16.1. Conceptual Overview

The make program generates a sequence of commands for execution by the Unix shell. It uses a table of file dependencies provided by the programmer, and with this information, can perform updating tasks automatically for the user. It can keep track of the sequence of commands that create certain files, and the list of files or programs that require other files to be current before they can be rebuilt correctly. When a program is changed, make can create the proper files with a minimum of effort.

Each statement of a dependency is called a rule. Rules define one or more targets, which are the files to be generated, and the files they depend upon, the prerequisites or dependencies. For example, prog.o would be a target that depends upon prog.c; each time you update prog.c, prog.o must be regenerated. It is this task that make automates, and it is a critical one for large programs that have many pieces.

The file containing all the rules is termed a makefile; for GNU make, it may be named GNUmakefile, makefile or Makefile, in which case make will read it automatically, or you may use a file with a different name and tell make about it with the -f option.

Over the years, different enhancements to make have been made by many vendors, often in incompatible ways. POSIX standardizes how make is supposed to work. Today, GNU make is the most popular version in the Unix world. It has (or can emulate) the features of just about every other version of make, and many Open Source programs require it.

This chapter covers GNU make. Commercial Unix systems come with versions derived from the original System V version; these can be used for bootstrapping GNU make if need be. On the x86 versions of Solaris 10, you can find GNU make in /usr/sfw/bin/gmake. It isn't available on the Sparc version, although it can be easily bootstrapped with the standard version of make in /usr/ccs/bin.