Unix Shell Programming (3rd Edition) (Paperback)

After a quick review of the basics of UNIX, the authors give a purely descriptive explanation of the UNIX shell in chapter 3. Emphasizing that it is an interpretive language, the most commonly used shell commands are discussed in chapter 4, starting with a discussion of regular expressions. The cut, paste, sed, tr, grep, uniq, and sort commands are treated in detail.

In chapter 5, one begins the actual task of creating shell programs using shell variables. There is no data typing in the shell, so values can be assigned to variables without noting their type as integer, float, etc. The authors only briefly discuss the mechanism in shell programming. The method by which the shell interprets quotation characters is covered in the next chapter. The single, double, backslash, and back quote characters are discussed in detail. Noting that arithmetic operations are done on values stored in variables in the shell, the authors show to proceed with these operations using the expr program.

The mechanisms for passing arguments to shell programs is treated in chapter 7, the authors showing how to write shell programs that take arguments typed on the command line. The role of positional variables for delaying assignment after normal command line processing is discussed. The $#, and $* variables are discussed briefly, with $# getting set to the number of arguments typed on the command line and $* used for programs taking a variable number of arguments. The shift command is explained well as a method to allow one to use more than nine arguments to a program.

The ability of shell programs to execute decision blocks is treated in chapter 8, via the if statement. The role of the test and exit commands are in if blocks are discussed in good detail. In addition, the case command, familiar to C programmers is introduced as a technique to allow a single value to be compared against other values. The && and || constructs are used to show the reader how to execute commands that are contingent on the success or failure of the preceding command.

Program loops, via the for, while, and until commands are discussed in chapter 9, followed in the next chapter by a discussion of how to read data from the terminal or from a file using the read command. The ability to perform I/O redirection is discussed also.

Local and export variables are discussed in the next chapter on the user environment, and the authors give a good summary of how these work in shell programming. More discussion on parameter passing is done in chapter 12, with the different methods of parameter substitution given detailed treatment. The authors show how to use the $0 variable to check whether two or more programs have been executed, and how to use the set command to set shell options and to reassign positional parameters. This is followed in the next chapter by a discussion of the eval command, which makes the shell scan the command line twice before executing it, and the wait command, which will allow serialization in program execution. The trap and type commands are discussed also.

The Korn shell is discussed in chapter 15, with emphasis on the features added to Korn shell that cannot be found in the Bourne shell. The vi and emacs capability of this shell is briefly discussed in this chapter. The differences between Korn shell functions and Bourne shell functions are discussed in detail by the authors. Most importantly, the ability of the Korn shell to do integer arithmetic without using the expr command is discussed via the let command, which is built-in to the Korn shell. Also, the capability of the Korn shell to support data typing is discussed, along with its pattern matching capabilities. Pattern matching is done most efficiently now using Perl however.

BTW, I have been looking for a better book all these years and haven't found one. It's not that I think this book is perfect, rather there just aren't many good books on this subject.