Tools and Concepts

1
Chapter 9

• Tools and Concepts

2
find

• Very powerful command used to locate files
• conditions are things like name, or attributes
• if the attribute is numeric, it follows these
  rules
• unsigned attribute matches exactly that number
• attribute matches greater than that number
• -attribute matches less than that number
• can code -exec command \ to execute the
  command on any file that matches the condition
• careful construction allows great efficiency

3
pipe

• First implemented in 1972
• Allows user to construct powerful command sets
• Commands in a pipe are in pseudo-simultaneous
  execution
• To be used in a pipe, each process except the
  endmost two must be filters

4
The Filter Program Concept

• A filter is a process with the following
  characteristics
• it reads from standard input
• it writes good data to standard output
• while not required, it should send errors to
  standard error
• Most Unix programs are filters, that allows them
  to be used in pipes.

5
tee

• You can insert a tee into a pipe to capture
  intermediate results
• the data can only be redirected into a file
• the tee does not affect the flow of data
• you CANNOT create parallel pipes with a tee
• they are very useful for debugging pipes

6
redirection (lt, gt, gtgt)

• Besides piping the data, it is possible to send
  the output of a process or pipe to a file, or to
  take input from a file.
• This is possible because a file is a stream of
  bytes.
• The keyboard creates a stream of bytes
• The screen sees a stream of bytes

7
redirecting standard output (gt)

• Used to store the output of a command or pipe in
  a file.
• Remember the rule
• When a file is the target of a redirection, the
  first thing that happens is the next byte pointer
  is set to byte zero
• You can redirect any filter and any command that
  sends output to standard out

8
Appending gtgt

• If you choose the append operator gtgt you will add
  the current output to the end of the existing
  file.
• Some flavors require that the file you are
  appending to exist, others will create the file
  if it doesnt exist.

9
redirecting standard error

• Error messages are important clues to problems.
• You should always evaluate any error messages you
  are sent.
• You should never allow users to see error
  messages!
• The two shells handle this situation differently

10
redirecting standard errorC shell version

• Sending both error and out to the same file is
  straight forward

11
redirecting standard errorC shell (cont.)

• First redirect just output
• Then using gt redirect both output and error to
  the same file.

12
redirecting standard error C shell (cont.)

• Sending standard out to one file and error to
  another is a little tricky in csh.

13
redirecting standard error C shell (cont.)

• Create a sub-shell to run the command.
• Redirect standard out to one file in the
  sub-shell.
• Redirect standard error to another file outside
  the sub-shell.

Pay close attention to the difference in the
figure between parentheses and redirection.
14
Redirecting standard output Bourne Shell

• The Bourne shell references the three standard
  files by number
• File 0 - standard input
• File 1 - standard output
• File 2 - standard error
• Just redirecting standard out is the same as in
  the C shell

15
Redirecting standard error Bourne Shell

• Sending output to one file and error to another
  is simple in the Bourne shell

16
Redirecting standard error Bourne Shell (cont.)

• Sending output and error to the same file is al
  little more intricate
• Send file1 to filename, and send file 2 to the
  same place.

17
Redirecting standard input (lt)

• Most commands have evolved to take files on the
  command line and do their own input redirection.
• The only command that requires redirection of
  standard input is translate (tr).
• You can redirect standard input into most
  commands but it is a waste of a character.

18
Here document ltlt

• You can create an input stream within a script by
  using a here document.
• Bourne added this feature in response to faster
  computers that read ahead.
• Useful way to produce preformatted output in
  scripts.

19
alias

• Used to create new commands
• Some aliases protect the users, like alias cp
  cp -i
• It is dangerous to depend on aliases because
• they arent always there
• they arent always the same
• you can forget how the real command works

20
alias (cont.)

• Some rules for building aliases
• be sure you know what you are creating
• dont use one character aliases
• dont create aliases using other OS commands
• dont use questionable language
• wait until you understand a command before you
  alias it

21
unalias

• Used to remove an alias from the alias table
• Be careful about the aliases you unalias, you may
  want them later!

22
diff

• Used to find differences between two files
• Useful to track changes in a file
• can create output in ed format so you can make
  the two files look the same
• neither file is changed by diff

23
uniq

• Used to remove duplicate, concurrent lines in a
  file.
• The -c (count) option tells you how many
  duplicate lines there were
• if used with two files, the second file will be
  overwritten with the uniq output from the first
  file

24
File Name Completion

• Really nice feature of many shells
• usually invoked by ltESCgt or ltTABgt depending upon
  the shell
• making file names unique near the beginning
  facilitates file name completion
• in some shells D shows you a partial file list
  that matches your pattern
• yet another way to be more efficient in Unix

25
Shell Variables

• A variable is a mnemonic for a place.
• A variable name is the identifier for a location
  in memory.
• A variable value is the datum stored at the
  location specified by the variable name.
• Initializing variables varies between shells
• C shell use set variable value
• Bourne shells use variablevalue

26
set

• To create variables in the C shell use the set
  command
• you can assign values to variables
• you can set particular variables like filec as
  switches
• shell variables exist only in the instance of the
  shell that created them unless you make them
  available to subsequent shells

27
setenv

• Used to make a variable available to subsequent
  shells.
• The variable becomes part of the environment
  passed to the child process
• Normally environmental variables are coded in
  upper case
• In most systems environmental variables are
  available to all descendent instances of the shell

28
unsetenv

• Used to remove a variable from the environment of
  newly created child processes
• dont try to unsetenv any variable you dont
  create
• once a variable is unsetenv it will no longer be
  available to subsequent shells

29
export

• In the Bourne shell you export variables to make
  them available to child processes
• the child process receives the variable and value
  as part of its environment
• any change to the variable will NOT be reflected
  in the parent process

30
Startup files .login

• Run once when you log into the system
• Use this file to set up the environment that is
  common to all shells
• This is a good place to set your backspace
  key stty erase H
• Also a good place to set up an X-windows
  environment

31
Startup files .cshrc

• Run each time a new instance of the C shell is
  forked
• Use this file to set up your C shell variables
  and C shell dependent elements
• A good place to set aliases
• I always set filec here
• The rc in the file name stands for runnable
  commands

32
Startup files .bashrc

• Used to initialize the environment for the bash
  shell.
• This is the place to setup your bash aliases and
  the like

33
Setup files .profile

• This setup file is for the Bourne shell.
• It predates the idea of runnable commands so
  doesnt end with rc.
• Usually this is a small files as the Bourne shell
  doesnt have all that stuff to set up like
  aliases.

34
Other Setup files

• Most shells have a setup file you can use
• T shell - .tcshrc
• Z shell - .zcshrc
• The use of theses files is optional, but a wise
  user will utilize the efficiency available.