- The Unix Operating System
- The Linux Operating System
- The GNU Project
- Summary
The GNU Project
The GNU Project was created in 1984 to create a free Unix-like operating system. It is responsible for maintaining open source versions of many common Unix utilities. Without the GNU Project, the Linux operating system would not be very exciting. Most of the core pieces of the Linux operating system are products of the GNU Project. This section describes three programs that are crucial to the operation of the mail server: the bash shell, the gcc compiler, and the make utility.
GNU bash
The kernel requires some kind of macro processor to enable a user to execute commands (programs) on the system. In the Unix world, that macro processor is called the shell. The most common shell in the Unix environment is the Bourne shell, named after its creator, Stephen Bourne. The Bourne shell is a program that runs as a process on the system and has an interactive session that enables the user to enter commands at a command prompt. The commands can be executable programs, internal shell commands, or a program file that contains shell commands (called a script file). The shell launches executable programs by creating a new process and running the program within that new process. This allows every program that runs from the shell to have its own process on the system.
The GNU Project developers knew that it was crucial to have a good open source shell to use with an open source Unix-like operating system. The shell program they developed was called bash, for Bourne-Again SHell. The bash shell is compatible with the original Bourne shell (called sh). The bash shell also includes features from other shells that have been developed in the Unix environment: the C shell (csh) and the Korn shell (ksh). bash has become the default shell for Linux systems. The current version of bash at the time of this writing is version 2.03.
The shell a user utilizes after logging in to a Linux system is determined by the user's entry in the /etc/passwd file. A typical record in this file looks like this:
riley:x:504:506:Riley M.:/home/riley:/bin/bash
Colons are used to separate the fields in the record. The first field identifies the user login name. The second field is a placeholder for the user password since this particular Linux system uses shadow passwords. As a result, the real password is encrypted and placed in a separate file. The third and fourth fields are the user id and group id for the user. The fifth field is the text identifier for the user. The sixth field identifies the user's default, or home, directory when he logs in to the system. The last field identifies the default shell for the user. This points to the location of the bash shell executable file on the server.
The bash shell has several different configuration files that you can use to modify the features of the shell as a user logs in. When bash is invoked as a shell from a login process, any commands present in the /etc/profile file are executed. This occurs for all users who specify the bash shell as the default login shell in the password file. Listing 3.8 shows the default /etc/profile file from a Mandrake 6.0 Linux system.
Listing 3.8 Sample /etc/profile File
# /etc/profile
# System wide environment and startup programs
# Functions and aliases go in /etc/bashrc
PATH="$PATH:/usr/X11R6/bin"
PS1="[\u@\h \W]\\$ "
# In bash2 we can't define a ulimit for user :-(
[ "$UID" = "0" ] && {
ulimit -c 1000000
}
if [ ´id -gn´ = ´id -un´ -a ´id -u´ -gt 14 ]; then
umask 002
else
umask 022
fi
USER=´id -un´
LOGNAME=$USER
MAIL="/var/spool/mail/$USER"
HOSTNAME=´/bin/hostname´
HISTSIZE=1000
HISTFILESIZE=1000
export PATH PS1 HOSTNAME HISTSIZE HISTFILESIZE USER LOGNAME MAIL
for i in /etc/profile.d/*.sh ; do
if [ -x $i ]; then
. $i
fi
done
unset i
The main thing the /etc/profile file does is create new environment variables for the shell to identify special characteristics for the session that application programs running in the shell can use. The MAIL environment variable is of special interest to the mail administrator. It points the user's mail program to the proper mailbox for the user. If the Postfix server is configured to use mailbox files located in the user's $HOME directory, this environment variable must be changed.
After the common /etc/profile program executes, bash looks for three more configuration files in the user's default (home) directory. If they exist, the .bash_profile, .bash_login, and .profile files execute, in order. Each of these files should be located in the user's home directory, so these files can be specific for a particular user. One final configuration file is available for use: .bash_logout. This script file executes when the user logs out of the interactive session. By using a combination of script files, the system administrator can fine-tune the bash shell for each user on the system.
GNU gcc
If you plan to install software programs that are distributed in source code, you must be able to compile the code to create an executable file. To do this, you need the proper compiler. All the programs described in this book are written in the C programming language. This requires that a C compiler be installed on your Unix server. The most common C compiler package for Linux is the GNU Compiler Collection (gcc).
The gcc package has itself had quite an interesting past. The GNU Project team developed gcc and released version 1 in early 1990. Back then the gcc actually stood for GNU C Compiler, because it only supported the C language. The GNU Project continued development of gcc, creating version 2.0 and continuing with improvements until version 2.8 was released in 1997. At the same time, another group of developers was working on a C++ compiler called egcs (pronounced "eggs"). After gcc 2.8, both projects were combined into the egcs project and egcs 1.0 was released. egcs 1.0 combined both the C and C++ compilers into one package.
Unfortunately, the egcs project was short-lived (only getting up to version 1.1). Now both the gcc and egcs projects have been rolled into the gcc project again (thus the name change for gcc). At the time of this writing, the current version of gcc is version 2.95.2, although version 3.0 is rumored to be ready for release soon. This version supports C, C++, and Java compilers.
To determine the version of gcc that your Linux distribution uses, you can use the --version option as follows:
[rich@shadrach]$ gcc --version 2.95.2
The sample Mandrake 7.0 Linux system shown is using the current gcc version 2.95.2.
GNU make
Large C and C++ projects often become complicated. There are several different source code files, each with several different header files. Compiling individual source files creates multiple object files that must be linked together in specific combinations to create executable files.
Maintaining the source, object, and executable files is often a difficult job. To simplify this task, most C and C++ compilers utilize the make program. The job of the make program is to control the creation of executable files, based on changes made to the source code files or to variables in a standard make configuration file.
The GNU Project has a version of make that is compatible with the gcc compiler. At the time of this writing, the current version available is version 3.79.1.
The meat and potatoes of the make utility is the Makefile. The Makefile is a text file that specifies how the make utility should compile the source code to create the executable program(s). A sample Makefile is shown in Listing 3.9.
Listing 3.9 Sample Makefile
1 # Makefile -- Make file for test program 2 # 3 4 # Edit the following for your installation 5 6 CC = gcc 7 #=================================== 8 9 # Compiler and linker flags 10 11 CFLAGS = -O 12 LFLAGS = -O 13 14 # This program's object code files 15 16 OBJS = test.o 17 18 # File dependencies 19 20 all: test 21 22 objs: $(OBJS) 23 24 clean: 25 rm -f $(OBJS) 26 rm -f test 27 28 test: $(OBJS) 29 $(CC) -o $@ $(LFLAGS) $(OBJS) $(LIBS) 30 31 test.o: test.c 32 $(CC) -c $(CFLAGS) -o $@ $<
The line numbers are not part of the original Makefile, but were added to aid in the discussion. Lines 6, 11, 12, and 16 show the use of variables within the Makefile. The user can change these values to the appropriate values for the system. Line 18 declares the make targets for the system. You can build each target individually by specifying the target name as a parameter on the make command line. For example, to run the clean target, which removes any old object and executable files, you can type
make clean
To create just the object files for the test program, you can type
make objs
If you type only make at the command line, the all target will be executed, which builds the executable file test.