia entry.

File Information


A utility for identifying file types. The command file file-name will return a file specification for file-name, such as ascii text or data. It references the magic numbers found in /usr/share/magic, /etc/magic, or /usr/lib/magic, depending on the Linux/UNIX distribution.

The -f option causes file to run in batch mode, to read from a designated file a list of filenames to analyze. The -z option, when used on a compressed target file, forces an attempt to analyze the uncompressed file type.

bash$ file test.tar.gz
test.tar.gz: gzip compressed data, deflated,
 last modified: Sun Sep 16 13:34:51 2001, os: Unix

bash file -z test.tar.gz
test.tar.gz: GNU tar archive (gzip compressed data, deflated,
 last modified: Sun Sep 16 13:34:51 2001, os: Unix)

# Find sh and Bash scripts in a given directory:

# Bourne and Bourne-Again shell scripts

file $DIRECTORY/* | fgrep $KEYWORD

# Output:

# /usr/local/bin/burn-cd:          Bourne-Again shell script text executable
# /usr/local/bin/burnit:           Bourne-Again shell script text executable
# /usr/local/bin/      Bourne shell script text executable
# /usr/local/bin/copy-cd:          Bourne-Again shell script text executable
# . . .

Example 16-32. Stripping comments from C program files

# Strips out the comments (/* COMMENT */) in a C program.


if [ $# -eq "$E_NOARGS" ]
  echo "Usage: `basename $0` C-program-file" >&2 # Error message to stderr.
  exit $E_ARGERROR

# Test for correct file type.
type=`file $1 | awk '{ print $2, $3, $4, $5 }'`
# "file $1" echoes file type . . .
# Then awk removes the first field, the filename . . .
# Then the result is fed into the variable "type."
correct_type="ASCII C program text"

if [ "$type" != "$correct_type" ]
  echo "This script works on C program files only."

# Rather cryptic sed script:
sed '
' $1
# Easy to understand if you take several hours to learn sed fundamentals.

#  Need to add one more line to the sed script to deal with
#+ case where line of code has a comment following it on same line.
#  This is left as a non-trivial exercise.

#  Also, the above code deletes non-comment lines with a "*/" . . .
#+ not a desirable result.

exit 0

# ----------------------------------------------------------------
# Code below this line will not execute because of 'exit 0' above.

# Stephane Chazelas suggests the following alternative:

usage() {
  echo "Usage: `basename $0` C-program-file" >&2
  exit 1

WEIRD=`echo -n -e '\377'`   # or WEIRD=$'\377'
[[ $# -eq 1 ]] || usage
case `file "$1"` in
  *"C program text"*) sed -e "s%/\*%${WEIRD}%g;s%\*/%${WEIRD}%g" "$1" \
     | tr '\377\n' '\n\377' \
     | sed -ne 'p;n' \
     | tr -d '\n' | tr '\377' '\n';;
  *) usage;;

#  This is still fooled by things like:
#  printf("/*");
#  or
#  /*  /* buggy embedded comment */
#  To handle all special cases (comments in strings, comments in string
#+ where there is a \", \\" ...),
#+ the only way is to write a C parser (using lex or yacc perhaps?).

exit 0

which command gives the full path to "command." This is useful for finding out whether a particular command or utility is installed on the system.

$bash which rm

For an interesting use of this command, see Example 36-16.


Similar to which, above, whereis command gives the full path to "command," but also to its manpage.

$bash whereis rm
rm: /bin/rm /usr/share/man/man1/rm.1.bz2


whatis command looks up "command" in the whatis database. This is useful for identifying system commands and important configuration files. Consider it a simplified man command.

$bash whatis whatis
whatis               (1)  - search the whatis database for complete words

Example 16-33. Exploring /usr/X11R6/bin


# What are all those mysterious binaries in /usr/X11R6/bin?

# Try also "/bin", "/usr/bin", "/usr/local/bin", etc.

for file in $DIRECTORY/*
  whatis `basename $file`   # Echoes info about the binary.

exit 0

#  Note: For this to work, you must create a "whatis" database
#+ with /usr/sbin/makewhatis.
#  You may wish to redirect output of this script, like so:
#    ./ >>whatis.db
#  or view it a page at a time on stdout,
#    ./ | less

See also Example 11-3.


Show a detailed directory listing. The effect is similar to ls -lb.

This is one of the GNU fileutils.

bash$ vdir
total 10
 -rw-r--r--    1 bozo  bozo      4034 Jul 18 22:04 data1.xrolo
 -rw-r--r--    1 bozo  bozo      4602 May 25 13:58 data1.xrolo.bak
 -rw-r--r--    1 bozo  bozo       877 Dec 17  2000 employment.xrolo

bash ls -l
total 10
 -rw-r--r--    1 bozo  bozo      4034 Jul 18 22:04 data1.xrolo
 -rw-r--r--    1 bozo  bozo      4602 May 25 13:58 data1.xrolo.bak
 -rw-r--r--    1 bozo  bozo       877 Dec 17  2000 employment.xrolo

locate, slocate

The locate command searches for files using a database stored for just that purpose. The slocate command is the secure version of locate (which may be aliased to slocate).

$bash locate hickson

getfacl, setfacl

These commands retrieve or set the file access control list -- the owner, group, and file permissions.

bash$ getfacl *
# file: test1.txt
 # owner: bozo
 # group: bozgrp

 # file: test2.txt
 # owner: bozo
 # group: bozgrp

bash$ setfacl -m u:bozo:rw yearly_budget.csv
bash$ getfacl yearly_budget.csv
# file: yearly_budget.csv
 # owner: accountant
 # group: budgetgrp


Disclose the file that a symbolic link points to.

bash$ readlink /usr/bin/awk


Use the strings command to find printable strings in a binary or data file. It will list sequences of printable characters found in the target file. This might be handy for a quick 'n dirty examination of a core dump or for looking at an unknown graphic image file (strings image-file | more might show something like JFIF, which would identify the file as a jpeg graphic). In a script, you would probably parse the output of strings with grep or sed. See Example 11-8 and Example 11-10.

Example 16-34. An "improved" strings command

# "word-strings" (enhanced "strings" command)
#  This script filters the output of "strings" by checking it
#+ against a standard word list file.
#  This effectively eliminates gibberish and noise,
#+ and outputs only recognized words.

# ===========================================================
#                 Standard Check for Script Argument(s)

if [ $# -ne $ARGS ]
  echo "Usage: `basename $0` filename"
  exit $E_BADARGS

if [ ! -f "$1" ]                      # Check if file exists.
    echo "File \"$1\" does not exist."
    exit $E_NOFILE
# ===========================================================

MINSTRLEN=3                           #  Minimum string length.
WORDFILE=/usr/share/dict/linux.words  #  Dictionary file.
#  May specify a different word list file
#+ of one-word-per-line format.
#  For example, the "yawl" word-list package,

wlist=`strings "$1" | tr A-Z a-z | tr '[:space:]' Z | \
       tr -cs '[:alpha:]' Z | tr -s '\173-\377' Z | tr Z ' '`

# Translate output of 'strings' command with multiple passes of 'tr'.
#  "tr A-Z a-z"  converts to lowercase.
#  "tr '[:space:]'"  converts whitespace characters to Z's.
#  "tr -cs '[:alpha:]' Z"  converts non-alphabetic characters to Z's,
#+ and squeezes multiple consecutive Z's.
#  "tr -s '\173-\377' Z"  converts all characters past 'z' to Z's
#+ and squeezes multiple consecutive Z's,
#+ which gets rid of all the weird characters that the previous
#+ translation failed to deal with.
#  Finally, "tr Z ' '" converts all those Z's to whitespace,
#+ which will be seen as word separators in the loop below.

#  ***********************************************************************
#  Note the technique of feeding/piping the output of 'tr' back to itself,
#+ but with different arguments and/or options on each successive pass.
#  ***********************************************************************

for word in $wlist                    #  Important:
                                      #  $wlist must not be quoted here.
                                      # "$wlist" does not work.
                                      #  Why not?
  strlen=${#word}                     #  String length.
  if [ "$strlen" -lt "$MINSTRLEN" ]   #  Skip over short strings.

  grep -Fw $word "$WORDFILE"          #   Match whole words only.
#      ^^^                            #  "Fixed strings" and
                                      #+ "whole words" options. 

exit $?


diff, patch

diff: flexible file comparison utility. It compares the target files line-by-line sequentially. In some applications, such as comparing word dictionaries, it may be helpful to filter the files through sort and uniq before piping them to diff. diff file-1 file-2 outputs the lines in the files that differ, with carets showing which file each particular line belongs to.

The --side-by-side option to diff outputs each compared file, line by line, in separate columns, with non-matching lines marked. The -c and -u options likewise make the output of the command easier to interpret.

There are available various fancy frontends for diff, such as sdiff, wdiff, xdiff, and mgdiff.


The diff command returns an exit status of 0 if the compared files are identical, and 1 if they differ (or 2 when binary files are being compared). This permits use of diff in a test construct within a shell script (see below).

A common use for diff is generating difference files to be used with patch The -e option outputs files suitable for ed or ex scripts.

patch: flexible versioning utility. Given a difference file generated by diff, patch can upgrade a previous version of a package to a newer version. It is much more convenient to distribute a relatively small "diff" file than the entire body of a newly revised package. Kernel "patches" have become the preferred method of distributing the frequent releases of the Linux kernel.

patch -p1 <patch-file
# Takes all the changes listed in 'patch-file'
# and applies them to the files referenced therein.
# This upgrades to a newer version of the package.

Patching the kernel:

cd /usr/src
gzip -cd patchXX.gz | patch -p0
# Upgrading kernel source using 'patch'.
# From the Linux kernel docs "README",
# by anonymous author (Alan Cox?).


The diff command can also recursively compare directories (for the filenames present).

bash$ diff -r ~/notes1 ~/notes2
Only in /home/bozo/notes1: file02
 Only in /home/bozo/notes1: file03
 Only in /home/bozo/notes2: file04


Use zdiff to compare gzipped files.


Use diffstat to create a histogram (point-distribution graph) of output from diff.

diff3, merge

An extended version of diff that compares three files at a time. This command returns an exit value of 0 upon successful execution, but unfortunately this gives no information about the results of the comparison.

bash$ diff3 file-1 file-2 file-3
   This is line 1 of "file-1".
   This is line 1 of "file-2".
   This is line 1 of "file-3"

The merge (3-way file merge) command is an interesting adjunct to diff3. Its syntax is merge Mergefile file1 file2. The result is to output to Mergefile the changes that lead from file1 to file2. Consider this command a stripped-down version of patch.


Compare and/or edit two files in order to merge them into an output file. Because of its interactive nature, this command would find little use in a script.


The cmp command is a simpler version of diff, above. Whereas diff reports the differences between two files, cmp merely shows at what point they differ.


Like diff, cmp returns an exit status of 0 if the compared files are identical, and 1 if they differ. This permits use in a test construct within a shell script.

Example 16-35. Using cmp to compare two files within a script.


ARGS=2  # Two args to script expected.

if [ $# -ne "$ARGS" ]
  echo "Usage: `basename $0` file1 file2"
  exit $E_BADARGS

if [[ ! -r "$1" || ! -r "$2" ]]
  echo "Both files to be compared must exist and be readable."

cmp $1 $2 &> /dev/null
#   Redirection to /dev/null buries the output of the "cmp" command.
#   cmp -s $1 $2  has same result ("-s" silent flag to "cmp")
#   Thank you  Anders Gustavsson for pointing this out.
#  Also works with 'diff', i.e.,
#+ diff $1 $2 &> /dev/null

if [ $? -eq 0 ]         # Test exit status of "cmp" command.
  echo "File \"$1\" is identical to file \"$2\"."
  echo "File \"$1\" differs from file \"$2\"."

exit 0


Use zcmp on gzipped files.


Versatile file comparison utility. The files must be sorted for this to be useful.

comm -options first-file second-file

comm file-1 file-2 outputs three columns:

  • column 1 = lines unique to file-1

  • column 2 = lines unique to file-2

  • column 3 = lines common to both.

The options allow suppressing output of one or more columns.

  • -1 suppresses column 1

  • -2 suppresses column 2

  • -3 suppresses column 3

  • -12 suppresses both columns 1 and 2, etc.

This command is useful for comparing "dictionaries" or word lists -- sorted text files with one word per line.



Strips the path information from a file name, printing only the file name. The construction basename $0 lets the script know its name, that is, the name it was invoked by. This can be used for "usage" messages if, for example a script is called with missing arguments:
echo "Usage: `basename $0` arg1 arg2 ... argn"


Strips the basename from a filename, printing only the path information.


basename and dirname can operate on any arbitrary string. The argument does not need to refer to an existing file, or even be a filename for that matter (see Example A-7).

Example 16-36. basename and dirname



echo "Basename of /home/bozo/daily-journal.txt = `basename $address`"
echo "Dirname of /home/bozo/daily-journal.txt = `dirname $address`"
echo "My own home is `basename ~/`."         # `basename ~` also works.
echo "The home of my home is `dirname ~/`."  # `dirname ~`  also works.

exit 0
split, csplit

These are utilities for splitting a file into smaller chunks. Their usual use is for splitting up large files in order to back them up on floppies or preparatory to e-mailing or uploading them.

The csplit command splits a file according to context, the split occuring where patterns are matched.

Example 16-37. A script that copies itself in sections


#  A script that splits itself into chunks,
#+ then reassembles the chunks into an exact copy
#+ of the original script.

CHUNKSIZE=4    #  Size of first chunk of split files.
OUTPREFIX=xx   #  csplit prefixes, by default,
               #+ files with "xx" ...

csplit "$0" "$CHUNKSIZE"

# Some comment lines for padding . . .
# Line 15
# Line 16
# Line 17
# Line 18
# Line 19
# Line 20

cat "$OUTPREFIX"* > "$0.copy"  # Concatenate the chunks.
rm "$OUTPREFIX"*               # Get rid of the chunks.

exit $?

Encoding and Encryption

sum, cksum, md5sum, sha1sum

These are utilities for generating checksums. A checksum is a number [77] mathematically calculated from the contents of a file, for the purpose of checking its integrity. A script might refer to a list of checksums for security purposes, such as ensuring that the contents of key system files have not been altered or corrupted. For security applications, use the md5sum (message digest 5 checksum) command, or better yet, the newer sha1sum (Secure Hash Algorithm). [78]

bash$ cksum /boot/vmlinuz
1670054224 804083 /boot/vmlinuz

bash$ echo -n "Top Secret" | cksum
3391003827 10

bash$ md5sum /boot/vmlinuz
0f43eccea8f09e0a0b2b5cf1dcf333ba  /boot/vmlinuz

bash$ echo -n "Top Secret" | md5sum
8babc97a6f62a4649716f4df8d61728f  -


The cksum command shows the size, in bytes, of its target, whether file or stdout.

The md5sum and sha1sum commands display a dash when they receive their input from stdout.

Example 16-38. Checking file integrity

# Checking whether files in a given directory
#                    have been tampered with.


# Filename for storing records (database file).

set_up_database ()
  echo ""$directory"" > "$dbfile"
  # Write directory name to first line of file.
  md5sum "$directory"/* >> "$dbfile"
  # Append md5 checksums and filenames.

check_database ()
  local n=0
  local filename
  local checksum

  # ------------------------------------------- #
  #  This file check should be unnecessary,
  #+ but better safe than sorry.

  if [ ! -r "$dbfile" ]
    echo "Unable to read checksum database file!"
    exit $E_BAD_DBFILE
  # ------------------------------------------- #

  while read record[n]

    if [ "$directory_checked" != "$directory" ]
      echo "Directories do not match up!"
      # Tried to use file for a different directory.
      exit $E_DIR_NOMATCH

    if [ "$n" -gt 0 ]   # Not directory name.
      filename[n]=$( echo ${record[$n]} | awk '{ print $2 }' )
      #  md5sum writes records backwards,
      #+ checksum first, then filename.
      checksum[n]=$( md5sum "${filename[n]}" )

      if [ "${record[n]}" = "${checksum[n]}" ]
        echo "${filename[n]} unchanged."

        elif [ "`basename ${filename[n]}`" != "$dbfile" ]
               #  Skip over checksum database file,
               #+ as it will change with each invocation of script.
               #  ---
               #  This unfortunately means that when running
               #+ this script on $PWD, tampering with the
               #+ checksum database file will not be detected.
               #  Exercise: Fix this.
          echo "${filename[n]} : CHECKSUM ERROR!"
        # File has been changed since last checked.


    let "n+=1"
  done <"$dbfile"       # Read from checksum database file. 


# =================================================== #
# main ()

if [ -z  "$1" ]
  directory="$PWD"      #  If not specified,
else                    #+ use current working directory.

clear                   # Clear screen.
echo " Running file integrity check on $directory"

# ------------------------------------------------------------------ #
  if [ ! -r "$dbfile" ] # Need to create database file?
    echo "Setting up database file, \""$directory"/"$dbfile"\"."; echo
# ------------------------------------------------------------------ #

check_database          # Do the actual work.


#  You may wish to redirect the stdout of this script to a file,
#+ especially if the directory checked has many files in it.

exit 0

#  For a much more thorough file integrity check,
#+ consider the "Tripwire" package,

Also see Example A-19, Example 36-16, and Example 10-2 for creative uses of the md5sum command.


There have been reports that the 128-bit md5sum can be cracked, so the more secure 160-bit sha1sum is a welcome new addition to the checksum toolkit.

bash$ md5sum testfile
e181e2c8720c60522c4c4c981108e367  testfile

bash$ sha1sum testfile
5d7425a9c08a66c3177f1e31286fa40986ffc996  testfile

Security consultants have demonstrated that even sha1sum can be compromised. Fortunately, newer Linux distros include longer bit-length sha224sum, sha256sum, sha384sum, and sha512sum commands.


This utility encodes binary files (images, sound files, compressed files, etc.) into ASCII characters, making them suitable for transmission in the body of an e-mail message or in a newsgroup posting. This is especially useful where MIME (multimedia) encoding is not available.


This reverses the encoding, decoding uuencoded files back into the original binaries.

Example 16-39. Uudecoding encoded files

# Uudecodes all uuencoded files in current working directory.

lines=35        # Allow 35 lines for the header (very generous).

for File in *   # Test all the files in $PWD.
  search1=`head -n $lines $File | grep begin | wc -w`
  search2=`tail -n $lines $File | grep end | wc -w`
  #  Uuencoded files have a "begin" near the beginning,
  #+ and an "end" near the end.
  if [ "$search1" -gt 0 ]
    if [ "$search2" -gt 0 ]
      echo "uudecoding - $File -"
      uudecode $File

#  Note that running this script upon itself fools it
#+ into thinking it is a uuencoded file,
#+ because it contains both "begin" and "end".

#  Exercise:
#  --------
#  Modify this script to check each file for a newsgroup header,
#+ and skip to next if not found.

exit 0


The fold -s command may be useful (possibly in a pipe) to process long uudecoded text messages downloaded from Usenet newsgroups.

mimencode, mmencode

The mimencode and mmencode commands process multimedia-encoded e-mail attachments. Although mail user agents (such as pine or kmail) normally handle this automatically, these particular utilities permit manipulating such attachments manually from the command-line or in batch processing mode by means of a shell script.


At one time, this was the standard UNIX file encryption utility. [79] Politically-motivated government regulations prohibiting the export of encryption software resulted in the disappearance of crypt from much of the UNIX world, and it is still missing from most Linux distributions. Fortunately, programmers have come up with a number of decent alternatives to it, among them the author's very own cruft (see Example A-4).


This is an Open Source implementation of Secure Sockets Layer encryption.
# To encrypt a file:
openssl aes-128-ecb -salt -in file.txt -out file.encrypted \
-pass pass:my_password
#          ^^^^^^^^^^^   User-selected password.
#       aes-128-ecb      is the encryption method chosen.

# To decrypt an openssl-encrypted file:
openssl aes-128-ecb -d -salt -in file.encrypted -out file.txt \
-pass pass:my_password
#          ^^^^^^^^^^^   User-selected password.

Piping openssl to/from tar makes it possible to encrypt an entire directory tree.
# To encrypt a directory:


tar czvf - "$sourcedir" |
openssl des3 -salt -out "$encrfile" -pass pass:"$password"
#       ^^^^   Uses des3 encryption.
# Writes encrypted file "encr-dir.tar.gz" in current working directory.

# To decrypt the resulting tarball:
openssl des3 -d -salt -in "$encrfile" -pass pass:"$password" |
tar -xzv
# Decrypts and unpacks into current working directory.

Of course, openssl has many other uses, such as obtaining signed certificates for Web sites. See the info page.


Securely erase a file by overwriting it multiple times with random bit patterns before deleting it. This command has the same effect as Example 16-61, but does it in a more thorough and elegant manner.

This is one of the GNU fileutils.


Advanced forensic technology may still be able to recover the contents of a file, even after application of shred.



Create a temporary file [80] with a "unique" filename. When invoked from the command-line without additional arguments, it creates a zero-length file in the /tmp directory.

bash$ mktemp

tempfile=`mktemp $PREFIX.XXXXXX`
#                        ^^^^^^ Need at least 6 placeholders
#+                              in the filename template.
#   If no filename template supplied,
#+ "tmp.XXXXXXXXXX" is the default.

echo "tempfile name = $tempfile"
# tempfile name = filename.QA2ZpY
#                 or something similar...

#  Creates a file of that name in the current working directory
#+ with 600 file permissions.
#  A "umask 177" is therefore unnecessary,
#+ but it's good programming practice nevertheless.


Utility for building and compiling binary packages. This can also be used for any set of operations triggered by incremental changes in source files.

The make command checks a Makefile, a list of file dependencies and operations to be carried out.

The make utility is, in effect, a powerful scripting language similar in many ways to Bash, but with the capability of recognizing dependencies. For in-depth coverage of this useful tool set, see the GNU software documentation site.


Special purpose file copying command, similar to cp, but capable of setting permissions and attributes of the copied files. This command seems tailormade for installing software packages, and as such it shows up frequently in Makefiles (in the make install : section). It could likewise prove useful in installation scripts.


This utility, written by Benjamin Lin and collaborators, converts DOS-formatted text files (lines terminated by CR-LF) to UNIX format (lines terminated by LF only), and vice-versa.


The ptx [targetfile] command outputs a permuted index (cross-reference list) of the targetfile. This may be further filtered and formatted in a pipe, if necessary.

more, less

Pagers that display a text file or stream to stdout, one screenful at a time. These may be used to filter the output of stdout . . . or of a script.

An interesting application of more is to "test drive" a command sequence, to forestall potentially unpleasant consequences.
ls /home/bozo | awk '{print "rm -rf " $1}' | more
#                                            ^^^^
# Testing the effect of the following (disastrous) command-line:
#      ls /home/bozo | awk '{print "rm -rf " $1}' | sh
#      Hand off to the shell to execute . . .       ^^

The less pager has the interesting property of doing a formatted display of man page source. See Example A-39.

16.6. Communications Commands

Certain of the following commands find use in network data transfer and analysis, as well as in chasing spammers.

Information and Statistics


Searches for information about an Internet host by name or IP address, using DNS.

bash$ host has address


Displays IP information for a host. With the -h option, ipcalc does a reverse DNS lookup, finding the name of the host (server) from the IP address.

bash$ ipcalc -h


Do an Internet "name server lookup" on a host by IP address. This is essentially equivalent to ipcalc -h or dig -x . The command may be run either interactively or noninteractively, i.e., from within a script.

The nslookup command has allegedly been "deprecated," but it is still useful.

bash$ nslookup -sil

 Non-authoritative answer:


Domain Information Groper. Similar to nslookup, dig does an Internet name server lookup on a host. May be run from the command-line or from within a script.

Some interesting options to dig are +time=N for setting a query timeout to N seconds, +nofail for continuing to query servers until a reply is received, and -x for doing a reverse address lookup.

Compare the output of dig -x with ipcalc -h and nslookup.

bash$ dig -x
;; Got answer:
 ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 11649
 ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0

 ;         IN      PTR

 ;; AUTHORITY SECTION:    3600    IN      SOA
 2002031705 900 600 86400 3600

 ;; Query time: 537 msec
 ;; WHEN: Wed Jun 26 08:35:24 2002
 ;; MSG SIZE  rcvd: 91

Example 16-40. Finding out where to report a spammer

# Look up abuse contact to report a spammer.
# Thanks, Michael Zick.

# Check for command-line arg.
if [ $# -ne "$ARGCOUNT" ]
  echo "Usage: `basename $0` domain-name"

dig +short $ -c in -t txt
# Also try:
#     dig +nssearch $1
#     Tries to find "authoritative name servers" and display SOA records.

# The following also works:
#     whois -h $1
#           ^^ ^^^^^^^^^^^^^^^  Specify host.  
#     Can even lookup multiple spammers with this, i.e."
#     whois -h $spamdomain1 $spamdomain2 . . .

#  Exercise:
#  --------
#  Expand the functionality of this script
#+ so that it automatically e-mails a notification
#+ to the responsible ISP's contact address(es).
#  Hint: use the "mail" command.

exit $?

#                A known spam domain.

# "[email protected]"
# "[email protected]"
# "[email protected]"

#  For a more elaborate version of this script,
#+ see the SpamViz home page,

Example 16-41. Analyzing a spam domain

#! /bin/bash
# Identifying spam domains

# $Id: is-spammer, v 1.4 2004/09/01 19:37:52 mszick Exp $
# Above line is RCS ID info.
#  This is a simplified version of the "is_spammer.bash
#+ script in the Contributed Scripts appendix.

# is-spammer <>

# Uses an external program: 'dig'
# Tested with version: 9.2.4rc5

# Uses functions.
# Uses IFS to parse strings by assignment into arrays.
# And even does something useful: checks e-mail blacklists.

# Use the from the text body:
#                       ^^^^^^^^^^^
# Or the from any e-mail address:
# [email protected]
# as the only argument to this script.
#(PS: have your Inet connection running)
# So, to invoke this script in the above two instances:

# Whitespace == :Space:Tab:Line Feed:Carriage Return:

# No Whitespace == Line Feed:Carriage Return

# Field separator for dotted decimal ip addresses

# Get the dns text resource record.
# get_txt <error_code> <list_query>
get_txt() {

    # Parse $1 by assignment at the dots.
    local -a dns
    dns=( $1 )
    if [ "${dns[0]}" == '127' ]
        # See if there is a reason.
        echo $(dig +short $2 -t txt)

# Get the dns address resource record.
# chk_adr <rev_dns> <list_server>
chk_adr() {
    local reply
    local server
    local reason

    reply=$( dig +short ${server} )

    # If reply might be an error code . . .
    if [ ${#reply} -gt 6 ]
        reason=$(get_txt ${reply} ${server} )
    echo ${reason:-' not blacklisted.'}

# Need to get the IP address from the name.
echo 'Get address of: '$1
ip_adr=$(dig +short $1)
dns_reply=${ip_adr:-' no answer '}
echo ' Found address: '${dns_reply}

# A valid reply is at least 4 digits plus 3 dots.
if [ ${#ip_adr} -gt 6 ]
    declare query

    # Parse by assignment at the dots.
    declare -a dns
    dns=( ${ip_adr} )

    # Reorder octets into dns query order.

# See: (Conservative, well maintained)
    echo -n ' says: '
    echo $(chk_adr ${rev_dns} '')

# See: (Open mail relays)
    echo -n '  says: '
    echo $(chk_adr ${rev_dns} '')

# See: (You can report spammers here)
    echo -n ' says: '
    echo $(chk_adr ${rev_dns} '')

# # # other blacklist operations # # #

# See:
    echo -n ' says: '
    echo $(chk_adr ${rev_dns} '')

# See: (Various mail relays)
    echo 'Distributed Server Listings'
    echo -n ' says: '
    echo $(chk_adr ${rev_dns} '')

    echo -n ' says: '
    echo $(chk_adr ${rev_dns} '')

    echo -n ' says: '
    echo $(chk_adr ${rev_dns} '')

    echo 'Could not use that address.'

exit 0

# Exercises:
# --------

# 1) Check arguments to script,
#    and exit with appropriate error message if necessary.

# 2) Check if on-line at invocation of script,
#    and exit with appropriate error message if necessary.

# 3) Substitute generic variables for "hard-coded" BHL domains.

# 4) Set a time-out for the script using the "+time=" option
     to the 'dig' command.

For a much more elaborate version of the above script, see Example A-28.


Trace the route taken by packets sent to a remote host. This command works within a LAN, WAN, or over the Internet. The remote host may be specified by an IP address. The output of this command may be filtered by grep or sed in a pipe.

bash$ traceroute
traceroute to (, 30 hops max, 38 byte packets
 1 (  191.303 ms  179.400 ms  179.767 ms
 2 (  179.536 ms  179.534 ms  169.685 ms
 3 (  189.471 ms  189.556 ms *


Broadcast an ICMP ECHO_REQUEST packet to another machine, either on a local or remote network. This is a diagnostic tool for testing network connections, and it should be used with caution.

bash$ ping localhost
PING localhost.localdomain ( from : 56(84) bytes of data.
 64 bytes from localhost.localdomain ( icmp_seq=0 ttl=255 time=709 usec
 64 bytes from localhost.localdomain ( icmp_seq=1 ttl=255 time=286 usec

 --- localhost.localdomain ping statistics ---
 2 packets transmitted, 2 packets received, 0% packet loss
 round-trip min/avg/max/mdev = 0.286/0.497/0.709/0.212 ms

A successful ping returns an exit status of 0. This can be tested for in a script.  # Notorious spammer.
# HNAME=$HOST     # Debug: test for localhost.
  count=2  # Send only two pings.

if [[ `ping -c $count "$HNAME"` ]]
  echo ""$HNAME" still up and broadcasting spam your way."
  echo ""$HNAME" seems to be down. Pity."


Perform a DNS (Domain Name System) lookup. The -h option permits specifying which particular whois server to query. See Example 4-6 and Example 16-40.


Retrieve information about users on a network. Optionally, this command can display a user's ~/.plan, ~/.project, and ~/.forward files, if present.

bash$ finger
Login  Name           Tty      Idle  Login Time   Office     Office Phone
 bozo   Bozo Bozeman   tty1        8  Jun 25 16:59                (:0)
 bozo   Bozo Bozeman   ttyp0          Jun 25 16:59                (:0.0)
 bozo   Bozo Bozeman   ttyp1          Jun 25 17:07                (:0.0)

bash$ finger bozo
Login: bozo                             Name: Bozo Bozeman
 Directory: /home/bozo                   Shell: /bin/bash
 Office: 2355 Clown St., 543-1234
 On since Fri Aug 31 20:13 (MST) on tty1    1 hour 38 minutes idle
 On since Fri Aug 31 20:13 (MST) on pts/0   12 seconds idle
 On since Fri Aug 31 20:13 (MST) on pts/1
 On since Fri Aug 31 20:31 (MST) on pts/2   1 hour 16 minutes idle
 Mail last read Tue Jul  3 10:08 2007 (MST) 
 No Plan.

Out of security considerations, many networks disable finger and its associated daemon. [81]


Change information disclosed by the finger command.


Verify an Internet e-mail address.

This command seems to be missing from newer Linux distros.

Remote Host Access

sx, rx

The sx and rx command set serves to transfer files to and from a remote host using the xmodem protocol. These are generally part of a communications package, such as minicom.

sz, rz

The sz and rz command set serves to transfer files to and from a remote host using the zmodem protocol. Zmodem has certain advantages over xmodem, such as faster transmission rate and resumption of interrupted file transfers. Like sx and rx, these are generally part of a communications package.


Utility and protocol for uploading / downloading files to or from a remote host. An ftp session can be automated in a script (see Example 19-6 and Example A-4).

uucp, uux, cu

uucp: UNIX to UNIX copy. This is a communications package for transferring files between UNIX servers. A shell script is an effective way to handle a uucp command sequence.

Since the advent of the Internet and e-mail, uucp seems to have faded into obscurity, but it still exists and remains perfectly workable in situations where an Internet connection is not available or appropriate. The advantage of uucp is that it is fault-tolerant, so even if there is a service interruption the copy operation will resume where it left off when the connection is restored.


uux: UNIX to UNIX execute. Execute a command on a remote system. This command is part of the uucp package.


cu: Call Up a remote system and connect as a simple terminal. It is a sort of dumbed-down version of telnet. This command is part of the uucp package.


Utility and protocol for connecting to a remote host.


The telnet protocol contains security holes and should therefore probably be avoided. Its use within a shell script is not recommended.


The wget utility noninteractively retrieves or downloads files from a Web or ftp site. It works well in a script.

wget -p
#  The -p or --page-requisite option causes wget to fetch all files
#+ required to display the specified page.

wget -r -O $SAVEFILE
#  The -r option recursively follows and retrieves all links
#+ on the specified site.

wget -c
#  The -c option lets wget resume an interrupted download.
#  This works with ftp servers and many HTTP sites.

Example 16-42. Getting a stock quote

# Download a stock quote.


if [ -z "$1" ]  # Must specify a stock (symbol) to fetch.
  then echo "Usage: `basename $0` stock-symbol"
  exit $E_NOPARAMS


# Fetches an HTML file, so name it appropriately.
# Yahoo finance board, with stock query suffix.

# -----------------------------------------------------------
wget -O ${stock_symbol}${file_suffix} "${URL}${stock_symbol}"
# -----------------------------------------------------------

# To look up stuff on
# -----------------------------------------------------------
# URL="${query}"
# wget -O "$savefilename" "${URL}"
# -----------------------------------------------------------
# Saves a list of relevant URLs.

exit $?

# Exercises:
# ---------
# 1) Add a test to ensure the user running the script is on-line.
#    (Hint: parse the output of 'ps -ax' for "ppp" or "connect."
# 2) Modify this script to fetch the local weather report,
#+   taking the user's zip code as an argument.

See also Example A-30 and Example A-31.


The lynx Web and file browser can be used inside a script (with the -dump option) to retrieve a file from a Web or ftp site noninteractively.

lynx -dump >$SAVEFILE

With the -traversal option, lynx starts at the HTTP URL specified as an argument, then "crawls" through all links located on that particular server. Used together with the -crawl option, outputs page text to a log file.


Remote login, initates a session on a remote host. This command has security issues, so use ssh instead.


Remote shell, executes command(s) on a remote host. This has security issues, so use ssh instead.


Remote copy, copies files between two different networked machines.


Remote synchronize, updates (synchronizes) files between two different networked machines.

bash$ rsync -a ~/sourcedir/*txt /node1/subdirectory/

Example 16-43. Updating FC4


# Script author: Frank Wang.
# Slight stylistic modifications by ABS Guide author.
# Used in ABS Guide with permission.

#  Download Fedora Core 4 update from mirror site using rsync. 
#  Should also work for newer Fedora Cores -- 5, 6, . . .
#  Only download latest package if multiple versions exist,
#+ to save space.

# URL=rsync://
# URL=rsync://

LOG=/tmp/repo-update-$(/bin/date +%Y-%m-%d).txt

E_RETURN=85        # Something unexpected happened.

# General rsync options
# -r: recursive download
# -t: reserve time
# -v: verbose

OPTS="-rtv --delete-excluded --delete-after --partial"

# rsync include pattern
# Leading slash causes absolute path name match.
#   ^                         ^
# Quoting is necessary to prevent globbing.

# rsync exclude pattern
# Temporarily comment out unwanted pkgs using "#" . . .
#  "/4/i386/kernel-xen*" 
#  "/4/i386/xen-*" 

init () {
    # Let pipe command return possible rsync error, e.g., stalled network.
    set -o pipefail                  # Newly introduced in Bash, version 3.

    TMP=${TMPDIR:-/tmp}/${0##*/}.$$  # Store refined download list.
    trap "{
        rm -f $TMP 2>/dev/null
    }" EXIT                          # Clear temporary file on exit.

check_pid () {
# Check if process exists.
    if [ -s "$PID_FILE" ]; then
        echo "PID file exists. Checking ..."
        PID=$(/bin/egrep -o "^[[:digit:]]+" $PID_FILE)
        if /bin/ps --pid $PID &>/dev/null; then
            echo "Process $PID found. ${0##*/} seems to be running!"
           /usr/bin/logger -t ${0##*/} \
                 "Process $PID found. ${0##*/} seems to be running!"
            exit $E_RETURN
        echo "Process $PID not found. Start new process . . ."

#  Set overall file update range starting from root or $URL,
#+ according to above patterns.
set_range () {
    for p in "${INCLUDE[@]}"; do
        include="$include --include \"$p\""

    for p in "${EXCLUDE[@]}"; do
        exclude="$exclude --exclude \"$p\""

# Retrieve and refine rsync update list.
get_list () {
    echo $$ > $PID_FILE || {
        echo "Can't write to pid file $PID_FILE"
        exit $E_RETURN

    echo -n "Retrieving and refining update list . . ."

    # Retrieve list -- 'eval' is needed to run rsync as a single command.
    # $3 and $4 is the date and time of file creation.
    # $5 is the full package name.
    eval /bin/nice /usr/bin/rsync \
        -r $include $exclude $URL | \
        egrep '^dr.x|^-r' | \
        awk '{print $3, $4, $5}' | \
        sort -k3 | \
        { while read line; do
            # Get seconds since epoch, to filter out obsolete pkgs.
            cur_date=$(date -d "$(echo $line | awk '{print $1, $2}')" +%s)
            #  echo $cur_date

            # Get file name.
            cur_file=$(echo $line | awk '{print $3}')
            #  echo $cur_file

            # Get rpm pkg name from file name, if possible.
            if [[ $cur_file == *rpm ]]; then
                pkg_name=$(echo $cur_file | sed -r -e \
            # echo $pkg_name

            if [ -z "$pkg_name" ]; then   #  If not a rpm file,
                echo $cur_file >> $TMP    #+ then append to download list.
            elif [ "$pkg_name" != "$previous" ]; then   # A new pkg found.
                echo $pre_file >> $TMP                  # Output latest file.
                previous=$pkg_name                      # Save current.
            elif [ "$cur_date" -gt "$pre_date" ]; then
                                                #  If same pkg, but newer,
                pre_date=$cur_date              #+ then update latest pointer.
            echo $pre_file >> $TMP              #  TMP contains ALL
                                                #+ of refined list now.
            # echo "subshell=$BASH_SUBSHELL"

    }       # Bracket required here to let final "echo $pre_file >> $TMP" 
            # Remained in the same subshell ( 1 ) with the entire loop.

    RET=$?  # Get return code of the pipe command.

    [ "$RET" -ne 0 ] && {
        echo "List retrieving failed with code $RET"
        exit $E_RETURN

    echo "done"; echo

# Real rsync download part.
get_file () {

    echo "Downloading..."
    /bin/nice /usr/bin/rsync \
        $OPTS \
        --filter "merge,+/ $TMP" \
        --exclude '*'  \
        $URL $DEST     \
        | /usr/bin/tee $LOG


   #  --filter merge,+/ is crucial for the intention. 
   #  + modifier means include and / means absolute path.
   #  Then sorted list in $TMP will contain ascending dir name and 
   #+ prevent the following --exclude '*' from "shortcutting the circuit." 

    echo "Done"

    rm -f $PID_FILE 2>/dev/null

    return $RET

# -------
# Main
# -------

if [ "$RET" -eq 0 ]; then
    /usr/bin/logger -t ${0##*/} "Fedora update mirrored successfully."
    /usr/bin/logger -t ${0##*/} \
    "Fedora update mirrored with failure code: $RET"

exit $RET

See also Example A-32.


Using rcp, rsync, and similar utilities with security implications in a shell script may not be advisable. Consider, instead, using ssh, scp, or an expect script.


Secure shell, logs onto a remote host and executes commands there. This secure replacement for telnet, rlogin, rcp, and rsh uses identity authentication and encryption. See its manpage for details.

Example 16-44. Using ssh

# remote.bash: Using ssh.

# This example by Michael Zick.
# Used with permission.

#   Presumptions:
#   ------------
#   fd-2 isn't being captured ( '2>/dev/null' ).
#   ssh/sshd presumes stderr ('2') will display to user.
#   sshd is running on your machine.
#   For any 'standard' distribution, it probably is,
#+  and without any funky ssh-keygen having been done.

# Try ssh to your machine from the command-line:
# $ ssh $HOSTNAME
# Without extra set-up you'll be asked for your password.
#   enter password
#   when done,  $ exit
# Did that work? If so, you're ready for more fun.

# Try ssh to your machine as 'root':
#   $  ssh -l root $HOSTNAME
#   When asked for password, enter root's, not yours.
#          Last login: Tue Aug 10 20:25:49 2004 from localhost.localdomain
#   Enter 'exit' when done.

#  The above gives you an interactive shell.
#  It is possible for sshd to be set up in a 'single command' mode,
#+ but that is beyond the scope of this example.
#  The only thing to note is that the following will work in
#+ 'single command' mode.

# A basic, write stdout (local) command.

ls -l

# Now the same basic command on a remote machine.
# Pass a different 'USERNAME' 'HOSTNAME' if desired:

#  Now excute the above command-line on the remote host,
#+ with all transmissions encrypted.

ssh -l ${USER} ${HOST} " ls -l "

#  The expected result is a listing of your username's home
#+ directory on the remote machine.
#  To see any difference, run this script from somewhere
#+ other than your home directory.

#  In other words, the Bash command is passed as a quoted line
#+ to the remote shell, which executes it on the remote machine.
#  In this case, sshd does  ' bash -c "ls -l" '   on your behalf.

#  For information on topics such as not having to enter a
#+ password/passphrase for every command-line, see
#+    man ssh
#+    man ssh-keygen
#+    man sshd_config.

exit 0


Within a loop, ssh may cause unexpected behavior. According to a Usenet post in the comp.unix shell archives, ssh inherits the loop's stdin. To remedy this, pass ssh either the -n or -f option.

Thanks, Jason Bechtel, for pointing this out.


Secure copy, similar in function to rcp, copies files between two different networked machines, but does so using authentication, and with a security level similar to ssh.

Local Network


This is a utility for terminal-to-terminal communication. It allows sending lines from your terminal (console or xterm) to that of another user. The mesg command may, of course, be used to disable write access to a terminal

Since write is interactive, it would not normally find use in a script.


A command-line utility for configuring a network adapter (using DHCP). This command is native to Red Hat centric Linux distros.



Send or read e-mail messages.

This stripped-down command-line mail client works fine as a command embedded in a script.

Example 16-45. A script that mails itself

# Self-mailing script

adr=${1:-`whoami`}     # Default to current user, if not specified.
#  Typing ' [email protected]'
#+ sends this script to that addressee.
#  Just '' (no argument) sends the script
#+ to the person invoking it, for example, [email protected].
#  For more on the ${parameter:-default} construct,
#+ see the "Parameter Substitution" section
#+ of the "Variables Revisited" chapter.

# ============================================================================
  cat $0 | mail -s "Script \"`basename $0`\" has mailed itself to you." "$adr"
# ============================================================================

# --------------------------------------------
#  Greetings from the self-mailing script.
#  A mischievous person has run this script,
#+ which has caused it to mail itself to you.
#  Apparently, some people have nothing better
#+ to do with their time.
# --------------------------------------------

echo "At `date`, script \"`basename $0`\" mailed to "$adr"."

exit 0

#  Note that the "mailx" command (in "send" mode) may be substituted
#+ for "mail" ... but with somewhat different options.

Similar to the mail command, mailto sends e-mail messages from the command-line or in a script. However, mailto also permits sending MIME (multimedia) messages.


Show mail statistics. This command may be invoked only by root.

root# mailstats
Statistics from Tue Jan  1 20:32:08 2008
  M   msgsfr  bytes_from   msgsto    bytes_to  msgsrej msgsdis msgsqur  Mailer
  4     1682      24118K        0          0K        0       0       0  esmtp
  9      212        640K     1894      25131K        0       0       0  local
  T     1894      24758K     1894      25131K        0       0       0
  C      414                    0


This utility automatically replies to e-mails that the intended recipient is on vacation and temporarily unavailable. It runs on a network, in conjunction with sendmail, and is not applicable to a dial-up POPmail account.

16.7. Terminal Control Commands

Command affecting the console or terminal


Initialize terminal and/or fetch information about it from terminfo data. Various options permit certain terminal operations: tput clear is the equivalent of clear; tput reset is the equivalent of reset.

bash$ tput longname
xterm terminal emulator (X Window System)

Issuing a tput cup X Y moves the cursor to the (X,Y) coordinates in the current terminal. A clear to erase the terminal screen would normally precede this.

Some interesting options to tput are:

  • bold, for high-intensity text

  • smul, to underline text in the terminal

  • smso, to render text in reverse

  • sgr0, to reset the terminal parameters (to normal), without clearing the screen

Example scripts using tput:

  1. Example 36-15

  2. Example 36-13

  3. Example A-44

  4. Example A-42

  5. Example 27-2

Note that stty offers a more powerful command set for controlling a terminal.


This command prints out extensive information about the current terminal. It references the terminfo database.

bash$ infocmp
#       Reconstructed via infocmp from file:
 rxvt|rxvt terminal emulator (X Window System), 
         am, bce, eo, km, mir, msgr, xenl, xon, 
         colors#8, cols#80, it#8, lines#24, pairs#64, 
         bel=^G, blink=\E[5m, bold=\E[1m,
         clear=\E[H\E[2J, cnorm=\E[?25h, cr=^M, 


Reset terminal parameters and clear text screen. As with clear, the cursor and prompt reappear in the upper lefthand corner of the terminal.


The clear command simply clears the text screen at the console or in an xterm. The prompt and cursor reappear at the upper lefthand corner of the screen or xterm window. This command may be used either at the command line or in a script. See Example 11-26.


Echoes commands necessary to set $TERM and $TERMCAP to duplicate the size (dimensions) of the current terminal.

bash$ resize
set noglob;
 setenv COLUMNS '80';
 setenv LINES '24';
 unset noglob;


This utility records (saves to a file) all the user keystrokes at the command-line in a console or an xterm window. This, in effect, creates a record of a session.

16.8. Math Commands

"Doing the numbers"


Decompose an integer into prime factors.

bash$ factor 27417
27417: 3 13 19 37

Example 16-46. Generating prime numbers


#  Generating prime numbers the quick-and-easy way,
#+ without resorting to fancy algorithms.

CEILING=10000   # 1 to 10000

is_prime ()
  local factors
  factors=( $(factor $1) )  # Load output of `factor` into array.

if [ -z "${factors[2]}" ]
#  Third element of "factors" array:
#+ ${factors[2]} is 2nd factor of argument.
#  If it is blank, then there is no 2nd factor,
#+ and the argument is therefore prime.
  return $PRIME             # 0
  return $E_NOTPRIME        # null

for n in $(seq $CEILING)
  if is_prime $n
    printf %5d $n
  fi   #    ^  Five positions per number suffices.
done   #       For a higher $CEILING, adjust upward, as necessary.



Bash can't handle floating point calculations, and it lacks operators for certain important mathematical functions. Fortunately, bc gallops to the rescue.

Not just a versatile, arbitrary precision calculation utility, bc offers many of the facilities of a programming language. It has a syntax vaguely resembling C.

Since it is a fairly well-behaved UNIX utility, and may therefore be used in a pipe, bc comes in handy in scripts.

Here is a simple template for using bc to calculate a script variable. This uses command substitution.

	      variable=$(echo "OPTIONS; OPERATIONS" | bc)

Example 16-47. Monthly Payment on a Mortgage

# Calculates monthly payment on a mortgage.

#  This is a modification of code in the
#+ "mcalc" (mortgage calculator) package,
#+ by Jeff Schmidt
#+ and
#+ Mendel Cooper (yours truly, the ABS Guide author).

echo "Given the principal, interest rate, and term of a mortgage,"
echo "calculate the monthly payment."


echo -n "Enter principal (no commas) "
read principal
echo -n "Enter interest rate (percent) "  # If 12%, enter "12", not ".12".
read interest_r
echo -n "Enter term (months) "
read term

 interest_r=$(echo "scale=9; $interest_r/100.0" | bc) # Convert to decimal.
                 #           ^^^^^^^^^^^^^^^^^  Divide by 100. 
                 # "scale" determines how many decimal places.

 interest_rate=$(echo "scale=9; $interest_r/12 + 1.0" | bc)

 top=$(echo "scale=9; $principal*$interest_rate^$term" | bc)
          #           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
          #           Standard formula for figuring interest.

 echo; echo "Please be patient. This may take a while."

 let "months = $term - 1"
# ==================================================================== 
 for ((x=$months; x > 0; x--))
   bot=$(echo "scale=9; $interest_rate^$x" | bc)
   bottom=$(echo "scale=9; $bottom+$bot" | bc)
#  bottom = $(($bottom + $bot"))
# ==================================================================== 

# -------------------------------------------------------------------- 
#  Rick Boivie pointed out a more efficient implementation
#+ of the above loop, which decreases computation time by 2/3.

# for ((x=1; x <= $months; x++))
# do
#   bottom=$(echo "scale=9; $bottom * $interest_rate + 1" | bc)
# done

#  And then he came up with an even more efficient alternative,
#+ one that cuts down the run time by about 95%!

# bottom=`{
#     echo "scale=9; bottom=$bottom; interest_rate=$interest_rate"
#     for ((x=1; x <= $months; x++))
#     do
#          echo 'bottom = bottom * interest_rate + 1'
#     done
#     echo 'bottom'
#     } | bc`       # Embeds a 'for loop' within command substitution.
# --------------------------------------------------------------------------
#  On the other hand, Frank Wang suggests:
#  bottom=$(echo "scale=9; ($interest_rate^$term-1)/($interest_rate-1)" | bc)

#  Because . . .
#  The algorithm behind the loop
#+ is actually a sum of geometric proportion series.
#  The sum formula is e0(1-q^n)/(1-q),
#+ where e0 is the first element and q=e(n+1)/e(n)
#+ and n is the number of elements.
# --------------------------------------------------------------------------

 # let "payment = $top/$bottom"
 payment=$(echo "scale=2; $top/$bottom" | bc)
 # Use two decimal places for dollars and cents.
 echo "monthly payment = \$$payment"  # Echo a dollar sign in front of amount.

 exit 0

 # Exercises:
 #   1) Filter input to permit commas in principal amount.
 #   2) Filter input to permit interest to be entered as percent or decimal.
 #   3) If you are really ambitious,
 #+     expand this script to print complete amortization tables.

Example 16-48. Base Conversion

# Shellscript: - print number to different bases (Bourne Shell)
# Author     :	Heiner Steven ([email protected])
# Date       :	07-03-95
# Category   :	Desktop
# $Id:,v 1.2 2000/02/06 19:55:35 heiner Exp $
# ==> Above line is RCS ID info.
# Description
# Changes
# 21-03-95 stv	fixed error occuring with 0xb as input (0.2)

# ==> Used in ABS Guide with the script author's permission.
# ==> Comments added by ABS Guide author.

PN=`basename "$0"`			       # Program name
VER=`echo '$Revision: 1.2 $' | cut -d' ' -f2`  # ==> VER=1.2

Usage () {
    echo "$PN - print number to different bases, $VER (stv '95)
usage: $PN [number ...]

If no number is given, the numbers are read from standard input.
A number may be
    binary (base 2)		starting with 0b (i.e. 0b1100)
    octal (base 8)		starting with 0  (i.e. 014)
    hexadecimal (base 16)	starting with 0x (i.e. 0xc)
    decimal			otherwise (i.e. 12)" >&2
    exit $NOARGS 
}   # ==> Prints usage message.

Msg () {
    for i   # ==> in [list] missing. Why?
    do echo "$PN: $i" >&2

Fatal () { Msg "$@"; exit 66; }

PrintBases () {
    # Determine base of the number
    for i      # ==> in [list] missing...
    do         # ==> so operates on command-line arg(s).
	case "$i" in
	    0b*)		ibase=2;;	# binary
	    0x*|[a-f]*|[A-F]*)	ibase=16;;	# hexadecimal
	    0*)			ibase=8;;	# octal
	    [1-9]*)		ibase=10;;	# decimal
		Msg "illegal number $i - ignored"

	# Remove prefix, convert hex digits to uppercase (bc needs this).
	number=`echo "$i" | sed -e 's:^0[bBxX]::' | tr '[a-f]' '[A-F]'`
	# ==> Uses ":" as sed separator, rather than "/".

	# Convert number to decimal
	dec=`echo "ibase=$ibase; $number" | bc`  # ==> 'bc' is calculator utility.
	case "$dec" in
	    [0-9]*)	;;			 # number ok
	    *)		continue;;		 # error: ignore

	# Print all conversions in one line.
	# ==> 'here document' feeds command list to 'bc'.
	echo `bc <<!
	    obase=16; "hex="; $dec
	    obase=10; "dec="; $dec
	    obase=8;  "oct="; $dec
	    obase=2;  "bin="; $dec
    ` | sed -e 's: :	:g'


while [ $# -gt 0 ]
# ==>  Is a "while loop" really necessary here,
# ==>+ since all the cases either break out of the loop
# ==>+ or terminate the script.
# ==> (Above comment by Paulo Marcel Coelho Aragao.)
    case "$1" in
	--)     shift; break;;
	-h)     Usage;;                 # ==> Help message.
	-*)     Usage;;
         *)     break;;                 # First number
    esac   # ==> Error checking for illegal input might be appropriate.

if [ $# -gt 0 ]
    PrintBases "$@"
else					# Read from stdin.
    while read line
	PrintBases $line


An alternate method of invoking bc involves using a here document embedded within a command substitution block. This is especially appropriate when a script needs to pass a list of options and commands to bc.

variable=`bc << LIMIT_STRING


variable=$(bc << LIMIT_STRING

Example 16-49. Invoking bc using a here document

# Invoking 'bc' using command substitution
# in combination with a 'here document'.

var1=`bc << EOF
18.33 * 19.78
echo $var1       # 362.56

#  $( ... ) notation also works.

var2=$(bc << EOF
scale = 4
a = ( $v1 + $v2 )
b = ( $v3 * $v4 )
a * b + 15.35
echo $var2       # 593487.8452

var3=$(bc -l << EOF
scale = 9
s ( 1.7 )
# Returns the sine of 1.7 radians.
# The "-l" option calls the 'bc' math library.
echo $var3       # .991664810

# Now, try it in a function...
hypotenuse ()    # Calculate hypotenuse of a right triangle.
{                # c = sqrt( a^2 + b^2 )
hyp=$(bc -l << EOF
scale = 9
sqrt ( $1 * $1 + $2 * $2 )
# Can't directly return floating point values from a Bash function.
# But, can echo-and-capture:
echo "$hyp"

hyp=$(hypotenuse 3.68 7.31)
echo "hypotenuse = $hyp"    # 8.184039344

exit 0

Example 16-50. Calculating PI

# Approximating PI by firing cannonballs.

# Author: Mendel Cooper
# License: Public Domain
# Version 2.2, reldate 13oct08.

# This is a very simple instance of a "Monte Carlo" simulation:
#+ a mathematical model of a real-life event,
#+ using pseudorandom numbers to emulate random chance.

#  Consider a perfectly square plot of land, 10000 units on a side.
#  This land has a perfectly circular lake in its center,
#+ with a diameter of 10000 units.
#  The plot is actually mostly water, except for land in the four corners.
#  (Think of it as a square with an inscribed circle.)
#  We will fire iron cannonballs from an old-style cannon
#+ at the square.
#  All the shots impact somewhere on the square,
#+ either in the lake or on the dry corners.
#  Since the lake takes up most of the area,
#+ most of the shots will SPLASH! into the water.
#  Just a few shots will THUD! into solid ground
#+ in the four corners of the square.
#  If we take enough random, unaimed shots at the square,
#+ Then the ratio of SPLASHES to total shots will approximate
#+ the value of PI/4.
#  The simplified explanation is that the cannon is actually
#+ shooting only at the upper right-hand quadrant of the square,
#+ i.e., Quadrant I of the Cartesian coordinate plane.
#  Theoretically, the more shots taken, the better the fit.
#  However, a shell script, as opposed to a compiled language
#+ with floating-point math built in, requires some compromises.
#  This decreases the accuracy of the simulation.

DIMENSION=10000  # Length of each side of the plot.
                 # Also sets ceiling for random integers generated.

MAXSHOTS=1000    # Fire this many shots.
                 # 10000 or more would be better, but would take too long.
PMULTIPLIER=4.0  # Scaling factor.

declare -r M_PI=3.141592654
                 # Actual 9-place value of PI, for comparison purposes.

get_random ()
SEED=$(head -n 1 /dev/urandom | od -N 1 | awk '{ print $2 }')
RANDOM=$SEED                                  #  From ""
                                              #+ example script.
let "rnum = $RANDOM % $DIMENSION"             #  Range less than 10000.
echo $rnum

distance=        # Declare global variable.
hypotenuse ()    # Calculate hypotenuse of a right triangle.
{                # From "" example.
distance=$(bc -l << EOF
scale = 0
sqrt ( $1 * $1 + $2 * $2 )
#  Setting "scale" to zero rounds down result to integer value,
#+ a necessary compromise in this script.
#  It decreases the accuracy of this simulation.

# ==========================================================
# main() {
# "Main" code block, mimicking a C-language main() function.

# Initialize variables.

while [ "$shots" -lt  "$MAXSHOTS" ]           # Main loop.

  xCoord=$(get_random)                        # Get random X and Y coords.
  hypotenuse $xCoord $yCoord                  #  Hypotenuse of
                                              #+ right-triangle = distance.

  printf "#%4d   " $shots
  printf "Xc = %4d  " $xCoord
  printf "Yc = %4d  " $yCoord
  printf "Distance = %5d  " $distance         #   Distance from
                                              #+  center of lake
                                              #+  -- the "origin" --
                                              #+  coordinate (0,0).

  if [ "$distance" -le "$DIMENSION" ]
    echo -n "SPLASH!  "
    echo -n "THUD!    "

  Pi=$(echo "scale=9; $PMULTIPLIER*$splashes/$shots" | bc)
  # Multiply ratio by 4.0.
  echo -n "PI ~ $Pi"


echo "After $shots shots, PI looks like approximately   $Pi"
#  Tends to run a bit high,
#+ possibly due to round-off error and imperfect randomness of $RANDOM.
#  But still usually within plus-or-minus 5% . . .
#+ a pretty fair rough approximation.
error=$(echo "scale=9; $Pi - $M_PI" | bc)
pct_error=$(echo "scale=2; 100.0 * $error / $M_PI" | bc)
echo -n "Deviation from mathematical value of PI =        $error"
echo " ($pct_error% error)"

# End of "main" code block.
# }
# ==========================================================

exit 0

#  One might well wonder whether a shell script is appropriate for
#+ an application as complex and computation-intensive as a simulation.
#  There are at least two justifications.
#  1) As a proof of concept: to show it can be done.
#  2) To prototype and test the algorithms before rewriting
#+    it in a compiled high-level language.

See also Example A-37.


The dc (desk calculator) utility is stack-oriented and uses RPN (Reverse Polish Notation). Like bc, it has much of the power of a programming language.

Similar to the procedure with bc, echo a command-string to dc.

echo "[Printing a string ... ]P" | dc
# The P command prints the string between the preceding brackets.

# And now for some simple arithmetic.
echo "7 8 * p" | dc     # 56
#  Pushes 7, then 8 onto the stack,
#+ multiplies ("*" operator), then prints the result ("p" operator).

Most persons avoid dc, because of its non-intuitive input and rather cryptic operators. Yet, it has its uses.

Example 16-51. Converting a decimal number to hexadecimal

# Convert a decimal number to hexadecimal.

E_NOARGS=85 # Command-line arg missing.
BASE=16     # Hexadecimal.

if [ -z "$1" ]
then        # Need a command-line argument.
  echo "Usage: $0 number"
  exit $E_NOARGS
fi          # Exercise: add argument validity checking.

hexcvt ()
if [ -z "$1" ]
  echo 0
  return    # "Return" 0 if no arg passed to function.

echo ""$1" "$BASE" o p" | dc
#                  o    sets radix (numerical base) of output.
#                    p  prints the top of stack.
# For other options: 'man dc' ...

hexcvt "$1"


Studying the info page for dc is a painful path to understanding its intricacies. There seems to be a small, select group of dc wizards who delight in showing off their mastery of this powerful, but arcane utility.

bash$ echo "16i[q]sa[ln0=aln100%Pln100/snlbx]sbA0D68736142snlbxq" | dc

dc <<< 10k5v1+2/p # 1.6180339887
#  ^^^            Feed operations to dc using a Here String.
#      ^^^        Pushes 10 and sets that as the precision (10k).
#         ^^      Pushes 5 and takes its square root
#                 (5v, v = square root).
#           ^^    Pushes 1 and adds it to the running total (1+).
#             ^^  Pushes 2 and divides the running total by that (2/).
#               ^ Pops and prints the result (p)
#  The result is  1.6180339887 ...
#  ... which happens to be the Pythagorean Golden Ratio, to 10 places.

Example 16-52. Factoring

# Factor a number

MIN=2       # Will not work for number smaller than this.

if [ -z $1 ]
  echo "Usage: $0 number"
  exit $E_NOARGS

if [ "$1" -lt "$MIN" ]
  echo "Number to factor must be $MIN or greater."
  exit $E_TOOSMALL

# Exercise: Add type checking (to reject non-integer arg).

echo "Factors of $1:"
# -------------------------------------------------------
echo  "$1[p]s2[lip/dli%0=1dvsr]s12sid2%0=13sidvsr[dli%0=\
1lrli2+dsi!>.]ds.xd1<2" | dc
# -------------------------------------------------------
#  Above code written by Michel Charpentier <[email protected]>
#  (as a one-liner, here broken into two lines for display purposes).
#  Used in ABS Guide with permission (thanks!).


 # $ sh 270138
 # 2
 # 3
 # 11
 # 4093

Yet another way of doing floating point math in a script is using awk's built-in math functions in a shell wrapper.

Example 16-53. Calculating the hypotenuse of a triangle

# Returns the "hypotenuse" of a right triangle.
#                (square root of sum of squares of the "legs")

ARGS=2                # Script needs sides of triangle passed.
E_BADARGS=85          # Wrong number of arguments.

if [ $# -ne "$ARGS" ] # Test number of arguments to script.
  echo "Usage: `basename $0` side_1 side_2"
  exit $E_BADARGS

AWKSCRIPT=' { printf( "%3.7f\n", sqrt($1*$1 + $2*$2) ) } '
#             command(s) / parameters passed to awk

# Now, pipe the parameters to awk.
    echo -n "Hypotenuse of $1 and $2 = "
    echo $1 $2 | awk "$AWKSCRIPT"
#   ^^^^^^^^^^^^
# An echo-and-pipe is an easy way of passing shell parameters to awk.


# Exercise: Rewrite this script using 'bc' rather than awk.
#           Which method is more intuitive?

16.9. Miscellaneous Commands

Command that fit in no special category

jot, seq

These utilities emit a sequence of integers, with a user-selectable increment.

The default separator character between each integer is a newline, but this can be changed with the -s option.

bash$ seq 5

bash$ seq -s : 5

Both jot and seq come in handy in a for loop.

Example 16-54. Using seq to generate loop arguments

# Using "seq"


for a in `seq 80`  # or   for a in $( seq 80 )
# Same as   for a in 1 2 3 4 5 ... 80   (saves much typing!).
# May also use 'jot' (if present on system).
  echo -n "$a "
done      # 1 2 3 4 5 ... 80
# Example of using the output of a command to generate 
# the [list] in a "for" loop.

echo; echo

COUNT=80  # Yes, 'seq' also accepts a replaceable parameter.

for a in `seq $COUNT`  # or   for a in $( seq $COUNT )
  echo -n "$a "
done      # 1 2 3 4 5 ... 80

echo; echo


for a in `seq $BEGIN $END`
#  Giving "seq" two arguments starts the count at the first one,
#+ and continues until it reaches the second.
  echo -n "$a "
done      # 75 76 77 78 79 80

echo; echo


for a in `seq $BEGIN $INTERVAL $END`
#  Giving "seq" three arguments starts the count at the first one,
#+ uses the second for a step interval,
#+ and continues until it reaches the third.
  echo -n "$a "
done      # 45 50 55 60 65 70 75 80

echo; echo

exit 0

A simpler example:

#  Create a set of 10 files,
#+ named file.1, file.2 . . . file.10.

for filename in `seq $COUNT`
  touch $PREFIX.$filename
  #  Or, can do other operations,
  #+ such as rm, grep, etc.

Example 16-55. Letter Count"

# Counting letter occurrences in a text file.
# Written by Stefano Palmeri.
# Used in ABS Guide with permission.
# Slightly modified by document author.

MINARGS=2          # Script requires at least two arguments.

let LETTERS=$#-1   # How many letters specified (as command-line args).
                   # (Subtract 1 from number of command-line args.)

           echo Usage: `basename $0` file letters  
           echo Note: `basename $0` arguments are case sensitive.
           echo Example: `basename $0` foobar.txt G n U L i N U x.

# Checks number of arguments.
if [ $# -lt $MINARGS ]; then
   echo "Not enough arguments."
   exit $E_BADARGS

# Checks if file exists.
if [ ! -f $FILE ]; then
    echo "File \"$FILE\" does not exist."
    exit $E_BADARGS

# Counts letter occurrences .
for n in `seq $LETTERS`; do
      if [[ `echo -n "$1" | wc -c` -eq 1 ]]; then             #  Checks arg.
             echo "$1" -\> `cat $FILE | tr -cd  "$1" | wc -c` #  Counting.
             echo "$1 is not a  single char."

exit $?

#  This script has exactly the same functionality as,
#+ but executes faster.
#  Why?


Somewhat more capable than seq, jot is a classic UNIX utility that is not normally included in a standard Linux distro. However, the source rpm is available for download from the MIT repository.

Unlike seq, jot can generate a sequence of random numbers, using the -r option.

bash$ jot -r 3 999


The getopt command parses command-line options preceded by a dash. This external command corresponds to the getopts Bash builtin. Using getopt permits handling long options by means of the -l flag, and this also allows parameter reshuffling.

Example 16-56. Using getopt to parse command-line options

# Using getopt

# Try the following when invoking this script:
#   sh -a
#   sh -abc
#   sh -a -b -c
#   sh -d
#   sh -dXYZ
#   sh -d XYZ
#   sh -abcd
#   sh -abcdZ
#   sh -z
#   sh a
# Explain the results of each of the above.


if [ "$#" -eq 0 ]
then   # Script needs at least one command-line argument.
  echo "Usage $0 -[options a,b,c]"
  exit $E_OPTERR

set -- `getopt "abcd:" "$@"`
# Sets positional parameters to command-line arguments.
# What happens if you use "$*" instead of "$@"?

while [ ! -z "$1" ]
  case "$1" in
    -a) echo "Option \"a\"";;
    -b) echo "Option \"b\"";;
    -c) echo "Option \"c\"";;
    -d) echo "Option \"d\" $2";;
     *) break;;


#  It is usually better to use the 'getopts' builtin in a script.
#  See ""

exit 0


As Peggy Russell points out:

It is often necessary to include an eval to correctly process whitespace and quotes.
args=$(getopt -o a:bc:d -- "$@")
eval set -- "$args"

See Example 10-5 for a simplified emulation of getopt.


The run-parts command [82] executes all the scripts in a target directory, sequentially in ASCII-sorted filename order. Of course, the scripts need to have execute permission.

The cron daemon invokes run-parts to run the scripts in the /etc/cron.* directories.


In its default behavior the yes command feeds a continuous string of the character y followed by a line feed to stdout. A control-C terminates the run. A different output string may be specified, as in yes different string, which would continually output different string to stdout.

One might well ask the purpose of this. From the command-line or in a script, the output of yes can be redirected or piped into a program expecting user input. In effect, this becomes a sort of poor man's version of expect.

yes | fsck /dev/hda1 runs fsck non-interactively (careful!).

yes | rm -r dirname has same effect as rm -rf dirname (careful!).


Caution advised when piping yes to a potentially dangerous system command, such as fsck or fdisk. It might have unintended consequences.


The yes command parses variables, or more accurately, it echoes parsed variables. For example:

bash$ yes $BASH_VERSION
 . . .

This particular "feature" may be used to create a very large ASCII file on the fly:
bash$ yes $PATH > huge_file.txt
Hit Ctl-C very quickly, or you just might get more than you bargained for. . . .

The yes command may be emulated in a very simple script function.

yes ()
{ # Trivial emulation of "yes" ...
  local DEFAULT_TEXT="y"
  while [ true ]   # Endless loop.
    if [ -z "$1" ]
      echo "$DEFAULT_TEXT"
    else           # If argument ...
      echo "$1"    # ... expand and echo it.
  done             #  The only things missing are the
}                  #+ --help and --version options.


Prints arguments as a large vertical banner to stdout, using an ASCII character (default '#'). This may be redirected to a printer for hardcopy.

Note that banner has been dropped from many Linux distros, presumably because it is no longer considered useful.


Show all the environmental variables set for a particular user.

bash$ printenv | grep HOME


The lp and lpr commands send file(s) to the print queue, to be printed as hard copy. [83] These commands trace the origin of their names to the line printers of another era. [84]

bash$ lp file1.txt or bash lp <file1.txt

It is often useful to pipe the formatted output from pr to lp.

bash$ pr -options file1.txt | lp

Formatting packages, such as groff and Ghostscript may send their output directly to lp.

bash$ groff -Tascii | lp

bash$ gs -options | lp

Related commands are lpq, for viewing the print queue, and lprm, for removing jobs from the print queue.


[UNIX borrows an idea from the plumbing trade.]

This is a redirection operator, but with a difference. Like the plumber's tee, it permits "siphoning off" to a file the output of a command or commands within a pipe, but without affecting the result. This is useful for printing an ongoing process to a file or paper, perhaps to keep track of it for debugging purposes.

                            |----> to file
  command ---> command ---> |tee ---> command ---> ---> output of pipe

cat listfile* | sort | tee check.file | uniq > result.file
#                      ^^^^^^^^^^^^^^   ^^^^    

#  The file "check.file" contains the concatenated sorted "listfiles,"
#+ before the duplicate lines are removed by 'uniq.'


This obscure command creates a named pipe, a temporary first-in-first-out buffer for transferring data between processes. [85] Typically, one process writes to the FIFO, and the other reads from it. See Example A-14.

# This short script by Omair Eshkenazi.
# Used in ABS Guide with permission (thanks!).

mkfifo pipe1   # Yes, pipes can be given names.
mkfifo pipe2   # Hence the designation "named pipe."

(cut -d' ' -f1 | tr "a-z" "A-Z") >pipe2 <pipe1 &
ls -l | tr -s ' ' | cut -d' ' -f3,9- | tee pipe1 |
cut -d' ' -f2 | paste - pipe2

rm -f pipe1
rm -f pipe2

# No need to kill background processes when script terminates (why not?).

exit $?

Now, invoke the script and explain the output:

4830.tar.gz          BOZO
pipe1   BOZO
pipe2   BOZO    BOZO
Mixed.msg BOZO


This command checks the validity of a filename. If the filename exceeds the maximum allowable length (255 characters) or one or more of the directories in its path is not searchable, then an error message results.

Unfortunately, pathchk does not return a recognizable error code, and it is therefore pretty much useless in a script. Consider instead the file test operators.


Though this somewhat obscure and much feared data duplicator command originated as a utility for exchanging data on magnetic tapes between UNIX minicomputers and IBM mainframes, it still has its uses. The dd command simply copies a file (or stdin/stdout), but with conversions. Possible conversions include ASCII/EBCDIC, [86] upper/lower case, swapping of byte pairs between input and output, and skipping and/or truncating the head or tail of the input file.

# Converting a file to all uppercase:

dd if=$filename conv=ucase > $filename.uppercase
#                    lcase   # For lower case conversion

Some basic options to dd are:

  • if=INFILE

    INFILE is the source file.

  • of=OUTFILE

    OUTFILE is the target file, the file that will have the data written to it.


    This is the size of each block of data being read and written, usually a power of 2.

  • skip=BLOCKS

    How many blocks of data to skip in INFILE before starting to copy. This is useful when the INFILE has "garbage" or garbled data in its header or when it is desirable to copy only a portion of the INFILE.

  • seek=BLOCKS

    How many blocks of data to skip in OUTFILE before starting to copy, leaving blank data at beginning of OUTFILE.

  • count=BLOCKS

    Copy only this many blocks of data, rather than the entire INFILE.


    Type of conversion to be applied to INFILE data before copying operation.

A dd --help lists all the options this powerful utility takes.

Example 16-57. A script that copies itself


# This script copies itself.


dd if=$0 of=$0.$file_subscript 2>/dev/null
# Suppress messages from dd:   ^^^^^^^^^^^

exit $?

#  A program whose only output is its own source code
#+ is called a "quine" per Willard Quine.
#  Does this script qualify as a quine?

Example 16-58. Exercising dd


# Script by Stephane Chazelas.
# Somewhat modified by ABS Guide author.

infile=$0           # This script.
outfile=log.txt     # Output file left behind.

dd if=$infile of=$outfile bs=1 skip=$((n-1)) count=$((p-n+1)) 2> /dev/null
# Extracts characters n to p (8 to 11) from this script ("bash").

# ----------------------------------------------------------------

echo -n "hello vertical world" | dd cbs=1 conv=unblock 2> /dev/null
# Echoes "hello vertical world" vertically downward.
# Why? A newline follows each character dd emits.

exit $?

To demonstrate just how versatile dd is, let's use it to capture keystrokes.

Example 16-59. Capturing Keystrokes

# Capture keystrokes without needing to press ENTER.

keypresses=4                      # Number of keypresses to capture.

old_tty_setting=$(stty -g)        # Save old terminal settings.

echo "Press $keypresses keys."
stty -icanon -echo                # Disable canonical mode.
                                  # Disable local echo.
keys=$(dd bs=1 count=$keypresses 2> /dev/null)
# 'dd' uses stdin, if "if" (input file) not specified.

stty "$old_tty_setting"           # Restore old terminal settings.

echo "You pressed the \"$keys\" keys."

# Thanks, Stephane Chazelas, for showing the way.
exit 0

The dd command can do random access on a data stream.
echo -n . | dd bs=1 seek=4 of=file conv=notrunc
#  The "conv=notrunc" option means that the output file
#+ will not be truncated.

# Thanks, S.C.

The dd command can copy raw data and disk images to and from devices, such as floppies and tape drives (Example A-5). A common use is creating boot floppies.

dd if=kernel-image of=/dev/fd0H1440

Similarly, dd can copy the entire contents of a floppy, even one formatted with a "foreign" OS, to the hard drive as an image file.

dd if=/dev/fd0 of=/home/bozo/projects/floppy.img

Likewise, dd can create bootable flash drives and SD cards.

dd if=image.iso of=/dev/sdb

Example 16-60. Preparing a bootable SD card for the Raspberry Pi

# Preparing an SD card with a bootable image for the Raspberry Pi.

# $1 = imagefile name
# $2 = sdcard (device file)
# Otherwise defaults to the defaults, see below.

DEFAULTbs=4M                                 # Block size, 4 mb default.
DEFAULTif="2013-07-26-wheezy-raspbian.img"   # Commonly used distro.
DEFAULTsdcard="/dev/mmcblk0"                 # May be different. Check!
ROOTUSER_NAME=root                           # Must run as root!

username=$(id -nu)                           # Who is running this script?
if [ "$username" != "$ROOTUSER_NAME" ]
  echo "This script must run as root or with root privileges."
  exit $E_NOTROOT

if [ -n "$1" ]

if [ -n "$2" ]

if [ ! -e $imagefile ]
  echo "Image file \"$imagefile\" not found!"
  exit $E_NOIMAGE

echo "Last chance to change your mind!"; echo
read -s -n1 -p "Hit a key to write $imagefile to $sdcard [Ctl-c to exit]."
echo; echo

echo "Writing $imagefile to $sdcard ..."
dd bs=$DEFAULTbs if=$imagefile of=$sdcard

exit $?

# Exercises:
# ---------
# 1) Provide additional error checking.
# 2) Have script autodetect device file for SD card (difficult!).
# 3) Have script sutodetect image file (*img) in $PWD.

Other applications of dd include initializing temporary swap files (Example 31-2) and ramdisks (Example 31-3). It can even do a low-level copy of an entire hard drive partition, although this is not necessarily recommended.

People (with presumably nothing better to do with their time) are constantly thinking of interesting applications of dd.

Example 16-61. Securely deleting a file

# Erase "all" traces of a file.

#  This script overwrites a target file alternately
#+ with random bytes, then zeros before finally deleting it.
#  After that, even examining the raw disk sectors by conventional methods
#+ will not reveal the original file data.

PASSES=7         #  Number of file-shredding passes.
                 #  Increasing this slows script execution,
                 #+ especially on large target files.
BLOCKSIZE=1      #  I/O with /dev/urandom requires unit block size,
                 #+ otherwise you get weird results.
E_BADARGS=70     #  Various error exit codes.

if [ -z "$1" ]   # No filename specified.
  echo "Usage: `basename $0` filename"
  exit $E_BADARGS


if [ ! -e "$file" ]
  echo "File \"$file\" not found."
  exit $E_NOT_FOUND

echo; echo -n "Are you absolutely sure you want to blot out \"$file\" (y/n)? "
read answer
case "$answer" in
[nN]) echo "Changed your mind, huh?"
      exit $E_CHANGED_MIND
*)    echo "Blotting out file \"$file\".";;

flength=$(ls -l "$file" | awk '{print $5}')  # Field 5 is file length.

chmod u+w "$file"   # Allow overwriting/deleting the file.


while [ "$pass_count" -le "$PASSES" ]
  echo "Pass #$pass_count"
  sync         # Flush buffers.
  dd if=/dev/urandom of=$file bs=$BLOCKSIZE count=$flength
               # Fill with random bytes.
  sync         # Flush buffers again.
  dd if=/dev/zero of=$file bs=$BLOCKSIZE count=$flength
               # Fill with zeros.
  sync         # Flush buffers yet again.
  let "pass_count += 1"

rm -f $file    # Finally, delete scrambled and shredded file.
sync           # Flush buffers a final time.

echo "File \"$file\" blotted out and deleted."; echo

exit 0

#  This is a fairly secure, if inefficient and slow method
#+ of thoroughly "shredding" a file.
#  The "shred" command, part of the GNU "fileutils" package,
#+ does the same thing, although more efficiently.

#  The file cannot not be "undeleted" or retrieved by normal methods.
#  However . . .
#+ this simple method would *not* likely withstand
#+ sophisticated forensic analysis.

#  This script may not play well with a journaled file system.
#  Exercise (difficult): Fix it so it does.

#  Tom Vier's "wipe" file-deletion package does a much more thorough job
#+ of file shredding than this simple script.

#  For an in-depth analysis on the topic of file deletion and security,
#+ see Peter Gutmann's paper,
#+     "Secure Deletion of Data From Magnetic and Solid-State Memory".

See also the dd thread entry in the bibliography.


The od, or octal dump filter converts input (or files) to octal (base-8) or other bases. This is useful for viewing or processing binary data files or otherwise unreadable system device files, such as /dev/urandom, and as a filter for binary data.

head -c4 /dev/urandom | od -N4 -tu4 | sed -ne '1s/.* //p'
# Sample output: 1324725719, 3918166450, 2989231420, etc.

# From example script, by St�phane Chazelas

See also Example 9-16 and Example A-36.


Performs a hexadecimal, octal, decimal, or ASCII dump of a binary file. This command is the rough equivalent of od, above, but not nearly as useful. May be used to view the contents of a binary file, in combination with dd and less.

dd if=/bin/ls | hexdump -C | less
# The -C option nicely formats the output in tabular form.


Displays information about an object file or binary executable in either hexadecimal form or as a disassembled listing (with the -d option).

bash$ objdump -d /bin/ls
/bin/ls:     file format elf32-i386

 Disassembly of section .init:

 080490bc <.init>:
  80490bc:       55                      push   %ebp
  80490bd:       89 e5                   mov    %esp,%ebp
  . . .


This command generates a "magic cookie," a 128-bit (32-character) pseudorandom hexadecimal number, normally used as an authorization "signature" by the X server. This also available for use in a script as a "quick 'n dirty" random number.


Of course, a script could use md5sum for the same purpose.

# Generate md5 checksum on the script itself.
random001=`md5sum $0 | awk '{print $1}'`
# Uses 'awk' to strip off the filename.

The mcookie command gives yet another way to generate a "unique" filename.

Example 16-62. Filename generator

#  temp filename generator

BASE_STR=`mcookie`   # 32-character magic cookie.
POS=11               # Arbitrary position in magic cookie string.
LEN=5                # Get $LEN consecutive characters.

prefix=temp          #  This is, after all, a "temp" file.
                     #  For more "uniqueness," generate the
                     #+ filename prefix using the same method
                     #+ as the suffix, below.

                     #  Extract a 5-character string,
                     #+ starting at position 11.

                     # Construct the filename.

echo "Temp filename = "$temp_filename""

# sh
# Temp filename = temp.e19ea

#  Compare this method of generating "unique" filenames
#+ with the 'date' method in

exit 0

This utility converts between different units of measure. While normally invoked in interactive mode, units may find use in a script.

Example 16-63. Converting meters to miles

# Must have 'units' utility installed.

convert_units ()  # Takes as arguments the units to convert.
  cf=$(units "$1" "$2" | sed --silent -e '1p' | awk '{print $2}')
  # Strip off everything except the actual conversion factor.
  echo "$cf"

cfactor=`convert_units $Unit1 $Unit2`

result=$(echo $quantity*$cfactor | bc)

echo "There are $result $Unit2 in $quantity $Unit1."

#  What happens if you pass incompatible units,
#+ such as "acres" and "miles" to the function?

exit 0

# Exercise: Edit this script to accept command-line parameters,
#           with appropriate error checking, of course.

A hidden treasure, m4 is a powerful macro [87] processing filter, virtually a complete language. Although originally written as a pre-processor for RatFor, m4 turned out to be useful as a stand-alone utility. In fact, m4 combines some of the functionality of eval, tr, and awk, in addition to its extensive macro expansion facilities.

The April, 2002 issue of Linux Journal has a very nice article on m4 and its uses.

Example 16-64. Using m4

# Using the m4 macro processor

# Strings
echo "len($string)" | m4                            #   7
echo "substr($string,4)" | m4                       # A01
echo "regexp($string,[0-1][0-1],\&Z)" | m4      # 01Z

# Arithmetic
echo "incr($var)" | m4                              #  100
echo "eval($var / 3)" | m4                          #   33


This X-based variant of echo pops up a message/query window on the desktop.

xmessage Left click to continue -button okay


The zenity utility is adept at displaying GTK+ dialog widgets and very suitable for scripting purposes.


The doexec command enables passing an arbitrary list of arguments to a binary executable. In particular, passing argv[0] (which corresponds to $0 in a script) lets the executable be invoked by various names, and it can then carry out different sets of actions, according to the name by which it was called. What this amounts to is roundabout way of passing options to an executable.

For example, the /usr/local/bin directory might contain a binary called "aaa". Invoking doexec /usr/local/bin/aaa list would list all those files in the current working directory beginning with an "a", while invoking (the same executable with) doexec /usr/local/bin/aaa delete would delete those files.


The various behaviors of the executable must be defined within the code of the executable itself, analogous to something like the following in a shell script:
case `basename $0` in
"name1" ) do_something;;
"name2" ) do_something_else;;
"name3" ) do_yet_another_thing;;
*       ) bail_out;;


The dialog family of tools provide a method of calling interactive "dialog" boxes from a script. The more elaborate variations of dialog -- gdialog, Xdialog, and kdialog -- actually invoke X-Windows widgets.


The sox, or "sound exchange" command plays and performs transformations on sound files. In fact, the /usr/bin/play executable (now deprecated) is nothing but a shell wrapper for sox.

For example, sox soundfile.wav changes a WAV sound file into a (Sun audio format) AU sound file.

Shell scripts are ideally suited for batch-processing sox operations on sound files. For examples, see the Linux Radio Timeshift HOWTO and the MP3do Project.

Chapter 17. System and Administrative Commands

The startup and shutdown scripts in /etc/rc.d illustrate the uses (and usefulness) of many of these comands. These are usually invoked by root and used for system maintenance or emergency filesystem repairs. Use with caution, as some of these commands may damage your system if misused.

Users and Groups


Show all logged on users. This is the approximate equivalent of who -q.


Lists the current user and the groups she belongs to. This corresponds to the $GROUPS internal variable, but gives the group names, rather than the numbers.

bash$ groups
bozita cdrom cdwriter audio xgrp

bash$ echo $GROUPS
chown, chgrp

The chown command changes the ownership of a file or files. This command is a useful method that root can use to shift file ownership from one user to another. An ordinary user may not change the ownership of files, not even her own files. [88]

root# chown bozo *.txt


The chgrp command changes the group ownership of a file or files. You must be owner of the file(s) as well as a member of the destination group (or root) to use this operation.
chgrp --recursive dunderheads *.data
#  The "dunderheads" group will now own all the "*.data" files
#+ all the way down the $PWD directory tree (that's what "recursive" means).

useradd, userdel

The useradd administrative command adds a user account to the system and creates a home directory for that particular user, if so specified. The corresponding userdel command removes a user account from the system [89] and deletes associated files.


The adduser command is a synonym for useradd and is usually a symbolic link to it.


Modify a user account. Changes may be made to the password, group membership, expiration date, and other attributes of a given user's account. With this command, a user's password may be locked, which has the effect of disabling the account.


Modify a given group. The group name and/or ID number may be changed using this command.


The id command lists the real and effective user IDs and the group IDs of the user associated with the current process. This is the counterpart to the $UID, $EUID, and $GROUPS internal Bash variables.

bash$ id
uid=501(bozo) gid=501(bozo) groups=501(bozo),22(cdrom),80(cdwriter),81(audio)

bash$ echo $UID


The id command shows the effective IDs only when they differ from the real ones.

Also see Example 9-5.


The lid (list ID) command shows the group(s) that a given user belongs to, or alternately, the users belonging to a given group. May be invoked only by root.

root# lid bozo

root# lid daemon


Show all users logged on to the system.

bash$ who
bozo  tty1     Apr 27 17:45
 bozo  pts/0    Apr 27 17:46
 bozo  pts/1    Apr 27 17:47
 bozo  pts/2    Apr 27 17:49

The -m gives detailed information about only the current user. Passing any two arguments to who is the equivalent of who -m, as in who am i or who The Man.

bash$ who -m
localhost.localdomain!bozo  pts/2    Apr 27 17:49

whoami is similar to who -m, but only lists the user name.

bash$ whoami


Show all logged on users and the processes belonging to them. This is an extended version of who. The output of w may be piped to grep to find a specific user and/or process.

bash$ w | grep startx
bozo  tty1     -                 4:22pm  6:41   4.47s  0.45s  startx

Show current user's login name (as found in /var/run/utmp). This is a near-equivalent to whoami, above.

bash$ logname

bash$ whoami

However . . .

bash$ su
Password: ......

bash# whoami
bash# logname


While logname prints the name of the logged in user, whoami gives the name of the user attached to the current process. As we have just seen, sometimes these are not the same.


Runs a program or script as a substitute user. su rjones starts a shell as user rjones. A naked su defaults to root. See Example A-14.


Runs a command as root (or another user). This may be used in a script, thus permitting a regular user to run the script.


# Some commands.
sudo cp /root/secretfile /home/bozo/secret
# Some more commands.

The file /etc/sudoers holds the names of users permitted to invoke sudo.


Sets, changes, or manages a user's password.

The passwd command can be used in a script, but probably should not be.

Example 17-1. Setting a new password

# For demonstration purposes only.
#                      Not a good idea to actually run this script.
#  This script must be run as root.

ROOT_UID=0         # Root has $UID 0.
E_WRONG_USER=65    # Not root?


if [ "$UID" -ne "$ROOT_UID" ]
  echo; echo "Only root can run this script."; echo
  exit $E_WRONG_USER
  echo "You should know better than to run this script, root."
  echo "Even root users get the blues... "


# Check if bozo lives here.
grep -q "$username" /etc/passwd
if [ $? -ne $SUCCESS ]
  echo "User $username does not exist."
  echo "No password changed."

echo "$NEWPASSWORD" | passwd --stdin "$username"
#  The '--stdin' option to 'passwd' permits
#+ getting a new password from stdin (or a pipe).

echo; echo "User $username's password changed!"

# Using the 'passwd' command in a script is dangerous.

exit 0

The passwd command's -l, -u, and -d options permit locking, unlocking, and deleting a user's password. Only root may use these options.


Show users' logged in time, as read from /var/log/wtmp. This is one of the GNU accounting utilities.

bash$ ac
        total       68.08

List last logged in users, as read from /var/log/wtmp. This command can also show remote logins.

For example, to show the last few times the system rebooted:

bash$ last reboot
reboot   system boot  2.6.9-1.667      Fri Feb  4 18:18          (00:02)    
 reboot   system boot  2.6.9-1.667      Fri Feb  4 15:20          (01:27)    
 reboot   system boot  2.6.9-1.667      Fri Feb  4 12:56          (00:49)    
 reboot   system boot  2.6.9-1.667      Thu Feb  3 21:08          (02:17)    
 . . .

 wtmp begins Tue Feb  1 12:50:09 2005

Change user's group ID without logging out. This permits access to the new group's files. Since users may be members of multiple groups simultaneously, this command finds only limited use.


Kurt Glaesemann points out that the newgrp command could prove helpful in setting the default group permissions for files a user writes. However, the chgrp command might be more convenient for this purpose.



Echoes the name (filename) of the current user's terminal. Note that each separate xterm window counts as a different terminal.

bash$ tty

Shows and/or changes terminal settings. This complex command, used in a script, can control terminal behavior and the way output displays. See the info page, and study it carefully.

Example 17-2. Setting an erase character

# Using "stty" to set an erase character when reading input.

echo -n "What is your name? "
read name                      #  Try to backspace
                               #+ to erase characters of input.
                               #  Problems?
echo "Your name is $name."

stty erase '#'                 #  Set "hashmark" (#) as erase character.
echo -n "What is your name? "
read name                      #  Use # to erase last character typed.
echo "Your name is $name."

exit 0

# Even after the script exits, the new key value remains set.
# Exercise: How would you reset the erase character to the default value?

Example 17-3. secret password: Turning off terminal echoing

# secret password

echo -n "Enter password "
read passwd
echo "password is $passwd"
echo -n "If someone had been looking over your shoulder, "
echo "your password would have been compromised."

echo && echo  # Two line-feeds in an "and list."

stty -echo    # Turns off screen echo.
#   May also be done with
#   read -sp passwd
#   A big Thank You to Leigh James for pointing this out.

echo -n "Enter password again "
read passwd
echo "password is $passwd"

stty echo     # Restores screen echo.

exit 0

# Do an 'info stty' for more on this useful-but-tricky command.

A creative use of stty is detecting a user keypress (without hitting ENTER).

Example 17-4. Keypress detection

# Detect a user keypress ("hot keys").


old_tty_settings=$(stty -g)   # Save old settings (why?).
stty -icanon
Keypress=$(head -c1)          # or $(dd bs=1 count=1 2> /dev/null)
                              # on non-GNU systems

echo "Key pressed was \""$Keypress"\"."

stty "$old_tty_settings"      # Restore old settings.

# Thanks, Stephane Chazelas.

exit 0

Also see Example 9-3 and Example A-43.


Set certain terminal attributes. This command writes to its terminal's stdout a string that changes the behavior of that terminal.

bash$ setterm -cursor off

The setterm command can be used within a script to change the appearance of text written to stdout, although there are certainly better tools available for this purpose.

setterm -bold on
echo bold hello

setterm -bold off
echo normal hello


Show or initialize terminal settings. This is a less capable version of stty.

bash$ tset -r
Terminal type is xterm-xfree86.
 Kill is control-U (^U).
 Interrupt is control-C (^C).


Set or display serial port parameters. This command must be run by root and is usually found in a system setup script.

# From /etc/pcmcia/serial script:

IRQ=`setserial /dev/$DEVICE | sed -e 's/.*IRQ: //'`
setserial /dev/$DEVICE irq 0 ; setserial /dev/$DEVICE irq $IRQ

getty, agetty

The initialization process for a terminal uses getty or agetty to set it up for login by a user. These commands are not used within user shell scripts. Their scripting counterpart is stty.


Enables or disables write access to the current user's terminal. Disabling access would prevent another user on the network to write to the terminal.


It can be quite annoying to have a message about ordering pizza suddenly appear in the middle of the text file you are editing. On a multi-user network, you might therefore wish to disable write access to your terminal when you need to avoid interruptions.


This is an acronym for "write all," i.e., sending a message to all users at every terminal logged into the network. It is primarily a system administrator's tool, useful, for example, when warning everyone that the system will shortly go down due to a problem (see Example 19-1).

bash$ wall System going down for maintenance in 5 minutes!
Broadcast message from bozo (pts/1) Sun Jul  8 13:53:27 2001...

 System going down for maintenance in 5 minutes!


If write access to a particular terminal has been disabled with mesg, then wall cannot send a message to that terminal.

Information and Statistics


Output system specifications (OS, kernel version, etc.) to stdout. Invoked with the -a option, gives verbose system info (see Example 16-5). The -s option shows only the OS type.

bash$ uname

bash$ uname -s

bash$ uname -a
Linux iron.bozo 2.6.15-1.2054_FC5 #1 Tue Mar 14 15:48:33 EST 2006
 i686 i686 i386 GNU/Linux

Show system architecture. Equivalent to uname -m. See Example 11-27.

bash$ arch

bash$ uname -m

Gives information about previous commands, as stored in the /var/account/pacct file. Command name and user name can be specified by options. This is one of the GNU accounting utilities.


List the last login time of all system users. This references the /var/log/lastlog file.

bash$ lastlog
root          tty1                      Fri Dec  7 18:43:21 -0700 2001
 bin                                     **Never logged in**
 daemon                                  **Never logged in**
 bozo          tty1                      Sat Dec  8 21:14:29 -0700 2001

bash$ lastlog | grep root
root          tty1                      Fri Dec  7 18:43:21 -0700 2001


This command will fail if the user invoking it does not have read permission for the /var/log/lastlog file.


List open files. This command outputs a detailed table of all currently open files and gives information about their owner, size, the processes associated with them, and more. Of course, lsof may be piped to grep and/or awk to parse and analyze its results.

bash$ lsof
 init         1    root  mem    REG        3,5   30748    30303 /sbin/init
 init         1    root  mem    REG        3,5   73120     8069 /lib/
 init         1    root  mem    REG        3,5  931668     8075 /lib/
 cardmgr    213    root  mem    REG        3,5   36956    30357 /sbin/cardmgr

The lsof command is a useful, if complex administrative tool. If you are unable to dismount a filesystem and get an error message that it is still in use, then running lsof helps determine which files are still open on that filesystem. The -i option lists open network socket files, and this can help trace intrusion or hack attempts.

bash$ lsof -an -i tcp
 firefox 2330 bozo  32u IPv4   9956       TCP> ...
 firefox 2330 bozo  38u IPv4  10535       TCP> ...

See Example 30-2 for an effective use of lsof.


System trace: diagnostic and debugging tool for tracing system calls and signals. This command and ltrace, following, are useful for diagnosing why a given program or package fails to run . . . perhaps due to missing libraries or related causes.

bash$ strace df
execve("/bin/df", ["df"], [/* 45 vars */]) = 0
 uname({sys="Linux", node="bozo.localdomain", ...}) = 0
 brk(0)                                  = 0x804f5e4


This is the Linux equivalent of the Solaris truss command.


Library trace: diagnostic and debugging tool that traces library calls invoked by a given command.

bash$ ltrace df
__libc_start_main(0x804a910, 1, 0xbfb589a4, 0x804fb70, 0x804fb68 <unfinished ...>:
 setlocale(6, "")                                 = "en_US.UTF-8"
bindtextdomain("coreutils", "/usr/share/locale") = "/usr/share/locale"
textdomain("coreutils")                          = "coreutils"
__cxa_atexit(0x804b650, 0, 0, 0x8052bf0, 0xbfb58908) = 0
getenv("DF_BLOCK_SIZE")                          = NULL



The nc (netcat) utility is a complete toolkit for connecting to and listening to TCP and UDP ports. It is useful as a diagnostic and testing tool and as a component in simple script-based HTTP clients and servers.

bash$ nc localhost.localdomain 25
220 localhost.localdomain ESMTP Sendmail 8.13.1/8.13.1;
 Thu, 31 Mar 2005 15:41:35 -0700

A real-life usage example.

Example 17-5. Checking a remote server for identd

#! /bin/sh
## Duplicate DaveG's ident-scan thingie using netcat. Oooh, he'll be p*ssed.
## Args: target port [port port port ...]
## Hose stdout _and_ stderr together.
##  Advantages: runs slower than ident-scan, giving remote inetd less cause
##+ for alarm, and only hits the few known daemon ports you specify.
##  Disadvantages: requires numeric-only port args, the output sleazitude,
##+ and won't work for r-services when coming from high source ports.
# Script author: Hobbit <[email protected]>
# Used in ABS Guide with permission.

# ---------------------------------------------------
E_BADARGS=65       # Need at least two args.
TWO_WINKS=2        # How long to sleep.
IDPORT=113         # Authentication "tap ident" port.
# ---------------------------------------------------

case "${2}" in
  "" ) echo "Need HOST and at least one PORT." ; exit $E_BADARGS ;;

# Ping 'em once and see if they *are* running identd.
nc -z -w $TIMEOUT0 "$1" $IDPORT || \
{ echo "Oops, $1 isn't running identd." ; exit 0 ; }
#  -z scans for listening daemons.
#     -w $TIMEOUT = How long to try to connect.

# Generate a randomish base port.
RP=`expr $$ % $RAND1 + $RAND2`


while test "$1" ; do
  nc -v -w $TIMEOUT1 -p ${RP} "$TRG" ${1} < /dev/null > /dev/null &
  sleep $THREE_WINKS
  echo "${1},${RP}" | nc -w $TIMEOUT2 -r "$TRG" $IDPORT 2>&1
  sleep $TWO_WINKS

# Does this look like a lamer script or what . . . ?
# ABS Guide author comments: "Ain't really all that bad . . .
#+                            kinda clever, actually."

  kill -HUP $PROC
  RP=`expr ${RP} + 1`

exit $?

#  Notes:
#  -----

#  Try commenting out line 30 and running this script
#+ with "localhost.localdomain 25" as arguments.

#  For more of Hobbit's 'nc' example scripts,
#+ look in the documentation:
#+ the /usr/share/doc/nc-X.XX/scripts directory.

And, of course, there's Dr. Andrew Tridgell's notorious one-line script in the BitKeeper Affair:
echo clone | nc 5000 > e2fsprogs.dat


Shows memory and cache usage in tabular form. The output of this command lends itself to parsing, using grep, awk or Perl. The procinfo command shows all the information that free does, and much more.

bash$ free
                total       used       free     shared    buffers     cached
   Mem:         30504      28624       1880      15820       1608       16376
   -/+ buffers/cache:      10640      19864
   Swap:        68540       3128      65412

To show unused RAM memory:

bash$ free | grep Mem | awk '{ print $4 }'

Extract and list information and statistics from the /proc pseudo-filesystem. This gives a very extensive and detailed listing.

bash$ procinfo | grep Bootup
Bootup: Wed Mar 21 15:15:50 2001    Load average: 0.04 0.21 0.34 3/47 6829

List devices, that is, show installed hardware.

bash$ lsdev
Device            DMA   IRQ  I/O Ports
 cascade             4     2 
 dma                          0080-008f
 dma1                         0000-001f
 dma2                         00c0-00df
 fpu                          00f0-00ff
 ide0                     14  01f0-01f7 03f6-03f6


Show (disk) file usage, recursively. Defaults to current working directory, unless otherwise specified.

bash$ du -ach
1.0k    ./
 1.0k    ./
 1.0k    ./random.file
 6.0k    .
 6.0k    total

Shows filesystem usage in tabular form.

bash$ df
Filesystem           1k-blocks      Used Available Use% Mounted on
 /dev/hda5               273262     92607    166547  36% /
 /dev/hda8               222525    123951     87085  59% /home
 /dev/hda7              1408796   1075744    261488  80% /usr

Lists all system bootup messages to stdout. Handy for debugging and ascertaining which device drivers were installed and which system interrupts in use. The output of dmesg may, of course, be parsed with grep, sed, or awk from within a script.

bash$ dmesg | grep hda
Kernel command line: ro root=/dev/hda2
 hda: IBM-DLGA-23080, ATA DISK drive
 hda: 6015744 sectors (3080 MB) w/96KiB Cache, CHS=746/128/63
 hda: hda1 hda2 hda3 < hda5 hda6 hda7 > hda4


Gives detailed and verbose statistics on a given file (even a directory or device file) or set of files.

bash$ stat test.cru
  File: "test.cru"
   Size: 49970        Allocated Blocks: 100          Filetype: Regular File
   Mode: (0664/-rw-rw-r--)         Uid: (  501/ bozo)  Gid: (  501/ bozo)
 Device:  3,8   Inode: 18185     Links: 1    
 Access: Sat Jun  2 16:40:24 2001
 Modify: Sat Jun  2 16:40:24 2001
 Change: Sat Jun  2 16:40:24 2001

If the target file does not exist, stat returns an error message.

bash$ stat nonexistent-file
nonexistent-file: No such file or directory

In a script, you can use stat to extract information about files (and filesystems) and set variables accordingly.


# Per suggestion of Jo�l Bourquard and . . .

file_name=$(stat -c%n "$FILENAME")   # Same as "$FILENAME" of course.
file_owner=$(stat -c%U "$FILENAME")
file_size=$(stat -c%s "$FILENAME")
#  Certainly easier than using "ls -l $FILENAME"
#+ and then parsing with sed.
file_inode=$(stat -c%i "$FILENAME")
file_type=$(stat -c%F "$FILENAME")
file_access_rights=$(stat -c%A "$FILENAME")

echo "File name:          $file_name"
echo "File owner:         $file_owner"
echo "File size:          $file_size"
echo "File inode:         $file_inode"
echo "File type:          $file_type"
echo "File access rights: $file_access_rights"

exit 0


File name:          testfile.txt
File owner:         bozo
File size:          418
File inode:         1730378
File type:          regular file
File access rights: -rw-rw-r--


Display virtual memory statistics.

bash$ vmstat
   procs                      memory    swap          io system         cpu
 r  b  w   swpd   free   buff  cache  si  so    bi    bo   in    cs  us  sy id
 0  0  0      0  11040   2636  38952   0   0    33     7  271    88   8   3 89


Shows how long the system has been running, along with associated statistics.

bash$ uptime
10:28pm  up  1:57,  3 users,  load average: 0.17, 0.34, 0.27


A load average of 1 or less indicates that the system handles processes immediately. A load average greater than 1 means that processes are being queued. When the load average gets above 3 (on a single-core processor), then system performance is significantly degraded.


Lists the system's host name. This command sets the host name in an /etc/rc.d setup script (/etc/rc.d/rc.sysinit or similar). It is equivalent to uname -n, and a counterpart to the $HOSTNAME internal variable.

bash$ hostname

bash$ echo $HOSTNAME

Similar to the hostname command are the domainname, dnsdomainname, nisdomainname, and ypdomainname commands. Use these to display or set the system DNS or NIS/YP domain name. Various options to hostname also perform these functions.


Echo a 32-bit hexadecimal numerical identifier for the host machine.

bash$ hostid


This command allegedly fetches a "unique" serial number for a particular system. Certain product registration procedures use this number to brand a particular user license. Unfortunately, hostid only returns the machine network address in hexadecimal, with pairs of bytes transposed.

The network address of a typical non-networked Linux machine, is found in /etc/hosts.

bash$ cat /etc/hosts               localhost.localdomain localhost

As it happens, transposing the bytes of, we get, which translates in hex to 007f0100, the exact equivalent of what hostid returns, above. There exist only a few million other Linux machines with this identical hostid.


Invoking sar (System Activity Reporter) gives a very detailed rundown on system statistics. The Santa Cruz Operation ("Old" SCO) released sar as Open Source in June, 1999.

This command is not part of the base Linux distribution, but may be obtained as part of the sysstat utilities package, written by Sebastien Godard.

bash$ sar
Linux 2.4.9 ( 	09/26/03

10:30:00          CPU     %user     %nice   %system   %iowait     %idle
10:40:00          all      2.21     10.90     65.48      0.00     21.41
10:50:00          all      3.36      0.00     72.36      0.00     24.28
11:00:00          all      1.12      0.00     80.77      0.00     18.11
Average:          all      2.23      3.63     72.87      0.00     21.27

14:32:30          LINUX RESTART

15:00:00          CPU     %user     %nice   %system   %iowait     %idle
15:10:00          all      8.59      2.40     17.47      0.00     71.54
15:20:00          all      4.07      1.00     11.95      0.00     82.98
15:30:00          all      0.79      2.94      7.56      0.00     88.71
Average:          all      6.33      1.70     14.71      0.00     77.26

Show information and statistics about a designated elf binary. This is part of the binutils package.

bash$ readelf -h /bin/bash
ELF Header:
   Magic:   7f 45 4c 46 01 01 01 00 00 00 00 00 00 00 00 00 
   Class:                             ELF32
   Data:                              2's complement, little endian
   Version:                           1 (current)
   OS/ABI:                            UNIX - System V
   ABI Version:                       0
   Type:                              EXEC (Executable file)
   . . .

The size [/path/to/binary] command gives the segment sizes of a binary executable or archive file. This is mainly of use to programmers.

bash$ size /bin/bash
   text    data     bss     dec     hex filename
  495971   22496   17392  535859   82d33 /bin/bash

System Logs


Appends a user-generated message to the system log (/var/log/messages). You do not have to be root to invoke logger.

logger Experiencing instability in network connection at 23:10, 05/21.
# Now, do a 'tail /var/log/messages'.

By embedding a logger command in a script, it is possible to write debugging information to /var/log/messages.

logger -t $0 -i Logging at line "$LINENO".
# The "-t" option specifies the tag for the logger entry.
# The "-i" option records the process ID.

# tail /var/log/message
# ...
# Jul  7 20:48:58 localhost ./[1712]: Logging at line 3.


This utility manages the system log files, rotating, compressing, deleting, and/or e-mailing them, as appropriate. This keeps the /var/log from getting cluttered with old log files. Usually cron runs logrotate on a daily basis.

Adding an appropriate entry to /etc/logrotate.conf makes it possible to manage personal log files, as well as system-wide ones.


Stefano Falsetto has created rottlog, which he considers to be an improved version of logrotate.

Job Control


Process Statistics: lists currently executing processes by owner and PID (process ID). This is usually invoked with ax or aux options, and may be piped to grep or sed to search for a specific process (see Example 15-14 and Example 29-3).

bash$  ps ax | grep sendmail
295 ?	   S	  0:00 sendmail: accepting connections on port 25

To display system processes in graphical "tree" format: ps afjx or ps ax --forest.

pgrep, pkill

Combining the ps command with grep or kill.

bash$ ps a | grep mingetty
2212 tty2     Ss+    0:00 /sbin/mingetty tty2
 2213 tty3     Ss+    0:00 /sbin/mingetty tty3
 2214 tty4     Ss+    0:00 /sbin/mingetty tty4
 2215 tty5     Ss+    0:00 /sbin/mingetty tty5
 2216 tty6     Ss+    0:00 /sbin/mingetty tty6
 4849 pts/2    S+     0:00 grep mingetty

bash$ pgrep mingetty
2212 mingetty
 2213 mingetty
 2214 mingetty
 2215 mingetty
 2216 mingetty

Compare the action of pkill with killall.


Lists currently executing processes in "tree" format. The -p option shows the PIDs, as well as the process names.


Continuously updated display of most cpu-intensive processes. The -b option displays in text mode, so that the output may be parsed or accessed from a script.

bash$ top -b
  8:30pm  up 3 min,  3 users,  load average: 0.49, 0.32, 0.13
 45 processes: 44 sleeping, 1 running, 0 zombie, 0 stopped
 CPU states: 13.6% user,  7.3% system,  0.0% nice, 78.9% idle
 Mem:    78396K av,   65468K used,   12928K free,       0K shrd,    2352K buff
 Swap:  157208K av,       0K used,  157208K free                   37244K cached

   848 bozo      17   0   996  996   800 R     5.6  1.2   0:00 top
     1 root       8   0   512  512   444 S     0.0  0.6   0:04 init
     2 root       9   0     0    0     0 SW    0.0  0.0   0:00 keventd


Run a background job with an altered priority. Priorities run from 19 (lowest) to -20 (highest). Only root may set the negative (higher) priorities. Related commands are renice and snice, which change the priority of a running process or processes, and skill, which sends a kill signal to a process or processes.


Keeps a command running even after user logs off. The command will run as a foreground process unless followed by &. If you use nohup within a script, consider coupling it with a wait to avoid creating an orphan or zombie process.


Identifies process ID (PID) of a running job. Since job control commands, such as kill and renice act on the PID of a process (not its name), it is sometimes necessary to identify that PID. The pidof command is the approximate counterpart to the $PPID internal variable.

bash$ pidof xclock

Example 17-6. pidof helps kill a process



process=xxxyyyzzz  # Use nonexistent process.
# For demo purposes only...
# ... don't want to actually kill any actual process with this script.
# If, for example, you wanted to use this script to logoff the Internet,
#     process=pppd

t=`pidof $process`       # Find pid (process id) of $process.
# The pid is needed by 'kill' (can't 'kill' by program name).

if [ -z "$t" ]           # If process not present, 'pidof' returns null.
  echo "Process $process was not running."
  echo "Nothing killed."

kill $t                  # May need 'kill -9' for stubborn process.

# Need a check here to see if process allowed itself to be killed.
# Perhaps another " t=`pidof $process` " or ...

# This entire script could be replaced by
#        kill $(pidof -x process_name)
# or
#        killall process_name
# but it would not be as instructive.

exit 0

Identifies the processes (by PID) that are accessing a given file, set of files, or directory. May also be invoked with the -k option, which kills those processes. This has interesting implications for system security, especially in scripts preventing unauthorized users from accessing system services.

bash$ fuser -u /usr/bin/vim
/usr/bin/vim:         3207e(bozo)

bash$ fuser -u /dev/null
/dev/null:            3009(bozo)  3010(bozo)  3197(bozo)  3199(bozo)

One important application for fuser is when physically inserting or removing storage media, such as CD ROM disks or USB flash drives. Sometimes trying a umount fails with a device is busy error message. This means that some user(s) and/or process(es) are accessing the device. An fuser -um /dev/device_name will clear up the mystery, so you can kill any relevant processes.

bash$ umount /mnt/usbdrive
umount: /mnt/usbdrive: device is busy

bash$ fuser -um /dev/usbdrive
/mnt/usbdrive:        1772c(bozo)

bash$ kill -9 1772
bash$ umount /mnt/usbdrive

The fuser command, invoked with the -n option identifies the processes accessing a port. This is especially useful in combination with nmap.

root# nmap localhost.localdomain
 25/tcp   open  smtp

root# fuser -un tcp 25
25/tcp:               2095(root)

root# ps ax | grep 2095 | grep -v grep
2095 ?        Ss     0:00 sendmail: accepting connections


Administrative program scheduler, performing such duties as cleaning up and deleting system log files and updating the slocate database. This is the superuser version of at (although each user may have their own crontab file which can be changed with the crontab command). It runs as a daemon and executes scheduled entries from /etc/crontab.


Some flavors of Linux run crond, Matthew Dillon's version of cron.

Process Control and Booting


The init command is the parent of all processes. Called in the final step of a bootup, init determines the runlevel of the system from /etc/inittab. Invoked by its alias telinit, and by root only.


Symlinked to init, this is a means of changing the system runlevel, usually done for system maintenance or emergency filesystem repairs. Invoked only by root. This command can be dangerous -- be certain you understand it well before using!


Shows the current and last runlevel, that is, whether the system is halted (runlevel 0), in single-user mode (1), in multi-user mode (2 or 3), in X Windows (5), or rebooting (6). This command accesses the /var/run/utmp file.

halt, shutdown, reboot

Command set to shut the system down, usually just prior to a power down.


On some Linux distros, the halt command has 755 permissions, so it can be invoked by a non-root user. A careless halt in a terminal or a script may shut down the system!


Starts or stops a system service. The startup scripts in /etc/init.d and /etc/rc.d use this command to start services at bootup.

root# /sbin/service iptables stop
Flushing firewall rules:                                   [  OK  ]
 Setting chains to policy ACCEPT: filter                    [  OK  ]
 Unloading iptables modules:                                [  OK  ]



Network mapper and port scanner. This command scans a server to locate open ports and the services associated with those ports. It can also report information about packet filters and firewalls. This is an important security tool for locking down a network against hacking attempts.


SERVER=$HOST                           # localhost.localdomain (
PORT_NUMBER=25                         # SMTP port.

nmap $SERVER | grep -w "$PORT_NUMBER"  # Is that particular port open?
#              grep -w matches whole words only,
#+             so this wouldn't match port 1025, for example.

exit 0

# 25/tcp     open        smtp


Network interface configuration and tuning utility.

bash$ ifconfig -a
lo        Link encap:Local Loopback
           inet addr:  Mask:
           UP LOOPBACK RUNNING  MTU:16436  Metric:1
           RX packets:10 errors:0 dropped:0 overruns:0 frame:0
           TX packets:10 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:0 
           RX bytes:700 (700.0 b)  TX bytes:700 (700.0 b)

The ifconfig command is most often used at bootup to set up the interfaces, or to shut them down when rebooting.

# Code snippets from /etc/rc.d/init.d/network

# ...

# Check that networking is up.
[ ${NETWORKING} = "no" ] && exit 0

[ -x /sbin/ifconfig ] || exit 0

# ...

for i in $interfaces ; do
  if ifconfig $i 2>/dev/null | grep -q "UP" >/dev/null 2>&1 ; then
    action "Shutting down interface $i: " ./ifdown $i boot
#  The GNU-specific "-q" option to "grep" means "quiet", i.e.,
#+ producing no output.
#  Redirecting output to /dev/null is therefore not strictly necessary.
# ...

echo "Currently active devices:"
echo `/sbin/ifconfig | grep ^[a-z] | awk '{print $1}'`
#                            ^^^^^  should be quoted to prevent globbing.
#  The following also work.
#    echo $(/sbin/ifconfig | awk '/^[a-z]/ { print $1 })'
#    echo $(/sbin/ifconfig | sed -e 's/ .*//')
#  Thanks, S.C., for additional comments.

See also Example 32-6.


Show current network statistics and information, such as routing tables and active connections. This utility accesses information in /proc/net (Chapter 29). See Example 29-4.

netstat -r is equivalent to route.

bash$ netstat
Active Internet connections (w/o servers)
 Proto Recv-Q Send-Q Local Address           Foreign Address         State      
 Active UNIX domain sockets (w/o servers)
 Proto RefCnt Flags       Type       State         I-Node Path
 unix  11     [ ]         DGRAM                    906    /dev/log
 unix  3      [ ]         STREAM     CONNECTED     4514   /tmp/.X11-unix/X0
 unix  3      [ ]         STREAM     CONNECTED     4513
 . . .


A netstat -lptu shows sockets that are listening to ports, and the associated processes. This can be useful for determining whether a computer has been hacked or compromised.


This is the command set for configuring a wireless network. It is the wireless equivalent of ifconfig, above.


General purpose utility for setting up, changing, and analyzing IP (Internet Protocol) networks and attached devices. This command is part of the iproute2 package.

bash$ ip link show
1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue 
     link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
 2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast qlen 1000
     link/ether 00:d0:59:ce:af:da brd ff:ff:ff:ff:ff:ff
 3: sit0: <NOARP> mtu 1480 qdisc noop 
     link/sit brd

bash$ ip route list dev lo  scope link

Or, in a script:

# Script by Juan Nicolas Ruiz
# Used with his kind permission.

# Setting up (and stopping) a GRE tunnel.

# --- ---


/sbin/ip tunnel add netb mode gre remote $REMOTE_IP \
  local $LOCAL_IP ttl 255
/sbin/ip addr add $OTHER_IFACE dev netb
/sbin/ip link set netb up
/sbin/ip route add $REMOTE_NET dev netb

exit 0  #############################################

# --- ---


/sbin/ip route del $REMOTE_NET dev netb
/sbin/ip link set netb down
/sbin/ip tunnel del netb

exit 0


Show info about or make changes to the kernel routing table.

bash$ route
Destination     Gateway         Genmask         Flags   MSS Window  irtt Iface
 pm3-67.bozosisp *      UH       40 0          0 ppp0       *            U        40 0          0 lo
 default         pm3-67.bozosisp         UG       40 0          0 ppp0


The iptables command set is a packet filtering tool used mainly for such security purposes as setting up network firewalls. This is a complex tool, and a detailed explanation of its use is beyond the scope of this document. Oskar Andreasson's tutorial is a reasonable starting point.

See also shutting down iptables and Example 30-2.


Check network and system configuration. This command lists and manages the network and system services started at bootup in the /etc/rc?.d directory.

Originally a port from IRIX to Red Hat Linux, chkconfig may not be part of the core installation of some Linux flavors.

bash$ chkconfig --list
atd             0:off   1:off   2:off   3:on    4:on    5:on    6:off
 rwhod           0:off   1:off   2:off   3:off   4:off   5:off   6:off


Network packet "sniffer." This is a tool for analyzing and troubleshooting traffic on a network by dumping packet headers that match specified criteria.

Dump ip packet traffic between hosts bozoville and caduceus:

bash$ tcpdump ip host bozoville and caduceus

Of course, the output of tcpdump can be parsed with certain of the previously discussed text processing utilities.



Mount a filesystem, usually on an external device, such as a floppy or CDROM. The file /etc/fstab provides a handy listing of available filesystems, partitions, and devices, including options, that may be automatically or manually mounted. The file /etc/mtab shows the currently mounted filesystems and partitions (including the virtual ones, such as /proc).

mount -a mounts all filesystems and partitions listed in /etc/fstab, except those with a noauto option. At bootup, a startup script in /etc/rc.d (rc.sysinit or something similar) invokes this to get everything mounted.

mount -t iso9660 /dev/cdrom /mnt/cdrom
# Mounts CD ROM. ISO 9660 is a standard CD ROM filesystem.
mount /mnt/cdrom
# Shortcut, if /mnt/cdrom listed in /etc/fstab

The versatile mount command can even mount an ordinary file on a block device, and the file will act as if it were a filesystem. Mount accomplishes that by associating the file with a loopback device. One application of this is to mount and examine an ISO9660 filesystem image before burning it onto a CDR. [90]

Example 17-7. Checking a CD image

# As root...

mkdir /mnt/cdtest  # Prepare a mount point, if not already there.

mount -r -t iso9660 -o loop cd-image.iso /mnt/cdtest   # Mount the image.
#                  "-o loop" option equivalent to "losetup /dev/loop0"
cd /mnt/cdtest     # Now, check the image.
ls -alR            # List the files in the directory tree there.
                   # And so forth.

Unmount a currently mounted filesystem. Before physically removing a previously mounted floppy or CDROM disk, the device must be umounted, else filesystem corruption may result.
umount /mnt/cdrom
# You may now press the eject button and safely remove the disk.


The automount utility, if properly installed, can mount and unmount floppies or CDROM disks as they are accessed or removed. On "multispindle" laptops with swappable floppy and optical drives, this can cause problems, however.


The newer Linux distros have deprecated mount and umount. The successor, for command-line mounting of removable storage devices, is gnome-mount. It can take the -d option to mount a device file by its listing in /dev.

For example, to mount a USB flash drive:

bash$ gnome-mount -d /dev/sda1
gnome-mount 0.4

bash$ df
. . .
 /dev/sda1                63584     12034     51550  19% /media/disk


Forces an immediate write of all updated data from buffers to hard drive (synchronize drive with buffers). While not strictly necessary, a sync assures the sys admin or user that the data just changed will survive a sudden power failure. In the olden days, a sync; sync (twice, just to make absolutely sure) was a useful precautionary measure before a system reboot.

At times, you may wish to force an immediate buffer flush, as when securely deleting a file (see Example 16-61) or when the lights begin to flicker.


Sets up and configures loopback devices.

Example 17-8. Creating a filesystem in a file

SIZE=1000000  # 1 meg

head -c $SIZE < /dev/zero > file  # Set up file of designated size.
losetup /dev/loop0 file           # Set it up as loopback device.
mke2fs /dev/loop0                 # Create filesystem.
mount -o loop /dev/loop0 /mnt     # Mount it.

# Thanks, S.C.

Creates a swap partition or file. The swap area must subsequently be enabled with swapon.

swapon, swapoff

Enable / disable swap partitition or file. These commands usually take effect at bootup and shutdown.


Create a Linux ext2 filesystem. This command must be invoked as root.

Example 17-9. Adding a new hard drive


# Adding a second hard drive to system.
# Software configuration. Assumes hardware already mounted.
# From an article by the author of the ABS Guide.
# In issue #38 of _Linux Gazette_,

ROOT_UID=0     # This script must be run as root.
E_NOTROOT=67   # Non-root exit error.

if [ "$UID" -ne "$ROOT_UID" ]
  echo "Must be root to run this script."
  exit $E_NOTROOT

# Use with extreme caution!
# If something goes wrong, you may wipe out your current filesystem.

NEWDISK=/dev/hdb         # Assumes /dev/hdb vacant. Check!
MOUNTPOINT=/mnt/newdisk  # Or choose another mount point.

fdisk $NEWDISK
mke2fs -cv $NEWDISK1   # Check for bad blocks (verbose output).
#  Note:           ^     /dev/hdb1, *not* /dev/hdb!
chmod 777 $MOUNTPOINT  # Makes new drive accessible to all users.

# Now, test ...
# mount -t ext2 /dev/hdb1 /mnt/newdisk
# Try creating a directory.
# If it works, umount it, and proceed.

# Final step:
# Add the following line to /etc/fstab.
# /dev/hdb1  /mnt/newdisk  ext2  defaults  1 1


See also Example 17-8 and Example 31-3.


Create a DOS FAT filesystem.


Tune ext2 filesystem. May be used to change filesystem parameters, such as maximum mount count. This must be invoked as root.


This is an extremely dangerous command. Use it at your own risk, as you may inadvertently destroy your filesystem.


Dump (list to stdout) very verbose filesystem info. This must be invoked as root.

root# dumpe2fs /dev/hda7 | grep 'ount count'
dumpe2fs 1.19, 13-Jul-2000 for EXT2 FS 0.5b, 95/08/09
 Mount count:              6
 Maximum mount count:      20

List or change hard disk parameters. This command must be invoked as root, and it may be dangerous if misused.


Create or change a partition table on a storage device, usually a hard drive. This command must be invoked as root.


Use this command with extreme caution. If something goes wrong, you may destroy an existing filesystem.

fsck, e2fsck, debugfs

Filesystem check, repair, and debug command set.

fsck: a front end for checking a UNIX filesystem (may invoke other utilities). The actual filesystem type generally defaults to ext2.

e2fsck: ext2 filesystem checker.

debugfs: ext2 filesystem debugger. One of the uses of this versatile, but dangerous command is to (attempt to) recover deleted files. For advanced users only!


All of these should be invoked as root, and they can damage or destroy a filesystem if misused.


Checks for bad blocks (physical media flaws) on a storage device. This command finds use when formatting a newly installed hard drive or testing the integrity of backup media. [91] As an example, badblocks /dev/fd0 tests a floppy disk.

The badblocks command may be invoked destructively (overwrite all data) or in non-destructive read-only mode. If root user owns the device to be tested, as is generally the case, then root must invoke this command.

lsusb, usbmodules

The lsusb command lists all USB (Universal Serial Bus) buses and the devices hooked up to them.

The usbmodules command outputs information about the driver modules for connected USB devices.

bash$ lsusb
Bus 001 Device 001: ID 0000:0000  
 Device Descriptor:
   bLength                18
   bDescriptorType         1
   bcdUSB               1.00
   bDeviceClass            9 Hub
   bDeviceSubClass         0 
   bDeviceProtocol         0 
   bMaxPacketSize0         8
   idVendor           0x0000 
   idProduct          0x0000

   . . .


Lists pci busses present.

bash$ lspci
00:00.0 Host bridge: Intel Corporation 82845 845
 (Brookdale) Chipset Host Bridge (rev 04)
 00:01.0 PCI bridge: Intel Corporation 82845 845
 (Brookdale) Chipset AGP Bridge (rev 04)
 00:1d.0 USB Controller: Intel Corporation 82801CA/CAM USB (Hub #1) (rev 02)
 00:1d.1 USB Controller: Intel Corporation 82801CA/CAM USB (Hub #2) (rev 02)
 00:1d.2 USB Controller: Intel Corporation 82801CA/CAM USB (Hub #3) (rev 02)
 00:1e.0 PCI bridge: Intel Corporation 82801 Mobile PCI Bridge (rev 42)

   . . .


Creates a boot floppy which can be used to bring up the system if, for example, the MBR (master boot record) becomes corrupted. Of special interest is the --iso option, which uses mkisofs to create a bootable ISO9660 filesystem image suitable for burning a bootable CDR.

The mkbootdisk command is actually a Bash script, written by Erik Troan, in the /sbin directory.


Creates an ISO9660 filesystem suitable for a CDR image.


CHange ROOT directory. Normally commands are fetched from $PATH, relative to /, the default root directory. This changes the root directory to a different one (and also changes the working directory to there). This is useful for security purposes, for instance when the system administrator wishes to restrict certain users, such as those telnetting in, to a secured portion of the filesystem (this is sometimes referred to as confining a guest user to a "chroot jail"). Note that after a chroot, the execution path for system binaries is no longer valid.

A chroot /opt would cause references to /usr/bin to be translated to /opt/usr/bin. Likewise, chroot /aaa/bbb /bin/ls would redirect future instances of ls to /aaa/bbb as the base directory, rather than / as is normally the case. An alias XX 'chroot /aaa/bbb ls' in a user's ~/.bashrc effectively restricts which portion of the filesystem she may run command "XX" on.

The chroot command is also handy when running from an emergency boot floppy (chroot to /dev/fd0), or as an option to lilo when recovering from a system crash. Other uses include installation from a different filesystem (an rpm option) or running a readonly filesystem from a CD ROM. Invoke only as root, and use with care.


It might be necessary to copy certain system files to a chrooted directory, since the normal $PATH can no longer be relied upon.


This utility is part of the procmail package ( It creates a lock file, a semaphore that controls access to a file, device, or resource.

The lock file serves as a flag that this particular file, device, or resource is in use by a process (and is therefore "busy"). The presence of a lock file permits only restricted access (or no access) to other processes.

lockfile /home/bozo/lockfiles/$0.lock
# Creates a write-protected lockfile prefixed with the name of the script.

lockfile /home/bozo/lockfiles/${0##*/}.lock
# A safer version of the above, as pointed out by E. Choroba.

Lock files are used in such applications as protecting system mail folders from simultaneously being changed by multiple users, indicating that a modem port is being accessed, and showing that an instance of Firefox is using its cache. Scripts may check for the existence of a lock file created by a certain process to check if that process is running. Note that if a script attempts to create a lock file that already exists, the script will likely hang.

Normally, applications create and check for lock files in the /var/lock directory. [92] A script can test for the presence of a lock file by something like the following.
# Application "xyzip" created lock file "/var/lock/xyzip.lock".

if [ -e "/var/lock/$appname.lock" ]
then   #+ Prevent other programs & scripts
       #  from accessing files/resources used by xyzip.


Much less useful than the lockfile command is flock. It sets an "advisory" lock on a file and then executes a command while the lock is on. This is to prevent any other process from setting a lock on that file until completion of the specified command.

flock $0 cat $0 > lockfile__$0
#  Set a lock on the script the above line appears in,
#+ while listing the script to stdout.


Unlike lockfile, flock does not automatically create a lock file.


Creates block or character device files (may be necessary when installing new hardware on the system). The MAKEDEV utility has virtually all of the functionality of mknod, and is easier to use.


Utility for creating device files. It must be run as root, and in the /dev directory. It is a sort of advanced version of mknod.


Automatically deletes files which have not been accessed within a specified period of time. Usually invoked by cron to remove stale log files.


dump, restore

The dump command is an elaborate filesystem backup utility, generally used on larger installations and networks. [93] It reads raw disk partitions and writes a backup file in a binary format. Files to be backed up may be saved to a variety of storage media, including disks and tape drives. The restore command restores backups made with dump.


Perform a low-level format on a floppy disk (/dev/fd0*).

System Resources


Sets an upper limit on use of system resources. Usually invoked with the -f option, which sets a limit on file size (ulimit -f 1000 limits files to 1 meg maximum). [94] The -t option limits the coredump size (ulimit -c 0 eliminates coredumps). Normally, the value of ulimit would be set in /etc/profile and/or ~/.bash_profile (see Appendix H).


Judicious use of ulimit can protect a system against the dreaded fork bomb.

# This script is for illustrative purposes only.
# Run it at your own peril -- it WILL freeze your system.

while true  #  Endless loop.
  $0 &      #  This script invokes itself . . .
            #+ forks an infinite number of times . . .
            #+ until the system freezes up because all resources exhausted.
done        #  This is the notorious "sorcerer's appentice" scenario.

exit 0      #  Will not exit here, because this script will never terminate.

A ulimit -Hu XX (where XX is the user process limit) in /etc/profile would abort this script when it exceeded the preset limit.


Display user or group disk quotas.


Set user or group disk quotas from the command-line.


User file creation permissions mask. Limit the default file attributes for a particular user. All files created by that user take on the attributes specified by umask. The (octal) value passed to umask defines the file permissions disabled. For example, umask 022 ensures that new files will have at most 755 permissions (777 NAND 022). [95] Of course, the user may later change the attributes of particular files with chmod. The usual practice is to set the value of umask in /etc/profile and/or ~/.bash_profile (see Appendix H).

Example 17-10. Using umask to hide an output file from prying eyes

# Same as "" script, but writes output to "secure" file.

# Usage: ./ filename
# or     ./ <filename
# or     ./ and supply keyboard input (stdin)

umask 177               #  File creation mask.
                        #  Files created by this script
                        #+ will have 600 permissions.

OUTFILE=decrypted.txt   #  Results output to file "decrypted.txt"
                        #+ which can only be read/written
                        #  by invoker of script (or root).

cat "$@" | tr 'a-zA-Z' 'n-za-mN-ZA-M' > $OUTFILE 
#    ^^ Input from stdin or a file.   ^^^^^^^^^^ Output redirected to file. 

exit 0

Get info about or make changes to root device, swap space, or video mode. The functionality of rdev has generally been taken over by lilo, but rdev remains useful for setting up a ram disk. This is a dangerous command, if misused.



List installed kernel modules.

bash$ lsmod
Module                  Size  Used by
 autofs                  9456   2 (autoclean)
 opl3                   11376   0
 serial_cs               5456   0 (unused)
 sb                     34752   0
 uart401                 6384   0 [sb]
 sound                  58368   0 [opl3 sb uart401]
 soundlow                 464   0 [sound]
 soundcore               2800   6 [sb sound]
 ds                      6448   2 [serial_cs]
 i82365                 22928   2
 pcmcia_core            45984   0 [serial_cs ds i82365]


Doing a cat /proc/modules gives the same information.


Force installation of a kernel module (use modprobe instead, when possible). Must be invoked as root.


Force unloading of a kernel module. Must be invoked as root.


Module loader that is normally invoked automatically in a startup script. Must be invoked as root.


Creates module dependency file. Usually invoked from a startup script.


Output information about a loadable module.

bash$ modinfo hid
filename:    /lib/modules/2.4.20-6/kernel/drivers/usb/hid.o
 description: "USB HID support drivers"
 author:      "Andreas Gal, Vojtech Pavlik <[email protected]>"
 license:     "GPL"



Runs a program or script with certain environmental variables set or changed (without changing the overall system environment). The [varname=xxx] permits changing the environmental variable varname for the duration of the script. With no options specified, this command lists all the environmental variable settings. [96]


The first line of a script (the "sha-bang" line) may use env when the path to the shell or interpreter is unknown.

#! /usr/bin/env perl

print "This Perl script will run,\n";
print "even when I don't know where to find Perl.\n";

# Good for portable cross-platform scripts,
# where the Perl binaries may not be in the expected place.
# Thanks, S.C.

Or even ...

#!/bin/env bash
# Queries the $PATH enviromental variable for the location of bash.
# Therefore ...
# This script will run where Bash is not in its usual place, in /bin.


Show shared lib dependencies for an executable file.

bash$ ldd /bin/ls => /lib/ (0x4000c000)
/lib/ => /lib/ (0x80000000)

Run a command repeatedly, at specified time intervals.

The default is two-second intervals, but this may be changed with the -n option.

watch -n 5 tail /var/log/messages
# Shows tail end of system log, /var/log/messages, every five seconds.


Unfortunately, piping the output of watch command to grep does not work.


Remove the debugging symbolic references from an executable binary. This decreases its size, but makes debugging it impossible.

This command often occurs in a Makefile, but rarely in a shell script.


List symbols in an unstripped compiled binary.


Command-line tool for manipulating the root window of the screen.

Example 17-11. Backlight: changes the brightness of the (laptop) screen backlight

# reldate 02dec2011

#  A bug in Fedora Core 16/17 messes up the keyboard backlight controls.
#  This script is a quick-n-dirty workaround, essentially a shell wrapper
#+ for xrandr. It gives more control than on-screen sliders and widgets.

OUTPUT=$(xrandr | grep LV | awk '{print $1}')   # Get display name!
INCR=.05      # For finer-grained control, set INCR to .03 or .02.

old_brightness=$(xrandr --verbose | grep rightness | awk '{ print $2 }')

if [ -z "$1" ]
  bright=1    # If no command-line arg, set brightness to 1.0 (default).

    if [ "$1" = "+" ]
      bright=$(echo "scale=2; $old_brightness + $INCR" | bc)   # +.05

    if [ "$1" = "-" ]
      bright=$(echo "scale=2; $old_brightness - $INCR" | bc)   # -.05

    if [ "$1" = "#" ]   # Echoes current brightness; does not change it.

    if [[ "$1" = "h" || "$1" = "H" ]]
      echo "Usage:"
      echo "$0 [No args]    Sets/resets brightness to default (1.0)."
      echo "$0 +            Increments brightness by 0.5."
      echo "$0 -            Decrements brightness by 0.5."
      echo "$0 #            Echoes current brightness without changing it."
      echo "$0 N