Mario Dantas Unix-Perl

Advanced Course on Bioinformatic and Comparative Genome Analysis
UFSC Florianopolis - June 30 - July 12, 2008
UFSC - CTC
Unix/perl
Prof. Mario Dantas
mario@inf.ufsc.br
http://www.inf.ufsc.br
Motivation
The large amount of computing resources
available in many organizations can be gather to
solve a large amount of problems from several
research areas.
Biology represents an example of an areas that
can improve experiments through the use of
these distributed resources.
Motivation
Workflow:
- represent a execution flow which

data are passed between some
tasks obeying rules previously
defined.
Ontology:
Ontology can be expressed as a
formal and explicit specification from
a shared concept.
GRID 2007 - September 20, 2007
Motivation
Grid Information Service
R2
Grid Resource Broker
R3
R5
R4
RN
R6
R1
Grid Information Service
Several
Virtual
Organizations
Each organization
develops your
ontology
It does not have an

unique ontology
Motivation
The Architecture Proposal
module Integration Portal

module Information
Provider
module Matchmaker
Motivation
Approach Semantic Integration
Motivation
Reference Ontology (RO)
Motivation
Graphic interface for editing
queries
Motivation
Case Study 1 (matching semantic)
Motivation
Case Study 2 (checking queries)
Motivation
SuMMIT
11
Motivation
SuMMIT Mobile GUI
Monitoring Interface
12
Motivation
SuMMIT Agent
13
Motivation
SuMMIT Workflow Manager
(4/7)
14
Motivation
SuMMIT Operation
Automation
and
Coordination
15
Motivation
SuMMIT Resource Selector
16
Objective
In this course will introduce the basics of
UNIX/perl programming, therefore in the
end participants will be able to write
simple scripts programms.
References
There are several books and sites that
can help in the task to develop and
improve your knowledge about Unix and
perl, some examples are:
Books
Available Free Online
http://proquest.safaribooksonline.com
Books
Interesting Sites
http://directory.fsf.org/project/emacs/
http://stein.cshl.org/genome_informatics/
http://www.cs.usfca.edu/~parrt/course/601/l
ectures/unix.util.html
Interesting Sites
http://people.genome.duke.edu/~jes12//co
urses/perl_duke_2005/
http://www.pasteur.fr/~tekaia/BCGA_useful
_links.html
http://google.com/linux
Course Outline
Introduction
Operating system overview

UNIX utilities
Scripting languages
Programming tools
Course Outline
Introduction

UNIX utilities
Scripting languages
Programming tools
In the Beginning
UNICS: 1969 PDP-7 minicomputer

PDP-7 goes away, rewritten on PDP-11 to
help patent lawyers
V1: 1971
V3: 1973 (pipes, C language)
V6: 1976 (rewritten in C, base for BSD)
V7: 1979 (Licensed, portable)
PDP-11
Big Reason for V6 Success
Commercial Success
AIX
SunOS, Solaris
Ultrix, Digital Unix
HP-UX
Irix
UnixWare -> Novell -> SCO -> Caldera ->SCO
Xenix:
-> SCO
Standardization (Posix, X/Open)
But Then The Feuding Began

Unix International vs. Open Software Foundation
(to compete with desktop PCs)
Battle of the Window Managers
Openlook
Motif
Threat of Windows NT resolves battle with CDE
Send in the Clones

Linux
Written in 1991 by Linus Torvalds
Most popular UNIX variant
Free with GNU license
BSD Lite
FreeBSD (1993, focus on PCs)

NetBSD (1993, focus on portability)
OpenBSD (1996, focus on security)
Free with BSD license
Development less centralized
Today: Unix is Big
Popular Success!
Linux at Google & Elsewhere
Darwin
Apple abandoned old Mac OS for UNIX
Purchased NeXT in December 1996

Unveiled in 2000
Based on 4.4BSD-Lite
Aqua UI written over Darwin
Open Source
Why did UNIX succeed?
Technical strengths!
Research, not commercial
PDP-11 was popular with an unusable OS
AT&Ts legal concerns
Not allowed to enter computer business but
needed to write software to help with switches
Licensed cheaply or free
The Open Source Movement

Has fueled much growth in UNIX
Keeps up with pace of change
More users, developers
More platforms, better
performance, better code
Many vendors switching to Linux
SCO vs. Linux

Jan 2002: SCO releases Ancient Unix : BSD style
licensing of V5/V6/V7/32V/System III
March 2003: SCO sues IBM for $3 billion. Alleges
contributions to Linux come from proprietary licensed
code
AIX is based on System V r4, now owned by SCO
Aug 2003: Evidence released

Code traced to Ancient UNIX
Isnt in 90% of all running Linux distributions
Already dropped from Linux in July
Aug 2005: Linux Kernel Code May Have Been in SCO
Does Linux borrow from ancient UNIX or System V R4?
UNIX vs. Linux
The UNIX Philosophy

Small is beautiful
Easy to understand
Easy to maintain
More efficient
Better for reuse
Make each program do one thing well

More complex functionality by combining
programs
Make every program a filter
The UNIX Philosophy

Portability over efficiency
..continued
Most efficient implementation is rarely portable

Portability better for rapidly changing hardware
Use flat ASCII files

Common, simple file format (yesterdays XML)
Example of portability over efficiency
Reusable code
Good programmers write good code;
great programmers borrow good code
The UNIX Philosophy

..continued
Scripting increases leverage and portability
print $(who | awk '{print $1}' | sort | uniq) | sed 's/ /,/g'
List the logins of a systems users on a single line.
Build prototypes quickly (high level

interpreted languages)
who
755
awk
3,412
sort
2,614
uniq
302
sed
2,093
9,176 lines
The UNIX Philosophy

..continued
Avoid captive interfaces

The user of a program isnt always human
Look nice, but code is big and ugly
Problems with scale
Silence is golden
Only report if something is wrong
Think hierarchically
UNIX Highlights / Contributions
Portability (variety of hardware; C implementation)

Hierarchical file system; the file abstraction
Multitasking and multiuser capability for minicomputer
Inter-process communication
Pipes: output of one programmed fed into input of another
Software tools
Development tools
Scripting languages
TCP/IP
The Operating System

The government of your computer
Kernel: Performs critical system functions
and interacts with the hardware
Systems utilities: Programs and libraries
that provide various functions through
systems calls to the kernel
Kernel Basics
The kernel is
a program loaded into memory during the
boot process, and always stays in physical
memory.
responsible for managing CPU and memory
for processes, managing file systems, and
interacting with devices.
UNIX Structural Layout

shell scripts
system calls
signal handler
device drivers
terminal
disk
User Space
utilities
C programs
compilers
Kernel
scheduler
Devices
printer
swapper
RAM
Kernel Subsystems
Process management
Schedule processes to run on CPU
Inter-process communication (IPC)
Memory management
Virtual memory
Paging and swapping
I/O system
File system
Device drivers
Buffer cache
System Calls
Interface to the kernel
Over 1,000 system calls available on Linux
3 main categories
File/device manipulation
e.g. mkdir(), unlink()
Process control
e.g. fork(), execve(), nice()
Information manipulation
e.g. getuid(), time()
Logging In
Need an account and password first
Enter at login: prompt
Password not echoed
After successful login, you will see a shell prompt
Entering commands
At the shell prompt, type in commands
Typical format: command options arguments
Examples: who, date, ls, cat myfile, ls l
Case sensitive
exit to log out
Remote Login
Use Secure Shell
(SSH)
Windows
e.g. PuTTY
UNIX-like OS
ssh name@access.cims.nyu.edu
UNIX on Windows
Two recommended UNIX emulation environments:
UWIN (AT&T)
http://www.research.att.com/sw/tools/uwin
Cygwin (GPL)
http://www.cygwin.com
Linux Distributions
Slackware the original

Debian collaboration of volunteers
Red Hat / Fedora commerical success
Ubuntu currently most popular, based on
Debian. Focus on desktop
Gentoo portability
Knoppix live distribution
Course Outline
Introduction

UNIX utilities
Scripting languages
Programming tools
Unix System Structure

user
c programs
scripts
shell and utilities
ls
ksh
gcc
find
open()
fork()
exec()
kernel
hardware
Kernel Subsystems
File system
Deals with all input and output
Includes files and terminals
Integration of storage devices
Process management
Deals with programs and program interaction
How processes share CPU, memory and signals

Scheduling
Interprocess Communication
Memory management
UNIX variants have different implementations

of different subsystems.
What is a shell?
The user interface to the operating system
Functionality:
Execute other programs
Manage files
Manage processes
A program like any other

Executed when you log on
Most Commonly Used Shells

/bin/sh The Bourne Shell / POSIX shell
/bin/csh
C shell
/bin/tcsh
Enhanced C Shell
/bin/ksh
Korn shell
/bin/bash
Free ksh clone
Basic form of shell:
while (read command) {
parse command
execute command
}
Shell Interactive Use

When you log in, you interactively use the shell:
Command history
Command line editing
File expansion (tab completion)
Command expansion
Key bindings
Spelling correction
Job control
Shell Scripting
A set of shell commands that
constitute an executable program
A shell script is a regular text file that
contains shell or UNIX commands
Very useful for automating repetitive task and
administrative tools and for storing
commands for later execution
Simple Commands
simple command: sequence of non blanks
arguments separated by blanks or tabs.
1st argument (numbered zero) usually specifies
the name of the command to be executed.
Any remaining arguments:
Are passed as arguments to that command.
Arguments may be filenames, pathnames, directories
or special options (up to command)
Special characters are interpreted by shell
A simple example
$ ls l /bin
-rwxr-xr-x 1 root
$
prompt
command
sys
43234 Sep 26
2001 date
arguments
Execute a basic command

Parsing into command in arguments is called
splitting
Types of Arguments
$ tar c v f archive.tar main.c main.h
Options/Flags
Convention: -X or --longname
Parameters
May be files, may be strings
Depends on command
Getting Help on UNIX

man: display entries from UNIX online
documentation
whatis, apropos
Manual entries organization:
1. Commands
2. System calls
3. Subroutines
4. Special files
5. File format and conventions
6. Games
http://en.wikipedia.org/wiki/Unix_manual
Example Man Page

ls ( 1 )
USER
COMMANDS
ls ( 1 )
NAME
ls - list files and/or directories
SYNOPSIS
ls [ options ] [ file ... ]
DESCRIPTION
For each directory argument ls lists the contents; for each file argument the name and requested information are listed. The
current directory is listed if no file arguments appear. The listing is sorted by file name by default, except that file arguments
are listed before directories.
.
OPTIONS
a, --all
List entries starting with .; turns off --almost-all.
-F, --classify
Append a character for typing each entry.
-l, --long|verbose
Use a long listing format.
-r, --reverse
Reverse order while sorting.
-R, --recursive
List subdirectories recursively.
-
SEE ALSO
chmod(1), find(1), getconf(1), tw(1)
Fundamentals of Security
UNIX systems have one or more users,
identified with a number and name.
A set of users can form a group. A user
can be a member of multiple groups.
A special user (id 0, name root) has
complete control.
Each user has a primary (default)
group.
How are Users & Groups used?

Used to determine if file or process
operations can be performed:
Can a given file be read? written to?
Can this program be run?
Can I use this piece of hardware?
Can I stop a particular process thats running?
A simple example
$ ls l /bin
-rwxr-xr-x 1 root
$
read
write
sys
execute
43234 Sep 26
2001 date
The UNIX File Hierarchy
Hierarchies are Ubiquitous
Definition: Filename
A sequence of characters other than slash.
Case sensitive.
tmp
usr
dmr
wm4
foo
etc
bin
who
date
foo
.profile
who
date
.profile
Definition: Directory
Holds a set of files or other directories.
Case sensitive.
tmp
usr
dmr
wm4
foo
etc
bin
who
date
etc
.profile
usr
dmr
bin
Definition: Pathname
A sequence of directory names followed by a simple
filename, each separated from the previous one by a /
/
tmp
usr
dmr
wm4
.profile
foo
etc
bin
who
date
/usr/wm4/.profile
Definition: Working Directory

A directory that file names refer to by default.
One per process.
/
tmp
usr
dmr
wm4
.profile
foo
etc
bin
who
date
Definition: Relative Pathname

A pathname relative to the working directory (as
opposed to absolute pathname)
/
tmp
usr
dmr
wm4
.profile
foo
etc
bin
who
date
.. refers to parent directory

. refers to current directory
.profile
./.profile
../wm4/.profile
Files and Directories

Files are just a sequence of bytes
No file types (data vs. executable)
No sections
Example of UNIX philosophy
Directories are a list of files and status of the

files:
Creation date
Attributes
etc.
Tilde Expansion
Each user has a home directory
Most shells (ksh, csh) support ~ operator:
~ expands to my home directory
~/myfile
/home/kornj/myfile
~user expands to users home directory

~unixtool/file2
/home/unixtool/file2
Useful because home directory locations

vary by machine
Mounting File Systems

When UNIX is started, the directory hierarchy
corresponds to the file system located on a
single disk called the root device.
Mounting allows root to splice the root
directory of a file system into the existing
directory hierarchy.
File systems created on other devices can be
attached to the original directory hierarchy
using the mount mechanism.
Commands mount and umount manage
Mounting File Systems

root device
external device
/
a
a
b
Device
Mount Point
/
/a/b
b
a
Mount table
Printing File Contents

The cat command can be used to copy the
contents of a file to the terminal. When invoked
with a list of file names, it concatenates them.
Some options:
-n
-v
number output lines (starting from 1)

display control-characters in visible
form (e.g. ^C)
Interactive commands more and less show a

page at a time
Common Utilities for Managing files and

directories
pwd
print process working dir
ed, vi, emacs
create/edit files
ls
list contents of directory
rm
remove file
mv
rename file
cp
copy a file
touch
create an empty file or update
mkdir and rmdir create and remove dir
wc
counts the words in a file
file
determine file contents
du
directory usage
File Permissions
UNIX provides a way to protect files based on
users and groups.
Three types of permissions:
read, process may read contents of file
write, process may write contents of file
execute, process may execute file
Three sets of permissions:

permissions for owner
permissions for group (1 group per file)
permissions for other
Directory permissions
Same types and sets of permissions as for
files:
read: process may a read the directory
contents (i.e., list files)
write: process may add/remove files in the
directory
execute: process may open files in directory
or subdirectories
Utilities for Manipulating file

attributes
chmod
change file permissions
chown
change file owner
chgrp
change file group
umask
user file creation mode mask
only owner or super-user can change file
attributes
upon creation, default permissions given
to file modified by process umask value
Chmod command
Symbolic access modes {u,g,o} /
{r,w,x}
example: chmod +r file
Octal access modes

octal
read
write
execute
0
1
2
3
4
5
6
7
no
no
no
no
yes
yes
yes
yes
no
no
yes
yes
no
no
yes
yes
no
yes
no
yes
no
yes
no
yes
File System Internals
The Open File Table

I/O operations are done on files by first
opening them, reading/writing/etc., then
closing them.
The kernel maintains a global table
containing information about each open file.
Inode
Mode
Count
Position
1023
1331
read
read/write
1
2
0
50
The File Descriptor Table

Each process contains a
table of files it has opened.
Inherits open files from
parent
Each open file is associated
with a number or handle,
called file descriptor, (fd).
Each entry of this table
points to an entry in the
open file table.
Always starts at 0
Why not directly use

the open file table?
Convenient for kernel
Indirection makes security easier
Numbering scheme can be local to

process (0 .. 128)
Extra information stored:
Should the open file be inherited by children?
(close-on-exec flag)
Standard in/out/err
The first three entries in the file descriptor
table are special by convention:
cat
Entry 0 is for input

Entry 1 is for output
Entry 2 is for error
messages
What about reading/writing to the

screen?
Devices
Besides files, input and output can go
from/to various hardware devices
UNIX innovation: Treat these just like files!

/dev/tty, /dev/lpr, /dev/modem
By default, standard in/out/err opened

with /dev/tty
Redirection
Before a command is executed, the input and
output can be changed from the default
(terminal) to a file
Shell modifies file descriptors in child process
The child program knows nothing about this
ls
ls
Redirection of input/ouput
Redirection of output: >
example:$ ls > my_files
Redirection of input: <

example: $ mail kornj <input.data
Append output: >>

example: $ date >> logfile
Bourne Shell derivatives: fd>

example: $ ls 2> error_log
Using Devices
Redirection works with devices (just like
files)
Special files in /dev directory
Example: /dev/tty
Example: /dev/lp
Example: /dev/null
cat big_file > /dev/lp
cat big_file > /dev/null
Links
Directories are a list of files and directories.
Each directory entry links to a file on the disk
mydir
hello
file2
subdir
Hello
World!
Two different directory entries can link to the same file

In same directory or across different directories
Moving a file does not actually move any data around.

Creates link in new location
Deletes link in old location
ln command
Symbolic links
Symbolic links are different than regular links (often
called hard links). Created with ln -s
Can be thought of as a directory entry that points to
the name of another file.
Does not change link count for file
When original deleted, symbolic link remains
They exist because:

Hard links dont work across file systems
Hard links only work for regular files, not directories
dirent
dirent
Contents of file
Hard link
symlink
dirent
Symbolic Link
Contents of file
Example
usr tmp etc bin
dmr wm4
.profile
etc
foo
who date
Hard Link
usr tmp etc bin
dmr wm4
.profile
etc
foo
who date
Symbolic Link
usr tmp etc bin
dmr wm4
foo
who date
.profile
etc
/usr/wm4/.profile
Can a file have no links?

usr tmp etc bin
dmr wm4
foo
who date
.profile
etc
cat
Tree Walking
How can do we find a set of files in the
hierarchy?
One possibility:
ls l R /
What about:
All files below a given directory in the hierarchy?
All files since Jan 1, 2001?
All files larger than 10K?
find utility
find pathlist expression
find recursively descends through pathlist
and applies expression to every file.
expression can be:
-name pattern
true if file name matches pattern. Pattern may
include shell patterns such as *, must be in quotes
to suppress shell interpretation.
Eg: find / -name '*.c'
find utility (continued)

-perm [+-]mode
Find files with given access mode, mode must be in octal.
Eg: find . 755
-type ch
Find files of type ch (c=character, b=block, f for plain file,
etc..). Eg: find /home type f
-user userid/username
Find by owner userid or username
-group groupid/groupname
Find by group groupid or groupname
-size size
File size is at least size
many more
find: logical operations

! expression
op1 -a op2
op1 -o op2
( )
returns the logical negation

of expression
matches both patterns
op1 and op2
matches either op1 or op2
group expressions together
find: actions
-print
prints out the name of the

current file (default)
-exec cmd
Executes cmd, where cmd must be terminated by an
escaped semicolon (\; or ';').
If you specify {} as a command line argument, it is
replaced by the name of the current file just found.
exec executes cmd once per file.
Example:
find -name "*.o" -exec rm "{}" ";"
find Examples
Find all files beneath home directory beginning with f
find ~ -name 'f*' -print
Find all files beneath home directory modified in last day

find ~ -mtime 1 -print
Find all files beneath home directory larger than 10K

find ~ -size 10k -print
Count words in files under home directory

find ~ -exec wc -w {} \; -print
Remove core files

find / -name core exec rm {} \;
diff: comparing two files

diff: compares two files and outputs a
description of their differences
Usage: diff [options] file1 file2
-i: ignore case
apples
oranges
walnuts
apples
oranges
grapes
$ diff test1 test2

3c3
< walnuts
--> grapes
Other file comparison utilities

cmp
Tests two files for equality
If equal, nothing returned. If different, location of
first differing byte returned
Faster than diff for checking equality
comm
Reads two files and outputs three columns:
Lines in first file only
Lines in second file only
Lines in both files
Must be sorted
Options: fields to suppress ( [-123] )
Course Outline
Introduction

UNIX utilities
Scripting languages
Programming tools
Kernel Data Structures

Information
about each
process.
Process table:
contains an
entry for every
process in the
system.
Open-file table:
contains at least
one entry for
every open file
in the system.
User Space
Code
Code
Code
Data
Data
Data
Process
Info
Process
Info
Process
Info
Open File
Table
Kernel Space
Process
Table
Unix Processes
Process: An entity of execution
Definitions
program: collection of bytes stored in a file that can
be run
image: computer execution environment of program
process: execution of an image
Unix can execute many processes

simultaneously.
Process Creation
Interesting trait of UNIX
fork system call clones the current process
A
exec system call replaces current process

A
A fork is typically followed by an exec
Process Setup
All of the per process information is copied
with the fork operation
Working directory
Open files
Copy-on-write makes this efficient

Before exec, these values can be
modified
fork and exec

Example: the shell
while(1){
display_prompt();
read_input(cmd,params);
pid=fork();
if(pid!=0)
waitpid(1,&stat,0);
else
execve(cmd,params,0);
}
/*createchild*/
/*parentwaits*/
/*childexecs*/
Unix process genealogy

P r o c e s s g e n e r a t io n
I n it p r o c e s s 1
f o r k s in it p r o c e s s e s
in it
execs
in it
execs
I n it
execs
g e tty
g e tty
execs
g e tty
lo g in
execs
/ b in / s h
Background Jobs
By default, executing a command in the
shell will wait for it to exit before printing
out the next prompt
Trailing a command with & allows the shell
and command to run simultaneously
$ /bin/sleep 10 &
[1] 3424
$
Program Arguments
When a process is started, it is sent a list
of strings
argv, argc
The process can use this list however it

wants to
Ending a process
When a process ends, there is a return
code associated with the process
This is a positive integer
0 means success
>0 represent various kinds of failure, up to
process
Process Information Maintained

Working directory
File descriptor table
Process id
number used to identify process
Process group id
number used to identify set of processes
Parent process id
process id of the process that created the process
Process Information Maintained

Umask
Default file permissions for new file
We havent talked about these yet:
Effective user and group id

The user and group this process is running with
permissions as
Real user and group id

The user and group that invoked the process
Environment variables
Setuid and Setgid Mechanisms

The kernel can set the effective user and
group ids of a process to something
different than the real user and group
Files executed with a setuid or setgid flag set
cause the these values to change
Make it possible to do privileged tasks:

Change your password
Open up a can of worms for security if

buggy
Environment of a Process
A set of name-value pairs associated with
a process
Keys and values are strings
Passed to children processes
Cannot be passed back up
Common examples:
PATH: Where to search for programs
TERM: Terminal type
The PATH environment variable

Colon-separated list of directories.
Non-absolute pathnames of executables
are only executed if found in the list.
Searched left to right
Example:
$ myprogram
sh: myprogram not found
$ PATH=/bin:/usr/bin:/home/kornj/bin
$ myprogram
hello!
Having . In Your Path

$ ls
foo
$ foo
sh: foo: not found
$ ./foo
Hello, foo.
What not to do:

$ PATH=.:/bin
$ ls
foo
$ cd /usr/badguy
$ ls
Congratulations, your files have been removed
and you have just sent email to Prof. Korn
challenging him to a fight.
Shell Variables
Shells have several mechanisms for creating
variables. A variable is a name representing a
string value. Example: PATH
Shell variables can save time and reduce typing
errors
Allow you to store and manipulate information

Eg: ls $DIR > $FILE
Two types: local and environmental

local are set by the user or by the shell itself
environmental come from the operating system
and are passed to children
Variables (cont)
Syntax varies by shell
varname=value
set varname = value
# sh, ksh
# csh
To access the value: $varname

Turn local variable into environment:
export varname
# sh, ksh
setenv varname value # csh
Environmental Variables
NAMEMEANING
$HOME Absolute pathname of your home
directory
$PATH A list of directories to search for
$MAIL Absolute pathname to mailbox
$USER Your user id
$SHELL Absolute pathname of login shell
$TERM Type of your terminal
$PS1
Prompt
Inter-process Communication
Ways in which processes communicate:
Passing arguments, environment
Read/write regular files
Exit values
Signals
Pipes
Signals
Signal: A message a process can send to a process
or process group, if it has appropriate permissions.
Message type represented by a symbolic name
For each signal, the receiving process can:
Explicitly ignore signal
Specify action to be taken upron receipt (signal handler)
Otherwise, default action takes place (usually process is
killed)
Common signals:
SIGKILL, SIGTERM, SIGINT
SIGSTOP, SIGCONT
SIGSEGV, SIGBUS
An Example of Signals
When a child exists, it sends a
SIGCHLD signal to its parent.
If a parent wants to wait for a child to
exit, it tells the system it wants to catch
the SIGCHLD signal
When a parent does not issue a wait, it
ignores the SIGCHLD signal
Process Subsystem utilities

ps
kill
wait
nohup
the
signal
sleep
nice
monitors status of processes

send a signal to a pid
parent process wait for one of its
children to terminate
makes a command immune to
hangup and terminate
sleep in seconds
run processes at low priority
Pipes
One of the cornerstones of UNIX
Pipes
General idea: The input of one program
is the output of the other, and vice versa
Both programs run at the same time
Pipes (2)
Often, only one end of the pipe is used
standard out
standard in
Could this be done with files?
File Approach
Run first program, save output into file
Run second program, using file as input
process 1
Unnecessary use of the disk

Slower
Can take up a lot of space
Makes no use of multi-tasking
process 2
More about pipes

What if a process tries to read data but nothing
is available?
UNIX puts the reader to sleep until data available
What if a process cant keep up reading from

the process thats writing?
UNIX keeps a buffer of unread data
This is referred to as the pipe size.
If the pipe fills up, UNIX puts the writer to sleep until
the reader frees up space (by doing a read)
Multiple readers and writers possible with pipes.
More about Pipes

Pipes are often chained together
Called filters
standard out
standard in
Interprocess Communication
For Unrelated Processes
FIFO (named pipes)
A special file that when opened represents
pipe
System V IPC
p1
message queues
semaphores
shared memory
Sockets (client/server model)
p2
Pipelines
Output of one program becomes input to
another
Uses concept of UNIX pipes
Example: $ who | wc -l
counts the number of users logged in
Pipelines can be long
Whats the difference?

Both of these commands send input to command from a
file instead of the terminal:
$ cat file | command

vs.
$ command < file
An Extra Process
$ cat file | command
cat
command
$ command < file

command
Introduction to Filters
A class of Unix tools called filters.
Utilities that read from standard input,
transform the file, and write to standard out
Using filters can be thought of as data

oriented programming.
Each step of the computation transforms data
stream.
Examples of Filters
Sort
Input: lines from a file
Output: lines from the file sorted
Grep
Input: lines from a file
Output: lines that match the argument
Awk
Programmable filter
cat: The simplest filter

The cat command copies its input to output
unchanged (identity filter). When supplied a list
of file names, it concatenates them onto stdout.
Some options:
-n
-v
^C)
number output lines (starting from 1)

display control-characters in visible form (e.g.
cat file*
ls | cat -n
head
Display the first few lines of a specified file
Syntax: head [-n] [filename...]
-n - number of lines to display, default is 10
filename... - list of filenames to display
When more than one filename is

specified, the start of each files listing
displays
==>filename<==
tail
Displays the last part of a file
Syntax: tail +|-number [lbc] [f] [filename]
or: tail +|-number [l] [rf] [filename]
+number - begins copying at distance number from
beginning of file, if number isnt given, defaults to 10
-number - begins from end of file
l,b,c - number is in units of lines/block/characters
r - print in reverse order (lines only)
f - if input is not a pipe, do not terminate after end of
file has been copied but loop. This is useful to
monitor a file being written by another process
head and tail examples

head /etc/passwd
head *.c
tail +20 /etc/passwd
ls -lt | tail -3
head 100 /etc/passwd | tail -5
tail f /usr/local/httpd/access_log
tee
Unix Command
Standard output
file-list
Copy standard input to standard output

and one or more files
Captures intermediate results from a filter
in the pipeline
tee cont
Syntax: tee [ -ai ] file-list
-a - append to output file rather than overwrite,
default is to overwrite (replace) the output file
-i - ignore interrupts
file-list - one or more file names for capturing
output
Examples
ls | head 10 | tee first_10 | tail 5
who | tee user_list | wc
Unix Text Files: Delimited Data

Tab Separated
John
Anne
Andrew
Tim
Arun
Sowmya
99
75
50
95
33
76
Pipe-separated
COMP1011|2252424|Abbot, Andrew John |3727|1|M
COMP2011|2211222|Abdurjh, Saeed |3640|2|M
COMP1011|2250631|Accent, Aac-Ek-Murhg |3640|1|M
COMP1021|2250127|Addison, Blair |3971|1|F
COMP4012|2190705|Allen, David Peter |3645|4|M
COMP4910|2190705|Allen, David Pater |3645|4|M
Colon-separated
root:ZHolHAHZw8As2:0:0:root:/root:/bin/ksh
jas:nJz3ru5a/44Ko:100:100:John Shepherd:/home/jas:/bin/ksh
cs1021:iZ3sO90O5eZY6:101:101:COMP1021:/home/cs1021:/bin/bash
cs2041:rX9KwSSPqkLyA:102:102:COMP2041:/home/cs2041:/bin/csh
cs3311:mLRiCIvmtI9O2:103:103:COMP3311:/home/cs3311:/bin/sh
cut: select columns

The cut command prints selected parts of input
lines.
can select columns (assumes tab-separated input)
can select a range of character positions
Some options:
-f listOfCols: print only the specified columns (tabseparated) on output
-c listOfPos: print only chars in the specified positions
-d c: use character c as the column separator
Lists are specified as ranges (e.g. 1-5) or commaseparated (e.g. 2,4,5).
cut examples
cut -f 1 < data
cut -f 1-3 < data
cut -f 1,4 < data
cut -f 4- < data
cut -d'|' -f 1-3 < data
cut -c 1-4 < data
Unfortunately, there's no way to refer to "last
column" without counting the columns.
paste: join columns

The paste command displays several text files "in
parallel" on output.
1
3
5
If the inputs are files a, b, c
2
4
6
the first line of output is composed
of the first lines of a, b, c
the second line of output is composed
of the second lines of a, b, c
1
2
3
4
5
6
Lines from each file are separated by a tab character.

If files are different lengths, output has all lines from
longest file, with empty strings for missing lines.
paste example
cut -f 1 < data > data1
paste data1 data3 data2 > newdata
sort: Sort lines of a file

The sort command copies input to output but
ensures that the output is arranged in
ascending order of lines.
By default, sorting is based on ASCII comparisons
of the whole line.
Other features of sort:

understands text data that occurs in columns.
(can also sort on a column other than the first)
can distinguish numbers and sort appropriately
can sort files "in place" as well as behaving like a
filter
capable of sorting very large files
sort: Options
Syntax: sort [-dftnr] [-o filename] [filename(s)]
-dDictionary order, only letters, digits, and whitespace
are significant in determining sort order
-f Ignore case (fold into lower case)
-t Specify delimiter
-n Numeric order, sort by arithmetic value instead of
first digit
-r Sort in reverse order
-ofilename - write output to filename, filename can be
the same as one of the input files
Lots of more options
sort: Specifying fields

Delimiter : -td
Old way:
+f[.c][options] [-f[.c][options]
+2.1 3 +0 2 +3n
Exclusive
Start from 0 (unlike cut, which starts at 1)
New way:
-k f[.c][options][,f[.c][options]]
-k2.1 k0,1 k3n
Inclusive
Start from 1
sort Examples
sort +2nr < data
sort k2nr data
sort -t: +4 /etc/passwd
sort -o mydata mydata
uniq: list UNIQue items

Remove or report adjacent duplicate lines
Syntax: uniq [ -cdu] [input-file] [ output-file]
-c Supersede the -u and -d options and
generate an output
report with each line
preceded by an occurrence count
-d Write only the duplicated lines
-u Write only those lines which are not
duplicated
The default output is the union (combination) of
-d and -u
wc: Counting results

The word count utility, wc, counts the
number of lines, characters or words
Options:
-l
-w
-c
Count lines
Count words
Count characters
Default: count lines, words and chars
wc and uniq Examples

who | sort | uniq d
wc my_essay
who | wc
sort file | uniq | wc l
sort file | uniq d | wc l
sort file | uniq u | wc -l
tr: TRanslate Characters

Copies standard input to standard output with
substitution or deletion of selected characters
Syntax: tr [ -cds ] [ string1 ] [ string2 ]
-d delete all input characters contained in string1
-c complements the characters in string1 with respect
to the entire ASCII character set
-s squeeze all strings of repeated output characters
in the last operand to single characters
tr (continued)
tr reads from standard input.
Any character that does not match a character in
string1 is passed to standard output unchanged
Any character that does match a character in string1
is translated into the corresponding character in
string2 and then passed to standard output
Examples
tr s z replaces all instances of s with z
tr so zx replaces all instances of s with z and o with
x
tr a-z A-Z
replaces all lower case characters
with
upper case characters
tr d a-c
deletes all a-c characters
tr uses
Change delimiter
tr | :
Rewrite numbers
tr ,. .,
Import DOS files

tr d \r < dos_file
Find printable ASCII in a binary file

tr cd \na-zA-Z0-9 < binary_file
xargs
Unix limits the size of arguments and
environment that can be passed down to child
What happens when we have a list of 10,000
files to send to a command?
xargs solves this problem
Reads arguments as standard input
Sends them to commands that take file lists
May invoke program several times depending on size
of arguments
cmd a1 a2
a1 a300
xargs
cmd
cmd a100 a101

cmd a200 a201
find utility and xargs

find . -type f -print | xargs wc -l
-type f for files
-print to print them
out
xargs invokes wc 1 or more times
wc -l a b c d e f g
wc -l h i j k l m n o
Compare to: find . -type f exec wc -l {} \;
Next Time
Regular Expressions
Allow you to search for text in files
grep command
We will soon learn how to write scripts that

use these utilities in interesting ways.
Previously
Basic UNIX Commands
Files: rm, cp, mv, ls, ln
Processes: ps, kill
Unix Filters
cat, head, tail, tee, wc

cut, paste
find
sort, uniq
comm, diff, cmp
tr
Subtleties of commands
Executing commands with find

Specification of columns in cut
Specification of columns in sort
Methods of input
Standard in
File name arguments
Special "-" filename
Options for uniq
Today
Regular Expressions
Allow you to search for text in files
grep command
Stream manipulation:
sed
But first, a command we didnt cover last time
xargs
Unix limits the size of arguments and
environment that can be passed down to child
What happens when we have a list of 10,000
files to send to a command?
xargs handles this problem
Reads arguments as standard input
Sends them to commands that take file lists
May invoke program several times depending on size
of arguments
cmd a1 a2
a1 a300
xargs
cmd
cmd a100 a101

cmd a200 a201
find utility and xargs

find . -type f -print | xargs wc -l
for files
-print to print them out
xargs invokes wc 1 or more times
-type f
wc -l a b c d e f g
wc -l h i j k l m n o
Compare to: find . -type f exec wc -l {} \;
The -n option can be used to limit number of args
Regular Expressions
What Is a Regular Expression?

A regular expression (regex) describes a set
of possible input strings.
Regular expressions descend from a
fundamental concept in Computer Science
called finite automata theory
Regular expressions are endemic to Unix
vi, ed, sed, and emacs

awk, tcl, perl and Python
grep, egrep, fgrep
compilers
Regular Expressions
The simplest regular expressions are a
string of literal characters to match.
The string matches the regular
expression if it contains the substring.
regular expression
c k s
UNIX Tools rocks.

match
UNIX Tools sucks.

match
UNIX Tools is okay.

no match
Regular Expressions
A regular expression can match a string in
more than one place.
regular expression
a p p l e
Scrapple from the apple.

match 1
match 2
Regular Expressions
The . regular expression can be used to
match any character.
regular expression
o .
For me to poop on.

match 1
match 2
Character Classes
Character classes [] can be used to
match any specific set of characters.
regular expression
b [eor] a t
beat a brat on a boat

match 1
match 2
match 3
Negated Character Classes

Character classes can be negated with the
[^] syntax.
regular expression
b [êo] a t

match
More About Character Classes

[aeiou] will match any of the characters a, e, i, o,
or u
[kK]orn will match korn or Korn
Ranges can also be specified in character

classes
[1-9] is the same as [123456789]
[abcde] is equivalent to [a-e]
You can also combine multiple ranges
[abcde123456789] is equivalent to [a-e1-9]
Note that the - character has a special meaning in a
character class but only if it is used within a range,
[-123] would match the characters -, 1, 2, or 3
Named Character Classes

Commonly used character classes can be
referred to by name (alpha, lower, upper,
alnum, digit, punct, cntrl)
Syntax [:name:]
[a-zA-Z]
[a-zA-Z0-9]
[45a-z]
[[:alpha:]]
[[:alnum:]]
[45[:lower:]]
Important for portability across languages
Anchors
Anchors are used to match at the beginning or end of a line (or both).
^ means beginning of the line
$ means end of the line
^ b [eor] a t
regular expression

match
regular expression
b [eor] a t $

match
^word$
^$
Repetition
The * is used to define zero or more
occurrences of the single regular
expression preceding it.
y a * y
regular expression
I got mail, yaaaaaaaaaay!

match
regular expression
o a * o
For me to poop on.

match
.*
Match length
A match will be the longest string that
satisfies the regular expression.
regular expression
a . * e
Scrapple from the apple.

no
no
yes
Repetition Ranges
Ranges can also be specified
{ } notation can specify a range of
repetitions for the immediately preceding
regex
{n} means exactly n occurrences
{n,} means at least n occurrences
{n,m} means at least n occurrences but no
more than m occurrences
Example:
.{0,} same as .*
a{2,} same as aaa*
Subexpressions
If you want to group part of an expression so
that * or { } applies to more than just the
previous character, use ( ) notation
Subexpresssions are treated like a single
character
a* matches 0 or more occurrences of a
abc* matches ab, abc, abcc, abccc,
(abc)* matches abc, abcabc, abcabcabc,
(abc){2,3} matches abcabc or abcabcabc
grep
grep comes from the ed (Unix text editor)
search command global regular expression
print or g/re/p
This was such a useful command that it was
written as a standalone utility
There are two other variants, egrep and fgrep
that comprise the grep family
grep is the answer to the moments where you
know you want the file that contains a specific
phrase but you cant remember its name
Family Differences
grep - uses regular expressions for pattern
matching
fgrep - file grep, does not use regular
expressions, only matches fixed strings but can
get search strings from a file
egrep - extended grep, uses a more powerful
set of regular expressions but does not support
backreferencing, generally the fastest member
of the grep family
agrep approximate grep; not standard
Syntax
Regular expression concepts we have seen
so far are common to grep and egrep.
grep and egrep have slightly different syntax
grep: BREs
egrep: EREs (enhanced features we will discuss)
Major syntax differences:

grep: $ and $, \{ and \}
egrep: ( and ), { and }
Protecting Regex
Metacharacters
Since many of the special characters used
in regexs also have special meaning to the
shell, its a good idea to get in the habit of
single quoting your regexs
This will protect any special characters from
being operated on by the shell
If you habitually do it, you wont have to worry
about when it is necessary
Escaping Special Characters

Even though we are single quoting our regexs so
the shell wont interpret the special characters,
some characters are special to grep (eg * and .)
To get literal characters, we escape the character
with a \ (backslash)
Suppose we want to search for the character
sequence a*b*
Unless we do something special, this will match zero or
more as followed by zero or more bs, not what we
want
a\*b\* will fix this - now the asterisks are treated as
regular characters
Egrep: Alternation
Regex also provides an alternation character |
for matching one or another subexpression
(T|Fl)an will match Tan or Flan
^(From|Subject): will match the From and Subject
lines of a typical email message
It matches a beginning of line followed by either the characters
From or Subject followed by a :
Subexpressions are used to limit the scope of the

alternation
At(ten|nine)tion then matches Attention or
Atninetion, not Atten or ninetion as would happen
without the parenthesis - Atten|ninetion
Egrep: Repetition Shorthands

The * (star) has already been seen to specify
zero or more occurrences of the immediately
preceding character
+ (plus) means one or more
abc+d will match abcd, abccd, or abccccccd
but will not match abd
Equivalent to {1,}
Egrep: Repetition Shorthands

cont
The ? (question mark) specifies an optional
character, the single character that immediately
precedes it
July? will match Jul or July
Equivalent to {0,1}
Also equivalent to (Jul|July)
The *, ?, and + are known as quantifiers because
they specify the quantity of a match
Quantifiers can also be used with subexpressions
(a*c)+ will match c, ac, aac or aacaacac but will not
match a or a blank line
Grep: Backreferences
Sometimes it is handy to be able to refer to a
match that was made earlier in a regex
This is done using backreferences
\n is the backreference specifier, where n is a number
Looks for nth subexpression

For example, to find if the first word of a line is
the same as the last:
^$[[:alpha:]]\{1,\}$ .* \1$
The $[[:alpha:]]\{1,\}$ matches 1 or more letters
Practical Regex Examples

Variable names in C
[a-zA-Z_][a-zA-Z_0-9]*
Dollar amount with optional cents

\$[0-9]+(\.[0-9][0-9])?
Time of day
(1[012]|[1-9]):[0-5][0-9] (am|pm)
HTML headers <h1> <H1> <h2>

<[hH][1-4]>
grep Family
Syntax
grep [-hilnv] [-e expression] [filename]
egrep [-hilnv] [-e expression] [-f filename] [expression]
[filename]
fgrep [-hilnxv] [-e string] [-f filename] [string] [filename]
-h Do not display filenames
-i Ignore case
-l List only filenames containing matching lines
-n Precede each matching line with its line number
-v Negate matches
-x Match whole line only (fgrep only)
-e expression Specify expression as option
-f filename
Take the regular expression (egrep)
or a list of strings (fgrep) from filename
grep Examples
grep 'men' GrepMe

grep 'fo*' GrepMe
egrep 'fo+' GrepMe
egrep -n '[Tt]he' GrepMe
fgrep 'The' GrepMe
egrep 'NC+[0-9]*A?' GrepMe
fgrep -f expfile GrepMe
Find all lines with signed numbers

$ egrep[-+][0-9]+\.?[0-9]* *.c
bsearch. c: return -1;
compile. c: strchr("+1-2*3", t-> op)[1] - 0, dst,
convert. c: Print integers in a given base 2-16 (default 10)
convert. c: sscanf( argv[ i+1], "% d", &base);
strcmp. c: return -1;
strcmp. c: return +1;
egrep has its limits: For example, it cannot match all lines that
contain a number divisible by 7.
Fun with the Dictionary

/usr/dict/words contains about
egrep hh /usr/dict/words
beachhead
highhanded
withheld
withhold
25,000 words
egrep as a simple spelling checker: Specify plausible

alternatives you know
egrep "n(ie|ei)ther" /usr/dict/words
neither
How many words have 3 as one letter apart?

egrep a.a.a /usr/dict/words | wc l
54
egrep u.u.u /usr/dict/words
cumulus
Other Notes
Use /dev/null as an extra file name
Will print the name of the file that matched
grep test bigfile
This is a test.
grep test /dev/null bigfile

bigfile:This is a test.
Return code of grep is useful
grep fred filename > /dev/null && rm filename
This is one line of text

o.*o
x
xyz
\m
^
$
.
[xy^$x]
[^xy^$z]
[a-z]
r*
r1r2
$r$
\n
\{n,m\}
r+
r?
r1|r2
(r1|r2)r3
(r1|r2)*
{n,m}
Ordinary characters match themselves

(NEWLINES and metacharacters excluded)
Ordinary strings match themselves
Matches literal character m
Start of line
End of line
Any single character
Any of x, y, ^, $, or z
Any one character other than x, y, ^, $, or z
Any single character in given range
zero or more occurrences of regex r
Matches r1 followed by r2
Tagged regular expression, matches r
Set to what matched the nth tagged expression
(n = 1-9)
Repetition
One or more occurrences of r
Zero or one occurrences of r
Either r1 or r2
Either r1r3 or r2r3
Zero or more occurrences of r1|r2, e.g., r1, r1r1,
r2r1, r1r1r2r1,)
Repetition
input line
regular expression
fgrep, grep, egrep
grep, egrep
grep
egrep
Quick
Reference
Sed: Stream-oriented, NonInteractive, Text Editor

Look for patterns one line at a time, like grep
Change lines of the file
Non-interactive text editor
Editing commands come in as script
There is an interactive editor ed which accepts the
same commands
A Unix filter
Superset of previously mentioned tools
Sed Architecture
Input
Input line
(Pattern Space)
Output
scriptfile
Commands in a sed script are applied

in order to each line.
If a command changes the input,
subsequent command will be applied to
the modified line in the pattern space,
not the original input line.
The input file is unchanged (sed is a
filter).
Results are sent to standard output
Scripts
A script is nothing more than a file of commands
Each command consists of up to two addresses
and an action, where the address can be a
regular expression or line number.
address
action
address
action
address
action
address
action
address
action
command
script
Sed Flow of Control

sed then reads the next line in the input file and
restarts from the beginning of the script file
All commands in the script file are compared to,
and potentially act on, all lines in the input file
script
cmd 1
cmd 2
...
cmd n
Executed if line matches address
print command
input
output
output
only without -n
sed Syntax
Syntax: sed [-n] [-e] [command] [file]
sed [-n] [-f scriptfile] [file]
-n - only print lines specified with the print
command (or the p flag of the substitute (s)
command)
-f scriptfile - next argument is a filename
containing editing commands
-e command - the next argument is an editing
command rather than a filename, useful if multiple
commands are specified
If the first line of a scriptfile is #n, sed acts as
though -n had been specified
sed Commands
sed commands have the general form
[address[, address]][!]command [arguments]
sed copies each input line into a pattern space

If the address of the command matches the line in the
pattern space, the command is applied to that line
If the command has no address, it is applied to each
line as it enters pattern space
If a command changes the line in pattern space,
subsequent commands operate on the modified line
When all commands have been read, the line in

pattern space is written to standard output and a
new line is read into pattern space
Addressing
An address can be either a line number or
a pattern, enclosed in slashes (
/pattern/ )
A pattern is described using regular
expressions (BREs, as in grep)
If no pattern is specified, the command will
be applied to all lines of the input file
To refer to the last line: $
Addressing (continued)
Most commands will accept two addresses
If only one address is given, the command operates
only on that line
If two comma separated addresses are given, then
the command operates on a range of lines between
the first and second address, inclusively
The ! operator can be used to negate an

address, ie; address!command causes
command to be applied to all lines that do not
match address
Commands
command is a single letter
Example: Deletion: d
[address1][,address2]d
Delete the addressed line(s) from the pattern
space; line(s) not passed to standard output.
A new line of input is read and editing
resumes with the first command of the script.
Address and Command

Examples
deletes the all lines
deletes line 6
/^$/d
deletes all blank lines
1,10d
deletes lines 1 through 10
1,/^$/d
deletes from line 1 through the first blank line
/^$/,$d
deletes from the first blank line through
the last line of the file
/^$/,10d
deletes from the first blank line through line 10
/^ya*y/,/[0-9]$/d deletes from the first line that begins
with yay, yaay, yaaay, etc. through
the first line that ends with a digit
d
6d
Multiple Commands
Braces {} can be used to apply multiple commands to
an address
[/pattern/[,/pattern/]]{
command1
command2
command3
}
Strange syntax:
The opening brace must be the last character on a line
The closing brace must be on a line by itself
Make sure there are no spaces following the braces
Sed Commands
Although sed contains many editing commands,
we are only going to cover the following subset:
s - substitute
a - append
i - insert
c - change
d - delete
p - print
y - transform
q - quit
Print
The Print command (p) can be used to force
the pattern space to be output, useful if the -n
option has been specified
Syntax: [address1[,address2]]p
Note: if the -n option has not been specified, p
will cause the line to be output twice!
Examples:
1,5p will display lines 1 through 5
/^$/,$p will display the lines from the first
blank line through the last line of the file
Substitute
Syntax: [address(es)]s/pattern/replacement/
[flags]
pattern - search pattern
replacement - replacement string for pattern
flags - optionally any of the following
a number from 1 to 512 indicating which

occurrence of pattern should be
replaced
global, replace all occurrences of pattern
in pattern space
print contents of pattern space
Substitute Examples
s/Puff Daddy/P. Diddy/
Substitute P. Diddy for the first occurrence of Puff
Daddy in pattern space
s/Tom/Dick/2
Substitutes Dick for the second occurrence of Tom in
the pattern space
s/wood/plastic/p
Substitutes plastic for the first occurrence of wood and
outputs (prints) pattern space
Replacement Patterns
Substitute can use several special
characters in the replacement string
& - replaced by the entire string matched in
the regular expression for pattern
\n - replaced by the nth substring (or
subexpression) previously specified using $
and $
\ - used to escape the ampersand (&) and
the backslash (\)
Replacement Pattern Examples

"the UNIX operating system "
s/.NI./wonderful &/
"the wonderful UNIX operating system "
cat test1
first:second
one:two
sed 's/$.*$:$.*$/\2:\1/' test1
second:first
two:one
sed 's/$[[:alpha:]]$$[^ \n]*$/\2\1ay/g'
Pig Latin ("unix is fun" -> "nixuay siay unfay")
Append, Insert, and Change

Syntax for these commands is a little strange
because they must be specified on multiple
lines
append
[address]a\
text
insert
[address]i\
text
change
[address(es)]c\
text
append/insert for single lines only, not range
Append and Insert

Append places text after the current line in pattern space
Insert places text before the current line in pattern space
Each of these commands requires a \ following it.
text must begin on the next line.
If text begins with whitespace, sed will discard it
unless you start the line with a \
Example:
/<Insert Text Here>/i\

Line 1 of inserted text\
\
Line 2 of inserted text
would leave the following in the pattern space

<Insert Text Here>
Change
Unlike Insert and Append, Change can be
applied to either a single line address or a range
of addresses
When applied to a range, the entire range is
replaced by text specified with change, not each
line
Exception: If the Change command is executed with
other commands enclosed in { } that act on a range
of lines, each line will be replaced with text
No subsequent editing allowed
Change Examples
Remove mail headers, ie;
/^From /,/^$/c\
the address specifies a
<Mail Headers Removed>
range of lines beginning
with a line that begins with
From until the first blank
/^From /,/^$/{
line.
s/^From //p
The first example replaces
all lines with a single
occurrence of <Mail Header
Removed>.
The second example
replaces each line with
<Mail Header Removed>
c\
<Mail Header Removed>
}
Using !
If an address is followed by an exclamation point
(!), the associated command is applied to all lines
that dont match the address or address range
Examples:
1,5!d would delete all lines except 1 through 5
/black/!s/cow/horse/ would substitute
horse for cow on all lines except those that
contained black
The brown cow -> The brown horse
The black cow -> The black cow
Transform
The Transform command (y) operates like tr, it
does a one-to-one or character-to-character
replacement
Transform accepts zero, one or two addresses
[address[,address]]y/abc/xyz/
every a within the specified address(es) is
transformed to an x. The same is true for b to y and
c to z
y/abcdefghijklmnopqrstuvwxyz/ABCDEFGHIJKLMNOPQR
STUVWXYZ/ changes all lower case characters on the
addressed line to upper case

If you only want to transform specific characters (or a
word) in the line, it is much more difficult and requires
use of the hold space
Quit
Quit causes sed to stop reading new input lines
and stop sending them to standard output
It takes at most a single line address
Once a line matching the address is reached, the
script will be terminated
This can be used to save time when you only want
to process some portion of the beginning of a file
Example: to print the first 100 lines of a file (like

head) use:
sed '100q' filename
sed will, by default, send the first 100 lines of
filename to standard output and then quit processing
Pattern and Hold spaces

Pattern space: Workspace or temporary
buffer where a single line of input is held
while the editing commands are applied
Hold space: Secondary temporary buffer
for temporary storage only
in
h, H, g, G, x
Pattern
Hold
out
Sed Advantages
Regular expressions
Fast
Concise
Sed Drawbacks
Hard to remember text from one line to
another
Not possible to go backward in the file
No way to do forward references
like /..../+1
No facilities to manipulate numbers
Cumbersome syntax
Course Outline
Introduction

UNIX utilities
Scripting languages
Programming tools
What is a shell?
The user interface to the operating system
Functionality:
Execute other programs
Manage files
Manage processes
Full programming language

A program like any other
This is why there are so many shells
Shell History
There are
many choices
for shells
Shell features
evolved as
UNIX grew
Most Commonly Used Shells

/bin/csh
/bin/tcsh
/bin/sh
/bin/ksh
/bin/bash
C shell
Enhanced C Shell
The Bourne Shell / POSIX shell
Korn shell
Korn shell clone, from GNU
Ways to use the shell

Interactively
When you log in, you interactively use the
shell
Scripting
A set of shell commands that constitute an
executable program
Review: UNIX Programs

Means of input:
Program arguments
[control information]
Environment variables
[state information]
Standard input [data]
Means of output:
Return status code [control information]
Standard out [data]
Standard error [error messages]
Shell Scripts
A shell script is a regular text file that
contains shell or UNIX commands
Before running it, it must have execute
permission:
chmod +x filename
A script can be invoked as:

sh name [ arg ]
sh < name [ args ]
name [ arg ]
Shell Scripts
When a script is run, the kernel determines
which shell it is written for by examining the first
line of the script
If 1st line starts with #!pathname-ofshell, then it invokes pathname and sends
the script as an argument to be interpreted
If #! is not specified, the current shell
assumes it is a script in its own language
leads to problems
Simple Example
#!/bin/sh
echo Hello World
Scripting vs. C Programming

Advantages of shell scripts
Easy to work with other programs
Easy to work with files
Easy to work with strings
Great for prototyping. No compilation
Disadvantages of shell scripts

Slower
Not well suited for algorithms & data
structures
The C Shell
C-like syntax (uses { }'s)
Inadequate for scripting
Poor control over file descriptors

Difficult quoting "I say \"hello\"" doesn't work
Can only trap SIGINT
Can't mix flow control and commands
Survives mostly because of interactive features.

Job control
Command history
Command line editing, with arrow keys (tcsh)
http://www.faqs.org/faqs/unix-faq/shell/csh-whynot
The Bourne Shell
Slight differences on various systems

Evolved into standardized POSIX shell
Scripts will also run with ksh, bash
Influenced by ALGOL
Simple Commands
simple command: sequence of non blanks
arguments separated by blanks or tabs.
1st argument (numbered zero) usually specifies
the name of the command to be executed.
Any remaining arguments:
Are passed as arguments to that command.
Arguments may be filenames, pathnames, directories
or special options
ls l /
/bin/ls
-l
/
Background Commands
Any command ending with "&" is run in the
background.
firefox &
wait will block until the command finishes
Complex Commands
The shell's power is in its ability to hook
commands together
We've seen one example of this so far
with pipelines:
cut d: -f2 /etc/passwd | sort | uniq
We will see others
Redirection of input/ouput
Redirection of output: >
example:$ ls -l > my_files
Redirection of input: <

example: $ cat <input.data
Append output: >>

example: $ date >> logfile
Arbitrary file descriptor redirection: fd>

example: $ ls l 2> error_log
Multiple Redirection
cmd 2>file
send standard error to file
standard output remains the same
cmd > file 2>&1

send both standard error and standard output to
file
cmd > file1 2>file2

send standard output to file1
send standard error to file2
Here Documents
Shell provides alternative ways of supplying
standard input to commands (an anonymous file)
Shell allows in-line input redirection using <<
called here documents
Syntax:
command [arg(s)] << arbitrary-delimiter
command input
:
:
arbitrary-delimiter
arbitrary-delimiter should be a string that does
not appear in text
Here Document Example

#!/bin/sh
mail steinbrenner@yankees.com <<EOT
Sorry, I really blew it this
year. Thanks for not firing me.
Yours,
Joe
EOT
Shell Variables
To set:
name=value
Read: $var
Variables can be local or environment.
Environment variables are part of UNIX
and can be accessed by child processes.
Turn local variable into environment:
export variable
Variable Example
#!/bin/sh
MESSAGE="Hello World"
echo $MESSAGE
Environmental Variables
NAMEMEANING
$HOME Absolute pathname of your home
directory
$PATH A list of directories to search for
$MAIL Absolute pathname to mailbox
$USER Your login name
$SHELL Absolute pathname of login shell
$TERM Type of your terminal
$PS1
Prompt
Here Documents Expand Vars

#!/bin/sh
mail steinbrenner@yankees.com <<EOT
Sorry, I really blew it this
year. Thanks for not firing me.
Yours,
$USERS
EOT
Parameters
A parameter is one of the following:
A variable
A positional parameter, starting from 1
A special parameter
To get the value of a parameter: ${param}
Can be part of a word (abc${foo}def)
Works within double quotes
The {} can be omitted for simple variables,

special parameters, and single digit positional
parameters.
Positional Parameters
The arguments to a shell script
$1, $2, $3
The arguments to a shell function

Arguments to the set built-in command
set this is a test
$1=this, $2=is, $3=a, $4=test
Manipulated with shift

shift 2
$1=a, $2=test
Parameter 0 is the name of the shell or the shell

script.
Example with Parameters

#!/bin/sh
# Parameter 1: word
# Parameter 2: file
grep $1 $2 | wc l
$ countlines ing /usr/dict/words
3277
Special Parameters
$#
$$?
$$
$!
process
$*
"$@"
Number of positional parameters

Options currently in effect
Exit value of last executed command
Process number of current process
Process number of background
All arguments on command line
All arguments on command line
individually quoted "$1" "$2" ...
Command Substitution
Used to turn the output of a command into a string
Used to create arguments or variables
Command is placed with grave accents ` ` to
capture the output of command
$ date
Wed Sep 25 14:40:56 EDT 2001
$ NOW=`date`
$ grep `generate_regexp` myfile.c
$ sed "s/oldtext/`ls | head -1`/g"
$ PATH=`myscript`:$PATH
File name expansion

Used to generate a set of arguments from files
Wildcards (patterns)
* matches any string of characters
? matches any single character
[list] matches any character in list
[lower-upper] matches any character in range lowerupper inclusive
[!list] matches any character not in list
This is the same syntax that find uses
File Expansion
If multiple matches, all are returned $ /bin/ls
and treated as separate arguments: file1 file2
Handled by the shell (programs

argv[0]: /bin/cat
argv[1]: file1
argv[2]: file2
NOT
$ cat file1
a
$ cat file2
b
$ cat file*
a
dont bsee the wildcards)
argv[0]: /bin/cat
argv[1]: file*
Compound Commands
Multiple commands
Separated by semicolon or newline
Command groupings
pipelines
Subshell
( command1; command2 ) > file
Boolean operators
Control structures
Boolean Operators
Exit value of a program (exit system call) is a
number
0 means success
anything else is a failure code
cmd1 && cmd2

executes cmd2 if cmd1 is successful
cmd1 || cmd2
executes
cmd2 if cmd1
is not successful
$ ls bad_file
> /dev/null
&& date
$ ls bad_file > /dev/null || date
Wed Sep 26 07:43:23 2006
Control Structures
if expression
then
command1
else
command2
fi
What is an expression?
Any UNIX command. Evaluates to true if the
exit code is 0, false if the exit code > 0
Special command /bin/test exists that does
most common expressions
String compare
Numeric comparison
Check file properties
[ often a builtin version of /bin/test for

syntactic sugar
Good example UNIX tools working together
Examples
if test "$USER" = "kornj"
then
echo "I know you"
else
echo "I dont know you"
fi
if [ -f /tmp/stuff ] && [ `wc l < /tmp/stuff` -gt 10 ]
then
echo "The file has more than 10 lines in it"
else
echo "The file is nonexistent or small"
fi
test Summary
String based tests

-z string
-n string
string1 = string2
string1 != string2
string
Numeric tests
int1 eq int2
int1 ne int2
-gt, -ge, -lt, -le
File tests
-r
-w
-f
-d
-s
file
file
file
file
file
Logic
!
-a, -o
( expr )
Length of string is 0
Length of string is not 0
Strings are identical
Strings differ
String is not NULL
First int equal to second
First int not equal to second
greater, greater/equal, less, less/equal
File exists and is readable
File exists and is writable
File is regular file
File is directory
file exists and is not empty
Negate result of expression
and operator, or operator
groups an expression
Arithmetic
No arithmetic built in to /bin/sh
Use external command /bin/expr
expr expression
Evaluates expression and sends the result to
standard output.
Yields a numeric or string result
expr 4 "*" 12
Particularly
expr "("
4 + with
3 ")"
"*" 2
useful
command
substitution
X=èxpr $X + 2`
Control Structures Summary

if then fi
while done
until do done
for do done
case in esac
for loops
Different than C:
for var in list
do
command
done
Typically used with positional parameters or a

list of files:
sum=0
for var in "$@"
do
sum=èxpr $sum + $var`
done
echo The sum is $sum
for file in *.c ; do echo "We have $file"
done
Case statement
Like a C switch statement for strings:
case $var in
opt1) command1
command2
;;
opt2) command
;;
*)
command
;;
esac
* is a catch all condition
Case Example
#!/bin/sh
for INPUT in "$@"
do
case $INPUT in
hello)
echo "Hello there."
;;
bye)
echo "See ya later."
;;
*)
echo "I'm sorry?"
;;
esac
done
echo "Take care."
Case Options
opt can be a shell pattern, or a list of shell
patterns delimited by |
Example:
case $name in
*[0-9]*)
echo "That doesn't seem like a name."
;;
J*|K*)
echo "Your name starts with J or K, cool."
;;
*)
echo "You're not special."
;;
esac
Types of Commands
All behave the same way
Programs
Most that are part of the OS in /bin
Built-in commands
Functions
Aliases
Built-in Commands
Built-in commands are internal to the shell and
do not create a separate process. Commands
are built-in because:
They are intrinsic to the language (exit)
They produce side effects on the current process (cd)
They perform faster
No fork/exec
Special built-ins
: . break continue eval exec export exit
readonly return set shift trap unset
Important Built-in Commands

exec
cd :
shift
set:
wait
umask
exit
eval
time
export
trap
:
replaces shell with program
change working directory
:
rearrange positional parameters
set positional parameters
:
wait for background proc. to exit
:
change default file permissions
:
quit the shell
:
parse and execute string
:
run command and print times
:
put variable into environment
:
set signal handlers
Important Built-in Commands

continue
break
return
:
.
:
:
:
:
:
continue in loop
break in loop
return from function
true
read file of commands into
current shell; like #include
Functions
Functions are similar to scripts and other
commands except:
They can produce side effects in the callers script.
Variables are shared between caller and callee.
The positional parameters are saved and restored
when invoking a function.
Syntax:
name ()
{
commands
}
Aliases
Like macros (#define in C)
Shorter to define than functions, but more
limited
Not recommended for scripts
Example:
alias rm='rm i'
Command Search Rules

Special built-ins
Functions
command bypasses search for functions
Built-ins not associated with PATH

PATH search
Built-ins associated with PATH
Parsing and Quoting
How the Shell Parses

Part 1: Read the command:
Read one or more lines a needed
Separate into tokens using space/tabs
Form commands based on token types
Part 2: Evaluate a command:

Expand word tokens (command substitution, parameter
expansion)
Split words into fields

File expansion
Setup redirections, environment
Run command with arguments
Useful Program for Testing

/home/unixtool/bin/showargs
#include <stdio.h>
int main(int argc, char *argv[])
{
int i;
for (i=0; i < argc; i++) {
printf("Arg %d: %s\n", i, argv[i]);
}
return(0);
}
Shell Comments
Comments begin with an unquoted #

Comments end at the end of the line
Comments can begin whenever a token begins
Examples
# This is a comment
# and so is this
grep foo bar # this is a comment
grep foo bar# this is not a comment
Special Characters
The shell processes the following characters specially
unless quoted:
| & ( ) < > ; " ' $ ` space tab newline
The following are special whenever patterns are

processed:
* ? [ ]
The following are special at the beginning of a word:

# ~
The following is special when processing assignments:

=
Token Types
The shell uses spaces and tabs to split the
line or lines into the following types of
tokens:
Control operators (||)
Redirection operators (<)
Reserved words (if)
Assignment tokens
Word tokens
Operator Tokens
Operator tokens are recognized everywhere
unless quoted. Spaces are optional before and
after operator tokens.
I/O Redirection Operators:
> >> >| >& < << <<- <&
Each I/O operator can be immediately preceded by a
single digit
Control Operators:
| & ; ( ) || && ;;
Shell Quoting
Quoting causes characters to loose special
meaning.
\ Unless quoted, \ causes next character to
be quoted. In front of new-line causes lines to
be joined.
''
Literal quotes. Cannot contain '
""
Removes special meaning of all
characters except $, ", \ and `. The \ is only
special before one of these characters and newline.
Quoting Examples
$ cat file*
a
b
$ cat "file*"
cat: file* not found
$ cat file1 > /dev/null
$ cat file1 ">" /dev/null
a
cat: >: cannot open
FILES="file1 file2"
$ cat "$FILES"
cat: file1 file2 not found
Simple Commands
A simple command consists of three types of
tokens:
Assignments (must come first)

Command word tokens
Redirections: redirection-op + word-op
The first token must not be a reserved word
Command terminated by new-line or ;
Example:
foo=bar z=`date`
echo $HOME
x=foobar > q$$ $xyz z=3
Word Splitting
After parameter expansion, command
substitution, and arithmetic expansion,
the characters that are generated as a
result of these expansions that are not
inside double quotes are checked for split
characters
Default split character is space or tab
Split characters are defined by the value
of the IFS variable (IFS="" disables)
Word Splitting Examples

FILES="file1 file2"
cat $FILES
a
b
IFS=
cat $FILES
cat: file1 file2: cannot open
IFS=x v=exit
echo exit $v "$v"
exit e it exit
Pathname Expansion
After word splitting, each field that
contains pattern characters is replaced by
the pathnames that match
Quoting prevents expansion
set o noglob disables
Not in original Bourne shell, but in POSIX
Parsing Example
DATE=`date` echo $foo > \
/dev/null
DATE=`date` echo $foo > /dev/null
assignment
word
param
redirection
echo hello there

/bin/echo
PATH expansion
hello there
split by IFS
/dev/null
/dev/null
Script Examples
Rename files to lower case
Strip CR from files
Emit HTML for directory contents
Rename files
#!/bin/sh
for file in *
do
lfile=ècho $file | tr A-Z a-z`
if [ $file != $lfile ]
then
mv $file $lfile
fi
done
Remove DOS Carriage Returns

#!/bin/sh
TMPFILE=/tmp/file$$
if [ "$1" = "" ]
then
tr -d '\r'
exit 0
fi
trap 'rm -f $TMPFILE' 1 2 3 6 15
for file in "$@"
do
if tr -d '\r' < $file > $TMPFILE
then
mv $TMPFILE $file
fi
done
Generate HTML
$ dir2html.sh > dir.html
The Script
#!/bin/sh
[ "$1" != "" ] && cd "$1"
cat <<HUP
<html>
<h1> Directory listing for $PWD </h1>
<table border=1>
<tr>
HUP
num=0
for file in *
do
genhtml $file
# this function is on next
page
done
cat <<HUP
</tr>
</table>
</html>
Function genhtml
genhtml()
{
file=$1
echo "<td><tt>"
if [ -f $file ]
then
echo "<font color=blue>$file</font>"
elif [ -d $file ]
then
echo "<font color=red>$file</font>"
else
echo "$file"
fi
echo "</tt></td>"
num=èxpr $num + 1`
if [ $num -gt 4 ]
then
echo "</tr><tr>"
num=0
fi
}
Korn Shell / bash Features
Better syntax with $(command)
Allows nesting
x=$(cat $(generate_file_list))
Backward compatible with ` ` notation
Expressions
Expressions are built-in with the [[ ]] operator
if [[ $var = "" ]]
Gets around parsing quirks of /bin/test, allows checking

strings against patterns
Operations:
string == pattern
string != pattern
string1 < string2
file1 nt file2
file1 ot file2
file1 ef file2
&&, ||
Patterns
Can be used to do string matching:
if [[ $foo = *a* ]]
if [[ $foo = [abc]* ]]
Similar to regular expressions, but

different syntax
Additonal Parameter Expansion
${#param} Length of param

${param#pattern} Left strip min pattern
${param##pattern} Left strip max pattern
${param%pattern} Right strip min pattern
${param%%pattern} Right strip max pattern
${param-value} Default value if param not
set
Variables
Variables can be arrays
foo[3]=test
echo ${foo[3]}
Indexed by number
${#arr} is length of the array
Multiple array elements can be set at once:
set A foo a b c d
echo ${foo[1]}
Set command can also be used for positional
params: set a b c d; print $2
Printing
Built-in print command to replace echo
Much faster
Allows options:
-u#
print to specific file descriptor
Functions
Alternative function syntax:
function name {
commands
}
Allows for local variables

$0 is set to the name of the function
Additional Features
Built-in arithmetic: Using $((expression ))
e.g., print $(( 1 + 1 * 8 / x ))
Tilde file expansion

~
~user
~+
~-
$HOME
home directory of user
$PWD
$OLDPWD
Course Outline
Introduction

UNIX utilities
Scripting languages
Programming tools
Parsing and Quoting
Shell Quoting
Quoting causes characters to loose special
meaning.
\
Unless quoted, \ causes next character

to be quoted. In front of new-line
causes lines to be joined.
''
Literal quotes. Cannot contain '
""
Removes special meaning of all
characters except $, ", \ and `.
The \ is only special before one of
these characters and new-line.
Quoting Examples
$ cat file*
a
b
$ cat "file*"
cat: file* not found
$ cat file1 > /dev/null
$ cat file1 ">" /dev/null
a
cat: >: cannot open
FILES="file1 file2"
$ cat "$FILES"
cat: file1 file2 not found
Shell Comments
Comments begin with an unquoted #

Comments end at the end of the line
Comments can begin whenever a token begins
Examples
# This is a comment
# and so is this
grep foo bar # this is a comment
grep foo bar# this is not a comment
How the Shell Parses

Part 1: Read the command:
Read one or more lines a needed
Separate into tokens using space/tabs
Form commands based on token types
Part 2: Evaluate a command:

Expand word tokens (command substitution, parameter
expansion)
Split words into fields

File expansion
Setup redirections, environment
Run command with arguments
Useful Program for Testing

/home/unixtool/bin/showargs
#include <stdio.h>
int main(int argc, char *argv[])
{
int i;
for (i=0; i < argc; i++) {
printf("Arg %d: %s\n", i, argv[i]);
}
return(0);
}
Special Characters
The shell processes the following characters specially
unless quoted:
| & ( ) < > ; " ' $ ` space tab newline
The following are special whenever patterns are

processed:
* ? [ ]
The following are special at the beginning of a word:

# ~
The following is special when processing assignments:

=
Token Types
The shell uses spaces and tabs to split the
line or lines into the following types of
tokens:
Control operators (||)
Redirection operators (<)
Reserved words (if)
Assignment tokens
Word tokens
Operator Tokens
Operator tokens are recognized everywhere
unless quoted. Spaces are optional before and
after operator tokens.
I/O Redirection Operators:
> >> >| >& < << <<- <&
Each I/O operator can be immediately preceded by a
single digit
Control Operators:
| & ; ( ) || && ;;
Simple Commands
A simple command consists of three types of
tokens:
Assignments (must come first)
Command word tokens
Redirections: redirection-op + word-op
The first token must not be a reserved word

Command terminated by new-line or ;
Examples:
foo=bar z=`date`
echo $HOME
x=foobar > q$$ $xyz z=3
Word Splitting
After parameter expansion, command
substitution, and arithmetic expansion,
the characters that are generated as a
result of these expansions that are not
inside double quotes are checked for split
characters
Default split character is space or tab
Split characters are defined by the value
of the IFS variable (IFS="" disables)
Word Splitting Examples

FILES="file1 file2"
cat $FILES
a
b
IFS=
cat $FILES
cat: file1 file2: cannot open
IFS=x v=exit
echo exit $v "$v"
exit e it exit
Pathname Expansion
After word splitting, each field that
contains pattern characters is replaced by
the pathnames that match
Quoting prevents expansion
set o noglob disables
Not in original Bourne shell, but in POSIX
Parsing Example
DATE=`date` echo $foo > \
/dev/null
DATE=`date` echo $foo > /dev/null
assignment
word
param
redirection
echo hello there

/bin/echo
PATH expansion
hello there
split by IFS
/dev/null
/dev/null
The eval built-in

eval arg
Causes all the tokenizing and expansions to
be performed again
trap command
trap specifies command that should be evaled
when the shell receives a signal of a particular
value.
trap [ [command] {signal}+]
If command is omitted, signals are ignored
Especially useful for cleaning up temporary files
trap 'echo "please, dont interrupt!"' SIGINT
trap 'rm /tmp/tmpfile' EXIT
Reading Lines
read is used to read a line from a file and
to store the result into shell variables
read r prevents special processing
Uses IFS to split into words
If no variable specified, uses REPLY
read
read r NAME
read FIRSTNAME LASTNAME
Script Examples
Rename files to lower case
Strip CR from files
Emit HTML for directory contents
Rename files
#!/bin/sh
for file in *
do
lfile=ècho $file | tr A-Z a-z`
if [ $file != $lfile ]
then
mv $file $lfile
fi
done
Remove DOS Carriage Returns

#!/bin/sh
TMPFILE=/tmp/file$$
if [ "$1" = "" ]
then
tr -d '\r'
exit 0
fi
trap 'rm -f $TMPFILE' 1 2 3 6 15
for file in "$@"
do
if tr -d '\r' < $file > $TMPFILE
then
mv $TMPFILE $file
fi
done
Generate HTML
$ dir2html.sh > dir.html
The Script
#!/bin/sh
[ "$1" != "" ] && cd "$1"
cat <<HUP
<html>
<h1> Directory listing for $PWD </h1>
<table border=1>
<tr>
HUP
num=0
for file in *
do
genhtml $file
# this function is on next
page
done
cat <<HUP
</tr>
</table>
</html>
Function genhtml
genhtml()
{
file=$1
echo "<td><tt>"
if [ -f $file ]
then
echo "<font color=blue>$file</font>"
elif [ -d $file ]
then
echo "<font color=red>$file</font>"
else
echo "$file"
fi
echo "</tt></td>"
num=èxpr $num + 1`
if [ $num -gt 4 ]
then
echo "</tr><tr>"
num=0
fi
}
Korn Shell / bash Features
Better syntax with $(command)
Allows nesting
x=$(cat $(generate_file_list))
Backward compatible with ` ` notation
Expressions
Expressions are built-in with the [[ ]] operator
if [[ $var = "" ]]
Gets around parsing quirks of /bin/test, allows checking

strings against patterns
Operations:
string == pattern
string != pattern
string1 < string2
file1 nt file2
file1 ot file2
file1 ef file2
&&, ||
Patterns
Can be used to do string matching:
if [[ $foo = *a* ]]
if [[ $foo = [abc]* ]]
Note: patterns are like a subset of regular

expressions, but different syntax
Additonal Parameter Expansion
${#param} Length of param

${param#pattern} Left strip min pattern
${param##pattern} Left strip max pattern
${param%pattern} Right strip min pattern
${param%%pattern} Right strip max pattern
${param-value} Default value if param not
set
Variables
Variables can be arrays
foo[3]=test
echo ${foo[3]}
Indexed by number
${#arr} is length of the array
Multiple array elements can be set at once:
set A foo a b c d
echo ${foo[1]}
Set command can also be used for positional
params: set a b c d; print $2
Functions
Alternative function syntax:
function name {
commands
}
Allows for local variables

$0 is set to the name of the function
Additional Features
Built-in arithmetic: Using $((expression ))
e.g., print $(( 1 + 1 * 8 / x ))
Tilde file expansion

~
~user
~+
~-
$HOME
home directory of user
$PWD
$OLDPWD
KornShell 93
Variable Attributes
By default attributes hold strings of unlimited length
Attributes can be set with typeset:
readonly (-r) cannot be changed

export (-x) value will be exported to env
upper (-u) letters will be converted to upper case
lower (-l) letters will be converted to lower case
ljust (-L width) left justify to given width
rjust (-R width) right justify to given width
zfill (-Z width) justify, fill with leading zeros
integer (-I [base]) value stored as integer
float (-E [prec]) value stored as C double

nameref (-n) a name reference
Name References
A name reference is a type of variable that
references another variable.
nameref is an alias for typeset -n
Example:
user1="jeff"
user2="adam"
typeset n name="user1"
print $name
jeff
New Parameter Expansion

${param/pattern/str} Replace first pattern
with str
${param//pattern/str} Replace all
patterns with str
${param:offset:len} Substring with offset
Patterns Extended
Additional pattern
types so that shell
patterns are
equally
expressive as
regular
expressions
Used for:
file expansion
[[ ]]
case statements
parameter
expansion
Patterns
Regular Expressions
ANSI C Quoting
$'' Uses C escape sequences
$'\t'
$'Hello\nthere'
printf added that supports C like printing:

printf "You have %d apples" $x
Extensions
%b ANSI escape sequences

%q Quote argument for reinput
\E Escape character (033)
%P convert ERE to shell pattern
%H convert using HTML conventions
%T date conversions using date formats
Associative Arrays
Arrays can be indexed by string

Declared with typeset A
Set: name["foo"]="bar"
Reference ${name["foo"]}
Subscripts: ${!name[@]}
Networking, HTTP, CGI
Network Application
Client application and server application
communicate via a network protocol
A protocol is a set of rules on how the client and
server communicate
web
client
HTTP
web
server
kernel
user
TCP/IP Suite
application
layer
client
TCP/UDP
transport
layer
server
TCP/UDP
internet
layer
IP
drivers/
hardware
IP
drivers/
hardware
network access layer
(ethernet)
Data Encapsulation
Data
ApplicationLayer
H1
Data
H2
H1
Data
H2
H1
Data
TransportLayer
InternetLayer
Network
Access
Layer
H3
Network Access/Internet Layers

Network Access Layer
Deliver data to devices on the same physical network
Ethernet
Internet Layer
Internet Protocol (IP)

Determines routing of datagram
IPv4 uses 32-bit addresses (e.g. 128.122.20.15)
Datagram fragmentation and reassembly
Transport Layer
Transport Layer
Host-host layer
Provides error-free, point-to-point connection between
hosts
User Datagram Protocol (UDP)

Unreliable, connectionless
Transmission Control Protocol (TCP)

Reliable, connection-oriented
Acknowledgements, sequencing, retransmission
Ports
Both TCP and UDP use 16-bit port numbers
A server application listen to a specific port for
connections
Ports used by popular applications are well-defined
SSH (22), SMTP (25), HTTP (80)
1-1023 are reserved (well-known)
Clients use ephemeral ports (OS dependent)
Name Service
Every node on the network normally has a
hostname in addition to an IP address
Domain Name System (DNS) maps IP
addresses to names
e.g. 128.122.81.155 is access1.cims.nyu.edu
DNS lookup utilities: nslookup, dig

Local name address mappings stored
in /etc/hosts
Sockets
Sockets provide access to TCP/IP on UNIX
systems
Sockets are communications endpoints
Invented in Berkeley UNIX
Allows a network connection to be opened as a
file (returns a file descriptor)
machine 1
machine 2
Major Network Services

Telnet (Port 23)
Provides virtual terminal for remote user
The telnet program can also be used to connect to
other ports
FTP (Port 20/21)

Used to transfer files from one machine to another
Uses port 20 for data, 21 for control
SSH (Port 22)

For logging in and executing commands on
remote machines
Data is encrypted
Major Network Services cont.

SMTP (Port 25)
Host-to-host mail transport
Used by mail transfer agents (MTAs)
IMAP (Port 143)

Allow clients to access and manipulate emails
on the server
HTTP (Port 80)

Protocol for WWW
Ksh93: /dev/tcp
Files in the form
/dev/tcp/hostname/port result in a
socket connection to the given service:

exec 3<>/dev/tcp/smtp.cs.nyu.edu/25 #SMTP
print u3 EHLO cs.nyu.edu"
print u3 QUIT"
while IFS= read u3
do
print r "$REPLY"
done
HTTP
Hypertext Transfer Protocol
Use port 80
Language used by web browsers (IE,

Netscape, Firefox) to communicate with
web servers (Apache, IIS)
QuickTime and a
TIFF (LZW) decompressor
are needed to see this picture.
HTTPrequest:
Getmethisdocument
HTTPresponse:
Hereisyourdocument
Resources
Web servers host web resources, including
HTML files, PDF files, GIF files, MPEG movies,
etc.
Each web object has an associated MIME type
HTML document has type text/html
JPEG image has type image/jpeg
Web resource is accessed using a Uniform

Resource Locator (URL)
http://www.cs.nyu.edu:80/courses/fall06/G22.2245-001/index.html
protocol
host
port
resource
HTTP Transactions
HTTP request to web server
GET/v40images/nyu.gifHTTP/1.1
Host:www.nyu.edu
HTTP response to web client

HTTP/1.1200OK
Contenttype:image/gif
Contentlength:3210
Sample HTTP Session

GET/HTTP/1.1
HOST:www.cs.nyu.edu
request
HTTP/1.1200OK
Date:Wed,19Oct200506:59:49GMT
Server:Apache/2.0.49(Unix)mod_perl/1.99_14Perl/v5.8.4
mod_ssl/2.0.49OpenSSL/0.9.7emod_auth_kerb/4.13PHP/5.0.0RC3
LastModified:Thu,12Sep200217:09:03GMT
ContentLength:163
ContentType:text/html;charset=ISO88591
response
<!DOCTYPEHTMLPUBLIC"//IETF//DTDHTML//EN">
<html>
<head>
<title></title>
<metaHTTPEQUIV="Refresh"CONTENT="0;URL=csweb/index.html">
<body>
</body>
</html>
Status Codes
Status code in the HTTP response indicates if a
request is successful
Some typical status codes:
200
302
OK
Found; Resource in different
URI
401
403
404
Authorization required
Forbidden
Not Found
Gateways
Interface between resource and a web
server
QuickTime and a
http WebServer
Gateway
resource
CGI
Common Gateway Interface is a standard interface for
running helper applications to generate dynamic
contents
Specify the encoding of data passed to programs
Allow HTML documents to be created on the fly

Transparent to clients
Client sends regular HTTP request
Web server receives HTTP request, runs CGI program, and
sends contents back in HTTP responses
CGI programs can be written in any language
CGI Diagram
HTTPrequest
QuickTime and a
WebServer
HTTPresponse
spawnprocess
Document
Script
HTML
Document format used on the web
<html>
<head>
<title>SomeDocument</title>
</head>
<body>
<h2>SomeTopics</h2>
ThisisanHTMLdocument
<p>
Thisisanotherparagraph
</body>
</html>
HTML
HTML is a file format that describes a web
page.
These files can be made by hand, or
generated by a program
A good way to generate an HTML file is by
writing a shell script
Forms
HTML forms are used to collect user input
Data sent via HTTP request
Server launches CGI script to process data
<formmethod=POST
action=http://www.cs.nyu.edu/~unixtool/cgi
bin/search.cgi>
Enteryourquery:<inputtype=textname=Search>
<inputtype=submit>
</form>
Input Types
Text Field
<inputtype=textname=zipcode>
Radio Buttons
<inputtype=radioname=sizevalue=S>Small
<inputtype=radioname=sizevalue=M>Medium
<inputtype=radioname=sizevalue=L>Large
Checkboxes
<inputtype=checkboxname=extrasvalue=lettuce>Lettuce
<inputtype=checkboxname=extrasvalue=tomato>Tomato
Text Area
<textareaname=addresscols=50rows=4>
</textarea>
Submit Button
Submits the form for processing by the
CGI script specified in the form tag
<inputtype=submitvalue=SubmitOrder>
HTTP Methods
Determine how form data are sent to web
server
Two methods:
GET
Form variables stored in URL
POST
Form variables sent as content of HTTP request
Encoding Form Values

Browser sends form variable as name-value
pairs
name1=value1&name2=value2&name3=value3
Names are defined in form elements

<inputtype=textname=ssnmaxlength=9>
Special characters are replaced with %## (2digit hex number), spaces replaced with +
e.g. 10/20Wed is encoded as 10%2F20+Wed
GET/POST examples
GET:
GET/cgibin/myscript.pl?name=Bill%20Gates&
company=MicrosoftHTTP/1.1
HOST:www.cs.nyu.edu
POST:
POST/cgibin/myscript.plHTTP/1.1
HOST:www.cs.nyu.edu
other headers
name=Bill%20Gates&company=Microsoft
GET or POST?
GET method is useful for
Retrieving information, e.g. from a database
Embedding data in URL without form element
POST method should be used for forms with

Many fields or long fields
Sensitive information
Data for updating database
GET requests may be cached by clients

browsers or proxies, but not POST requests
Parsing Form Input

Method stored in HTTP_METHOD
GET: Data encoded into QUERY_STRING
POST: Data in standard input (from body
of request)
Most scripts parse input into an
associative array
You can parse it yourself
Or use available libraries (better)
CGI Environment Variables
DOCUMENT_ROOT
HTTP_HOST
HTTP_REFERER
HTTP_USER_AGENT
HTTP_COOKIE
REMOTE_ADDR
REMOTE_HOST
REMOTE_USER
REQUEST_METHOD
SERVER_NAME
SERVER_PORT
CGI Script: Example
Part 1: HTML Form

<html>
<center>
<H1>Anonymous Comment Submission</H1>
</center>
Please enter your comment below which will
be sent anonymously to <tt>kornj@cs.nyu.edu</tt>.
If you want to be extra cautious, access this
page through <a
href="http://www.anonymizer.com">Anonymizer</a>.
<p>
<form action=cgi-bin/comment.cgi method=post>
<textarea name=comment rows=20 cols=80>
</textarea>
<input type=submit value="Submit Comment">
</form>
</html>
Part 2: CGI Script (ksh)

#!/home/unixtool/bin/ksh
. cgi-lib.ksh
ReadParse
PrintHeader
# Read special functions to help parse
print -r -- "${Cgi.comment}" | /bin/mailx -s "COMMENT" kornj

print
print
print
print
"<H2>You submitted the comment</H2>"

"<pre>"
-r -- "${Cgi.comment}"
"</pre>"
Debugging
Debugging can be tricky, since error
messages don't always print well as HTML
One method: run interactively
$ QUERY_STRING='birthday=10/15/03'
$ ./birthday.cgi
Content-type: text/html
<html>
Your birthday is <tt>10/15/02</tt>.
</html>
How to get your script run

This can vary by web server type
http://www.cims.nyu.edu/systems/resources/webhosting/index.html
Typically, you give your script a name that

ends with .cgi
Give the script execute permission
Specify the location of that script in the
URL
CGI Security Risks

Sometimes CGI scripts run as owner of the
scripts
Never trust user input - sanity-check
everything
If a shell command contains user input, run
without shell escapes
Always encode sensitive information, e.g.
passwords
Also use HTTPS
Clean up - dont leave sensitive data around
CGI Benefits
Simple
Language independent
UNIX tools are good for this because
Work well with text
Integrate programs well
Easy to prototype
No compilation (CGI scripts)
Example: Find words in

Dictionary
<form action=dict.cgi>
Regular expression: <input type=entry
name=re value=".*">
<input type=submit>
</form>
Example: Find words in

Dictionary
PATH=$PATH:.
. cgi-lib.ksh
ReadParse
PrintHeader
print "<H1> Words matching <tt>${Cgi.re}</tt> in the dictionary
</H1>\n";
print "<OL>"
grep "${Cgi.re}" /usr/dict/words | while read word
do
print "<LI> $word"
done
print "</OL>"
What is Perl?
Practical Extraction and Report Language
Scripting language created by Larry Wall in the
mid-80s
Functionality and speed somewhere between
low-level languages (like C) and high-level
ones (like shell)
Influence from awk, sed, and C Shell
Easy to write (after you learn it), but sometimes
hard to read
Widely used in CGI scripting
A Simple Perl Script

turnsonwarnings
hello:
#!/usr/bin/perl -w
print Hello, world!\n;
$ chmod a+x hello

$ ./hello
Hello, world!
$ perl -e print Hello, world!\n;
Hello, world!
Another Perl Script

$;=$_;$/='0#](.+,a()$=(\}$+_c2$sdl[h*du,(1ri)b$2](n} /
1)1tfz),}0(o{=4s)1rs(2u;2(u",bw-2b $
hc7s"tlio,tx[{ls9r11$e(1(9]q($,$2)=)_5{4*s{[9$,lh$2,_.
(ia]7[11f=*2308t$$)]4,;d/
{}83f,)s,65o@*ui),rt$bn;5(=_stf*0l[t(o$.o$rsrt.c!(i([$a]
$n$2ql/d(l])t2,$.+{i)$_.$zm+n[6t(e1+26[$;)+]61_l*,*)],
(41${/@20)/z1_0+=)(2,,4c*2)\5,h$4;$91r_,pa,)$[4r)
$=_$6i}tc}!,n}[h$]$t 0rd)_$';open(eval$/);
$_=<0>;for($x=2;$x<666;$a.=++$x){s}{{.|.}};push@@,$&;
$x==5?$z=$a:++$}}for(++$/..substr($a,1885))
{$p+=7;$;.=$@[$p%substr($a,$!,3)+11]}eval$;
Data Types
Basic types: scalar, lists, hashes
Support OO programming and userdefined types
What Type?
Type of variable determined by special
leading character
$foo
scalar
@foo
list
%foo
hash
&foo
function
Data types have separate name spaces
Scalars
Can be numbers
$num = 100;
$num = 223.45;
$num = -1.3e38;
Can be strings
$str
$str
$str
$str
=
=
=
=
unix tools;
Who\s there?;
good evening\n;
one\ttwo;
Backslash escapes and variable names are interpreted

inside double quotes
Special Scalar Variables

$0
$_
$$
$?
$!
$/
$.
undef
Name of script
Default variable
Current PID
Status of last pipe or system call
System error message
Input record separator
Input record number
Acts like 0 or empty string
Operators
Numeric: + - * / % **
String concatenation: .
$state = New . York;
# NewYork
String repetition: x
print bla x 3;
# blablabla
Binary assignments:
$val = 2; $val *= 3;
$state .= City;
# $val is 6
# NewYorkCity
Comparison Operators
Comparison
Numeri
c
String
Equal
==
eq
Not Equal
!=
ne
Greater than
>
gt
Less than
<
lt
Less than or equal to
<=
le
Greater than or equal

to
>=
ge
Boolean Values
if ($ostype eq unix) { }
if ($val) { }
No boolean data type
undef is false
0 is false; Non-zero numbers are true
and 0 are false; other strings are true
The unary not (!) negates the boolean
value
undef and defined

$f = 1;
while ($n < 10) {
# $n is undef at 1st iteration
$f *= ++$n;
}
Use defined to check if a value is undef

if (defined($val)) { }
Lists and Arrays
List: ordered collection of scalars

Array: Variable containing a list
Each element is a scalar variable
Indices are integers starting at 0
Array/List Assignment
@teams=(Knicks,Nets,Lakers);
print $teams[0];
# print Knicks
$teams[3]=Celtics;# add new elt
@foo = ();
# empty list
@nums = (1..100);
# list of 1-100
@arr = ($x, $y*6);
($a, $b) = (apple, orange);
($a, $b) = ($b, $a); # swap $a $b
@arr1 = @arr2;
More About Arrays and Lists

Quoted words - qw
@planets = qw/ earth mars jupiter /;
@planets = qw{ earth mars jupiter };
Last elements index: $#planets

Not the same as number of elements in array!
Last element: $planets[-1]
Scalar and List Context

@colors = qw< red green blue >;
Array interpolated as string:
print My favorite colors are @colors\n;
Prints My favorite colors are red green blue
Array in scalar context returns the number of

elements in the list
$num = @colors + 5;
# $num gets 8
Scalar expression in list context

@num = 88;
# a one-element list (88)
pop and push

push and pop: arrays used as stacks
push adds element to end of array
@colors = qw# red green blue #;
push(@colors, yellow);
# same as
@colors = (@colors, yellow);
push @colors, @more_colors;
pop removes last element of array and returns

it
$lastcolor = pop(@colors);
shift and unshift

shift and unshift: similar to push and pop on
the left side of an array
unshift adds elements to the beginning
@colors = qw# red green blue #;
unshift @colors, orange;
First element is now orange
shift removes element from beginning

$c = shift(@colors); # $c gets orange
sort and reverse

reverse returns list with elements in reverse
order
@list1 = qw# NY NJ CT #;
@list2 = reverse(@list1); # (CT,NJ,NY)
sort returns list with elements in ASCII order

@day = qw/ tues wed thurs /;
@sorted = sort(@day); #(thurs,tues,wed)
@nums = sort 1..10; # 1 10 2 3 8 9
reverse and sort do not modify their

arguments
Iterate over a List

foreach loops through a list of values
@teams = qw# Knicks Nets Lakers #;
foreach $team (@teams) {
print $team win\n;
}
Value of control variable restored at end of loop

Synonym for the for keyword
$_ is the default
foreach (@teams) {
$_ .= win\n;
print; # print $_
}
Hashes
Associative arrays - indexed by strings (keys)
$cap{Hawaii} = Honolulu;
%cap = ( New York, Albany, New Jersey,
Trenton, Delaware, Dover );
Can use => (big arrow or comma arrow) in

place of , (comma)
%cap = ( New York
=> Albany,
New Jersey => Trenton,
Delaware
=> Dover );
Hash Element Access

$hash{$key}
print $cap{New York};
print $cap{New . York};
Unwinding the hash

@cap_arr = %cap;
Gets unordered list of key-value pairs
Assigning one hash to another

%cap2 = %cap;
%cap_of = reverse %cap;
print $cap_of{Trenton};
# New Jersey
Hash Functions
keys returns a list of keys
@state = keys %cap;
values returns a list of values

@city = values %cap;
Use each to iterate over all (key, value) pairs

while ( ($state, $city) = each %cap )
{
print Capital of $state is $city\n;
}
Hash Element Interpolation

Unlike a list, entire hash cannot be interpolated
print %cap\n;
Prints %cap followed by a newline
Individual elements can

foreach $state (sort keys %cap) {
print Capital of $state is
$cap{$state}\n;
}
More Hash Functions

exists checks if a hash element has ever been
initialized
print Exists\n if exists $cap{Utah};
Can be used for array elements
A hash or array element can only be defined if it
exists
delete removes a key from the hash

delete $cap{New York};
Merging Hashes
Method 1: Treat them as lists
%h3 = (%h1, %h2);
Method 2 (save memory): Build a new hash by

looping over all elements
%h3 = ();
while ((%k,$v) = each(%h1)) {
$h3{$k} = $v;
}
while ((%k,$v) = each(%h2)) {
$h3{$k} = $v;
}
Subroutines
sub myfunc { }
$name=Jane;
sub print_hello {
print Hello $name\n; # global $name
}
&print_hello;
# print Hello Jane
print_hello;
# print Hello Jane
print_hello();
# print Hello Jane
Arguments
Parameters are assigned to the special array @_
Individual parameter can be accessed as $_[0],
$_[1],
sub sum {
my $x;
# private variable $x
foreach (@_) { # iterate over params
$x += $_;
}
return $x;
}
$n = &sum(3, 10, 22); # n gets 35
More on Parameter Passing

Any number of scalars, lists, and hashes can be
passed to a subroutine
Lists and hashes are flattened
func($x, @y, %z);
Inside func:
$_[0] is $x
$_[1] is $y[0]
$_[2] is $y[1], etc.
Scalars in @_ are implicit aliases (not copies) of

the ones passed changing values of $_[0],
etc. changes the original variables
Return Values
The return value of a subroutine is the last
expression evaluated, or the value returned by
the return operator
sub myfunc {
my $x = 1;
$x + 2; #returns 3
}
sub myfunc {
my $x = 1;
return $x + 2;
} @somelist;
Can also return a list: return
If return is used without an expression (failure),

undef or () is returned depending on context
Lexical Variables
Variables can be scoped to the enclosing block
with the my operator
sub myfunc {
my $x;
my($a, $b) = @_;
# copy params
Can be used in any block, such as if block or

while block
Without enclosing block, the scope is the source file
use strict
The use strict pragma enforces some
good programming rules
All new variables need to be declared with
my
#!/usr/bin/perl -w
use strict;
$n = 1;
# <-- perl will complain
Another Subroutine Example

@nums = (1, 2, 3);
$num = 4;
@res = dec_by_one(@nums, $num);
minus_one(@nums, $num);
sub dec_by_one {
my @ret = @_;
for my $n (@ret) { $n-- }
return @ret;
}
sub minus_one {
for (@_) { $_-- }
}
# @res=(0, 1, 2, 3)
# (@nums,$num)=(1, 2, 3, 4)
# (@nums,$num)=(0, 1, 2, 3)
# make a copy
Reading from STDIN

STDIN is the builtin filehandle to the std input
Use the line input operator around a file handle
to read from it
$line = <STDIN>;
chomp($line);
# read next line
chomp removes trailing string that corresponds

to the value of $/ (usually the newline character)
Reading from STDIN example

while (<STDIN>) {
chomp;
print Line $. ==> $_\n;
}
Line 1 ==> [Contents of line 1]
Line 2 ==> [Contents of line 2]
<>
Diamond operator < > helps Perl programs
behave like standard Unix utilities (cut, sed, )
Lines are read from list of files given as command
line arguments (@ARGV), otherwise from stdin
while (<>) {
chomp;
print Line $. from $ARGV is $_\n;
}
./myprog file1 file2 Read from file1, then file2, then standard input
$ARGV is the current filename
Filehandles
Use open to open a file for reading/writing
open
open
open
open
LOG,
LOG,
LOG,
LOG,
syslog;
<syslog;
>syslog;
>>syslog;
# read
# read
# write
# append
When youre done with a filehandle, close

it
close LOG;
Errors
When a fatal error is encountered, use die to
print out error message and exit program
die Something bad happened\n if .;
Always check return value of open
open LOG, >>syslog
or die Cannot open log: $!;
For non-fatal errors, use warn instead

warn Temperature is below 0!
if $temp < 0;
Reading from a File

open MSG, /var/log/messages
or die Cannot open messages: $!\n;
while (<MSG>) {
chomp;
# do something with $_
}
close MSG;
Reading Whole File

In scalar context, <FH> reads the next line
$line = <LOG>;
In list context, <FH> read all remaining lines

@lines = <LOG>;
Undefine $/ to read the rest of file as a string

undef $/;
$all_lines = <LOG>;
Writing to a File
open LOG, >/tmp/log
or die Cannot create log: $!;
print LOG Some log messages\n
printf LOG %d entries processed.\n,
$num;
close LOG;
no comma after filehandle
File Tests examples

die The file $filename is not readable
if ! -r $filename;
warn The file $filename is not owned by
you unless -o $filename;
print This file is old if -M $filename >
365;
File Tests list

-r
-w
-x
-o
File or directory is readable

File or directory is writable
File or directory is executable
File or directory is owned by this
user
-e
-z
-s
File or directory exists

File exists and has zero size
File or directory exists and has
nonzero size (value in bytes)
File Tests list

-f
-d
-l
-M
-A
Entry if a plain file

Entry is a directory
Entry is a symbolic link
Modification age (in days)
Access age (in days)
$_isthedefaultoperand
Manipulating Files and Dirs
unlink removes files

unlink file1, file2
or warn failed to remove file: $!;
rename renames a file

rename file1, file2;
link creates a new (hard) link

link file1, file2
or warn cant create link: $!;
symlink creates a soft link

link file1, file2 or warn ;
Manipulating Files and Dirs cont.

mkdir creates directory
mkdir mydir, 0755
or warn Cannot create mydir: $!;
rmdir removes empty directories
rmdir dir1, dir2, dir3;
chmod modifies permissions on file or directory
chmod 0600, file1, file2;
if - elsif - else
if elsif else
if ( $x
print
}
elsif (
print
}
else {
print
}
> 0 ) {
x is positive\n;
$x < 0 ) {
x is negative\n;
x is zero\n;
unless
Like the opposite of if
unless ($x < 0) {
print $x is non-negative\n;
}
unlink $file unless -A $file < 100;
while and until

while ($x < 100) {
$y += $x++;
}
until is like the opposite of while
until ($x >= 100) {
$y += $x++;
}
for
for (init; test; incr) { }
# sum of squares of 1 to 5
for ($i = 1; $i <= 5; $i++) {
$sum += $i*$i;
}
next
next skips the remaining of the current
iteration (like continue in C)
# only print non-blank lines
while (<>) {
if ( $_ eq \n) { next; }
else { print; }
}
last
last exits loop immediately (like break in
C)
# print up to first blank line
while (<>) {
if ( $_ eq \n) { last; }
else { print; }
}
Logical AND/OR
Logical AND : &&
if (($x > 0) && ($x < 10)) { }
Logical OR : ||
if
($x < 0) || ($x > 0)) { }
Both are short-circuit second

expression evaluated only if necessary
Ternary Operator
Same as the ternary operator (?:) in C
expr1 ? expr2 : expr3
Like if-then-else: If expr1 is true, expr2 is
used; otherwise expr3 is used
$weather=($temp>50)?warm:cold;
Regular Expressions
Use EREs (egrep style)
Plus the following character classes
\w
word characters: [A-Za-z0-9_]
\d
digits: [0-9]
\s
whitespaces: [\f\t\n\r ]
\b
word boundaries
\W, \D, \S, \B are complements of the corresponding
classes above
Can use \t to denote a tab
Backreferences
Support backreferences
Subexpressions are referred to using \1, \
2, etc. in the RE and $1, $2, etc. outside
RE
if (/^this (red|blue|green) (bat|ball) is \1/)
{
($color, $object) = ($1, $2);
}
Matching
Pattern match operator: /RE/ is shortcut of m/RE/
Returns true if there is a match
Match against $_
Can also use m(RE), m<RE>, m!RE!, etc.
if (/^\/usr\/local\//) { }
if (m%/usr/local/%) { }
Case-insensitive match
if (/new york/i) { };
Matching cont.
To match an RE against something other than
$_, use the binding operator =~
if ($s =~ /\bblah/i) {
print Found blah!
}
!~ negates the match
while (<STDIN> !~ /^#/) { }
Variables are interpolated inside REs

if (/^$word/) { }
\Substitutions
Sed-like search and replace with s///
s/red/blue/;
$x =~ s/\w+$/$`/;
m/// does not modify variable; s/// does
Global replacement with /g

s/(.)\1/$1/g;
Transliteration operator: tr/// or y///

tr/A-Z/a-z/;
RE Functions
split string using RE (whitespace by default)
@fields = split /:/, ::ab:cde:f;
# gets (,,ab,cde,f)
join strings into one

$str = join -, @fields;
# gets --ab-cde-f
grep something from a list

Similar to UNIX grep, but not limited to using RE
@selected = grep(!/^#/, @code);
@matched = grep { $_>100 && $_<150 } @nums;
Modifying elements in returned list actually modifies

the elements in the original list
Running Another program

Use the system function to run an external
program
With one argument, the shell is used to run the
command
Convenient when redirection is needed
$status = system(cmd1 args > file);
To avoid the shell, pass system a list

$status = system($prog, @args);
die $prog exited abnormally: $? unless $status
== 0;
Capturing Output
If output from another program needs to
be collected, use the backticks
my $files = `ls *.c`;
Collect all output lines into a single string
my @files = `ls *.c`;
Each element is an output line
The shell is invoked to run the command
Environment Variables
Environment variables are stored in the
special hash %ENV
$ENV{PATH} = /usr/local/bin:
$ENV{PATH};
Example: Word Frequency

#!/usr/bin/perl -w
# Read a list of words (one per line) and
# print the frequency of each word
use strict;
my(@words, %count, $word);
chomp(@words = <STDIN>); # read and chomp all lines
for $word (@words) {
$count{$word}++;
}
for $word (keys %count) {
print $word was seen $count{$word} times.\n;
}
Good Ways to Learn Perl

a2p
Translates an awk program to Perl
s2p
Translates a sed script to Perl
perldoc
$
$
$
$
Online Perl documentation

perldoc perldoc perldoc man page
perldoc perlintro Perl introduction
perldoc -f sort Perl sort function man page
perldoc CGI
CGI module man page
Modules
Perl modules are libraries of reusable
code with specific functionalities
Standard modules are distributed with
Perl, others can be obtained from
Include modules in your program with use,
e.g. use CGI incorporates the CGI module
Each module has its own namespace
CGI Programming
Forms
HTML forms are used to collect user input
Data sent via HTTP request
Server launches CGI script to process data
<form method=POST
action=http://www.cs.nyu.edu/~unixtool/cgibin/search.cgi>
Enter your query: <input type=text name=Search>
<input type=submit>
</form>
Input Types
Text Field
<input type=text name=zipcode>
Radio Buttons
<input type=radio name=size value=S> Small
<input type=radio name=size value=M> Medium
<input type=radio name=size value=L> Large
Checkboxes
<input type=checkbox name=extras value=lettuce> Lettuce
<input type=checkbox name=extras value=tomato> Tomato
Text Area
<textarea name=address cols=50 rows=4>
</textarea>
Submit Button
Submits the form for processing by the
CGI script specified in the form tag
<input type=submit value=Submit Order>
HTTP Methods
Determine how form data are sent to web
server
Two methods:
GET
Form variables stored in URL
POST
Form variables sent as content of HTTP request
Encoding Form Values

Browser sends form variable as name-value
pairs
name1=value1&name2=value2&name3=value3
Names are defined in form elements

<input type=text name=ssn maxlength=9>
Special characters are replaced with %## (2digit hex number), spaces replaced with +
e.g. 11/8 Wed is encoded as 11%2F8+Wed
HTTP GET/POST examples

GET:
GET /cgi-bin/myscript.pl?name=Bill%20Gates&
company=Microsoft HTTP/1.1
HOST: www.cs.nyu.edu
POST:
POST /cgi-bin/myscript.pl HTTP/1.1
HOST: www.cs.nyu.edu
other headers
name=Bill%20Gates&company=Microsoft
GET or POST?
GET method is useful for
Retrieving information, e.g. from a database
Embedding data in URL without form element
POST method should be used for forms with

Many fields or long fields
Sensitive information
Data for updating database
GET requests may be cached by clients

browsers or proxies, but not POST requests
Parsing Form Input

Method stored in HTTP_METHOD
GET: Data encoded into QUERY_STRING
POST: Data in standard input (from body
of request)
Most scripts parse input into an
associative array
You can parse it yourself
Or use available libraries (better)
CGI Script: Example
Part 1: HTML Form

<html>
<center>
<H1>Anonymous Comment Submission</H1>
</center>
Please enter your comment below which will
be sent anonymously to <tt>kornj@cs.nyu.edu</tt>.
If you want to be extra cautious, access this
page through <a
href="http://www.anonymizer.com">Anonymizer</a>.
<p>
<form action=cgi-bin/comment.cgi method=post>
<textarea name=comment rows=20 cols=80>
</textarea>
<input type=submit value="Submit Comment">
</form>
</html>
Part 2: CGI Script (ksh)

. cgi-lib.ksh
ReadParse
PrintHeader
# Read special functions to help parse
print -r -- "${Cgi.comment}" | /bin/mailx -s "COMMENT" kornj

print
print
print
print
"<H2>You submitted the comment</H2>"

"<pre>"
-r -- "${Cgi.comment}"
"</pre>"
Perl CGI Module

Interface for parsing and interpreting query
strings passed to CGI scripts
Methods for creating generating HTML
Methods to handle errors in CGI scripts
Two interfaces: procedural and OO
Ask for the procedural interface:
use CGI qw(:standard);
A Perl CGI Script

#!/usr/bin/perl -w
use strict;
use CGI qw(:standard);
my $bday = param("birthday");
# Print headers (text/html is the default)
print header(-type => 'text/html');
# Print <html>, <head>, <title>, <body> tags etc.
print start_html(Birthday);
# Your HTML body
print "Your birthday is $bday.\n";
# Print </body></html>
print end_html();
Debugging Perl CGI Scripts

Debugging CGI script is tricky - error messages
dont always come up on your browser
Check if the script compiles
$ perl -wc cgiScript
Run script with test data

$ perl -w cgiScript prod=MacBook price=1800
Content-Type: text/html
<html>
</html>
How to get your script run

This can vary by web server type
http://www.cims.nyu.edu/systems/resources/webhosting/index.html
Typically, you give your script a name that

ends with .cgi and/or put it in a special
directory (e.g. cgi-bin)
Give the script execute permission
Specify the location of that script in the
URL
CGI Security Risks

Sometimes CGI scripts run as owner of the
scripts
Never trust user input - sanity-check
everything
If a shell command contains user input, run
without shell escapes
Always encode sensitive information, e.g.
passwords
Also use HTTPS
Clean up - dont leave sensitive data around
CGI Benefits
Simple
Language independent
UNIX tools are good for this because
Work well with text
Integrate programs well
Easy to prototype
No compilation (CGI scripts)

Mario Dantas Unix-Perl

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mario Dantas Unix-Perl

Uploaded by

Copyright:

Available Formats

Advanced Course on Bioinformatic and Comparative Genome Analysis

UFSC Florianopolis - June 30 - July 12, 2008

- represent a execution flow which

Grid Information Service

It does not have an

module Integration Portal

Approach Semantic Integration

Reference Ontology (RO)

GRID 2007 - September 20, 2007

GRID 2007 - September 20, 2007

GRID 2007 - September 20, 2007

GRID 2007 - September 20, 2007

GRID 2007 - September 20, 2007

Available Free Online

Operating system overview

Operating system overview

UNICS: 1969 PDP-7 minicomputer

Big Reason for V6 Success

But Then The Feuding Began

Threat of Windows NT resolves battle with CDE

Send in the Clones

FreeBSD (1993, focus on PCs)

Today: Unix is Big

Linux at Google & Elsewhere

Purchased NeXT in December 1996

Why did UNIX succeed?

The Open Source Movement

Many vendors switching to Linux

SCO vs. Linux

Aug 2003: Evidence released

Aug 2005: Linux Kernel Code May Have Been in SCO

Does Linux borrow from ancient UNIX or System V R4?

UNIX vs. Linux

The UNIX Philosophy

Make each program do one thing well

The UNIX Philosophy

Most efficient implementation is rarely portable

Use flat ASCII files

The UNIX Philosophy

Scripting increases leverage and portability

List the logins of a systems users on a single line.

Build prototypes quickly (high level

The UNIX Philosophy

Avoid captive interfaces

UNIX Highlights / Contributions

Portability (variety of hardware; C implementation)

The Operating System

UNIX Structural Layout

Slackware the original

Operating system overview

Unix System Structure

shell and utilities

How processes share CPU, memory and signals

UNIX variants have different implementations

A program like any other

Most Commonly Used Shells

Shell Interactive Use

Execute a basic command

Getting Help on UNIX

Example Man Page

How are Users & Groups used?

The UNIX File Hierarchy

Hierarchies are Ubiquitous