GNU make

Short Table of Contents

Table of Contents

GNU make

This file documents the GNU make utility, which determines automatically which pieces of a large program need to be recompiled, and issues the commands to recompile them.

This is Edition 0.73, last updated 30 September 2014, of The GNU Make Manual, for GNU make version 4.1.

Copyright © 1988, 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014 Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, with the Front-Cover Texts being “A GNU Manual,” and with the Back-Cover Texts as in (a) below. A copy of the license is included in the section entitled “GNU Free Documentation License.”

(a) The FSF’s Back-Cover Text is: “You have the freedom to copy and modify this GNU manual. Buying copies from the FSF supports it in developing GNU and promoting software freedom.”

 — The Detailed Node Listing —

Overview of make

An Introduction to Makefiles

Writing Makefiles

What Makefiles Contain

Writing Rules

Using Wildcard Characters in File Names

Searching Directories for Prerequisites

Static Pattern Rules

Writing Recipes in Rules

Recipe Syntax

Recipe Execution

Parallel Execution

Recursive Use of make

How to Use Variables

Advanced Features for Reference to Variables

Conditional Parts of Makefiles

Functions for Transforming Text

How to Run make

Using Implicit Rules

Defining and Redefining Pattern Rules

Using make to Update Archive Files

Implicit Rule for Archive Member Targets

Extending GNU make

GNU Guile Integration

Loading Dynamic Objects

1 Overview of make

The make utility automatically determines which pieces of a large program need to be recompiled, and issues commands to recompile them. This manual describes GNU make, which was implemented by Richard Stallman and Roland McGrath. Development since Version 3.76 has been handled by Paul D. Smith.

GNU make conforms to section 6.2 of IEEE Standard 1003.2-1992 (POSIX.2).

Our examples show C programs, since they are most common, but you can use make with any programming language whose compiler can be run with a shell command. Indeed, make is not limited to programs. You can use it to describe any task where some files must be updated automatically from others whenever the others change.

Preparing

Preparing and Running Make

To prepare to use make, you must write a file called the makefile that describes the relationships among files in your program and provides commands for updating each file. In a program, typically, the executable file is updated from object files, which are in turn made by compiling source files.

Once a suitable makefile exists, each time you change some source files, this simple shell command:

make

suffices to perform all necessary recompilations. The make program uses the makefile data base and the last-modification times of the files to decide which of the files need to be updated. For each of those files, it issues the recipes recorded in the data base.

You can provide command line arguments to make to control which files should be recompiled, or how. See How to Run make.

1.1 How to Read This Manual

If you are new to make, or are looking for a general introduction, read the first few sections of each chapter, skipping the later sections. In each chapter, the first few sections contain introductory or general information and the later sections contain specialized or technical information. The exception is the second chapter, An Introduction to Makefiles, all of which is introductory.

If you are familiar with other make programs, see Features of GNU make, which lists the enhancements GNU make has, and Incompatibilities and Missing Features, which explains the few things GNU make lacks that others have.

For a quick summary, see Options Summary, Quick Reference, and Special Targets.

1.2 Problems and Bugs

If you have problems with GNU make or think you’ve found a bug, please report it to the developers; we cannot promise to do anything but we might well want to fix it.

Before reporting a bug, make sure you’ve actually found a real bug. Carefully reread the documentation and see if it really says you can do what you’re trying to do. If it’s not clear whether you should be able to do something or not, report that too; it’s a bug in the documentation!

Before reporting a bug or trying to fix it yourself, try to isolate it to the smallest possible makefile that reproduces the problem. Then send us the makefile and the exact results make gave you, including any error or warning messages. Please don’t paraphrase these messages: it’s best to cut and paste them into your report. When generating this small makefile, be sure to not use any non-free or unusual tools in your recipes: you can almost always emulate what such a tool would do with simple shell commands. Finally, be sure to explain what you expected to occur; this will help us decide whether the problem was really in the documentation.

Once you have a precise problem you can report it in one of two ways. Either send electronic mail to:

    bug-make@gnu.org

or use our Web-based project management tool, at:

    http://savannah.gnu.org/projects/make/

In addition to the information above, please be careful to include the version number of make you are using. You can get this information with the command ‘make --version’. Be sure also to include the type of machine and operating system you are using. One way to obtain this information is by looking at the final lines of output from the command ‘make --help’.

2 An Introduction to Makefiles

You need a file called a makefile to tell make what to do. Most often, the makefile tells make how to compile and link a program.

In this chapter, we will discuss a simple makefile that describes how to compile and link a text editor which consists of eight C source files and three header files. The makefile can also tell make how to run miscellaneous commands when explicitly asked (for example, to remove certain files as a clean-up operation). To see a more complex example of a makefile, see Complex Makefile.

When make recompiles the editor, each changed C source file must be recompiled. If a header file has changed, each C source file that includes the header file must be recompiled to be safe. Each compilation produces an object file corresponding to the source file. Finally, if any source file has been recompiled, all the object files, whether newly made or saved from previous compilations, must be linked together to produce the new executable editor.

2.1 What a Rule Looks Like

A simple makefile consists of “rules” with the following shape:

target … : prerequisites …
        recipe
        …
        …

A target is usually the name of a file that is generated by a program; examples of targets are executable or object files. A target can also be the name of an action to carry out, such as ‘clean’ (see Phony Targets).

A prerequisite is a file that is used as input to create the target. A target often depends on several files.

A recipe is an action that make carries out. A recipe may have more than one command, either on the same line or each on its own line. Please note: you need to put a tab character at the beginning of every recipe line! This is an obscurity that catches the unwary. If you prefer to prefix your recipes with a character other than tab, you can set the .RECIPEPREFIX variable to an alternate character (see Special Variables).

Usually a recipe is in a rule with prerequisites and serves to create a target file if any of the prerequisites change. However, the rule that specifies a recipe for the target need not have prerequisites. For example, the rule containing the delete command associated with the target ‘clean’ does not have prerequisites.

A rule, then, explains how and when to remake certain files which are the targets of the particular rule. make carries out the recipe on the prerequisites to create or update the target. A rule can also explain how and when to carry out an action. See Writing Rules.

A makefile may contain other text besides rules, but a simple makefile need only contain rules. Rules may look somewhat more complicated than shown in this template, but all fit the pattern more or less.

2.2 A Simple Makefile

Here is a straightforward makefile that describes the way an executable file called edit depends on eight object files which, in turn, depend on eight C source and three header files.

In this example, all the C files include defs.h, but only those defining editing commands include command.h, and only low level files that change the editor buffer include buffer.h.

edit : main.o kbd.o command.o display.o \
       insert.o search.o files.o utils.o
        cc -o edit main.o kbd.o command.o display.o \
                   insert.o search.o files.o utils.o

main.o : main.c defs.h
        cc -c main.c
kbd.o : kbd.c defs.h command.h
        cc -c kbd.c
command.o : command.c defs.h command.h
        cc -c command.c
display.o : display.c defs.h buffer.h
        cc -c display.c
insert.o : insert.c defs.h buffer.h
        cc -c insert.c
search.o : search.c defs.h buffer.h
        cc -c search.c
files.o : files.c defs.h buffer.h command.h
        cc -c files.c
utils.o : utils.c defs.h
        cc -c utils.c
clean :
        rm edit main.o kbd.o command.o display.o \
           insert.o search.o files.o utils.o

We split each long line into two lines using backslash/newline; this is like using one long line, but is easier to read. See Splitting Long Lines.

To use this makefile to create the executable file called edit, type:

make

To use this makefile to delete the executable file and all the object files from the directory, type:

make clean

In the example makefile, the targets include the executable file ‘edit’, and the object files ‘main.o’ and ‘kbd.o’. The prerequisites are files such as ‘main.c’ and ‘defs.h’. In fact, each ‘.o’ file is both a target and a prerequisite. Recipes include ‘cc -c main.c’ and ‘cc -c kbd.c’.

When a target is a file, it needs to be recompiled or relinked if any of its prerequisites change. In addition, any prerequisites that are themselves automatically generated should be updated first. In this example, edit depends on each of the eight object files; the object file main.o depends on the source file main.c and on the header file defs.h.

A recipe may follow each line that contains a target and prerequisites. These recipes say how to update the target file. A tab character (or whatever character is specified by the .RECIPEPREFIX variable; see Special Variables) must come at the beginning of every line in the recipe to distinguish recipes from other lines in the makefile. (Bear in mind that make does not know anything about how the recipes work. It is up to you to supply recipes that will update the target file properly. All make does is execute the recipe you have specified when the target file needs to be updated.)

The target ‘clean’ is not a file, but merely the name of an action. Since you normally do not want to carry out the actions in this rule, ‘clean’ is not a prerequisite of any other rule. Consequently, make never does anything with it unless you tell it specifically. Note that this rule not only is not a prerequisite, it also does not have any prerequisites, so the only purpose of the rule is to run the specified recipe. Targets that do not refer to files but are just actions are called phony targets. See Phony Targets, for information about this kind of target. See Errors in Recipes, to see how to cause make to ignore errors from rm or any other command.

2.3 How make Processes a Makefile

By default, make starts with the first target (not targets whose names start with ‘.’). This is called the default goal. (Goals are the targets that make strives ultimately to update. You can override this behavior using the command line (see Arguments to Specify the Goals) or with the .DEFAULT_GOAL special variable (see Other Special Variables).

In the simple example of the previous section, the default goal is to update the executable program edit; therefore, we put that rule first.

Thus, when you give the command:

make

make reads the makefile in the current directory and begins by processing the first rule. In the example, this rule is for relinking edit; but before make can fully process this rule, it must process the rules for the files that edit depends on, which in this case are the object files. Each of these files is processed according to its own rule. These rules say to update each ‘.o’ file by compiling its source file. The recompilation must be done if the source file, or any of the header files named as prerequisites, is more recent than the object file, or if the object file does not exist.

The other rules are processed because their targets appear as prerequisites of the goal. If some other rule is not depended on by the goal (or anything it depends on, etc.), that rule is not processed, unless you tell make to do so (with a command such as make clean).

Before recompiling an object file, make considers updating its prerequisites, the source file and header files. This makefile does not specify anything to be done for them—the ‘.c’ and ‘.h’ files are not the targets of any rules—so make does nothing for these files. But make would update automatically generated C programs, such as those made by Bison or Yacc, by their own rules at this time.

After recompiling whichever object files need it, make decides whether to relink edit. This must be done if the file edit does not exist, or if any of the object files are newer than it. If an object file was just recompiled, it is now newer than edit, so edit is relinked.

Thus, if we change the file insert.c and run make, make will compile that file to update insert.o, and then link edit. If we change the file command.h and run make, make will recompile the object files kbd.o, command.o and files.o and then link the file edit.

2.4 Variables Make Makefiles Simpler

In our example, we had to list all the object files twice in the rule for edit (repeated here):

edit : main.o kbd.o command.o display.o \
              insert.o search.o files.o utils.o
        cc -o edit main.o kbd.o command.o display.o \
                   insert.o search.o files.o utils.o

Such duplication is error-prone; if a new object file is added to the system, we might add it to one list and forget the other. We can eliminate the risk and simplify the makefile by using a variable. Variables allow a text string to be defined once and substituted in multiple places later (see How to Use Variables).

It is standard practice for every makefile to have a variable named objects, OBJECTS, objs, OBJS, obj, or OBJ which is a list of all object file names. We would define such a variable objects with a line like this in the makefile:

objects = main.o kbd.o command.o display.o \
          insert.o search.o files.o utils.o

Then, each place we want to put a list of the object file names, we can substitute the variable’s value by writing ‘$(objects)’ (see How to Use Variables).

Here is how the complete simple makefile looks when you use a variable for the object files:

objects = main.o kbd.o command.o display.o \
          insert.o search.o files.o utils.o

edit : $(objects)
        cc -o edit $(objects)
main.o : main.c defs.h
        cc -c main.c
kbd.o : kbd.c defs.h command.h
        cc -c kbd.c
command.o : command.c defs.h command.h
        cc -c command.c
display.o : display.c defs.h buffer.h
        cc -c display.c
insert.o : insert.c defs.h buffer.h
        cc -c insert.c
search.o : search.c defs.h buffer.h
        cc -c search.c
files.o : files.c defs.h buffer.h command.h
        cc -c files.c
utils.o : utils.c defs.h
        cc -c utils.c
clean :
        rm edit $(objects)

2.5 Letting make Deduce the Recipes

It is not necessary to spell out the recipes for compiling the individual C source files, because make can figure them out: it has an implicit rule for updating a ‘.o’ file from a correspondingly named ‘.c’ file using a ‘cc -c’ command. For example, it will use the recipe ‘cc -c main.c -o main.o’ to compile main.c into main.o. We can therefore omit the recipes from the rules for the object files. See Using Implicit Rules.

When a ‘.c’ file is used automatically in this way, it is also automatically added to the list of prerequisites. We can therefore omit the ‘.c’ files from the prerequisites, provided we omit the recipe.

Here is the entire example, with both of these changes, and a variable objects as suggested above:

objects = main.o kbd.o command.o display.o \
          insert.o search.o files.o utils.o

edit : $(objects)
        cc -o edit $(objects)

main.o : defs.h
kbd.o : defs.h command.h
command.o : defs.h command.h
display.o : defs.h buffer.h
insert.o : defs.h buffer.h
search.o : defs.h buffer.h
files.o : defs.h buffer.h command.h
utils.o : defs.h

.PHONY : clean
clean :
        rm edit $(objects)

This is how we would write the makefile in actual practice. (The complications associated with ‘clean’ are described elsewhere. See Phony Targets, and Errors in Recipes.)

Because implicit rules are so convenient, they are important. You will see them used frequently.

2.6 Another Style of Makefile

When the objects of a makefile are created only by implicit rules, an alternative style of makefile is possible. In this style of makefile, you group entries by their prerequisites instead of by their targets. Here is what one looks like:

objects = main.o kbd.o command.o display.o \
          insert.o search.o files.o utils.o

edit : $(objects)
        cc -o edit $(objects)

$(objects) : defs.h
kbd.o command.o files.o : command.h
display.o insert.o search.o files.o : buffer.h

Here defs.h is given as a prerequisite of all the object files; command.h and buffer.h are prerequisites of the specific object files listed for them.

Whether this is better is a matter of taste: it is more compact, but some people dislike it because they find it clearer to put all the information about each target in one place.

2.7 Rules for Cleaning the Directory

Compiling a program is not the only thing you might want to write rules for. Makefiles commonly tell how to do a few other things besides compiling a program: for example, how to delete all the object files and executables so that the directory is ‘clean’.

Here is how we could write a make rule for cleaning our example editor:

clean:
        rm edit $(objects)

In practice, we might want to write the rule in a somewhat more complicated manner to handle unanticipated situations. We would do this:

.PHONY : clean
clean :
        -rm edit $(objects)

This prevents make from getting confused by an actual file called clean and causes it to continue in spite of errors from rm. (See Phony Targets, and Errors in Recipes.)

A rule such as this should not be placed at the beginning of the makefile, because we do not want it to run by default! Thus, in the example makefile, we want the rule for edit, which recompiles the editor, to remain the default goal.

Since clean is not a prerequisite of edit, this rule will not run at all if we give the command ‘make’ with no arguments. In order to make the rule run, we have to type ‘make clean’. See How to Run make.

3 Writing Makefiles

The information that tells make how to recompile a system comes from reading a data base called the makefile.

3.1 What Makefiles Contain

Makefiles contain five kinds of things: explicit rules, implicit rules, variable definitions, directives, and comments. Rules, variables, and directives are described at length in later chapters.

• An explicit rule says when and how to remake one or more files, called the rule’s targets. It lists the other files that the targets depend on, called the prerequisites of the target, and may also give a recipe to use to create or update the targets. See Writing Rules.
• An implicit rule says when and how to remake a class of files based on their names. It describes how a target may depend on a file with a name similar to the target and gives a recipe to create or update such a target. See Using Implicit Rules.
• A variable definition is a line that specifies a text string value for a variable that can be substituted into the text later. The simple makefile example shows a variable definition for objects as a list of all object files (see Variables Make Makefiles Simpler).
• A directive is an instruction for make to do something special while reading the makefile. These include:
  • Reading another makefile (see Including Other Makefiles).
  • Deciding (based on the values of variables) whether to use or ignore a part of the makefile (see Conditional Parts of Makefiles).
  • Defining a variable from a verbatim string containing multiple lines (see Defining Multi-Line Variables).
• ‘#’ in a line of a makefile starts a comment. It and the rest of the line are ignored, except that a trailing backslash not escaped by another backslash will continue the comment across multiple lines. A line containing just a comment (with perhaps spaces before it) is effectively blank, and is ignored. If you want a literal #, escape it with a backslash (e.g., \#). Comments may appear on any line in the makefile, although they are treated specially in certain situations.

  You cannot use comments within variable references or function calls: any instance of # will be treated literally (rather than as the start of a comment) inside a variable reference or function call.

  Comments within a recipe are passed to the shell, just as with any other recipe text. The shell decides how to interpret it: whether or not this is a comment is up to the shell.

  Within a define directive, comments are not ignored during the definition of the variable, but rather kept intact in the value of the variable. When the variable is expanded they will either be treated as make comments or as recipe text, depending on the context in which the variable is evaluated.

3.1.1 Splitting Long Lines

Makefiles use a “line-based” syntax in which the newline character is special and marks the end of a statement. GNU make has no limit on the length of a statement line, up to the amount of memory in your computer.

However, it is difficult to read lines which are too long to display without wrapping or scrolling. So, you can format your makefiles for readability by adding newlines into the middle of a statement: you do this by escaping the internal newlines with a backslash (\) character. Where we need to make a distinction we will refer to “physical lines” as a single line ending with a newline (regardless of whether it is escaped) and a “logical line” being a complete statement including all escaped newlines up to the first non-escaped newline.

The way in which backslash/newline combinations are handled depends on whether the statement is a recipe line or a non-recipe line. Handling of backslash/newline in a recipe line is discussed later (see Splitting Recipe Lines).

Outside of recipe lines, backslash/newlines are converted into a single space character. Once that is done, all whitespace around the backslash/newline is condensed into a single space: this includes all whitespace preceding the backslash, all whitespace at the beginning of the line after the backslash/newline, and any consecutive backslash/newline combinations.

If the .POSIX special target is defined then backslash/newline handling is modified slightly to conform to POSIX.2: first, whitespace preceding a backslash is not removed and second, consecutive backslash/newlines are not condensed.

3.2 What Name to Give Your Makefile

By default, when make looks for the makefile, it tries the following names, in order: GNUmakefile, makefile and Makefile.

Normally you should call your makefile either makefile or Makefile. (We recommend Makefile because it appears prominently near the beginning of a directory listing, right near other important files such as README.) The first name checked, GNUmakefile, is not recommended for most makefiles. You should use this name if you have a makefile that is specific to GNU make, and will not be understood by other versions of make. Other make programs look for makefile and Makefile, but not GNUmakefile.

If make finds none of these names, it does not use any makefile. Then you must specify a goal with a command argument, and make will attempt to figure out how to remake it using only its built-in implicit rules. See Using Implicit Rules.

If you want to use a nonstandard name for your makefile, you can specify the makefile name with the ‘-f’ or ‘--file’ option. The arguments ‘-f name’ or ‘--file=name’ tell make to read the file name as the makefile. If you use more than one ‘-f’ or ‘--file’ option, you can specify several makefiles. All the makefiles are effectively concatenated in the order specified. The default makefile names GNUmakefile, makefile and Makefile are not checked automatically if you specify ‘-f’ or ‘--file’.

3.3 Including Other Makefiles

The include directive tells make to suspend reading the current makefile and read one or more other makefiles before continuing. The directive is a line in the makefile that looks like this:

include filenames…

filenames can contain shell file name patterns. If filenames is empty, nothing is included and no error is printed.

Extra spaces are allowed and ignored at the beginning of the line, but the first character must not be a tab (or the value of .RECIPEPREFIX)—if the line begins with a tab, it will be considered a recipe line. Whitespace is required between include and the file names, and between file names; extra whitespace is ignored there and at the end of the directive. A comment starting with ‘#’ is allowed at the end of the line. If the file names contain any variable or function references, they are expanded. See How to Use Variables.

For example, if you have three .mk files, a.mk, b.mk, and c.mk, and $(bar) expands to bish bash, then the following expression

include foo *.mk $(bar)

is equivalent to

include foo a.mk b.mk c.mk bish bash

When make processes an include directive, it suspends reading of the containing makefile and reads from each listed file in turn. When that is finished, make resumes reading the makefile in which the directive appears.

One occasion for using include directives is when several programs, handled by individual makefiles in various directories, need to use a common set of variable definitions (see Setting Variables) or pattern rules (see Defining and Redefining Pattern Rules).

Another such occasion is when you want to generate prerequisites from source files automatically; the prerequisites can be put in a file that is included by the main makefile. This practice is generally cleaner than that of somehow appending the prerequisites to the end of the main makefile as has been traditionally done with other versions of make. See Automatic Prerequisites.

If the specified name does not start with a slash, and the file is not found in the current directory, several other directories are searched. First, any directories you have specified with the ‘-I’ or ‘--include-dir’ option are searched (see Summary of Options). Then the following directories (if they exist) are searched, in this order: prefix/include (normally /usr/local/include ¹) /usr/gnu/include, /usr/local/include, /usr/include.

If an included makefile cannot be found in any of these directories, a warning message is generated, but it is not an immediately fatal error; processing of the makefile containing the include continues. Once it has finished reading makefiles, make will try to remake any that are out of date or don’t exist. See How Makefiles Are Remade. Only after it has tried to find a way to remake a makefile and failed, will make diagnose the missing makefile as a fatal error.

If you want make to simply ignore a makefile which does not exist or cannot be remade, with no error message, use the -include directive instead of include, like this:

-include filenames…

This acts like include in every way except that there is no error (not even a warning) if any of the filenames (or any prerequisites of any of the filenames) do not exist or cannot be remade.

For compatibility with some other make implementations, sinclude is another name for -include.

3.4 The Variable MAKEFILES

If the environment variable MAKEFILES is defined, make considers its value as a list of names (separated by whitespace) of additional makefiles to be read before the others. This works much like the include directive: various directories are searched for those files (see Including Other Makefiles). In addition, the default goal is never taken from one of these makefiles (or any makefile included by them) and it is not an error if the files listed in MAKEFILES are not found.

The main use of MAKEFILES is in communication between recursive invocations of make (see Recursive Use of make). It usually is not desirable to set the environment variable before a top-level invocation of make, because it is usually better not to mess with a makefile from outside. However, if you are running make without a specific makefile, a makefile in MAKEFILES can do useful things to help the built-in implicit rules work better, such as defining search paths (see Directory Search).

Some users are tempted to set MAKEFILES in the environment automatically on login, and program makefiles to expect this to be done. This is a very bad idea, because such makefiles will fail to work if run by anyone else. It is much better to write explicit include directives in the makefiles. See Including Other Makefiles.

3.5 How Makefiles Are Remade

Sometimes makefiles can be remade from other files, such as RCS or SCCS files. If a makefile can be remade from other files, you probably want make to get an up-to-date version of the makefile to read in.

To this end, after reading in all makefiles, make will consider each as a goal target and attempt to update it. If a makefile has a rule which says how to update it (found either in that very makefile or in another one) or if an implicit rule applies to it (see Using Implicit Rules), it will be updated if necessary. After all makefiles have been checked, if any have actually been changed, make starts with a clean slate and reads all the makefiles over again. (It will also attempt to update each of them over again, but normally this will not change them again, since they are already up to date.)

If you know that one or more of your makefiles cannot be remade and you want to keep make from performing an implicit rule search on them, perhaps for efficiency reasons, you can use any normal method of preventing implicit rule look-up to do so. For example, you can write an explicit rule with the makefile as the target, and an empty recipe (see Using Empty Recipes).

If the makefiles specify a double-colon rule to remake a file with a recipe but no prerequisites, that file will always be remade (see Double-Colon). In the case of makefiles, a makefile that has a double-colon rule with a recipe but no prerequisites will be remade every time make is run, and then again after make starts over and reads the makefiles in again. This would cause an infinite loop: make would constantly remake the makefile, and never do anything else. So, to avoid this, make will not attempt to remake makefiles which are specified as targets of a double-colon rule with a recipe but no prerequisites.

If you do not specify any makefiles to be read with ‘-f’ or ‘--file’ options, make will try the default makefile names; see What Name to Give Your Makefile. Unlike makefiles explicitly requested with ‘-f’ or ‘--file’ options, make is not certain that these makefiles should exist. However, if a default makefile does not exist but can be created by running make rules, you probably want the rules to be run so that the makefile can be used.

Therefore, if none of the default makefiles exists, make will try to make each of them in the same order in which they are searched for (see What Name to Give Your Makefile) until it succeeds in making one, or it runs out of names to try. Note that it is not an error if make cannot find or make any makefile; a makefile is not always necessary.

When you use the ‘-t’ or ‘--touch’ option (see Instead of Executing Recipes), you would not want to use an out-of-date makefile to decide which targets to touch. So the ‘-t’ option has no effect on updating makefiles; they are really updated even if ‘-t’ is specified. Likewise, ‘-q’ (or ‘--question’) and ‘-n’ (or ‘--just-print’) do not prevent updating of makefiles, because an out-of-date makefile would result in the wrong output for other targets. Thus, ‘make -f mfile -n foo’ will update mfile, read it in, and then print the recipe to update foo and its prerequisites without running it. The recipe printed for foo will be the one specified in the updated contents of mfile.

However, on occasion you might actually wish to prevent updating of even the makefiles. You can do this by specifying the makefiles as goals in the command line as well as specifying them as makefiles. When the makefile name is specified explicitly as a goal, the options ‘-t’ and so on do apply to them.

Thus, ‘make -f mfile -n mfile foo’ would read the makefile mfile, print the recipe needed to update it without actually running it, and then print the recipe needed to update foo without running that. The recipe for foo will be the one specified by the existing contents of mfile.

3.6 Overriding Part of Another Makefile

Sometimes it is useful to have a makefile that is mostly just like another makefile. You can often use the ‘include’ directive to include one in the other, and add more targets or variable definitions. However, it is invalid for two makefiles to give different recipes for the same target. But there is another way.

In the containing makefile (the one that wants to include the other), you can use a match-anything pattern rule to say that to remake any target that cannot be made from the information in the containing makefile, make should look in another makefile. See Pattern Rules, for more information on pattern rules.

For example, if you have a makefile called Makefile that says how to make the target ‘foo’ (and other targets), you can write a makefile called GNUmakefile that contains:

foo:
        frobnicate > foo

%: force
        @$(MAKE) -f Makefile $@
force: ;

If you say ‘make foo’, make will find GNUmakefile, read it, and see that to make foo, it needs to run the recipe ‘frobnicate > foo’. If you say ‘make bar’, make will find no way to make bar in GNUmakefile, so it will use the recipe from the pattern rule: ‘make -f Makefile bar’. If Makefile provides a rule for updating bar, make will apply the rule. And likewise for any other target that GNUmakefile does not say how to make.

The way this works is that the pattern rule has a pattern of just ‘%’, so it matches any target whatever. The rule specifies a prerequisite force, to guarantee that the recipe will be run even if the target file already exists. We give the force target an empty recipe to prevent make from searching for an implicit rule to build it—otherwise it would apply the same match-anything rule to force itself and create a prerequisite loop!

3.7 How make Reads a Makefile

GNU make does its work in two distinct phases. During the first phase it reads all the makefiles, included makefiles, etc. and internalizes all the variables and their values, implicit and explicit rules, and constructs a dependency graph of all the targets and their prerequisites. During the second phase, make uses these internal structures to determine what targets will need to be rebuilt and to invoke the rules necessary to do so.

It’s important to understand this two-phase approach because it has a direct impact on how variable and function expansion happens; this is often a source of some confusion when writing makefiles. Here we will present a summary of the phases in which expansion happens for different constructs within the makefile. We say that expansion is immediate if it happens during the first phase: in this case make will expand any variables or functions in that section of a construct as the makefile is parsed. We say that expansion is deferred if expansion is not performed immediately. Expansion of a deferred construct is not performed until either the construct appears later in an immediate context, or until the second phase.

You may not be familiar with some of these constructs yet. You can reference this section as you become familiar with them, in later chapters.

Variable Assignment

Variable definitions are parsed as follows:

immediate = deferred
immediate ?= deferred
immediate := immediate
immediate ::= immediate
immediate += deferred or immediate
immediate != immediate

define immediate
  deferred
endef

define immediate =
  deferred
endef

define immediate ?=
  deferred
endef

define immediate :=
  immediate
endef

define immediate ::=
  immediate
endef

define immediate +=
  deferred or immediate
endef

define immediate !=
  immediate
endef

For the append operator, ‘+=’, the right-hand side is considered immediate if the variable was previously set as a simple variable (‘:=’ or ‘::=’), and deferred otherwise.

For the shell assignment operator, ‘!=’, the right-hand side is evaluated immediately and handed to the shell. The result is stored in the variable named on the left, and that variable becomes a simple variable (and will thus be re-evaluated on each reference).

Conditional Directives

Conditional directives are parsed immediately. This means, for example, that automatic variables cannot be used in conditional directives, as automatic variables are not set until the recipe for that rule is invoked. If you need to use automatic variables in a conditional directive you must move the condition into the recipe and use shell conditional syntax instead.

Rule Definition

A rule is always expanded the same way, regardless of the form:

immediate : immediate ; deferred
        deferred

That is, the target and prerequisite sections are expanded immediately, and the recipe used to construct the target is always deferred. This general rule is true for explicit rules, pattern rules, suffix rules, static pattern rules, and simple prerequisite definitions.

3.8 Secondary Expansion

In the previous section we learned that GNU make works in two distinct phases: a read-in phase and a target-update phase (see How make Reads a Makefile). GNU make also has the ability to enable a second expansion of the prerequisites (only) for some or all targets defined in the makefile. In order for this second expansion to occur, the special target .SECONDEXPANSION must be defined before the first prerequisite list that makes use of this feature.

If that special target is defined then in between the two phases mentioned above, right at the end of the read-in phase, all the prerequisites of the targets defined after the special target are expanded a second time. In most circumstances this secondary expansion will have no effect, since all variable and function references will have been expanded during the initial parsing of the makefiles. In order to take advantage of the secondary expansion phase of the parser, then, it’s necessary to escape the variable or function reference in the makefile. In this case the first expansion merely un-escapes the reference but doesn’t expand it, and expansion is left to the secondary expansion phase. For example, consider this makefile:

.SECONDEXPANSION:
ONEVAR = onefile
TWOVAR = twofile
myfile: $(ONEVAR) $$(TWOVAR)

After the first expansion phase the prerequisites list of the myfile target will be onefile and $(TWOVAR); the first (unescaped) variable reference to ONEVAR is expanded, while the second (escaped) variable reference is simply unescaped, without being recognized as a variable reference. Now during the secondary expansion the first word is expanded again but since it contains no variable or function references it remains the value onefile, while the second word is now a normal reference to the variable TWOVAR, which is expanded to the value twofile. The final result is that there are two prerequisites, onefile and twofile.

Obviously, this is not a very interesting case since the same result could more easily have been achieved simply by having both variables appear, unescaped, in the prerequisites list. One difference becomes apparent if the variables are reset; consider this example:

.SECONDEXPANSION:
AVAR = top
onefile: $(AVAR)
twofile: $$(AVAR)
AVAR = bottom

Here the prerequisite of onefile will be expanded immediately, and resolve to the value top, while the prerequisite of twofile will not be full expanded until the secondary expansion and yield a value of bottom.

This is marginally more exciting, but the true power of this feature only becomes apparent when you discover that secondary expansions always take place within the scope of the automatic variables for that target. This means that you can use variables such as $@, $*, etc. during the second expansion and they will have their expected values, just as in the recipe. All you have to do is defer the expansion by escaping the $. Also, secondary expansion occurs for both explicit and implicit (pattern) rules. Knowing this, the possible uses for this feature increase dramatically. For example:

.SECONDEXPANSION:
main_OBJS := main.o try.o test.o
lib_OBJS := lib.o api.o

main lib: $$($$@_OBJS)

Here, after the initial expansion the prerequisites of both the main and lib targets will be $($@_OBJS). During the secondary expansion, the $@ variable is set to the name of the target and so the expansion for the main target will yield $(main_OBJS), or main.o try.o test.o, while the secondary expansion for the lib target will yield $(lib_OBJS), or lib.o api.o.

You can also mix in functions here, as long as they are properly escaped:

main_SRCS := main.c try.c test.c
lib_SRCS := lib.c api.c

.SECONDEXPANSION:
main lib: $$(patsubst %.c,%.o,$$($$@_SRCS))

This version allows users to specify source files rather than object files, but gives the same resulting prerequisites list as the previous example.

Evaluation of automatic variables during the secondary expansion phase, especially of the target name variable $$@, behaves similarly to evaluation within recipes. However, there are some subtle differences and “corner cases” which come into play for the different types of rule definitions that make understands. The subtleties of using the different automatic variables are described below.

Secondary Expansion of Explicit Rules

During the secondary expansion of explicit rules, $$@ and $$% evaluate, respectively, to the file name of the target and, when the target is an archive member, the target member name. The $$< variable evaluates to the first prerequisite in the first rule for this target. $$^ and $$+ evaluate to the list of all prerequisites of rules that have already appeared for the same target ($$+ with repetitions and $$^ without). The following example will help illustrate these behaviors:

.SECONDEXPANSION:

foo: foo.1 bar.1 $$< $$^ $$+    # line #1

foo: foo.2 bar.2 $$< $$^ $$+    # line #2

foo: foo.3 bar.3 $$< $$^ $$+    # line #3

In the first prerequisite list, all three variables ($$<, $$^, and $$+) expand to the empty string. In the second, they will have values foo.1, foo.1 bar.1, and foo.1 bar.1 respectively. In the third they will have values foo.1, foo.1 bar.1 foo.2 bar.2, and foo.1 bar.1 foo.2 bar.2 foo.1 foo.1 bar.1 foo.1 bar.1 respectively.

Rules undergo secondary expansion in makefile order, except that the rule with the recipe is always evaluated last.

The variables $$? and $$* are not available and expand to the empty string.

Secondary Expansion of Static Pattern Rules

Rules for secondary expansion of static pattern rules are identical to those for explicit rules, above, with one exception: for static pattern rules the $$* variable is set to the pattern stem. As with explicit rules, $$? is not available and expands to the empty string.

Secondary Expansion of Implicit Rules

As make searches for an implicit rule, it substitutes the stem and then performs secondary expansion for every rule with a matching target pattern. The value of the automatic variables is derived in the same fashion as for static pattern rules. As an example:

.SECONDEXPANSION:

foo: bar

foo foz: fo%: bo%

%oo: $$< $$^ $$+ $$*

When the implicit rule is tried for target foo, $$< expands to bar, $$^ expands to bar boo, $$+ also expands to bar boo, and $$* expands to f.

Note that the directory prefix (D), as described in Implicit Rule Search Algorithm, is appended (after expansion) to all the patterns in the prerequisites list. As an example:

.SECONDEXPANSION:

/tmp/foo.o:

%.o: $$(addsuffix /%.c,foo bar) foo.h
        @echo $^

The prerequisite list printed, after the secondary expansion and directory prefix reconstruction, will be /tmp/foo/foo.c /tmp/bar/foo.c foo.h. If you are not interested in this reconstruction, you can use $$* instead of % in the prerequisites list.

4 Writing Rules

A rule appears in the makefile and says when and how to remake certain files, called the rule’s targets (most often only one per rule). It lists the other files that are the prerequisites of the target, and the recipe to use to create or update the target.

The order of rules is not significant, except for determining the default goal: the target for make to consider, if you do not otherwise specify one. The default goal is the target of the first rule in the first makefile. If the first rule has multiple targets, only the first target is taken as the default. There are two exceptions: a target starting with a period is not a default unless it contains one or more slashes, ‘/’, as well; and, a target that defines a pattern rule has no effect on the default goal. (See Defining and Redefining Pattern Rules.)

Therefore, we usually write the makefile so that the first rule is the one for compiling the entire program or all the programs described by the makefile (often with a target called ‘all’). See Arguments to Specify the Goals.

4.1 Rule Example

Here is an example of a rule:

foo.o : foo.c defs.h       # module for twiddling the frobs
        cc -c -g foo.c

Its target is foo.o and its prerequisites are foo.c and defs.h. It has one command in the recipe: ‘cc -c -g foo.c’. The recipe starts with a tab to identify it as a recipe.

This rule says two things:

• How to decide whether foo.o is out of date: it is out of date if it does not exist, or if either foo.c or defs.h is more recent than it.
• How to update the file foo.o: by running cc as stated. The recipe does not explicitly mention defs.h, but we presume that foo.c includes it, and that that is why defs.h was added to the prerequisites.

4.2 Rule Syntax

In general, a rule looks like this:

targets : prerequisites
        recipe
        …

or like this:

targets : prerequisites ; recipe
        recipe
        …

The targets are file names, separated by spaces. Wildcard characters may be used (see Using Wildcard Characters in File Names) and a name of the form a(m) represents member m in archive file a (see Archive Members as Targets). Usually there is only one target per rule, but occasionally there is a reason to have more (see Multiple Targets in a Rule).

The recipe lines start with a tab character (or the first character in the value of the .RECIPEPREFIX variable; see Special Variables). The first recipe line may appear on the line after the prerequisites, with a tab character, or may appear on the same line, with a semicolon. Either way, the effect is the same. There are other differences in the syntax of recipes. See Writing Recipes in Rules.

Because dollar signs are used to start make variable references, if you really want a dollar sign in a target or prerequisite you must write two of them, ‘$$’ (see How to Use Variables). If you have enabled secondary expansion (see Secondary Expansion) and you want a literal dollar sign in the prerequisites list, you must actually write four dollar signs (‘$$$$’).

You may split a long line by inserting a backslash followed by a newline, but this is not required, as make places no limit on the length of a line in a makefile.

A rule tells make two things: when the targets are out of date, and how to update them when necessary.

The criterion for being out of date is specified in terms of the prerequisites, which consist of file names separated by spaces. (Wildcards and archive members (see Archives) are allowed here too.) A target is out of date if it does not exist or if it is older than any of the prerequisites (by comparison of last-modification times). The idea is that the contents of the target file are computed based on information in the prerequisites, so if any of the prerequisites changes, the contents of the existing target file are no longer necessarily valid.

How to update is specified by a recipe. This is one or more lines to be executed by the shell (normally ‘sh’), but with some extra features (see Writing Recipes in Rules).

4.3 Types of Prerequisites

There are actually two different types of prerequisites understood by GNU make: normal prerequisites such as described in the previous section, and order-only prerequisites. A normal prerequisite makes two statements: first, it imposes an order in which recipes will be invoked: the recipes for all prerequisites of a target will be completed before the recipe for the target is run. Second, it imposes a dependency relationship: if any prerequisite is newer than the target, then the target is considered out-of-date and must be rebuilt.

Normally, this is exactly what you want: if a target’s prerequisite is updated, then the target should also be updated.

Occasionally, however, you have a situation where you want to impose a specific ordering on the rules to be invoked without forcing the target to be updated if one of those rules is executed. In that case, you want to define order-only prerequisites. Order-only prerequisites can be specified by placing a pipe symbol (|) in the prerequisites list: any prerequisites to the left of the pipe symbol are normal; any prerequisites to the right are order-only:

targets : normal-prerequisites | order-only-prerequisites

The normal prerequisites section may of course be empty. Also, you may still declare multiple lines of prerequisites for the same target: they are appended appropriately (normal prerequisites are appended to the list of normal prerequisites; order-only prerequisites are appended to the list of order-only prerequisites). Note that if you declare the same file to be both a normal and an order-only prerequisite, the normal prerequisite takes precedence (since they have a strict superset of the behavior of an order-only prerequisite).

Consider an example where your targets are to be placed in a separate directory, and that directory might not exist before make is run. In this situation, you want the directory to be created before any targets are placed into it but, because the timestamps on directories change whenever a file is added, removed, or renamed, we certainly don’t want to rebuild all the targets whenever the directory’s timestamp changes. One way to manage this is with order-only prerequisites: make the directory an order-only prerequisite on all the targets:

OBJDIR := objdir
OBJS := $(addprefix $(OBJDIR)/,foo.o bar.o baz.o)

$(OBJDIR)/%.o : %.c
        $(COMPILE.c) $(OUTPUT_OPTION) $<

all: $(OBJS)

$(OBJS): | $(OBJDIR)

$(OBJDIR):
        mkdir $(OBJDIR)

Now the rule to create the objdir directory will be run, if needed, before any ‘.o’ is built, but no ‘.o’ will be built because the objdir directory timestamp changed.

4.4 Using Wildcard Characters in File Names

A single file name can specify many files using wildcard characters. The wildcard characters in make are ‘*’, ‘?’ and ‘[…]’, the same as in the Bourne shell. For example, *.c specifies a list of all the files (in the working directory) whose names end in ‘.c’.

The character ‘~’ at the beginning of a file name also has special significance. If alone, or followed by a slash, it represents your home directory. For example ~/bin expands to /home/you/bin. If the ‘~’ is followed by a word, the string represents the home directory of the user named by that word. For example ~john/bin expands to /home/john/bin. On systems which don’t have a home directory for each user (such as MS-DOS or MS-Windows), this functionality can be simulated by setting the environment variable HOME.

Wildcard expansion is performed by make automatically in targets and in prerequisites. In recipes, the shell is responsible for wildcard expansion. In other contexts, wildcard expansion happens only if you request it explicitly with the wildcard function.

The special significance of a wildcard character can be turned off by preceding it with a backslash. Thus, foo\*bar would refer to a specific file whose name consists of ‘foo’, an asterisk, and ‘bar’.

4.4.1 Wildcard Examples

Wildcards can be used in the recipe of a rule, where they are expanded by the shell. For example, here is a rule to delete all the object files:

clean:
        rm -f *.o

Wildcards are also useful in the prerequisites of a rule. With the following rule in the makefile, ‘make print’ will print all the ‘.c’ files that have changed since the last time you printed them:

print: *.c
        lpr -p $?
        touch print

This rule uses print as an empty target file; see Empty Target Files to Record Events. (The automatic variable ‘$?’ is used to print only those files that have changed; see Automatic Variables.)

Wildcard expansion does not happen when you define a variable. Thus, if you write this:

objects = *.o

then the value of the variable objects is the actual string ‘*.o’. However, if you use the value of objects in a target or prerequisite, wildcard expansion will take place there. If you use the value of objects in a recipe, the shell may perform wildcard expansion when the recipe runs. To set objects to the expansion, instead use:

objects := $(wildcard *.o)

See Wildcard Function.

4.4.2 Pitfalls of Using Wildcards

Now here is an example of a naive way of using wildcard expansion, that does not do what you would intend. Suppose you would like to say that the executable file foo is made from all the object files in the directory, and you write this:

objects = *.o

foo : $(objects)
        cc -o foo $(CFLAGS) $(objects)

The value of objects is the actual string ‘*.o’. Wildcard expansion happens in the rule for foo, so that each existing ‘.o’ file becomes a prerequisite of foo and will be recompiled if necessary.

But what if you delete all the ‘.o’ files? When a wildcard matches no files, it is left as it is, so then foo will depend on the oddly-named file *.o. Since no such file is likely to exist, make will give you an error saying it cannot figure out how to make *.o. This is not what you want!

Actually it is possible to obtain the desired result with wildcard expansion, but you need more sophisticated techniques, including the wildcard function and string substitution. See The Function wildcard.

Microsoft operating systems (MS-DOS and MS-Windows) use backslashes to separate directories in pathnames, like so:

  c:\foo\bar\baz.c

This is equivalent to the Unix-style c:/foo/bar/baz.c (the c: part is the so-called drive letter). When make runs on these systems, it supports backslashes as well as the Unix-style forward slashes in pathnames. However, this support does not include the wildcard expansion, where backslash is a quote character. Therefore, you must use Unix-style slashes in these cases.

4.4.3 The Function wildcard

Wildcard expansion happens automatically in rules. But wildcard expansion does not normally take place when a variable is set, or inside the arguments of a function. If you want to do wildcard expansion in such places, you need to use the wildcard function, like this:

$(wildcard pattern…)

This string, used anywhere in a makefile, is replaced by a space-separated list of names of existing files that match one of the given file name patterns. If no existing file name matches a pattern, then that pattern is omitted from the output of the wildcard function. Note that this is different from how unmatched wildcards behave in rules, where they are used verbatim rather than ignored (see Wildcard Pitfall).

One use of the wildcard function is to get a list of all the C source files in a directory, like this:

$(wildcard *.c)

We can change the list of C source files into a list of object files by replacing the ‘.c’ suffix with ‘.o’ in the result, like this:

$(patsubst %.c,%.o,$(wildcard *.c))

(Here we have used another function, patsubst. See Functions for String Substitution and Analysis.)

Thus, a makefile to compile all C source files in the directory and then link them together could be written as follows:

objects := $(patsubst %.c,%.o,$(wildcard *.c))

foo : $(objects)
        cc -o foo $(objects)

(This takes advantage of the implicit rule for compiling C programs, so there is no need to write explicit rules for compiling the files. See The Two Flavors of Variables, for an explanation of ‘:=’, which is a variant of ‘=’.)

4.5 Searching Directories for Prerequisites

For large systems, it is often desirable to put sources in a separate directory from the binaries. The directory search features of make facilitate this by searching several directories automatically to find a prerequisite. When you redistribute the files among directories, you do not need to change the individual rules, just the search paths.

4.5.1 VPATH: Search Path for All Prerequisites

The value of the make variable VPATH specifies a list of directories that make should search. Most often, the directories are expected to contain prerequisite files that are not in the current directory; however, make uses VPATH as a search list for both prerequisites and targets of rules.

Thus, if a file that is listed as a target or prerequisite does not exist in the current directory, make searches the directories listed in VPATH for a file with that name. If a file is found in one of them, that file may become the prerequisite (see below). Rules may then specify the names of files in the prerequisite list as if they all existed in the current directory. See Writing Recipes with Directory Search.

In the VPATH variable, directory names are separated by colons or blanks. The order in which directories are listed is the order followed by make in its search. (On MS-DOS and MS-Windows, semi-colons are used as separators of directory names in VPATH, since the colon can be used in the pathname itself, after the drive letter.)

For example,

VPATH = src:../headers

specifies a path containing two directories, src and ../headers, which make searches in that order.

With this value of VPATH, the following rule,

foo.o : foo.c

is interpreted as if it were written like this:

foo.o : src/foo.c

assuming the file foo.c does not exist in the current directory but is found in the directory src.

4.5.2 The vpath Directive

Similar to the VPATH variable, but more selective, is the vpath directive (note lower case), which allows you to specify a search path for a particular class of file names: those that match a particular pattern. Thus you can supply certain search directories for one class of file names and other directories (or none) for other file names.

There are three forms of the vpath directive:

vpath pattern directories

Specify the search path directories for file names that match pattern.

The search path, directories, is a list of directories to be searched, separated by colons (semi-colons on MS-DOS and MS-Windows) or blanks, just like the search path used in the VPATH variable.

vpath pattern

Clear out the search path associated with pattern.

vpath

Clear all search paths previously specified with vpath directives.

A vpath pattern is a string containing a ‘%’ character. The string must match the file name of a prerequisite that is being searched for, the ‘%’ character matching any sequence of zero or more characters (as in pattern rules; see Defining and Redefining Pattern Rules). For example, %.h matches files that end in .h. (If there is no ‘%’, the pattern must match the prerequisite exactly, which is not useful very often.)

‘%’ characters in a vpath directive’s pattern can be quoted with preceding backslashes (‘\’). Backslashes that would otherwise quote ‘%’ characters can be quoted with more backslashes. Backslashes that quote ‘%’ characters or other backslashes are removed from the pattern before it is compared to file names. Backslashes that are not in danger of quoting ‘%’ characters go unmolested.

When a prerequisite fails to exist in the current directory, if the pattern in a vpath directive matches the name of the prerequisite file, then the directories in that directive are searched just like (and before) the directories in the VPATH variable.

For example,

vpath %.h ../headers

tells make to look for any prerequisite whose name ends in .h in the directory ../headers if the file is not found in the current directory.

If several vpath patterns match the prerequisite file’s name, then make processes each matching vpath directive one by one, searching all the directories mentioned in each directive. make handles multiple vpath directives in the order in which they appear in the makefile; multiple directives with the same pattern are independent of each other.

Thus,

vpath %.c foo
vpath %   blish
vpath %.c bar

will look for a file ending in ‘.c’ in foo, then blish, then bar, while

vpath %.c foo:bar
vpath %   blish

will look for a file ending in ‘.c’ in foo, then bar, then blish.

4.5.3 How Directory Searches are Performed

When a prerequisite is found through directory search, regardless of type (general or selective), the pathname located may not be the one that make actually provides you in the prerequisite list. Sometimes the path discovered through directory search is thrown away.

The algorithm make uses to decide whether to keep or abandon a path found via directory search is as follows:

1. If a target file does not exist at the path specified in the makefile, directory search is performed.
2. If the directory search is successful, that path is kept and this file is tentatively stored as the target.
3. All prerequisites of this target are examined using this same method.
4. After processing the prerequisites, the target may or may not need to be rebuilt:
   1. If the target does not need to be rebuilt, the path to the file found during directory search is used for any prerequisite lists which contain this target. In short, if make doesn’t need to rebuild the target then you use the path found via directory search.
   2. If the target does need to be rebuilt (is out-of-date), the pathname found during directory search is thrown away, and the target is rebuilt using the file name specified in the makefile. In short, if make must rebuild, then the target is rebuilt locally, not in the directory found via directory search.

This algorithm may seem complex, but in practice it is quite often exactly what you want.

Other versions of make use a simpler algorithm: if the file does not exist, and it is found via directory search, then that pathname is always used whether or not the target needs to be built. Thus, if the target is rebuilt it is created at the pathname discovered during directory search.

If, in fact, this is the behavior you want for some or all of your directories, you can use the GPATH variable to indicate this to make.

GPATH has the same syntax and format as VPATH (that is, a space- or colon-delimited list of pathnames). If an out-of-date target is found by directory search in a directory that also appears in GPATH, then that pathname is not thrown away. The target is rebuilt using the expanded path.

4.5.4 Writing Recipes with Directory Search

When a prerequisite is found in another directory through directory search, this cannot change the recipe of the rule; they will execute as written. Therefore, you must write the recipe with care so that it will look for the prerequisite in the directory where make finds it.

This is done with the automatic variables such as ‘$^’ (see Automatic Variables). For instance, the value of ‘$^’ is a list of all the prerequisites of the rule, including the names of the directories in which they were found, and the value of ‘$@’ is the target. Thus:

foo.o : foo.c
        cc -c $(CFLAGS) $^ -o $@

(The variable CFLAGS exists so you can specify flags for C compilation by implicit rules; we use it here for consistency so it will affect all C compilations uniformly; see Variables Used by Implicit Rules.)

Often the prerequisites include header files as well, which you do not want to mention in the recipe. The automatic variable ‘$<’ is just the first prerequisite:

VPATH = src:../headers
foo.o : foo.c defs.h hack.h
        cc -c $(CFLAGS) $< -o $@

4.5.5 Directory Search and Implicit Rules

The search through the directories specified in VPATH or with vpath also happens during consideration of implicit rules (see Using Implicit Rules).

For example, when a file foo.o has no explicit rule, make considers implicit rules, such as the built-in rule to compile foo.c if that file exists. If such a file is lacking in the current directory, the appropriate directories are searched for it. If foo.c exists (or is mentioned in the makefile) in any of the directories, the implicit rule for C compilation is applied.

The recipes of implicit rules normally use automatic variables as a matter of necessity; consequently they will use the file names found by directory search with no extra effort.

4.5.6 Directory Search for Link Libraries

Directory search applies in a special way to libraries used with the linker. This special feature comes into play when you write a prerequisite whose name is of the form ‘-lname’. (You can tell something strange is going on here because the prerequisite is normally the name of a file, and the file name of a library generally looks like libname.a, not like ‘-lname’.)

When a prerequisite’s name has the form ‘-lname’, make handles it specially by searching for the file libname.so, and, if it is not found, for the file libname.a in the current directory, in directories specified by matching vpath search paths and the VPATH search path, and then in the directories /lib, /usr/lib, and prefix/lib (normally /usr/local/lib, but MS-DOS/MS-Windows versions of make behave as if prefix is defined to be the root of the DJGPP installation tree).

For example, if there is a /usr/lib/libcurses.a library on your system (and no /usr/lib/libcurses.so file), then

foo : foo.c -lcurses
        cc $^ -o $@

would cause the command ‘cc foo.c /usr/lib/libcurses.a -o foo’ to be executed when foo is older than foo.c or than /usr/lib/libcurses.a.

Although the default set of files to be searched for is libname.so and libname.a, this is customizable via the .LIBPATTERNS variable. Each word in the value of this variable is a pattern string. When a prerequisite like ‘-lname’ is seen, make will replace the percent in each pattern in the list with name and perform the above directory searches using each library file name.

The default value for .LIBPATTERNS is ‘lib%.so lib%.a’, which provides the default behavior described above.

You can turn off link library expansion completely by setting this variable to an empty value.

4.6 Phony Targets

A phony target is one that is not really the name of a file; rather it is just a name for a recipe to be executed when you make an explicit request. There are two reasons to use a phony target: to avoid a conflict with a file of the same name, and to improve performance.

If you write a rule whose recipe will not create the target file, the recipe will be executed every time the target comes up for remaking. Here is an example:

clean:
        rm *.o temp

Because the rm command does not create a file named clean, probably no such file will ever exist. Therefore, the rm command will be executed every time you say ‘make clean’.

In this example, the clean target will not work properly if a file named clean is ever created in this directory. Since it has no prerequisites, clean would always be considered up to date and its recipe would not be executed. To avoid this problem you can explicitly declare the target to be phony by making it a prerequisite of the special target .PHONY (see Special Built-in Target Names) as follows:

.PHONY: clean
clean:
        rm *.o temp

Once this is done, ‘make clean’ will run the recipe regardless of whether there is a file named clean.

Phony targets are also useful in conjunction with recursive invocations of make (see Recursive Use of make). In this situation the makefile will often contain a variable which lists a number of sub-directories to be built. A simplistic way to handle this is to define one rule with a recipe that loops over the sub-directories, like this:

SUBDIRS = foo bar baz

subdirs:
        for dir in $(SUBDIRS); do \
          $(MAKE) -C $$dir; \
        done

There are problems with this method, however. First, any error detected in a sub-make is ignored by this rule, so it will continue to build the rest of the directories even when one fails. This can be overcome by adding shell commands to note the error and exit, but then it will do so even if make is invoked with the -k option, which is unfortunate. Second, and perhaps more importantly, you cannot take advantage of make’s ability to build targets in parallel (see Parallel Execution), since there is only one rule.

By declaring the sub-directories as .PHONY targets (you must do this as the sub-directory obviously always exists; otherwise it won’t be built) you can remove these problems:

SUBDIRS = foo bar baz

.PHONY: subdirs $(SUBDIRS)

subdirs: $(SUBDIRS)

$(SUBDIRS):
        $(MAKE) -C $@

foo: baz

Here we’ve also declared that the foo sub-directory cannot be built until after the baz sub-directory is complete; this kind of relationship declaration is particularly important when attempting parallel builds.

The implicit rule search (see Implicit Rules) is skipped for .PHONY targets. This is why declaring a target as .PHONY is good for performance, even if you are not worried about the actual file existing.

A phony target should not be a prerequisite of a real target file; if it is, its recipe will be run every time make goes to update that file. As long as a phony target is never a prerequisite of a real target, the phony target recipe will be executed only when the phony target is a specified goal (see Arguments to Specify the Goals).

Phony targets can have prerequisites. When one directory contains multiple programs, it is most convenient to describe all of the programs in one makefile ./Makefile. Since the target remade by default will be the first one in the makefile, it is common to make this a phony target named ‘all’ and give it, as prerequisites, all the individual programs. For example:

all : prog1 prog2 prog3
.PHONY : all

prog1 : prog1.o utils.o
        cc -o prog1 prog1.o utils.o

prog2 : prog2.o
        cc -o prog2 prog2.o

prog3 : prog3.o sort.o utils.o
        cc -o prog3 prog3.o sort.o utils.o

Now you can say just ‘make’ to remake all three programs, or specify as arguments the ones to remake (as in ‘make prog1 prog3’). Phoniness is not inherited: the prerequisites of a phony target are not themselves phony, unless explicitly declared to be so.

When one phony target is a prerequisite of another, it serves as a subroutine of the other. For example, here ‘make cleanall’ will delete the object files, the difference files, and the file program:

.PHONY: cleanall cleanobj cleandiff

cleanall : cleanobj cleandiff
        rm program

cleanobj :
        rm *.o

cleandiff :
        rm *.diff

4.7 Rules without Recipes or Prerequisites

If a rule has no prerequisites or recipe, and the target of the rule is a nonexistent file, then make imagines this target to have been updated whenever its rule is run. This implies that all targets depending on this one will always have their recipe run.

An example will illustrate this:

clean: FORCE
        rm $(objects)
FORCE:

Here the target ‘FORCE’ satisfies the special conditions, so the target clean that depends on it is forced to run its recipe. There is nothing special about the name ‘FORCE’, but that is one name commonly used this way.

As you can see, using ‘FORCE’ this way has the same results as using ‘.PHONY: clean’.

Using ‘.PHONY’ is more explicit and more efficient. However, other versions of make do not support ‘.PHONY’; thus ‘FORCE’ appears in many makefiles. See Phony Targets.

4.8 Empty Target Files to Record Events

The empty target is a variant of the phony target; it is used to hold recipes for an action that you request explicitly from time to time. Unlike a phony target, this target file can really exist; but the file’s contents do not matter, and usually are empty.

The purpose of the empty target file is to record, with its last-modification time, when the rule’s recipe was last executed. It does so because one of the commands in the recipe is a touch command to update the target file.

The empty target file should have some prerequisites (otherwise it doesn’t make sense). When you ask to remake the empty target, the recipe is executed if any prerequisite is more recent than the target; in other words, if a prerequisite has changed since the last time you remade the target. Here is an example:

print: foo.c bar.c
        lpr -p $?
        touch print

With this rule, ‘make print’ will execute the lpr command if either source file has changed since the last ‘make print’. The automatic variable ‘$?’ is used to print only those files that have changed (see Automatic Variables).

4.9 Special Built-in Target Names

Certain names have special meanings if they appear as targets.

.PHONY

The prerequisites of the special target .PHONY are considered to be phony targets. When it is time to consider such a target, make will run its recipe unconditionally, regardless of whether a file with that name exists or what its last-modification time is. See Phony Targets.

.SUFFIXES

The prerequisites of the special target .SUFFIXES are the list of suffixes to be used in checking for suffix rules. See Old-Fashioned Suffix Rules.

.DEFAULT

The recipe specified for .DEFAULT is used for any target for which no rules are found (either explicit rules or implicit rules). See Last Resort. If a .DEFAULT recipe is specified, every file mentioned as a prerequisite, but not as a target in a rule, will have that recipe executed on its behalf. See Implicit Rule Search Algorithm.

.PRECIOUS

The targets which .PRECIOUS depends on are given the following special treatment: if make is killed or interrupted during the execution of their recipes, the target is not deleted. See Interrupting or Killing make. Also, if the target is an intermediate file, it will not be deleted after it is no longer needed, as is normally done. See Chains of Implicit Rules. In this latter respect it overlaps with the .SECONDARY special target.

You can also list the target pattern of an implicit rule (such as ‘%.o’) as a prerequisite file of the special target .PRECIOUS to preserve intermediate files created by rules whose target patterns match that file’s name.

.INTERMEDIATE

The targets which .INTERMEDIATE depends on are treated as intermediate files. See Chains of Implicit Rules. .INTERMEDIATE with no prerequisites has no effect.

.SECONDARY

The targets which .SECONDARY depends on are treated as intermediate files, except that they are never automatically deleted. See Chains of Implicit Rules.

.SECONDARY with no prerequisites causes all targets to be treated as secondary (i.e., no target is removed because it is considered intermediate).

.SECONDEXPANSION

If .SECONDEXPANSION is mentioned as a target anywhere in the makefile, then all prerequisite lists defined after it appears will be expanded a second time after all makefiles have been read in. See Secondary Expansion.

.DELETE_ON_ERROR

If .DELETE_ON_ERROR is mentioned as a target anywhere in the makefile, then make will delete the target of a rule if it has changed and its recipe exits with a nonzero exit status, just as it does when it receives a signal. See Errors in Recipes.

.IGNORE

If you specify prerequisites for .IGNORE, then make will ignore errors in execution of the recipe for those particular files. The recipe for .IGNORE (if any) is ignored.

If mentioned as a target with no prerequisites, .IGNORE says to ignore errors in execution of recipes for all files. This usage of ‘.IGNORE’ is supported only for historical compatibility. Since this affects every recipe in the makefile, it is not very useful; we recommend you use the more selective ways to ignore errors in specific recipes. See Errors in Recipes.

.LOW_RESOLUTION_TIME

If you specify prerequisites for .LOW_RESOLUTION_TIME, make assumes that these files are created by commands that generate low resolution time stamps. The recipe for the .LOW_RESOLUTION_TIME target are ignored.

The high resolution file time stamps of many modern file systems lessen the chance of make incorrectly concluding that a file is up to date. Unfortunately, some hosts do not provide a way to set a high resolution file time stamp, so commands like ‘cp -p’ that explicitly set a file’s time stamp must discard its sub-second part. If a file is created by such a command, you should list it as a prerequisite of .LOW_RESOLUTION_TIME so that make does not mistakenly conclude that the file is out of date. For example:

.LOW_RESOLUTION_TIME: dst
dst: src
        cp -p src dst

Since ‘cp -p’ discards the sub-second part of src’s time stamp, dst is typically slightly older than src even when it is up to date. The .LOW_RESOLUTION_TIME line causes make to consider dst to be up to date if its time stamp is at the start of the same second that src’s time stamp is in.

Due to a limitation of the archive format, archive member time stamps are always low resolution. You need not list archive members as prerequisites of .LOW_RESOLUTION_TIME, as make does this automatically.

.SILENT

If you specify prerequisites for .SILENT, then make will not print the recipe used to remake those particular files before executing them. The recipe for .SILENT is ignored.

If mentioned as a target with no prerequisites, .SILENT says not to print any recipes before executing them. This usage of ‘.SILENT’ is supported only for historical compatibility. We recommend you use the more selective ways to silence specific recipes. See Recipe Echoing. If you want to silence all recipes for a particular run of make, use the ‘-s’ or ‘--silent’ option (see Options Summary).

.EXPORT_ALL_VARIABLES

Simply by being mentioned as a target, this tells make to export all variables to child processes by default. See Communicating Variables to a Sub-make.

.NOTPARALLEL

If .NOTPARALLEL is mentioned as a target, then this invocation of make will be run serially, even if the ‘-j’ option is given. Any recursively invoked make command will still run recipes in parallel (unless its makefile also contains this target). Any prerequisites on this target are ignored.

.ONESHELL

If .ONESHELL is mentioned as a target, then when a target is built all lines of the recipe will be given to a single invocation of the shell rather than each line being invoked separately (see Recipe Execution).

.POSIX

If .POSIX is mentioned as a target, then the makefile will be parsed and run in POSIX-conforming mode. This does not mean that only POSIX-conforming makefiles will be accepted: all advanced GNU make features are still available. Rather, this target causes make to behave as required by POSIX in those areas where make’s default behavior differs.

In particular, if this target is mentioned then recipes will be invoked as if the shell had been passed the -e flag: the first failing command in a recipe will cause the recipe to fail immediately.

Any defined implicit rule suffix also counts as a special target if it appears as a target, and so does the concatenation of two suffixes, such as ‘.c.o’. These targets are suffix rules, an obsolete way of defining implicit rules (but a way still widely used). In principle, any target name could be special in this way if you break it in two and add both pieces to the suffix list. In practice, suffixes normally begin with ‘.’, so these special target names also begin with ‘.’. See Old-Fashioned Suffix Rules.

4.10 Multiple Targets in a Rule

A rule with multiple targets is equivalent to writing many rules, each with one target, and all identical aside from that. The same recipe applies to all the targets, but its effect may vary because you can substitute the actual target name into the recipe using ‘$@’. The rule contributes the same prerequisites to all the targets also.

This is useful in two cases.

• You want just prerequisites, no recipe. For example:

  kbd.o command.o files.o: command.h
  

  gives an additional prerequisite to each of the three object files mentioned.

• Similar recipes work for all the targets. The recipes do not need to be absolutely identical, since the automatic variable ‘$@’ can be used to substitute the particular target to be remade into the commands (see Automatic Variables). For example:

  bigoutput littleoutput : text.g
          generate text.g -$(subst output,,$@) > $@
  

  is equivalent to

  bigoutput : text.g
          generate text.g -big > bigoutput
  littleoutput : text.g
          generate text.g -little > littleoutput
  

  Here we assume the hypothetical program generate makes two types of output, one if given ‘-big’ and one if given ‘-little’. See Functions for String Substitution and Analysis, for an explanation of the subst function.

Suppose you would like to vary the prerequisites according to the target, much as the variable ‘$@’ allows you to vary the recipe. You cannot do this with multiple targets in an ordinary rule, but you can do it with a static pattern rule. See Static Pattern Rules.

4.11 Multiple Rules for One Target

One file can be the target of several rules. All the prerequisites mentioned in all the rules are merged into one list of prerequisites for the target. If the target is older than any prerequisite from any rule, the recipe is executed.

There can only be one recipe to be executed for a file. If more than one rule gives a recipe for the same file, make uses the last one given and prints an error message. (As a special case, if the file’s name begins with a dot, no error message is printed. This odd behavior is only for compatibility with other implementations of make… you should avoid using it). Occasionally it is useful to have the same target invoke multiple recipes which are defined in different parts of your makefile; you can use double-colon rules (see Double-Colon) for this.

An extra rule with just prerequisites can be used to give a few extra prerequisites to many files at once. For example, makefiles often have a variable, such as objects, containing a list of all the compiler output files in the system being made. An easy way to say that all of them must be recompiled if config.h changes is to write the following:

objects = foo.o bar.o
foo.o : defs.h
bar.o : defs.h test.h
$(objects) : config.h

This could be inserted or taken out without changing the rules that really specify how to make the object files, making it a convenient form to use if you wish to add the additional prerequisite intermittently.

Another wrinkle is that the additional prerequisites could be specified with a variable that you set with a command line argument to make (see Overriding Variables). For example,

extradeps=
$(objects) : $(extradeps)

means that the command ‘make extradeps=foo.h’ will consider foo.h as a prerequisite of each object file, but plain ‘make’ will not.

If none of the explicit rules for a target has a recipe, then make searches for an applicable implicit rule to find one see Using Implicit Rules).

4.12 Static Pattern Rules

Static pattern rules are rules which specify multiple targets and construct the prerequisite names for each target based on the target name. They are more general than ordinary rules with multiple targets because the targets do not have to have identical prerequisites. Their prerequisites must be analogous, but not necessarily identical.

4.12.1 Syntax of Static Pattern Rules

Here is the syntax of a static pattern rule:

targets …: target-pattern: prereq-patterns …
        recipe
        …

The targets list specifies the targets that the rule applies to. The targets can contain wildcard characters, just like the targets of ordinary rules (see Using Wildcard Characters in File Names).

The target-pattern and prereq-patterns say how to compute the prerequisites of each target. Each target is matched against the target-pattern to extract a part of the target name, called the stem. This stem is substituted into each of the prereq-patterns to make the prerequisite names (one from each prereq-pattern).

Each pattern normally contains the character ‘%’ just once. When the target-pattern matches a target, the ‘%’ can match any part of the target name; this part is called the stem. The rest of the pattern must match exactly. For example, the target foo.o matches the pattern ‘%.o’, with ‘foo’ as the stem. The targets foo.c and foo.out do not match that pattern.

The prerequisite names for each target are made by substituting the stem for the ‘%’ in each prerequisite pattern. For example, if one prerequisite pattern is %.c, then substitution of the stem ‘foo’ gives the prerequisite name foo.c. It is legitimate to write a prerequisite pattern that does not contain ‘%’; then this prerequisite is the same for all targets.

‘%’ characters in pattern rules can be quoted with preceding backslashes (‘\’). Backslashes that would otherwise quote ‘%’ characters can be quoted with more backslashes. Backslashes that quote ‘%’ characters or other backslashes are removed from the pattern before it is compared to file names or has a stem substituted into it. Backslashes that are not in danger of quoting ‘%’ characters go unmolested. For example, the pattern the\%weird\\%pattern\\ has ‘the%weird\’ preceding the operative ‘%’ character, and ‘pattern\\’ following it. The final two backslashes are left alone because they cannot affect any ‘%’ character.

Here is an example, which compiles each of foo.o and bar.o from the corresponding .c file:

objects = foo.o bar.o

all: $(objects)

$(objects): %.o: %.c
        $(CC) -c $(CFLAGS) $< -o $@

Here ‘$<’ is the automatic variable that holds the name of the prerequisite and ‘$@’ is the automatic variable that holds the name of the target; see Automatic Variables.

Each target specified must match the target pattern; a warning is issued for each target that does not. If you have a list of files, only some of which will match the pattern, you can use the filter function to remove non-matching file names (see Functions for String Substitution and Analysis):

files = foo.elc bar.o lose.o

$(filter %.o,$(files)): %.o: %.c
        $(CC) -c $(CFLAGS) $< -o $@
$(filter %.elc,$(files)): %.elc: %.el
        emacs -f batch-byte-compile $<

In this example the result of ‘$(filter %.o,$(files))’ is bar.o lose.o, and the first static pattern rule causes each of these object files to be updated by compiling the corresponding C source file. The result of ‘$(filter %.elc,$(files))’ is foo.elc, so that file is made from foo.el.

Another example shows how to use $* in static pattern rules:

bigoutput littleoutput : %output : text.g
        generate text.g -$* > $@

When the generate command is run, $* will expand to the stem, either ‘big’ or ‘little’.

4.12.2 Static Pattern Rules versus Implicit Rules

A static pattern rule has much in common with an implicit rule defined as a pattern rule (see Defining and Redefining Pattern Rules). Both have a pattern for the target and patterns for constructing the names of prerequisites. The difference is in how make decides when the rule applies.

An implicit rule can apply to any target that matches its pattern, but it does apply only when the target has no recipe otherwise specified, and only when the prerequisites can be found. If more than one implicit rule appears applicable, only one applies; the choice depends on the order of rules.

By contrast, a static pattern rule applies to the precise list of targets that you specify in the rule. It cannot apply to any other target and it invariably does apply to each of the targets specified. If two conflicting rules apply, and both have recipes, that’s an error.

The static pattern rule can be better than an implicit rule for these reasons:

• You may wish to override the usual implicit rule for a few files whose names cannot be categorized syntactically but can be given in an explicit list.
• If you cannot be sure of the precise contents of the directories you are using, you may not be sure which other irrelevant files might lead make to use the wrong implicit rule. The choice might depend on the order in which the implicit rule search is done. With static pattern rules, there is no uncertainty: each rule applies to precisely the targets specified.

4.13 Double-Colon Rules

Double-colon rules are explicit rules written with ‘::’ instead of ‘:’ after the target names. They are handled differently from ordinary rules when the same target appears in more than one rule. Pattern rules with double-colons have an entirely different meaning (see Match-Anything Rules).

When a target appears in multiple rules, all the rules must be the same type: all ordinary, or all double-colon. If they are double-colon, each of them is independent of the others. Each double-colon rule’s recipe is executed if the target is older than any prerequisites of that rule. If there are no prerequisites for that rule, its recipe is always executed (even if the target already exists). This can result in executing none, any, or all of the double-colon rules.

Double-colon rules with the same target are in fact completely separate from one another. Each double-colon rule is processed individually, just as rules with different targets are processed.

The double-colon rules for a target are executed in the order they appear in the makefile. However, the cases where double-colon rules really make sense are those where the order of executing the recipes would not matter.

Double-colon rules are somewhat obscure and not often very useful; they provide a mechanism for cases in which the method used to update a target differs depending on which prerequisite files caused the update, and such cases are rare.

Each double-colon rule should specify a recipe; if it does not, an implicit rule will be used if one applies. See Using Implicit Rules.

4.14 Generating Prerequisites Automatically

In the makefile for a program, many of the rules you need to write often say only that some object file depends on some header file. For example, if main.c uses defs.h via an #include, you would write:

main.o: defs.h

You need this rule so that make knows that it must remake main.o whenever defs.h changes. You can see that for a large program you would have to write dozens of such rules in your makefile. And, you must always be very careful to update the makefile every time you add or remove an #include.

To avoid this hassle, most modern C compilers can write these rules for you, by looking at the #include lines in the source files. Usually this is done with the ‘-M’ option to the compiler. For example, the command:

cc -M main.c

generates the output:

main.o : main.c defs.h

Thus you no longer have to write all those rules yourself. The compiler will do it for you.

Note that such a rule constitutes mentioning main.o in a makefile, so it can never be considered an intermediate file by implicit rule search. This means that make won’t ever remove the file after using it; see Chains of Implicit Rules.

With old make programs, it was traditional practice to use this compiler feature to generate prerequisites on demand with a command like ‘make depend’. That command would create a file depend containing all the automatically-generated prerequisites; then the makefile could use include to read them in (see Include).

In GNU make, the feature of remaking makefiles makes this practice obsolete—you need never tell make explicitly to regenerate the prerequisites, because it always regenerates any makefile that is out of date. See Remaking Makefiles.

The practice we recommend for automatic prerequisite generation is to have one makefile corresponding to each source file. For each source file name.c there is a makefile name.d which lists what files the object file name.o depends on. That way only the source files that have changed need to be rescanned to produce the new prerequisites.

Here is the pattern rule to generate a file of prerequisites (i.e., a makefile) called name.d from a C source file called name.c:

%.d: %.c
        @set -e; rm -f $@; \
         $(CC) -M $(CPPFLAGS) $< > $@.$$$$; \
         sed 's,\($*\)\.o[ :]*,\1.o $@ : ,g' < $@.$$$$ > $@; \
         rm -f $@.$$$$

See Pattern Rules, for information on defining pattern rules. The ‘-e’ flag to the shell causes it to exit immediately if the $(CC) command (or any other command) fails (exits with a nonzero status).

With the GNU C compiler, you may wish to use the ‘-MM’ flag instead of ‘-M’. This omits prerequisites on system header files. See Options Controlling the Preprocessor in Using GNU CC, for details.

The purpose of the sed command is to translate (for example):

main.o : main.c defs.h

into:

main.o main.d : main.c defs.h

This makes each ‘.d’ file depend on all the source and header files that the corresponding ‘.o’ file depends on. make then knows it must regenerate the prerequisites whenever any of the source or header files changes.

Once you’ve defined the rule to remake the ‘.d’ files, you then use the include directive to read them all in. See Include. For example:

sources = foo.c bar.c

include $(sources:.c=.d)

(This example uses a substitution variable reference to translate the list of source files ‘foo.c bar.c’ into a list of prerequisite makefiles, ‘foo.d bar.d’. See Substitution Refs, for full information on substitution references.) Since the ‘.d’ files are makefiles like any others, make will remake them as necessary with no further work from you. See Remaking Makefiles.

Note that the ‘.d’ files contain target definitions; you should be sure to place the include directive after the first, default goal in your makefiles or run the risk of having a random object file become the default goal. See How Make Works.

5 Writing Recipes in Rules

The recipe of a rule consists of one or more shell command lines to be executed, one at a time, in the order they appear. Typically, the result of executing these commands is that the target of the rule is brought up to date.

Users use many different shell programs, but recipes in makefiles are always interpreted by /bin/sh unless the makefile specifies otherwise. See Recipe Execution.

5.1 Recipe Syntax

Makefiles have the unusual property that there are really two distinct syntaxes in one file. Most of the makefile uses make syntax (see Writing Makefiles). However, recipes are meant to be interpreted by the shell and so they are written using shell syntax. The make program does not try to understand shell syntax: it performs only a very few specific translations on the content of the recipe before handing it to the shell.

Each line in the recipe must start with a tab (or the first character in the value of the .RECIPEPREFIX variable; see Special Variables), except that the first recipe line may be attached to the target-and-prerequisites line with a semicolon in between. Any line in the makefile that begins with a tab and appears in a “rule context” (that is, after a rule has been started until another rule or variable definition) will be considered part of a recipe for that rule. Blank lines and lines of just comments may appear among the recipe lines; they are ignored.

Some consequences of these rules include:

• A blank line that begins with a tab is not blank: it’s an empty recipe (see Empty Recipes).
• A comment in a recipe is not a make comment; it will be passed to the shell as-is. Whether the shell treats it as a comment or not depends on your shell.
• A variable definition in a “rule context” which is indented by a tab as the first character on the line, will be considered part of a recipe, not a make variable definition, and passed to the shell.
• A conditional expression (ifdef, ifeq, etc. see Syntax of Conditionals) in a “rule context” which is indented by a tab as the first character on the line, will be considered part of a recipe and be passed to the shell.

5.1.1 Splitting Recipe Lines

One of the few ways in which make does interpret recipes is checking for a backslash just before the newline. As in normal makefile syntax, a single logical recipe line can be split into multiple physical lines in the makefile by placing a backslash before each newline. A sequence of lines like this is considered a single recipe line, and one instance of the shell will be invoked to run it.

However, in contrast to how they are treated in other places in a makefile (see Splitting Long Lines), backslash/newline pairs are not removed from the recipe. Both the backslash and the newline characters are preserved and passed to the shell. How the backslash/newline is interpreted depends on your shell. If the first character of the next line after the backslash/newline is the recipe prefix character (a tab by default; see Special Variables), then that character (and only that character) is removed. Whitespace is never added to the recipe.

For example, the recipe for the all target in this makefile:

all :
        @echo no\
space
        @echo no\
        space
        @echo one \
        space
        @echo one\
         space

consists of four separate shell commands where the output is:

nospace
nospace
one space
one space

As a more complex example, this makefile:

all : ; @echo 'hello \
        world' ; echo "hello \
    world"

will invoke one shell with a command of:

echo 'hello \
world' ; echo "hello \
    world"

which, according to shell quoting rules, will yield the following output:

hello \
world
hello     world

Notice how the backslash/newline pair was removed inside the string quoted with double quotes ("…"), but not from the string quoted with single quotes ('…'). This is the way the default shell (/bin/sh) handles backslash/newline pairs. If you specify a different shell in your makefiles it may treat them differently.

Sometimes you want to split a long line inside of single quotes, but you don’t want the backslash/newline to appear in the quoted content. This is often the case when passing scripts to languages such as Perl, where extraneous backslashes inside the script can change its meaning or even be a syntax error. One simple way of handling this is to place the quoted string, or even the entire command, into a make variable then use the variable in the recipe. In this situation the newline quoting rules for makefiles will be used, and the backslash/newline will be removed. If we rewrite our example above using this method:

HELLO = 'hello \
world'

all : ; @echo $(HELLO)

we will get output like this:

hello world

If you like, you can also use target-specific variables (see Target-specific Variable Values) to obtain a tighter correspondence between the variable and the recipe that uses it.

5.1.2 Using Variables in Recipes

The other way in which make processes recipes is by expanding any variable references in them (see Basics of Variable References). This occurs after make has finished reading all the makefiles and the target is determined to be out of date; so, the recipes for targets which are not rebuilt are never expanded.

Variable and function references in recipes have identical syntax and semantics to references elsewhere in the makefile. They also have the same quoting rules: if you want a dollar sign to appear in your recipe, you must double it (‘$$’). For shells like the default shell, that use dollar signs to introduce variables, it’s important to keep clear in your mind whether the variable you want to reference is a make variable (use a single dollar sign) or a shell variable (use two dollar signs). For example:

LIST = one two three
all:
        for i in $(LIST); do \
            echo $$i; \
        done

results in the following command being passed to the shell:

for i in one two three; do \
    echo $i; \
done

which generates the expected result:

one
two
three

5.2 Recipe Echoing

Normally make prints each line of the recipe before it is executed. We call this echoing because it gives the appearance that you are typing the lines yourself.

When a line starts with ‘@’, the echoing of that line is suppressed. The ‘@’ is discarded before the line is passed to the shell. Typically you would use this for a command whose only effect is to print something, such as an echo command to indicate progress through the makefile:

@echo About to make distribution files

When make is given the flag ‘-n’ or ‘--just-print’ it only echoes most recipes, without executing them. See Summary of Options. In this case even the recipe lines starting with ‘@’ are printed. This flag is useful for finding out which recipes make thinks are necessary without actually doing them.

The ‘-s’ or ‘--silent’ flag to make prevents all echoing, as if all recipes started with ‘@’. A rule in the makefile for the special target .SILENT without prerequisites has the same effect (see Special Built-in Target Names). .SILENT is essentially obsolete since ‘@’ is more flexible.

5.3 Recipe Execution

When it is time to execute recipes to update a target, they are executed by invoking a new sub-shell for each line of the recipe, unless the .ONESHELL special target is in effect (see Using One Shell) (In practice, make may take shortcuts that do not affect the results.)

Please note: this implies that setting shell variables and invoking shell commands such as cd that set a context local to each process will not affect the following lines in the recipe.² If you want to use cd to affect the next statement, put both statements in a single recipe line. Then make will invoke one shell to run the entire line, and the shell will execute the statements in sequence. For example:

foo : bar/lose
        cd $(@D) && gobble $(@F) > ../$@

Here we use the shell AND operator (&&) so that if the cd command fails, the script will fail without trying to invoke the gobble command in the wrong directory, which could cause problems (in this case it would certainly cause ../foo to be truncated, at least).

5.3.1 Using One Shell

Sometimes you would prefer that all the lines in the recipe be passed to a single invocation of the shell. There are generally two situations where this is useful: first, it can improve performance in makefiles where recipes consist of many command lines, by avoiding extra processes. Second, you might want newlines to be included in your recipe command (for example perhaps you are using a very different interpreter as your SHELL). If the .ONESHELL special target appears anywhere in the makefile then all recipe lines for each target will be provided to a single invocation of the shell. Newlines between recipe lines will be preserved. For example:

.ONESHELL:
foo : bar/lose
        cd $(@D)
        gobble $(@F) > ../$@

would now work as expected even though the commands are on different recipe lines.

If .ONESHELL is provided, then only the first line of the recipe will be checked for the special prefix characters (‘@’, ‘-’, and ‘+’). Subsequent lines will include the special characters in the recipe line when the SHELL is invoked. If you want your recipe to start with one of these special characters you’ll need to arrange for them to not be the first characters on the first line, perhaps by adding a comment or similar. For example, this would be a syntax error in Perl because the first ‘@’ is removed by make:

.ONESHELL:
SHELL = /usr/bin/perl
.SHELLFLAGS = -e
show :
        @f = qw(a b c);
        print "@f\n";

However, either of these alternatives would work properly:

.ONESHELL:
SHELL = /usr/bin/perl
.SHELLFLAGS = -e
show :
        # Make sure "@" is not the first character on the first line
        @f = qw(a b c);
        print "@f\n";

or

.ONESHELL:
SHELL = /usr/bin/perl
.SHELLFLAGS = -e
show :
        my @f = qw(a b c);
        print "@f\n";

As a special feature, if SHELL is determined to be a POSIX-style shell, the special prefix characters in “internal” recipe lines will removed before the recipe is processed. This feature is intended to allow existing makefiles to add the .ONESHELL special target and still run properly without extensive modifications. Since the special prefix characters are not legal at the beginning of a line in a POSIX shell script this is not a loss in functionality. For example, this works as expected:

.ONESHELL:
foo : bar/lose
        @cd $(@D)
        @gobble $(@F) > ../$@

Even with this special feature, however, makefiles with .ONESHELL will behave differently in ways that could be noticeable. For example, normally if any line in the recipe fails, that causes the rule to fail and no more recipe lines are processed. Under .ONESHELL a failure of any but the final recipe line will not be noticed by make. You can modify .SHELLFLAGS to add the -e option to the shell which will cause any failure anywhere in the command line to cause the shell to fail, but this could itself cause your recipe to behave differently. Ultimately you may need to harden your recipe lines to allow them to work with .ONESHELL.

5.3.2 Choosing the Shell

The program used as the shell is taken from the variable SHELL. If this variable is not set in your makefile, the program /bin/sh is used as the shell. The argument(s) passed to the shell are taken from the variable .SHELLFLAGS. The default value of .SHELLFLAGS is -c normally, or -ec in POSIX-conforming mode.

Unlike most variables, the variable SHELL is never set from the environment. This is because the SHELL environment variable is used to specify your personal choice of shell program for interactive use. It would be very bad for personal choices like this to affect the functioning of makefiles. See Variables from the Environment.

Furthermore, when you do set SHELL in your makefile that value is not exported in the environment to recipe lines that make invokes. Instead, the value inherited from the user’s environment, if any, is exported. You can override this behavior by explicitly exporting SHELL (see Communicating Variables to a Sub-make), forcing it to be passed in the environment to recipe lines.

However, on MS-DOS and MS-Windows the value of SHELL in the environment is used, since on those systems most users do not set this variable, and therefore it is most likely set specifically to be used by make. On MS-DOS, if the setting of SHELL is not suitable for make, you can set the variable MAKESHELL to the shell that make should use; if set it will be used as the shell instead of the value of SHELL.

Choosing a Shell in DOS and Windows

Choosing a shell in MS-DOS and MS-Windows is much more complex than on other systems.

On MS-DOS, if SHELL is not set, the value of the variable COMSPEC (which is always set) is used instead.

The processing of lines that set the variable SHELL in Makefiles is different on MS-DOS. The stock shell, command.com, is ridiculously limited in its functionality and many users of make tend to install a replacement shell. Therefore, on MS-DOS, make examines the value of SHELL, and changes its behavior based on whether it points to a Unix-style or DOS-style shell. This allows reasonable functionality even if SHELL points to command.com.

If SHELL points to a Unix-style shell, make on MS-DOS additionally checks whether that shell can indeed be found; if not, it ignores the line that sets SHELL. In MS-DOS, GNU make searches for the shell in the following places:

1. In the precise place pointed to by the value of SHELL. For example, if the makefile specifies ‘SHELL = /bin/sh’, make will look in the directory /bin on the current drive.
2. In the current directory.
3. In each of the directories in the PATH variable, in order.

In every directory it examines, make will first look for the specific file (sh in the example above). If this is not found, it will also look in that directory for that file with one of the known extensions which identify executable files. For example .exe, .com, .bat, .btm, .sh, and some others.

If any of these attempts is successful, the value of SHELL will be set to the full pathname of the shell as found. However, if none of these is found, the value of SHELL will not be changed, and thus the line that sets it will be effectively ignored. This is so make will only support features specific to a Unix-style shell if such a shell is actually installed on the system where make runs.

Note that this extended search for the shell is limited to the cases where SHELL is set from the Makefile; if it is set in the environment or command line, you are expected to set it to the full pathname of the shell, exactly as things are on Unix.

The effect of the above DOS-specific processing is that a Makefile that contains ‘SHELL = /bin/sh’ (as many Unix makefiles do), will work on MS-DOS unaltered if you have e.g. sh.exe installed in some directory along your PATH.

5.4 Parallel Execution

GNU make knows how to execute several recipes at once. Normally, make will execute only one recipe at a time, waiting for it to finish before executing the next. However, the ‘-j’ or ‘--jobs’ option tells make to execute many recipes simultaneously. You can inhibit parallelism in a particular makefile with the .NOTPARALLEL pseudo-target (see Special Built-in Target Names).

On MS-DOS, the ‘-j’ option has no effect, since that system doesn’t support multi-processing.

If the ‘-j’ option is followed by an integer, this is the number of recipes to execute at once; this is called the number of job slots. If there is nothing looking like an integer after the ‘-j’ option, there is no limit on the number of job slots. The default number of job slots is one, which means serial execution (one thing at a time).

Handling recursive make invocations raises issues for parallel execution. For more information on this, see Communicating Options to a Sub-make.

If a recipe fails (is killed by a signal or exits with a nonzero status), and errors are not ignored for that recipe (see Errors in Recipes), the remaining recipe lines to remake the same target will not be run. If a recipe fails and the ‘-k’ or ‘--keep-going’ option was not given (see Summary of Options), make aborts execution. If make terminates for any reason (including a signal) with child processes running, it waits for them to finish before actually exiting.

When the system is heavily loaded, you will probably want to run fewer jobs than when it is lightly loaded. You can use the ‘-l’ option to tell make to limit the number of jobs to run at once, based on the load average. The ‘-l’ or ‘--max-load’ option is followed by a floating-point number. For example,

-l 2.5

will not let make start more than one job if the load average is above 2.5. The ‘-l’ option with no following number removes the load limit, if one was given with a previous ‘-l’ option.

More precisely, when make goes to start up a job, and it already has at least one job running, it checks the current load average; if it is not lower than the limit given with ‘-l’, make waits until the load average goes below that limit, or until all the other jobs finish.

By default, there is no load limit.

5.4.1 Output During Parallel Execution

When running several recipes in parallel the output from each recipe appears as soon as it is generated, with the result that messages from different recipes may be interspersed, sometimes even appearing on the same line. This can make reading the output very difficult.

To avoid this you can use the ‘--output-sync’ (‘-O’) option. This option instructs make to save the output from the commands it invokes and print it all once the commands are completed. Additionally, if there are multiple recursive make invocations running in parallel, they will communicate so that only one of them is generating output at a time.

If working directory printing is enabled (see The ‘--print-directory’ Option), the enter/leave messages are printed around each output grouping. If you prefer not to see these messages add the ‘--no-print-directory’ option to MAKEFLAGS.

There are four levels of granularity when synchronizing output, specified by giving an argument to the option (e.g., ‘-Oline’ or ‘--output-sync=recurse’).

none

This is the default: all output is sent directly as it is generated and no synchronization is performed.

line

Output from each individual line of the recipe is grouped and printed as soon as that line is complete. If a recipe consists of multiple lines, they may be interspersed with lines from other recipes.

target

Output from the entire recipe for each target is grouped and printed once the target is complete. This is the default if the --output-sync or -O option is given with no argument.

recurse

Output from each recursive invocation of make is grouped and printed once the recursive invocation is complete.

Regardless of the mode chosen, the total build time will be the same. The only difference is in how the output appears.

The ‘target’ and ‘recurse’ modes both collect the output of the entire recipe of a target and display it uninterrupted when the recipe completes. The difference between them is in how recipes that contain recursive invocations of make are treated (see Recursive Use of make). For all recipes which have no recursive lines, the ‘target’ and ‘recurse’ modes behave identically.

If the ‘recurse’ mode is chosen, recipes that contain recursive make invocations are treated the same as other targets: the output from the recipe, including the output from the recursive make, is saved and printed after the entire recipe is complete. This ensures output from all the targets built by a given recursive make instance are grouped together, which may make the output easier to understand. However it also leads to long periods of time during the build where no output is seen, followed by large bursts of output. If you are not watching the build as it proceeds, but instead viewing a log of the build after the fact, this may be the best option for you.

If you are watching the output, the long gaps of quiet during the build can be frustrating. The ‘target’ output synchronization mode detects when make is going to be invoked recursively, using the standard methods, and it will not synchronize the output of those lines. The recursive make will perform the synchronization for its targets and the output from each will be displayed immediately when it completes. Be aware that output from recursive lines of the recipe are not synchronized (for example if the recursive line prints a message before running make, that message will not be synchronized).

The ‘line’ mode can be useful for front-ends that are watching the output of make to track when recipes are started and completed.

Some programs invoked by make may behave differently if they determine they’re writing output to a terminal versus a file (often described as “interactive” vs. “non-interactive” modes). For example, many programs that can display colorized output will not do so if they determine they are not writing to a terminal. If your makefile invokes a program like this then using the output synchronization options will cause the program to believe it’s running in “non-interactive” mode even though the output will ultimately go to the terminal.

5.4.2 Input During Parallel Execution

Two processes cannot both take input from the same device at the same time. To make sure that only one recipe tries to take input from the terminal at once, make will invalidate the standard input streams of all but one running recipe. If another recipe attempts to read from standard input it will usually incur a fatal error (a ‘Broken pipe’ signal).

It is unpredictable which recipe will have a valid standard input stream (which will come from the terminal, or wherever you redirect the standard input of make). The first recipe run will always get it first, and the first recipe started after that one finishes will get it next, and so on.

We will change how this aspect of make works if we find a better alternative. In the mean time, you should not rely on any recipe using standard input at all if you are using the parallel execution feature; but if you are not using this feature, then standard input works normally in all recipes.

5.5 Errors in Recipes

After each shell invocation returns, make looks at its exit status. If the shell completed successfully (the exit status is zero), the next line in the recipe is executed in a new shell; after the last line is finished, the rule is finished.

If there is an error (the exit status is nonzero), make gives up on the current rule, and perhaps on all rules.

Sometimes the failure of a certain recipe line does not indicate a problem. For example, you may use the mkdir command to ensure that a directory exists. If the directory already exists, mkdir will report an error, but you probably want make to continue regardless.

To ignore errors in a recipe line, write a ‘-’ at the beginning of the line’s text (after the initial tab). The ‘-’ is discarded before the line is passed to the shell for execution.

For example,

clean:
        -rm -f *.o

This causes make to continue even if rm is unable to remove a file.

When you run make with the ‘-i’ or ‘--ignore-errors’ flag, errors are ignored in all recipes of all rules. A rule in the makefile for the special target .IGNORE has the same effect, if there are no prerequisites. These ways of ignoring errors are obsolete because ‘-’ is more flexible.

When errors are to be ignored, because of either a ‘-’ or the ‘-i’ flag, make treats an error return just like success, except that it prints out a message that tells you the status code the shell exited with, and says that the error has been ignored.

When an error happens that make has not been told to ignore, it implies that the current target cannot be correctly remade, and neither can any other that depends on it either directly or indirectly. No further recipes will be executed for these targets, since their preconditions have not been achieved.

Normally make gives up immediately in this circumstance, returning a nonzero status. However, if the ‘-k’ or ‘--keep-going’ flag is specified, make continues to consider the other prerequisites of the pending targets, remaking them if necessary, before it gives up and returns nonzero status. For example, after an error in compiling one object file, ‘make -k’ will continue compiling other object files even though it already knows that linking them will be impossible. See Summary of Options.

The usual behavior assumes that your purpose is to get the specified targets up to date; once make learns that this is impossible, it might as well report the failure immediately. The ‘-k’ option says that the real purpose is to test as many of the changes made in the program as possible, perhaps to find several independent problems so that you can correct them all before the next attempt to compile. This is why Emacs’ compile command passes the ‘-k’ flag by default.

Usually when a recipe line fails, if it has changed the target file at all, the file is corrupted and cannot be used—or at least it is not completely updated. Yet the file’s time stamp says that it is now up to date, so the next time make runs, it will not try to update that file. The situation is just the same as when the shell is killed by a signal; see Interrupts. So generally the right thing to do is to delete the target file if the recipe fails after beginning to change the file. make will do this if .DELETE_ON_ERROR appears as a target. This is almost always what you want make to do, but it is not historical practice; so for compatibility, you must explicitly request it.

5.6 Interrupting or Killing make

If make gets a fatal signal while a shell is executing, it may delete the target file that the recipe was supposed to update. This is done if the target file’s last-modification time has changed since make first checked it.

The purpose of deleting the target is to make sure that it is remade from scratch when make is next run. Why is this? Suppose you type Ctrl-c while a compiler is running, and it has begun to write an object file foo.o. The Ctrl-c kills the compiler, resulting in an incomplete file whose last-modification time is newer than the source file foo.c. But make also receives the Ctrl-c signal and deletes this incomplete file. If make did not do this, the next invocation of make would think that foo.o did not require updating—resulting in a strange error message from the linker when it tries to link an object file half of which is missing.

You can prevent the deletion of a target file in this way by making the special target .PRECIOUS depend on it. Before remaking a target, make checks to see whether it appears on the prerequisites of .PRECIOUS, and thereby decides whether the target should be deleted if a signal happens. Some reasons why you might do this are that the target is updated in some atomic fashion, or exists only to record a modification-time (its contents do not matter), or must exist at all times to prevent other sorts of trouble.

5.7 Recursive Use of make

Recursive use of make means using make as a command in a makefile. This technique is useful when you want separate makefiles for various subsystems that compose a larger system. For example, suppose you have a sub-directory subdir which has its own makefile, and you would like the containing directory’s makefile to run make on the sub-directory. You can do it by writing this:

subsystem:
        cd subdir && $(MAKE)

or, equivalently, this (see Summary of Options):

subsystem:
        $(MAKE) -C subdir

You can write recursive make commands just by copying this example, but there are many things to know about how they work and why, and about how the sub-make relates to the top-level make. You may also find it useful to declare targets that invoke recursive make commands as ‘.PHONY’ (for more discussion on when this is useful, see Phony Targets).

For your convenience, when GNU make starts (after it has processed any -C options) it sets the variable CURDIR to the pathname of the current working directory. This value is never touched by make again: in particular note that if you include files from other directories the value of CURDIR does not change. The value has the same precedence it would have if it were set in the makefile (by default, an environment variable CURDIR will not override this value). Note that setting this variable has no impact on the operation of make (it does not cause make to change its working directory, for example).

5.7.1 How the MAKE Variable Works

Recursive make commands should always use the variable MAKE, not the explicit command name ‘make’, as shown here:

subsystem:
        cd subdir && $(MAKE)

The value of this variable is the file name with which make was invoked. If this file name was /bin/make, then the recipe executed is ‘cd subdir && /bin/make’. If you use a special version of make to run the top-level makefile, the same special version will be executed for recursive invocations.

As a special feature, using the variable MAKE in the recipe of a rule alters the effects of the ‘-t’ (‘--touch’), ‘-n’ (‘--just-print’), or ‘-q’ (‘--question’) option. Using the MAKE variable has the same effect as using a ‘+’ character at the beginning of the recipe line. See Instead of Executing the Recipes. This special feature is only enabled if the MAKE variable appears directly in the recipe: it does not apply if the MAKE variable is referenced through expansion of another variable. In the latter case you must use the ‘+’ token to get these special effects.

Consider the command ‘make -t’ in the above example. (The ‘-t’ option marks targets as up to date without actually running any recipes; see Instead of Execution.) Following the usual definition of ‘-t’, a ‘make -t’ command in the example would create a file named subsystem and do nothing else. What you really want it to do is run ‘cd subdir && make -t’; but that would require executing the recipe, and ‘-t’ says not to execute recipes.

The special feature makes this do what you want: whenever a recipe line of a rule contains the variable MAKE, the flags ‘-t’, ‘-n’ and ‘-q’ do not apply to that line. Recipe lines containing MAKE are executed normally despite the presence of a flag that causes most recipes not to be run. The usual MAKEFLAGS mechanism passes the flags to the sub-make (see Communicating Options to a Sub-make), so your request to touch the files, or print the recipes, is propagated to the subsystem.

5.7.2 Communicating Variables to a Sub-make

Variable values of the top-level make can be passed to the sub-make through the environment by explicit request. These variables are defined in the sub-make as defaults, but they do not override variables defined in the makefile used by the sub-make unless you use the ‘-e’ switch (see Summary of Options).

To pass down, or export, a variable, make adds the variable and its value to the environment for running each line of the recipe. The sub-make, in turn, uses the environment to initialize its table of variable values. See Variables from the Environment.

Except by explicit request, make exports a variable only if it is either defined in the environment initially or set on the command line, and if its name consists only of letters, numbers, and underscores. Some shells cannot cope with environment variable names consisting of characters other than letters, numbers, and underscores.

The value of the make variable SHELL is not exported. Instead, the value of the SHELL variable from the invoking environment is passed to the sub-make. You can force make to export its value for SHELL by using the export directive, described below. See Choosing the Shell.

The special variable MAKEFLAGS is always exported (unless you unexport it). MAKEFILES is exported if you set it to anything.

make automatically passes down variable values that were defined on the command line, by putting them in the MAKEFLAGS variable. See Options/Recursion.

Variables are not normally passed down if they were created by default by make (see Variables Used by Implicit Rules). The sub-make will define these for itself.

If you want to export specific variables to a sub-make, use the export directive, like this:

export variable …

If you want to prevent a variable from being exported, use the unexport directive, like this:

unexport variable …

In both of these forms, the arguments to export and unexport are expanded, and so could be variables or functions which expand to a (list of) variable names to be (un)exported.

As a convenience, you can define a variable and export it at the same time by doing:

export variable = value

has the same result as:

variable = value
export variable

and

export variable := value

has the same result as:

variable := value
export variable

Likewise,

export variable += value

is just like:

variable += value
export variable

See Appending More Text to Variables.

You may notice that the export and unexport directives work in make in the same way they work in the shell, sh.

If you want all variables to be exported by default, you can use export by itself:

export

This tells make that variables which are not explicitly mentioned in an export or unexport directive should be exported. Any variable given in an unexport directive will still not be exported. If you use export by itself to export variables by default, variables whose names contain characters other than alphanumerics and underscores will not be exported unless specifically mentioned in an export directive.

The behavior elicited by an export directive by itself was the default in older versions of GNU make. If your makefiles depend on this behavior and you want to be compatible with old versions of make, you can write a rule for the special target .EXPORT_ALL_VARIABLES instead of using the export directive. This will be ignored by old makes, while the export directive will cause a syntax error.

Likewise, you can use unexport by itself to tell make not to export variables by default. Since this is the default behavior, you would only need to do this if export had been used by itself earlier (in an included makefile, perhaps). You cannot use export and unexport by themselves to have variables exported for some recipes and not for others. The last export or unexport directive that appears by itself determines the behavior for the entire run of make.

As a special feature, the variable MAKELEVEL is changed when it is passed down from level to level. This variable’s value is a string which is the depth of the level as a decimal number. The value is ‘0’ for the top-level make; ‘1’ for a sub-make, ‘2’ for a sub-sub-make, and so on. The incrementation happens when make sets up the environment for a recipe.

The main use of MAKELEVEL is to test it in a conditional directive (see Conditional Parts of Makefiles); this way you can write a makefile that behaves one way if run recursively and another way if run directly by you.

You can use the variable MAKEFILES to cause all sub-make commands to use additional makefiles. The value of MAKEFILES is a whitespace-separated list of file names. This variable, if defined in the outer-level makefile, is passed down through the environment; then it serves as a list of extra makefiles for the sub-make to read before the usual or specified ones. See The Variable MAKEFILES.

5.7.3 Communicating Options to a Sub-make

Flags such as ‘-s’ and ‘-k’ are passed automatically to the sub-make through the variable MAKEFLAGS. This variable is set up automatically by make to contain the flag letters that make received. Thus, if you do ‘make -ks’ then MAKEFLAGS gets the value ‘ks’.

As a consequence, every sub-make gets a value for MAKEFLAGS in its environment. In response, it takes the flags from that value and processes them as if they had been given as arguments. See Summary of Options.

Likewise variables defined on the command line are passed to the sub-make through MAKEFLAGS. Words in the value of MAKEFLAGS that contain ‘=’, make treats as variable definitions just as if they appeared on the command line. See Overriding Variables.

The options ‘-C’, ‘-f’, ‘-o’, and ‘-W’ are not put into MAKEFLAGS; these options are not passed down.

The ‘-j’ option is a special case (see Parallel Execution). If you set it to some numeric value ‘N’ and your operating system supports it (most any UNIX system will; others typically won’t), the parent make and all the sub-makes will communicate to ensure that there are only ‘N’ jobs running at the same time between them all. Note that any job that is marked recursive (see Instead of Executing Recipes) doesn’t count against the total jobs (otherwise we could get ‘N’ sub-makes running and have no slots left over for any real work!)

If your operating system doesn’t support the above communication, then ‘-j 1’ is always put into MAKEFLAGS instead of the value you specified. This is because if the ‘-j’ option were passed down to sub-makes, you would get many more jobs running in parallel than you asked for. If you give ‘-j’ with no numeric argument, meaning to run as many jobs as possible in parallel, this is passed down, since multiple infinities are no more than one.

If you do not want to pass the other flags down, you must change the value of MAKEFLAGS, like this:

subsystem:
        cd subdir && $(MAKE) MAKEFLAGS=

The command line variable definitions really appear in the variable MAKEOVERRIDES, and MAKEFLAGS contains a reference to this variable. If you do want to pass flags down normally, but don’t want to pass down the command line variable definitions, you can reset MAKEOVERRIDES to empty, like this:

MAKEOVERRIDES =

This is not usually useful to do. However, some systems have a small fixed limit on the size of the environment, and putting so much information into the value of MAKEFLAGS can exceed it. If you see the error message ‘Arg list too long’, this may be the problem. (For strict compliance with POSIX.2, changing MAKEOVERRIDES does not affect MAKEFLAGS if the special target ‘.POSIX’ appears in the makefile. You probably do not care about this.)

A similar variable MFLAGS exists also, for historical compatibility. It has the same value as MAKEFLAGS except that it does not contain the command line variable definitions, and it always begins with a hyphen unless it is empty (MAKEFLAGS begins with a hyphen only when it begins with an option that has no single-letter version, such as ‘--warn-undefined-variables’). MFLAGS was traditionally used explicitly in the recursive make command, like this:

subsystem:
        cd subdir && $(MAKE) $(MFLAGS)

but now MAKEFLAGS makes this usage redundant. If you want your makefiles to be compatible with old make programs, use this technique; it will work fine with more modern make versions too.

The MAKEFLAGS variable can also be useful if you want to have certain options, such as ‘-k’ (see Summary of Options), set each time you run make. You simply put a value for MAKEFLAGS in your environment. You can also set MAKEFLAGS in a makefile, to specify additional flags that should also be in effect for that makefile. (Note that you cannot use MFLAGS this way. That variable is set only for compatibility; make does not interpret a value you set for it in any way.)

When make interprets the value of MAKEFLAGS (either from the environment or from a makefile), it first prepends a hyphen if the value does not already begin with one. Then it chops the value into words separated by blanks, and parses these words as if they were options given on the command line (except that ‘-C’, ‘-f’, ‘-h’, ‘-o’, ‘-W’, and their long-named versions are ignored; and there is no error for an invalid option).

If you do put MAKEFLAGS in your environment, you should be sure not to include any options that will drastically affect the actions of make and undermine the purpose of makefiles and of make itself. For instance, the ‘-t’, ‘-n’, and ‘-q’ options, if put in one of these variables, could have disastrous consequences and would certainly have at least surprising and probably annoying effects.

If you’d like to run other implementations of make in addition to GNU make, and hence do not want to add GNU make-specific flags to the MAKEFLAGS variable, you can add them to the GNUMAKEFLAGS variable instead. This variable is parsed just before MAKEFLAGS, in the same way as MAKEFLAGS. When make constructs MAKEFLAGS to pass to a recursive make it will include all flags, even those taken from GNUMAKEFLAGS. As a result, after parsing GNUMAKEFLAGS GNU make sets this variable to the empty string to avoid duplicating flags during recursion.

It’s best to use GNUMAKEFLAGS only with flags which won’t materially change the behavior of your makefiles. If your makefiles require GNU make anyway then simply use MAKEFLAGS. Flags such as ‘--no-print-directory’ or ‘--output-sync’ may be appropriate for GNUMAKEFLAGS.

5.7.4 The ‘--print-directory’ Option

If you use several levels of recursive make invocations, the ‘-w’ or ‘--print-directory’ option can make the output a lot easier to understand by showing each directory as make starts processing it and as make finishes processing it. For example, if ‘make -w’ is run in the directory /u/gnu/make, make will print a line of the form:

make: Entering directory `/u/gnu/make'.

before doing anything else, and a line of the form:

make: Leaving directory `/u/gnu/make'.

when processing is completed.

Normally, you do not need to specify this option because ‘make’ does it for you: ‘-w’ is turned on automatically when you use the ‘-C’ option, and in sub-makes. make will not automatically turn on ‘-w’ if you also use ‘-s’, which says to be silent, or if you use ‘--no-print-directory’ to explicitly disable it.

5.8 Defining Canned Recipes

When the same sequence of commands is useful in making various targets, you can define it as a canned sequence with the define directive, and refer to the canned sequence from the recipes for those targets. The canned sequence is actually a variable, so the name must not conflict with other variable names.

Here is an example of defining a canned recipe:

define run-yacc =
yacc $(firstword $^)
mv y.tab.c $@
endef

Here run-yacc is the name of the variable being defined; endef marks the end of the definition; the lines in between are the commands. The define directive does not expand variable references and function calls in the canned sequence; the ‘$’ characters, parentheses, variable names, and so on, all become part of the value of the variable you are defining. See Defining Multi-Line Variables, for a complete explanation of define.

The first command in this example runs Yacc on the first prerequisite of whichever rule uses the canned sequence. The output file from Yacc is always named y.tab.c. The second command moves the output to the rule’s target file name.

To use the canned sequence, substitute the variable into the recipe of a rule. You can substitute it like any other variable (see Basics of Variable References). Because variables defined by define are recursively expanded variables, all the variable references you wrote inside the define are expanded now. For example:

foo.c : foo.y
        $(run-yacc)

‘foo.y’ will be substituted for the variable ‘$^’ when it occurs in run-yacc’s value, and ‘foo.c’ for ‘$@’.

This is a realistic example, but this particular one is not needed in practice because make has an implicit rule to figure out these commands based on the file names involved (see Using Implicit Rules).

In recipe execution, each line of a canned sequence is treated just as if the line appeared on its own in the rule, preceded by a tab. In particular, make invokes a separate sub-shell for each line. You can use the special prefix characters that affect command lines (‘@’, ‘-’, and ‘+’) on each line of a canned sequence. See Writing Recipes in Rules. For example, using this canned sequence:

define frobnicate =
@echo "frobnicating target $@"
frob-step-1 $< -o $@-step-1
frob-step-2 $@-step-1 -o $@
endef

make will not echo the first line, the echo command. But it will echo the following two recipe lines.

On the other hand, prefix characters on the recipe line that refers to a canned sequence apply to every line in the sequence. So the rule:

frob.out: frob.in
        @$(frobnicate)

does not echo any recipe lines. (See Recipe Echoing, for a full explanation of ‘@’.)

5.9 Using Empty Recipes

It is sometimes useful to define recipes which do nothing. This is done simply by giving a recipe that consists of nothing but whitespace. For example:

target: ;

defines an empty recipe for target. You could also use a line beginning with a recipe prefix character to define an empty recipe, but this would be confusing because such a line looks empty.

You may be wondering why you would want to define a recipe that does nothing. The only reason this is useful is to prevent a target from getting implicit recipes (from implicit rules or the .DEFAULT special target; see Implicit Rules and see Defining Last-Resort Default Rules).

You may be inclined to define empty recipes for targets that are not actual files, but only exist so that their prerequisites can be remade. However, this is not the best way to do that, because the prerequisites may not be remade properly if the target file actually does exist. See Phony Targets, for a better way to do this.

6 How to Use Variables

A variable is a name defined in a makefile to represent a string of text, called the variable’s value. These values are substituted by explicit request into targets, prerequisites, recipes, and other parts of the makefile. (In some other versions of make, variables are called macros.)

Variables and functions in all parts of a makefile are expanded when read, except for in recipes, the right-hand sides of variable definitions using ‘=’, and the bodies of variable definitions using the define directive.

Variables can represent lists of file names, options to pass to compilers, programs to run, directories to look in for source files, directories to write output in, or anything else you can imagine.

A variable name may be any sequence of characters not containing ‘:’, ‘#’, ‘=’, or whitespace. However, variable names containing characters other than letters, numbers, and underscores should be considered carefully, as in some shells they cannot be passed through the environment to a sub-make (see Communicating Variables to a Sub-make). Variable names beginning with ‘.’ and an uppercase letter may be given special meaning in future versions of make.

Variable names are case-sensitive. The names ‘foo’, ‘FOO’, and ‘Foo’ all refer to different variables.

It is traditional to use upper case letters in variable names, but we recommend using lower case letters for variable names that serve internal purposes in the makefile, and reserving upper case for parameters that control implicit rules or for parameters that the user should override with command options (see Overriding Variables).

A few variables have names that are a single punctuation character or just a few characters. These are the automatic variables, and they have particular specialized uses. See Automatic Variables.

6.1 Basics of Variable References

To substitute a variable’s value, write a dollar sign followed by the name of the variable in parentheses or braces: either ‘$(foo)’ or ‘${foo}’ is a valid reference to the variable foo. This special significance of ‘$’ is why you must write ‘$$’ to have the effect of a single dollar sign in a file name or recipe.

Variable references can be used in any context: targets, prerequisites, recipes, most directives, and new variable values. Here is an example of a common case, where a variable holds the names of all the object files in a program:

objects = program.o foo.o utils.o
program : $(objects)
        cc -o program $(objects)

$(objects) : defs.h

Variable references work by strict textual substitution. Thus, the rule

foo = c
prog.o : prog.$(foo)
        $(foo)$(foo) -$(foo) prog.$(foo)

could be used to compile a C program prog.c. Since spaces before the variable value are ignored in variable assignments, the value of foo is precisely ‘c’. (Don’t actually write your makefiles this way!)

A dollar sign followed by a character other than a dollar sign, open-parenthesis or open-brace treats that single character as the variable name. Thus, you could reference the variable x with ‘$x’. However, this practice is strongly discouraged, except in the case of the automatic variables (see Automatic Variables).

6.2 The Two Flavors of Variables

There are two ways that a variable in GNU make can have a value; we call them the two flavors of variables. The two flavors are distinguished in how they are defined and in what they do when expanded.

The first flavor of variable is a recursively expanded variable. Variables of this sort are defined by lines using ‘=’ (see Setting Variables) or by the define directive (see Defining Multi-Line Variables). The value you specify is installed verbatim; if it contains references to other variables, these references are expanded whenever this variable is substituted (in the course of expanding some other string). When this happens, it is called recursive expansion.

For example,

foo = $(bar)
bar = $(ugh)
ugh = Huh?

all:;echo $(foo)

will echo ‘Huh?’: ‘$(foo)’ expands to ‘$(bar)’ which expands to ‘$(ugh)’ which finally expands to ‘Huh?’.

This flavor of variable is the only sort supported by most other versions of make. It has its advantages and its disadvantages. An advantage (most would say) is that:

CFLAGS = $(include_dirs) -O
include_dirs = -Ifoo -Ibar

will do what was intended: when ‘CFLAGS’ is expanded in a recipe, it will expand to ‘-Ifoo -Ibar -O’. A major disadvantage is that you cannot append something on the end of a variable, as in

CFLAGS = $(CFLAGS) -O

because it will cause an infinite loop in the variable expansion. (Actually make detects the infinite loop and reports an error.)

Another disadvantage is that any functions (see Functions for Transforming Text) referenced in the definition will be executed every time the variable is expanded. This makes make run slower; worse, it causes the wildcard and shell functions to give unpredictable results because you cannot easily control when they are called, or even how many times.

To avoid all the problems and inconveniences of recursively expanded variables, there is another flavor: simply expanded variables.

Simply expanded variables are defined by lines using ‘:=’ or ‘::=’ (see Setting Variables). Both forms are equivalent in GNU make; however only the ‘::=’ form is described by the POSIX standard (support for ‘::=’ was added to the POSIX standard in 2012, so older versions of make won’t accept this form either).

The value of a simply expanded variable is scanned once and for all, expanding any references to other variables and functions, when the variable is defined. The actual value of the simply expanded variable is the result of expanding the text that you write. It does not contain any references to other variables; it contains their values as of the time this variable was defined. Therefore,

x := foo
y := $(x) bar
x := later

is equivalent to

y := foo bar
x := later

When a simply expanded variable is referenced, its value is substituted verbatim.

Here is a somewhat more complicated example, illustrating the use of ‘:=’ in conjunction with the shell function. (See The shell Function.) This example also shows use of the variable MAKELEVEL, which is changed when it is passed down from level to level. (See Communicating Variables to a Sub-make, for information about MAKELEVEL.)

ifeq (0,${MAKELEVEL})
whoami    := $(shell whoami)
host-type := $(shell arch)
MAKE := ${MAKE} host-type=${host-type} whoami=${whoami}
endif

An advantage of this use of ‘:=’ is that a typical ‘descend into a directory’ recipe then looks like this:

${subdirs}:
        ${MAKE} -C $@ all

Simply expanded variables generally make complicated makefile programming more predictable because they work like variables in most programming languages. They allow you to redefine a variable using its own value (or its value processed in some way by one of the expansion functions) and to use the expansion functions much more efficiently (see Functions for Transforming Text).

You can also use them to introduce controlled leading whitespace into variable values. Leading whitespace characters are discarded from your input before substitution of variable references and function calls; this means you can include leading spaces in a variable value by protecting them with variable references, like this:

nullstring :=
space := $(nullstring) # end of the line

Here the value of the variable space is precisely one space. The comment ‘# end of the line’ is included here just for clarity. Since trailing space characters are not stripped from variable values, just a space at the end of the line would have the same effect (but be rather hard to read). If you put whitespace at the end of a variable value, it is a good idea to put a comment like that at the end of the line to make your intent clear. Conversely, if you do not want any whitespace characters at the end of your variable value, you must remember not to put a random comment on the end of the line after some whitespace, such as this:

dir := /foo/bar    # directory to put the frobs in

Here the value of the variable dir is ‘/foo/bar    ’ (with four trailing spaces), which was probably not the intention. (Imagine something like ‘$(dir)/file’ with this definition!)

There is another assignment operator for variables, ‘?=’. This is called a conditional variable assignment operator, because it only has an effect if the variable is not yet defined. This statement:

FOO ?= bar

is exactly equivalent to this (see The origin Function):

ifeq ($(origin FOO), undefined)
  FOO = bar
endif

Note that a variable set to an empty value is still defined, so ‘?=’ will not set that variable.

6.3 Advanced Features for Reference to Variables

This section describes some advanced features you can use to reference variables in more flexible ways.

6.3.1 Substitution References

A substitution reference substitutes the value of a variable with alterations that you specify. It has the form ‘$(var:a=b)’ (or ‘${var:a=b}’) and its meaning is to take the value of the variable var, replace every a at the end of a word with b in that value, and substitute the resulting string.

When we say “at the end of a word”, we mean that a must appear either followed by whitespace or at the end of the value in order to be replaced; other occurrences of a in the value are unaltered. For example:

foo := a.o b.o c.o
bar := $(foo:.o=.c)

sets ‘bar’ to ‘a.c b.c c.c’. See Setting Variables.

A substitution reference is actually an abbreviation for use of the patsubst expansion function (see Functions for String Substitution and Analysis). We provide substitution references as well as patsubst for compatibility with other implementations of make.

Another type of substitution reference lets you use the full power of the patsubst function. It has the same form ‘$(var:a=b)’ described above, except that now a must contain a single ‘%’ character. This case is equivalent to ‘$(patsubst a,b,$(var))’. See Functions for String Substitution and Analysis, for a description of the patsubst function.

For example:

foo := a.o b.o c.o
bar := $(foo:%.o=%.c)

sets ‘bar’ to ‘a.c b.c c.c’.

6.3.2 Computed Variable Names

Computed variable names are a complicated concept needed only for sophisticated makefile programming. For most purposes you need not consider them, except to know that making a variable with a dollar sign in its name might have strange results. However, if you are the type that wants to understand everything, or you are actually interested in what they do, read on.

Variables may be referenced inside the name of a variable. This is called a computed variable name or a nested variable reference. For example,

x = y
y = z
a := $($(x))

defines a as ‘z’: the ‘$(x)’ inside ‘$($(x))’ expands to ‘y’, so ‘$($(x))’ expands to ‘$(y)’ which in turn expands to ‘z’. Here the name of the variable to reference is not stated explicitly; it is computed by expansion of ‘$(x)’. The reference ‘$(x)’ here is nested within the outer variable reference.

The previous example shows two levels of nesting, but any number of levels is possible. For example, here are three levels:

x = y
y = z
z = u
a := $($($(x)))

Here the innermost ‘$(x)’ expands to ‘y’, so ‘$($(x))’ expands to ‘$(y)’ which in turn expands to ‘z’; now we have ‘$(z)’, which becomes ‘u’.

References to recursively-expanded variables within a variable name are re-expanded in the usual fashion. For example:

x = $(y)
y = z
z = Hello
a := $($(x))

defines a as ‘Hello’: ‘$($(x))’ becomes ‘$($(y))’ which becomes ‘$(z)’ which becomes ‘Hello’.

Nested variable references can also contain modified references and function invocations (see Functions for Transforming Text), just like any other reference. For example, using the subst function (see Functions for String Substitution and Analysis):

x = variable1
variable2 := Hello
y = $(subst 1,2,$(x))
z = y
a := $($($(z)))

eventually defines a as ‘Hello’. It is doubtful that anyone would ever want to write a nested reference as convoluted as this one, but it works: ‘$($($(z)))’ expands to ‘$($(y))’ which becomes ‘$($(subst 1,2,$(x)))’. This gets the value ‘variable1’ from x and changes it by substitution to ‘variable2’, so that the entire string becomes ‘$(variable2)’, a simple variable reference whose value is ‘Hello’.

A computed variable name need not consist entirely of a single variable reference. It can contain several variable references, as well as some invariant text. For example,

a_dirs := dira dirb
1_dirs := dir1 dir2

a_files := filea fileb
1_files := file1 file2

ifeq "$(use_a)" "yes"
a1 := a
else
a1 := 1
endif

ifeq "$(use_dirs)" "yes"
df := dirs
else
df := files
endif

dirs := $($(a1)_$(df))

will give dirs the same value as a_dirs, 1_dirs, a_files or 1_files depending on the settings of use_a and use_dirs.

Computed variable names can also be used in substitution references:

a_objects := a.o b.o c.o
1_objects := 1.o 2.o 3.o

sources := $($(a1)_objects:.o=.c)

defines sources as either ‘a.c b.c c.c’ or ‘1.c 2.c 3.c’, depending on the value of a1.

The only restriction on this sort of use of nested variable references is that they cannot specify part of the name of a function to be called. This is because the test for a recognized function name is done before the expansion of nested references. For example,

ifdef do_sort
func := sort
else
func := strip
endif

bar := a d b g q c

foo := $($(func) $(bar))

attempts to give ‘foo’ the value of the variable ‘sort a d b g q c’ or ‘strip a d b g q c’, rather than giving ‘a d b g q c’ as the argument to either the sort or the strip function. This restriction could be removed in the future if that change is shown to be a good idea.

You can also use computed variable names in the left-hand side of a variable assignment, or in a define directive, as in:

dir = foo
$(dir)_sources := $(wildcard $(dir)/*.c)
define $(dir)_print =
lpr $($(dir)_sources)
endef

This example defines the variables ‘dir’, ‘foo_sources’, and ‘foo_print’.

Note that nested variable references are quite different from recursively expanded variables (see The Two Flavors of Variables), though both are used together in complex ways when doing makefile programming.

6.4 How Variables Get Their Values

Variables can get values in several different ways:

• You can specify an overriding value when you run make. See Overriding Variables.
• You can specify a value in the makefile, either with an assignment (see Setting Variables) or with a verbatim definition (see Defining Multi-Line Variables).
• Variables in the environment become make variables. See Variables from the Environment.
• Several automatic variables are given new values for each rule. Each of these has a single conventional use. See Automatic Variables.
• Several variables have constant initial values. See Variables Used by Implicit Rules.

6.5 Setting Variables

To set a variable from the makefile, write a line starting with the variable name followed by ‘=’ ‘:=’, or ‘::=’. Whatever follows the ‘=’, ‘:=’, or ‘::=’ on the line becomes the value. For example,

objects = main.o foo.o bar.o utils.o

defines a variable named objects. Whitespace around the variable name and immediately after the ‘=’ is ignored.

Variables defined with ‘=’ are recursively expanded variables. Variables defined with ‘:=’ or ‘::=’ are simply expanded variables; these definitions can contain variable references which will be expanded before the definition is made. See The Two Flavors of Variables.

The variable name may contain function and variable references, which are expanded when the line is read to find the actual variable name to use.

There is no limit on the length of the value of a variable except the amount of memory on the computer. You can split the value of a variable into multiple physical lines for readability (see Splitting Long Lines).

Most variable names are considered to have the empty string as a value if you have never set them. Several variables have built-in initial values that are not empty, but you can set them in the usual ways (see Variables Used by Implicit Rules). Several special variables are set automatically to a new value for each rule; these are called the automatic variables (see Automatic Variables).

If you’d like a variable to be set to a value only if it’s not already set, then you can use the shorthand operator ‘?=’ instead of ‘=’. These two settings of the variable ‘FOO’ are identical (see The origin Function):

FOO ?= bar

and

ifeq ($(origin FOO), undefined)
FOO = bar
endif

The shell assignment operator ‘!=’ can be used to execute a program and set a variable to its output. This operator first evaluates the right-hand side, then passes that result to the shell for execution. If the result of the execution ends in a newline, that one newline is removed; all other newlines are replaced by spaces. The resulting string is then placed into the named recursively-expanded variable. For example:

hash != printf '\043'
file_list != find . -name '*.c'

If the result of the execution could produce a $, and you don’t intend what follows that to be interpreted as a make variable or function reference, then you must replace every $ with $$ as part of the execution. Alternatively, you can set a simply expanded variable to the result of running a program using the shell function call. See The shell Function. For example:

hash := $(shell printf '\043')
var := $(shell find . -name "*.c")

6.6 Appending More Text to Variables

Often it is useful to add more text to the value of a variable already defined. You do this with a line containing ‘+=’, like this:

objects += another.o

This takes the value of the variable objects, and adds the text ‘another.o’ to it (preceded by a single space). Thus:

objects = main.o foo.o bar.o utils.o
objects += another.o

sets objects to ‘main.o foo.o bar.o utils.o another.o’.

Using ‘+=’ is similar to:

objects = main.o foo.o bar.o utils.o
objects := $(objects) another.o

but differs in ways that become important when you use more complex values.

When the variable in question has not been defined before, ‘+=’ acts just like normal ‘=’: it defines a recursively-expanded variable. However, when there is a previous definition, exactly what ‘+=’ does depends on what flavor of variable you defined originally. See The Two Flavors of Variables, for an explanation of the two flavors of variables.

When you add to a variable’s value with ‘+=’, make acts essentially as if you had included the extra text in the initial definition of the variable. If you defined it first with ‘:=’ or ‘::=’, making it a simply-expanded variable, ‘+=’ adds to that simply-expanded definition, and expands the new text before appending it to the old value just as ‘:=’ does (see Setting Variables, for a full explanation of ‘:=’ or ‘::=’). In fact,

variable := value
variable += more

is exactly equivalent to:

variable := value
variable := $(variable) more

On the other hand, when you use ‘+=’ with a variable that you defined first to be recursively-expanded using plain ‘=’, make does something a bit different. Recall that when you define a recursively-expanded variable, make does not expand the value you set for variable and function references immediately. Instead it stores the text verbatim, and saves these variable and function references to be expanded later, when you refer to the new variable (see The Two Flavors of Variables). When you use ‘+=’ on a recursively-expanded variable, it is this unexpanded text to which make appends the new text you specify.

variable = value
variable += more

is roughly equivalent to:

temp = value
variable = $(temp) more

except that of course it never defines a variable called temp. The importance of this comes when the variable’s old value contains variable references. Take this common example:

CFLAGS = $(includes) -O
…
CFLAGS += -pg # enable profiling

The first line defines the CFLAGS variable with a reference to another variable, includes. (CFLAGS is used by the rules for C compilation; see Catalogue of Built-In Rules.) Using ‘=’ for the definition makes CFLAGS a recursively-expanded variable, meaning ‘$(includes) -O’ is not expanded when make processes the definition of CFLAGS. Thus, includes need not be defined yet for its value to take effect. It only has to be defined before any reference to CFLAGS. If we tried to append to the value of CFLAGS without using ‘+=’, we might do it like this:

CFLAGS := $(CFLAGS) -pg # enable profiling

This is pretty close, but not quite what we want. Using ‘:=’ redefines CFLAGS as a simply-expanded variable; this means make expands the text ‘$(CFLAGS) -pg’ before setting the variable. If includes is not yet defined, we get ‘ -O -pg’, and a later definition of includes will have no effect. Conversely, by using ‘+=’ we set CFLAGS to the unexpanded value ‘$(includes) -O -pg’. Thus we preserve the reference to includes, so if that variable gets defined at any later point, a reference like ‘$(CFLAGS)’ still uses its value.

6.7 The override Directive

If a variable has been set with a command argument (see Overriding Variables), then ordinary assignments in the makefile are ignored. If you want to set the variable in the makefile even though it was set with a command argument, you can use an override directive, which is a line that looks like this:

override variable = value

or

override variable := value

To append more text to a variable defined on the command line, use:

override variable += more text

See Appending More Text to Variables.

Variable assignments marked with the override flag have a higher priority than all other assignments, except another override. Subsequent assignments or appends to this variable which are not marked override will be ignored.

The override directive was not invented for escalation in the war between makefiles and command arguments. It was invented so you can alter and add to values that the user specifies with command arguments.

For example, suppose you always want the ‘-g’ switch when you run the C compiler, but you would like to allow the user to specify the other switches with a command argument just as usual. You could use this override directive:

override CFLAGS += -g

You can also use override directives with define directives. This is done as you might expect:

override define foo =
bar
endef

See Defining Multi-Line Variables.

6.8 Defining Multi-Line Variables

Another way to set the value of a variable is to use the define directive. This directive has an unusual syntax which allows newline characters to be included in the value, which is convenient for defining both canned sequences of commands (see Defining Canned Recipes), and also sections of makefile syntax to use with eval (see Eval Function).

The define directive is followed on the same line by the name of the variable being defined and an (optional) assignment operator, and nothing more. The value to give the variable appears on the following lines. The end of the value is marked by a line containing just the word endef. Aside from this difference in syntax, define works just like any other variable definition. The variable name may contain function and variable references, which are expanded when the directive is read to find the actual variable name to use.

You may omit the variable assignment operator if you prefer. If omitted, make assumes it to be ‘=’ and creates a recursively-expanded variable (see The Two Flavors of Variables). When using a ‘+=’ operator, the value is appended to the previous value as with any other append operation: with a single space separating the old and new values.

You may nest define directives: make will keep track of nested directives and report an error if they are not all properly closed with endef. Note that lines beginning with the recipe prefix character are considered part of a recipe, so any define or endef strings appearing on such a line will not be considered make directives.

define two-lines =
echo foo
echo $(bar)
endef

The value in an ordinary assignment cannot contain a newline; but the newlines that separate the lines of the value in a define become part of the variable’s value (except for the final newline which precedes the endef and is not considered part of the value).

When used in a recipe, the previous example is functionally equivalent to this:

two-lines = echo foo; echo $(bar)

since two commands separated by semicolon behave much like two separate shell commands. However, note that using two separate lines means make will invoke the shell twice, running an independent sub-shell for each line. See Recipe Execution.

If you want variable definitions made with define to take precedence over command-line variable definitions, you can use the override directive together with define:

override define two-lines =
foo
$(bar)
endef

See The override Directive.

6.9 Undefining Variables

If you want to clear a variable, setting its value to empty is usually sufficient. Expanding such a variable will yield the same result (empty string) regardless of whether it was set or not. However, if you are using the flavor (see Flavor Function) and origin (see Origin Function) functions, there is a difference between a variable that was never set and a variable with an empty value. In such situations you may want to use the undefine directive to make a variable appear as if it was never set. For example:

foo := foo
bar = bar

undefine foo
undefine bar

$(info $(origin foo))
$(info $(flavor bar))

This example will print “undefined” for both variables.

If you want to undefine a command-line variable definition, you can use the override directive together with undefine, similar to how this is done for variable definitions:

override undefine CFLAGS

6.10 Variables from the Environment

Variables in make can come from the environment in which make is run. Every environment variable that make sees when it starts up is transformed into a make variable with the same name and value. However, an explicit assignment in the makefile, or with a command argument, overrides the environment. (If the ‘-e’ flag is specified, then values from the environment override assignments in the makefile. See Summary of Options. But this is not recommended practice.)

Thus, by setting the variable CFLAGS in your environment, you can cause all C compilations in most makefiles to use the compiler switches you prefer. This is safe for variables with standard or conventional meanings because you know that no makefile will use them for other things. (Note this is not totally reliable; some makefiles set CFLAGS explicitly and therefore are not affected by the value in the environment.)

When make runs a recipe, variables defined in the makefile are placed into the environment of each shell. This allows you to pass values to sub-make invocations (see Recursive Use of make). By default, only variables that came from the environment or the command line are passed to recursive invocations. You can use the export directive to pass other variables. See Communicating Variables to a Sub-make, for full details.

Other use of variables from the environment is not recommended. It is not wise for makefiles to depend for their functioning on environment variables set up outside their control, since this would cause different users to get different results from the same makefile. This is against the whole purpose of most makefiles.

Such problems would be especially likely with the variable SHELL, which is normally present in the environment to specify the user’s choice of interactive shell. It would be very undesirable for this choice to affect make; so, make handles the SHELL environment variable in a special way; see Choosing the Shell.

6.11 Target-specific Variable Values

Variable values in make are usually global; that is, they are the same regardless of where they are evaluated (unless they’re reset, of course). One exception to that is automatic variables (see Automatic Variables).

The other exception is target-specific variable values. This feature allows you to define different values for the same variable, based on the target that make is currently building. As with automatic variables, these values are only available within the context of a target’s recipe (and in other target-specific assignments).

Set a target-specific variable value like this:

target … : variable-assignment

Target-specific variable assignments can be prefixed with any or all of the special keywords export, override, or private; these apply their normal behavior to this instance of the variable only.

Multiple target values create a target-specific variable value for each member of the target list individually.

The variable-assignment can be any valid form of assignment; recursive (‘=’), simple (‘:=’ or ‘::=’), appending (‘+=’), or conditional (‘?=’). All variables that appear within the variable-assignment are evaluated within the context of the target: thus, any previously-defined target-specific variable values will be in effect. Note that this variable is actually distinct from any “global” value: the two variables do not have to have the same flavor (recursive vs. simple).

Target-specific variables have the same priority as any other makefile variable. Variables provided on the command line (and in the environment if the ‘-e’ option is in force) will take precedence. Specifying the override directive will allow the target-specific variable value to be preferred.

There is one more special feature of target-specific variables: when you define a target-specific variable that variable value is also in effect for all prerequisites of this target, and all their prerequisites, etc. (unless those prerequisites override that variable with their own target-specific variable value). So, for example, a statement like this:

prog : CFLAGS = -g
prog : prog.o foo.o bar.o

will set CFLAGS to ‘-g’ in the recipe for prog, but it will also set CFLAGS to ‘-g’ in the recipes that create prog.o, foo.o, and bar.o, and any recipes which create their prerequisites.

Be aware that a given prerequisite will only be built once per invocation of make, at most. If the same file is a prerequisite of multiple targets, and each of those targets has a different value for the same target-specific variable, then the first target to be built will cause that prerequisite to be built and the prerequisite will inherit the target-specific value from the first target. It will ignore the target-specific values from any other targets.

6.12 Pattern-specific Variable Values

In addition to target-specific variable values (see Target-specific Variable Values), GNU make supports pattern-specific variable values. In this form, the variable is defined for any target that matches the pattern specified.

Set a pattern-specific variable value like this:

pattern … : variable-assignment

where pattern is a %-pattern. As with target-specific variable values, multiple pattern values create a pattern-specific variable value for each pattern individually. The variable-assignment can be any valid form of assignment. Any command line variable setting will take precedence, unless override is specified.

For example:

%.o : CFLAGS = -O

will assign CFLAGS the value of ‘-O’ for all targets matching the pattern %.o.

If a target matches more than one pattern, the matching pattern-specific variables with longer stems are interpreted first. This results in more specific variables taking precedence over the more generic ones, for example:

%.o: %.c
        $(CC) -c $(CFLAGS) $(CPPFLAGS) $< -o $@

lib/%.o: CFLAGS := -fPIC -g
%.o: CFLAGS := -g

all: foo.o lib/bar.o

In this example the first definition of the CFLAGS variable will be used to update lib/bar.o even though the second one also applies to this target. Pattern-specific variables which result in the same stem length are considered in the order in which they were defined in the makefile.

Pattern-specific variables are searched after any target-specific variables defined explicitly for that target, and before target-specific variables defined for the parent target.

6.13 Suppressing Inheritance

As described in previous sections, make variables are inherited by prerequisites. This capability allows you to modify the behavior of a prerequisite based on which targets caused it to be rebuilt. For example, you might set a target-specific variable on a debug target, then running ‘make debug’ will cause that variable to be inherited by all prerequisites of debug, while just running ‘make all’ (for example) would not have that assignment.

Sometimes, however, you may not want a variable to be inherited. For these situations, make provides the private modifier. Although this modifier can be used with any variable assignment, it makes the most sense with target- and pattern-specific variables. Any variable marked private will be visible to its local target but will not be inherited by prerequisites of that target. A global variable marked private will be visible in the global scope but will not be inherited by any target, and hence will not be visible in any recipe.

As an example, consider this makefile:

EXTRA_CFLAGS =

prog: private EXTRA_CFLAGS = -L/usr/local/lib
prog: a.o b.o

Due to the private modifier, a.o and b.o will not inherit the EXTRA_CFLAGS variable assignment from the prog target.

6.14 Other Special Variables

GNU make supports some variables that have special properties.

MAKEFILE_LIST

Contains the name of each makefile that is parsed by make, in the order in which it was parsed. The name is appended just before make begins to parse the makefile. Thus, if the first thing a makefile does is examine the last word in this variable, it will be the name of the current makefile. Once the current makefile has used include, however, the last word will be the just-included makefile.

If a makefile named Makefile has this content:

name1 := $(lastword $(MAKEFILE_LIST))

include inc.mk

name2 := $(lastword $(MAKEFILE_LIST))

all:
        @echo name1 = $(name1)
        @echo name2 = $(name2)

then you would expect to see this output:

name1 = Makefile
name2 = inc.mk

.DEFAULT_GOAL

Sets the default goal to be used if no targets were specified on the command line (see Arguments to Specify the Goals). The .DEFAULT_GOAL variable allows you to discover the current default goal, restart the default goal selection algorithm by clearing its value, or to explicitly set the default goal. The following example illustrates these cases:

# Query the default goal.
ifeq ($(.DEFAULT_GOAL),)
  $(warning no default goal is set)
endif

.PHONY: foo
foo: ; @echo $@

$(warning default goal is $(.DEFAULT_GOAL))

# Reset the default goal.
.DEFAULT_GOAL :=

.PHONY: bar
bar: ; @echo $@

$(warning default goal is $(.DEFAULT_GOAL))

# Set our own.
.DEFAULT_GOAL := foo

This makefile prints:

no default goal is set
default goal is foo
default goal is bar
foo

Note that assigning more than one target name to .DEFAULT_GOAL is invalid and will result in an error.

MAKE_RESTARTS

This variable is set only if this instance of make has restarted (see How Makefiles Are Remade): it will contain the number of times this instance has restarted. Note this is not the same as recursion (counted by the MAKELEVEL variable). You should not set, modify, or export this variable.

MAKE_TERMOUT

MAKE_TERMERR

When make starts it will check whether stdout and stderr will show their output on a terminal. If so, it will set MAKE_TERMOUT and MAKE_TERMERR, respectively, to the name of the terminal device (or true if this cannot be determined). If set these variables will be marked for export. These variables will not be changed by make and they will not be modified if already set.

These values can be used (particularly in combination with output synchronization (see Output During Parallel Execution) to determine whether make itself is writing to a terminal; they can be tested to decide whether to force recipe commands to generate colorized output for example.

If you invoke a sub-make and redirect its stdout or stderr it is your responsibility to reset or unexport these variables as well, if your makefiles rely on them.

.RECIPEPREFIX

The first character of the value of this variable is used as the character make assumes is introducing a recipe line. If the variable is empty (as it is by default) that character is the standard tab character. For example, this is a valid makefile:

.RECIPEPREFIX = >
all:
> @echo Hello, world

The value of .RECIPEPREFIX can be changed multiple times; once set it stays in effect for all rules parsed until it is modified.

.VARIABLES

Expands to a list of the names of all global variables defined so far. This includes variables which have empty values, as well as built-in variables (see Variables Used by Implicit Rules), but does not include any variables which are only defined in a target-specific context. Note that any value you assign to this variable will be ignored; it will always return its special value.

.FEATURES

Expands to a list of special features supported by this version of make. Possible values include, but are not limited to:

‘archives’

Supports ar (archive) files using special file name syntax. See Using make to Update Archive Files.

‘check-symlink’

Supports the -L (--check-symlink-times) flag. See Summary of Options.

‘else-if’

Supports “else if” non-nested conditionals. See Syntax of Conditionals.

‘jobserver’

Supports “job server” enhanced parallel builds. See Parallel Execution.

‘oneshell’

Supports the .ONESHELL special target. See Using One Shell.

‘order-only’

Supports order-only prerequisites. See Types of Prerequisites.

‘second-expansion’

Supports secondary expansion of prerequisite lists.

‘shortest-stem’

Uses the “shortest stem” method of choosing which pattern, of multiple applicable options, will be used. See How Patterns Match.

‘target-specific’

Supports target-specific and pattern-specific variable assignments. See Target-specific Variable Values.

‘undefine’

Supports the undefine directive. See Undefine Directive.

‘guile’

Has GNU Guile available as an embedded extension language. See GNU Guile Integration.

‘load’

Supports dynamically loadable objects for creating custom extensions. See Loading Dynamic Objects.

.INCLUDE_DIRS

Expands to a list of directories that make searches for included makefiles (see Including Other Makefiles).

7 Conditional Parts of Makefiles

A conditional directive causes part of a makefile to be obeyed or ignored depending on the values of variables. Conditionals can compare the value of one variable to another, or the value of a variable to a constant string. Conditionals control what make actually “sees” in the makefile, so they cannot be used to control recipes at the time of execution.

7.1 Example of a Conditional

The following example of a conditional tells make to use one set of libraries if the CC variable is ‘gcc’, and a different set of libraries otherwise. It works by controlling which of two recipe lines will be used for the rule. The result is that ‘CC=gcc’ as an argument to make changes not only which compiler is used but also which libraries are linked.

libs_for_gcc = -lgnu
normal_libs =

foo: $(objects)
ifeq ($(CC),gcc)
        $(CC) -o foo $(objects) $(libs_for_gcc)
else
        $(CC) -o foo $(objects) $(normal_libs)
endif

This conditional uses three directives: one ifeq, one else and one endif.

The ifeq directive begins the conditional, and specifies the condition. It contains two arguments, separated by a comma and surrounded by parentheses. Variable substitution is performed on both arguments and then they are compared. The lines of the makefile following the ifeq are obeyed if the two arguments match; otherwise they are ignored.

The else directive causes the following lines to be obeyed if the previous conditional failed. In the example above, this means that the second alternative linking command is used whenever the first alternative is not used. It is optional to have an else in a conditional.

The endif directive ends the conditional. Every conditional must end with an endif. Unconditional makefile text follows.

As this example illustrates, conditionals work at the textual level: the lines of the conditional are treated as part of the makefile, or ignored, according to the condition. This is why the larger syntactic units of the makefile, such as rules, may cross the beginning or the end of the conditional.

When the variable CC has the value ‘gcc’, the above example has this effect:

foo: $(objects)
        $(CC) -o foo $(objects) $(libs_for_gcc)

When the variable CC has any other value, the effect is this:

foo: $(objects)
        $(CC) -o foo $(objects) $(normal_libs)

Equivalent results can be obtained in another way by conditionalizing a variable assignment and then using the variable unconditionally:

libs_for_gcc = -lgnu
normal_libs =

ifeq ($(CC),gcc)
  libs=$(libs_for_gcc)
else
  libs=$(normal_libs)
endif

foo: $(objects)
        $(CC) -o foo $(objects) $(libs)

7.2 Syntax of Conditionals

The syntax of a simple conditional with no else is as follows:

conditional-directive
text-if-true
endif

The text-if-true may be any lines of text, to be considered as part of the makefile if the condition is true. If the condition is false, no text is used instead.

The syntax of a complex conditional is as follows:

conditional-directive
text-if-true
else
text-if-false
endif

or:

conditional-directive-one
text-if-one-is-true
else conditional-directive-two
text-if-two-is-true
else
text-if-one-and-two-are-false
endif

There can be as many “else conditional-directive” clauses as necessary. Once a given condition is true, text-if-true is used and no other clause is used; if no condition is true then text-if-false is used. The text-if-true and text-if-false can be any number of lines of text.

The syntax of the conditional-directive is the same whether the conditional is simple or complex; after an else or not. There are four different directives that test different conditions. Here is a table of them:

ifeq (arg1, arg2)

ifeq 'arg1' 'arg2'

ifeq "arg1" "arg2"

ifeq "arg1" 'arg2'

ifeq 'arg1' "arg2"

Expand all variable references in arg1 and arg2 and compare them. If they are identical, the text-if-true is effective; otherwise, the text-if-false, if any, is effective.

Often you want to test if a variable has a non-empty value. When the value results from complex expansions of variables and functions, expansions you would consider empty may actually contain whitespace characters and thus are not seen as empty. However, you can use the strip function (see Text Functions) to avoid interpreting whitespace as a non-empty value. For example:

ifeq ($(strip $(foo)),)
text-if-empty
endif

will evaluate text-if-empty even if the expansion of $(foo) contains whitespace characters.

ifneq (arg1, arg2)

ifneq 'arg1' 'arg2'

ifneq "arg1" "arg2"

ifneq "arg1" 'arg2'

ifneq 'arg1' "arg2"

Expand all variable references in arg1 and arg2 and compare them. If they are different, the text-if-true is effective; otherwise, the text-if-false, if any, is effective.

ifdef variable-name

The ifdef form takes the name of a variable as its argument, not a reference to a variable. The value of that variable has a non-empty value, the text-if-true is effective; otherwise, the text-if-false, if any, is effective. Variables that have never been defined have an empty value. The text variable-name is expanded, so it could be a variable or function that expands to the name of a variable. For example:

bar = true
foo = bar
ifdef $(foo)
frobozz = yes
endif

The variable reference $(foo) is expanded, yielding bar, which is considered to be the name of a variable. The variable bar is not expanded, but its value is examined to determine if it is non-empty.

Note that ifdef only tests whether a variable has a value. It does not expand the variable to see if that value is nonempty. Consequently, tests using ifdef return true for all definitions except those like foo =. To test for an empty value, use ifeq ($(foo),). For example,

bar =
foo = $(bar)
ifdef foo
frobozz = yes
else
frobozz = no
endif

sets ‘frobozz’ to ‘yes’, while:

foo =
ifdef foo
frobozz = yes
else
frobozz = no
endif

sets ‘frobozz’ to ‘no’.

ifndef variable-name

If the variable variable-name has an empty value, the text-if-true is effective; otherwise, the text-if-false, if any, is effective. The rules for expansion and testing of variable-name are identical to the ifdef directive.

Extra spaces are allowed and ignored at the beginning of the conditional directive line, but a tab is not allowed. (If the line begins with a tab, it will be considered part of a recipe for a rule.) Aside from this, extra spaces or tabs may be inserted with no effect anywhere except within the directive name or within an argument. A comment starting with ‘#’ may appear at the end of the line.

The other two directives that play a part in a conditional are else and endif. Each of these directives is written as one word, with no arguments. Extra spaces are allowed and ignored at the beginning of the line, and spaces or tabs at the end. A comment starting with ‘#’ may appear at the end of the line.

Conditionals affect which lines of the makefile make uses. If the condition is true, make reads the lines of the text-if-true as part of the makefile; if the condition is false, make ignores those lines completely. It follows that syntactic units of the makefile, such as rules, may safely be split across the beginning or the end of the conditional.

make evaluates conditionals when it reads a makefile. Consequently, you cannot use automatic variables in the tests of conditionals because they are not defined until recipes are run (see Automatic Variables).

To prevent intolerable confusion, it is not permitted to start a conditional in one makefile and end it in another. However, you may write an include directive within a conditional, provided you do not attempt to terminate the conditional inside the included file.

7.3 Conditionals that Test Flags

You can write a conditional that tests make command flags such as ‘-t’ by using the variable MAKEFLAGS together with the findstring function (see Functions for String Substitution and Analysis). This is useful when touch is not enough to make a file appear up to date.

The findstring function determines whether one string appears as a substring of another. If you want to test for the ‘-t’ flag, use ‘t’ as the first string and the value of MAKEFLAGS as the other.

For example, here is how to arrange to use ‘ranlib -t’ to finish marking an archive file up to date:

archive.a: …
ifneq (,$(findstring t,$(MAKEFLAGS)))
        +touch archive.a
        +ranlib -t archive.a
else
        ranlib archive.a
endif

The ‘+’ prefix marks those recipe lines as “recursive” so that they will be executed despite use of the ‘-t’ flag. See Recursive Use of make.

8 Functions for Transforming Text

Functions allow you to do text processing in the makefile to compute the files to operate on or the commands to use in recipes. You use a function in a function call, where you give the name of the function and some text (the arguments) for the function to operate on. The result of the function’s processing is substituted into the makefile at the point of the call, just as a variable might be substituted.

8.1 Function Call Syntax

A function call resembles a variable reference. It can appear anywhere a variable reference can appear, and it is expanded using the same rules as variable references. A function call looks like this:

$(function arguments)

or like this:

${function arguments}

Here function is a function name; one of a short list of names that are part of make. You can also essentially create your own functions by using the call built-in function.

The arguments are the arguments of the function. They are separated from the function name by one or more spaces or tabs, and if there is more than one argument, then they are separated by commas. Such whitespace and commas are not part of an argument’s value. The delimiters which you use to surround the function call, whether parentheses or braces, can appear in an argument only in matching pairs; the other kind of delimiters may appear singly. If the arguments themselves contain other function calls or variable references, it is wisest to use the same kind of delimiters for all the references; write ‘$(subst a,b,$(x))’, not ‘$(subst a,b,${x})’. This is because it is clearer, and because only one type of delimiter is matched to find the end of the reference.

The text written for each argument is processed by substitution of variables and function calls to produce the argument value, which is the text on which the function acts. The substitution is done in the order in which the arguments appear.

Commas and unmatched parentheses or braces cannot appear in the text of an argument as written; leading spaces cannot appear in the text of the first argument as written. These characters can be put into the argument value by variable substitution. First define variables comma and space whose values are isolated comma and space characters, then substitute these variables where such characters are wanted, like this:

comma:= ,
empty:=
space:= $(empty) $(empty)
foo:= a b c
bar:= $(subst $(space),$(comma),$(foo))
# bar is now ‘a,b,c’.

Here the subst function replaces each space with a comma, through the value of foo, and substitutes the result.

8.2 Functions for String Substitution and Analysis

Here are some functions that operate on strings:

$(subst from,to,text)

Performs a textual replacement on the text text: each occurrence of from is replaced by to. The result is substituted for the function call. For example,

$(subst ee,EE,feet on the street)

substitutes the string ‘fEEt on the strEEt’.

$(patsubst pattern,replacement,text)

Finds whitespace-separated words in text that match pattern and replaces them with replacement. Here pattern may contain a ‘%’ which acts as a wildcard, matching any number of any characters within a word. If replacement also contains a ‘%’, the ‘%’ is replaced by the text that matched the ‘%’ in pattern. Only the first ‘%’ in the pattern and replacement is treated this way; any subsequent ‘%’ is unchanged.

‘%’ characters in patsubst function invocations can be quoted with preceding backslashes (‘\’). Backslashes that would otherwise quote ‘%’ characters can be quoted with more backslashes. Backslashes that quote ‘%’ characters or other backslashes are removed from the pattern before it is compared file names or has a stem substituted into it. Backslashes that are not in danger of quoting ‘%’ characters go unmolested. For example, the pattern the\%weird\\%pattern\\ has ‘the%weird\’ preceding the operative ‘%’ character, and ‘pattern\\’ following it. The final two backslashes are left alone because they cannot affect any ‘%’ character.

Whitespace between words is folded into single space characters; leading and trailing whitespace is discarded.

For example,

$(patsubst %.c,%.o,x.c.c bar.c)

produces the value ‘x.c.o bar.o’.

Substitution references (see Substitution References) are a simpler way to get the effect of the patsubst function:

$(var:pattern=replacement)

is equivalent to

$(patsubst pattern,replacement,$(var))

The second shorthand simplifies one of the most common uses of patsubst: replacing the suffix at the end of file names.

$(var:suffix=replacement)

is equivalent to

$(patsubst %suffix,%replacement,$(var))

For example, you might have a list of object files:

objects = foo.o bar.o baz.o

To get the list of corresponding source files, you could simply write:

$(objects:.o=.c)

instead of using the general form:

$(patsubst %.o,%.c,$(objects))

$(strip string)

Removes leading and trailing whitespace from string and replaces each internal sequence of one or more whitespace characters with a single space. Thus, ‘$(strip a b c )’ results in ‘a b c’.

The function strip can be very useful when used in conjunction with conditionals. When comparing something with the empty string ‘’ using ifeq or ifneq, you usually want a string of just whitespace to match the empty string (see Conditionals).

Thus, the following may fail to have the desired results:

.PHONY: all
ifneq   "$(needs_made)" ""
all: $(needs_made)
else
all:;@echo 'Nothing to make!'
endif

Replacing the variable reference ‘$(needs_made)’ with the function call ‘$(strip $(needs_made))’ in the ifneq directive would make it more robust.

$(findstring find,in)

Searches in for an occurrence of find. If it occurs, the value is find; otherwise, the value is empty. You can use this function in a conditional to test for the presence of a specific substring in a given string. Thus, the two examples,

$(findstring a,a b c)
$(findstring a,b c)

produce the values ‘a’ and ‘’ (the empty string), respectively. See Testing Flags, for a practical application of findstring.

$(filter pattern…,text)

Returns all whitespace-separated words in text that do match any of the pattern words, removing any words that do not match. The patterns are written using ‘%’, just like the patterns used in the patsubst function above.

The filter function can be used to separate out different types of strings (such as file names) in a variable. For example:

sources := foo.c bar.c baz.s ugh.h
foo: $(sources)
        cc $(filter %.c %.s,$(sources)) -o foo

says that foo depends of foo.c, bar.c, baz.s and ugh.h but only foo.c, bar.c and baz.s should be specified in the command to the compiler.

$(filter-out pattern…,text)

Returns all whitespace-separated words in text that do not match any of the pattern words, removing the words that do match one or more. This is the exact opposite of the filter function.

For example, given:

objects=main1.o foo.o main2.o bar.o
mains=main1.o main2.o

the following generates a list which contains all the object files not in ‘mains’:

$(filter-out $(mains),$(objects))

$(sort list)

Sorts the words of list in lexical order, removing duplicate words. The output is a list of words separated by single spaces. Thus,

$(sort foo bar lose)

returns the value ‘bar foo lose’.

Incidentally, since sort removes duplicate words, you can use it for this purpose even if you don’t care about the sort order.

$(word n,text)

Returns the nth word of text. The legitimate values of n start from 1. If n is bigger than the number of words in text, the value is empty. For example,

$(word 2, foo bar baz)

returns ‘bar’.

$(wordlist s,e,text)

Returns the list of words in text starting with word s and ending with word e (inclusive). The legitimate values of s start from 1; e may start from 0. If s is bigger than the number of words in text, the value is empty. If e is bigger than the number of words in text, words up to the end of text are returned. If s is greater than e, nothing is returned. For example,

$(wordlist 2, 3, foo bar baz)

returns ‘bar baz’.

$(words text)

Returns the number of words in text. Thus, the last word of text is $(word $(words text),text).

$(firstword names…)

The argument names is regarded as a series of names, separated by whitespace. The value is the first name in the series. The rest of the names are ignored.

For example,

$(firstword foo bar)

produces the result ‘foo’. Although $(firstword text) is the same as $(word 1,text), the firstword function is retained for its simplicity.

$(lastword names…)

The argument names is regarded as a series of names, separated by whitespace. The value is the last name in the series.

For example,

$(lastword foo bar)

produces the result ‘bar’. Although $(lastword text) is the same as $(word $(words text),text), the lastword function was added for its simplicity and better performance.

Here is a realistic example of the use of subst and patsubst. Suppose that a makefile uses the VPATH variable to specify a list of directories that make should search for prerequisite files (see VPATH Search Path for All Prerequisites). This example shows how to tell the C compiler to search for header files in the same list of directories.

The value of VPATH is a list of directories separated by colons, such as ‘src:../headers’. First, the subst function is used to change the colons to spaces:

$(subst :, ,$(VPATH))

This produces ‘src ../headers’. Then patsubst is used to turn each directory name into a ‘-I’ flag. These can be added to the value of the variable CFLAGS, which is passed automatically to the C compiler, like this:

override CFLAGS += $(patsubst %,-I%,$(subst :, ,$(VPATH)))

The effect is to append the text ‘-Isrc -I../headers’ to the previously given value of CFLAGS. The override directive is used so that the new value is assigned even if the previous value of CFLAGS was specified with a command argument (see The override Directive).

8.3 Functions for File Names

Several of the built-in expansion functions relate specifically to taking apart file names or lists of file names.

Each of the following functions performs a specific transformation on a file name. The argument of the function is regarded as a series of file names, separated by whitespace. (Leading and trailing whitespace is ignored.) Each file name in the series is transformed in the same way and the results are concatenated with single spaces between them.

$(dir names…)

Extracts the directory-part of each file name in names. The directory-part of the file name is everything up through (and including) the last slash in it. If the file name contains no slash, the directory part is the string ‘./’. For example,

$(dir src/foo.c hacks)

produces the result ‘src/ ./’.

$(notdir names…)

Extracts all but the directory-part of each file name in names. If the file name contains no slash, it is left unchanged. Otherwise, everything through the last slash is removed from it.

A file name that ends with a slash becomes an empty string. This is unfortunate, because it means that the result does not always have the same number of whitespace-separated file names as the argument had; but we do not see any other valid alternative.

For example,

$(notdir src/foo.c hacks)

produces the result ‘foo.c hacks’.

$(suffix names…)

Extracts the suffix of each file name in names. If the file name contains a period, the suffix is everything starting with the last period. Otherwise, the suffix is the empty string. This frequently means that the result will be empty when names is not, and if names contains multiple file names, the result may contain fewer file names.

For example,

$(suffix src/foo.c src-1.0/bar.c hacks)

produces the result ‘.c .c’.

$(basename names…)

Extracts all but the suffix of each file name in names. If the file name contains a period, the basename is everything starting up to (and not including) the last period. Periods in the directory part are ignored. If there is no period, the basename is the entire file name. For example,

$(basename src/foo.c src-1.0/bar hacks)

produces the result ‘src/foo src-1.0/bar hacks’.

$(addsuffix suffix,names…)

The argument names is regarded as a series of names, separated by whitespace; suffix is used as a unit. The value of suffix is appended to the end of each individual name and the resulting larger names are concatenated with single spaces between them. For example,

$(addsuffix .c,foo bar)

produces the result ‘foo.c bar.c’.

$(addprefix prefix,names…)

The argument names is regarded as a series of names, separated by whitespace; prefix is used as a unit. The value of prefix is prepended to the front of each individual name and the resulting larger names are concatenated with single spaces between them. For example,

$(addprefix src/,foo bar)

produces the result ‘src/foo src/bar’.

$(join list1,list2)

Concatenates the two arguments word by word: the two first words (one from each argument) concatenated form the first word of the result, the two second words form the second word of the result, and so on. So the nth word of the result comes from the nth word of each argument. If one argument has more words that the other, the extra words are copied unchanged into the result.

For example, ‘$(join a b,.c .o)’ produces ‘a.c b.o’.

Whitespace between the words in the lists is not preserved; it is replaced with a single space.

This function can merge the results of the dir and notdir functions, to produce the original list of files which was given to those two functions.

$(wildcard pattern)

The argument pattern is a file name pattern, typically containing wildcard characters (as in shell file name patterns). The result of wildcard is a space-separated list of the names of existing files that match the pattern. See Using Wildcard Characters in File Names.

$(realpath names…)

For each file name in names return the canonical absolute name. A canonical name does not contain any . or .. components, nor any repeated path separators (/) or symlinks. In case of a failure the empty string is returned. Consult the realpath(3) documentation for a list of possible failure causes.

$(abspath names…)

For each file name in names return an absolute name that does not contain any . or .. components, nor any repeated path separators (/). Note that, in contrast to realpath function, abspath does not resolve symlinks and does not require the file names to refer to an existing file or directory. Use the wildcard function to test for existence.

8.4 Functions for Conditionals

There are three functions that provide conditional expansion. A key aspect of these functions is that not all of the arguments are expanded initially. Only those arguments which need to be expanded, will be expanded.

$(if condition,then-part[,else-part])

The if function provides support for conditional expansion in a functional context (as opposed to the GNU make makefile conditionals such as ifeq (see Syntax of Conditionals).

The first argument, condition, first has all preceding and trailing whitespace stripped, then is expanded. If it expands to any non-empty string, then the condition is considered to be true. If it expands to an empty string, the condition is considered to be false.

If the condition is true then the second argument, then-part, is evaluated and this is used as the result of the evaluation of the entire if function.

If the condition is false then the third argument, else-part, is evaluated and this is the result of the if function. If there is no third argument, the if function evaluates to nothing (the empty string).

Note that only one of the then-part or the else-part will be evaluated, never both. Thus, either can contain side-effects (such as shell function calls, etc.)

$(or condition1[,condition2[,condition3…]])

The or function provides a “short-circuiting” OR operation. Each argument is expanded, in order. If an argument expands to a non-empty string the processing stops and the result of the expansion is that string. If, after all arguments are expanded, all of them are false (empty), then the result of the expansion is the empty string.

$(and condition1[,condition2[,condition3…]])

The and function provides a “short-circuiting” AND operation. Each argument is expanded, in order. If an argument expands to an empty string the processing stops and the result of the expansion is the empty string. If all arguments expand to a non-empty string then the result of the expansion is the expansion of the last argument.

8.5 The foreach Function

The foreach function is very different from other functions. It causes one piece of text to be used repeatedly, each time with a different substitution performed on it. It resembles the for command in the shell sh and the foreach command in the C-shell csh.

The syntax of the foreach function is:

$(foreach var,list,text)

The first two arguments, var and list, are expanded before anything else is done; note that the last argument, text, is not expanded at the same time. Then for each word of the expanded value of list, the variable named by the expanded value of var is set to that word, and text is expanded. Presumably text contains references to that variable, so its expansion will be different each time.

The result is that text is expanded as many times as there are whitespace-separated words in list. The multiple expansions of text are concatenated, with spaces between them, to make the result of foreach.

This simple example sets the variable ‘files’ to the list of all files in the directories in the list ‘dirs’:

dirs := a b c d
files := $(foreach dir,$(dirs),$(wildcard $(dir)/*))

Here text is ‘$(wildcard $(dir)/*)’. The first repetition finds the value ‘a’ for dir, so it produces the same result as ‘$(wildcard a/*)’; the second repetition produces the result of ‘$(wildcard b/*)’; and the third, that of ‘$(wildcard c/*)’.

This example has the same result (except for setting ‘dirs’) as the following example:

files := $(wildcard a/* b/* c/* d/*)

When text is complicated, you can improve readability by giving it a name, with an additional variable:

find_files = $(wildcard $(dir)/*)
dirs := a b c d
files := $(foreach dir,$(dirs),$(find_files))

Here we use the variable find_files this way. We use plain ‘=’ to define a recursively-expanding variable, so that its value contains an actual function call to be re-expanded under the control of foreach; a simply-expanded variable would not do, since wildcard would be called only once at the time of defining find_files.

The foreach function has no permanent effect on the variable var; its value and flavor after the foreach function call are the same as they were beforehand. The other values which are taken from list are in effect only temporarily, during the execution of foreach. The variable var is a simply-expanded variable during the execution of foreach. If var was undefined before the foreach function call, it is undefined after the call. See The Two Flavors of Variables.

You must take care when using complex variable expressions that result in variable names because many strange things are valid variable names, but are probably not what you intended. For example,

files := $(foreach Esta-escrito-en-espanol!,b c ch,$(find_files))

might be useful if the value of find_files references the variable whose name is ‘Esta-escrito-en-espanol!’ (es un nombre bastante largo, no?), but it is more likely to be a mistake.

8.6 The file Function

The file function allows the makefile to write to a file. Two modes of writing are supported: overwrite, where the text is written to the beginning of the file and any existing content is lost, and append, where the text is written to the end of the file, preserving the existing content. In all cases the file is created if it does not exist.

The syntax of the file function is:

$(file op filename[,text])

The operator op can be either > which indicates overwrite mode, or >> which indicates append mode. The filename indicates the file to be written to. There may optionally be whitespace between the operator and the file name.

When the file function is expanded all its arguments are expanded first, then the file indicated by filename will be opened in the mode described by op. Finally text will be written to the file. If text does not already end in a newline, even if empty, a final newline will be written. If the text argument is not given, nothing will be written. The result of evaluating the file function is always the empty string.

It is a fatal error if the file cannot be opened for writing, or if the write operation fails.

For example, the file function can be useful if your build system has a limited command line size and your recipe runs a command that can accept arguments from a file as well. Many commands use the convention that an argument prefixed with an @ specifies a file containing more arguments. Then you might write your recipe in this way:

program: $(OBJECTS)
        $(file >$@.in,$^)
        $(CMD) $(CMDFLAGS) @$@.in
        @rm $@.in

If the command required each argument to be on a separate line of the input file, you might write your recipe like this:

program: $(OBJECTS)
        $(file >$@.in) $(foreach O,$^,$(file >>$@.in,$O))
        $(CMD) $(CMDFLAGS) @$@.in
        @rm $@.in

8.7 The call Function

The call function is unique in that it can be used to create new parameterized functions. You can write a complex expression as the value of a variable, then use call to expand it with different values.

The syntax of the call function is:

$(call variable,param,param,…)

When make expands this function, it assigns each param to temporary variables $(1), $(2), etc. The variable $(0) will contain variable. There is no maximum number of parameter arguments. There is no minimum, either, but it doesn’t make sense to use call with no parameters.

Then variable is expanded as a make variable in the context of these temporary assignments. Thus, any reference to $(1) in the value of variable will resolve to the first param in the invocation of call.

Note that variable is the name of a variable, not a reference to that variable. Therefore you would not normally use a ‘$’ or parentheses when writing it. (You can, however, use a variable reference in the name if you want the name not to be a constant.)

If variable is the name of a built-in function, the built-in function is always invoked (even if a make variable by that name also exists).

The call function expands the param arguments before assigning them to temporary variables. This means that variable values containing references to built-in functions that have special expansion rules, like foreach or if, may not work as you expect.

Some examples may make this clearer.

This macro simply reverses its arguments:

reverse = $(2) $(1)

foo = $(call reverse,a,b)

Here foo will contain ‘b a’.

This one is slightly more interesting: it defines a macro to search for the first instance of a program in PATH:

pathsearch = $(firstword $(wildcard $(addsuffix /$(1),$(subst :, ,$(PATH)))))

LS := $(call pathsearch,ls)

Now the variable LS contains /bin/ls or similar.

The call function can be nested. Each recursive invocation gets its own local values for $(1), etc. that mask the values of higher-level call. For example, here is an implementation of a map function:

map = $(foreach a,$(2),$(call $(1),$(a)))

Now you can map a function that normally takes only one argument, such as origin, to multiple values in one step:

o = $(call map,origin,o map MAKE)

and end up with o containing something like ‘file file default’.

A final caution: be careful when adding whitespace to the arguments to call. As with other functions, any whitespace contained in the second and subsequent arguments is kept; this can cause strange effects. It’s generally safest to remove all extraneous whitespace when providing parameters to call.

8.8 The value Function

The value function provides a way for you to use the value of a variable without having it expanded. Please note that this does not undo expansions which have already occurred; for example if you create a simply expanded variable its value is expanded during the definition; in that case the value function will return the same result as using the variable directly.

The syntax of the value function is:

$(value variable)

Note that variable is the name of a variable, not a reference to that variable. Therefore you would not normally use a ‘$’ or parentheses when writing it. (You can, however, use a variable reference in the name if you want the name not to be a constant.)

The result of this function is a string containing the value of variable, without any expansion occurring. For example, in this makefile:

FOO = $PATH

all:
        @echo $(FOO)
        @echo $(value FOO)

The first output line would be ATH, since the “$P” would be expanded as a make variable, while the second output line would be the current value of your $PATH environment variable, since the value function avoided the expansion.

The value function is most often used in conjunction with the eval function (see Eval Function).

8.9 The eval Function

The eval function is very special: it allows you to define new makefile constructs that are not constant; which are the result of evaluating other variables and functions. The argument to the eval function is expanded, then the results of that expansion are parsed as makefile syntax. The expanded results can define new make variables, targets, implicit or explicit rules, etc.

The result of the eval function is always the empty string; thus, it can be placed virtually anywhere in a makefile without causing syntax errors.

It’s important to realize that the eval argument is expanded twice; first by the eval function, then the results of that expansion are expanded again when they are parsed as makefile syntax. This means you may need to provide extra levels of escaping for “$” characters when using eval. The value function (see Value Function) can sometimes be useful in these situations, to circumvent unwanted expansions.

Here is an example of how eval can be used; this example combines a number of concepts and other functions. Although it might seem overly complex to use eval in this example, rather than just writing out the rules, consider two things: first, the template definition (in PROGRAM_template) could need to be much more complex than it is here; and second, you might put the complex, “generic” part of this example into another makefile, then include it in all the individual makefiles. Now your individual makefiles are quite straightforward.

PROGRAMS    = server client

server_OBJS = server.o server_priv.o server_access.o
server_LIBS = priv protocol

client_OBJS = client.o client_api.o client_mem.o
client_LIBS = protocol

# Everything after this is generic

.PHONY: all
all: $(PROGRAMS)

define PROGRAM_template =
 $(1): $$($(1)_OBJS) $$($(1)_LIBS:%=-l%)
 ALL_OBJS   += $$($(1)_OBJS)
endef

$(foreach prog,$(PROGRAMS),$(eval $(call PROGRAM_template,$(prog))))

$(PROGRAMS):
        $(LINK.o) $^ $(LDLIBS) -o $@

clean:
        rm -f $(ALL_OBJS) $(PROGRAMS)

8.10 The origin Function

The origin function is unlike most other functions in that it does not operate on the values of variables; it tells you something about a variable. Specifically, it tells you where it came from.

The syntax of the origin function is:

$(origin variable)

Note that variable is the name of a variable to inquire about, not a reference to that variable. Therefore you would not normally use a ‘$’ or parentheses when writing it. (You can, however, use a variable reference in the name if you want the name not to be a constant.)

The result of this function is a string telling you how the variable variable was defined:

‘undefined’

if variable was never defined.

‘default’

if variable has a default definition, as is usual with CC and so on. See Variables Used by Implicit Rules. Note that if you have redefined a default variable, the origin function will return the origin of the later definition.

‘environment’

if variable was inherited from the environment provided to make.

‘environment override’

if variable was inherited from the environment provided to make, and is overriding a setting for variable in the makefile as a result of the ‘-e’ option (see Summary of Options).

‘file’

if variable was defined in a makefile.

‘command line’

if variable was defined on the command line.

‘override’

if variable was defined with an override directive in a makefile (see The override Directive).

‘automatic’

if variable is an automatic variable defined for the execution of the recipe for each rule (see Automatic Variables).

This information is primarily useful (other than for your curiosity) to determine if you want to believe the value of a variable. For example, suppose you have a makefile foo that includes another makefile bar. You want a variable bletch to be defined in bar if you run the command ‘make -f bar’, even if the environment contains a definition of bletch. However, if foo defined bletch before including bar, you do not want to override that definition. This could be done by using an override directive in foo, giving that definition precedence over the later definition in bar; unfortunately, the override directive would also override any command line definitions. So, bar could include:

ifdef bletch
ifeq "$(origin bletch)" "environment"
bletch = barf, gag, etc.
endif
endif

If bletch has been defined from the environment, this will redefine it.

If you want to override a previous definition of bletch if it came from the environment, even under ‘-e’, you could instead write:

ifneq "$(findstring environment,$(origin bletch))" ""
bletch = barf, gag, etc.
endif

Here the redefinition takes place if ‘$(origin bletch)’ returns either ‘environment’ or ‘environment override’. See Functions for String Substitution and Analysis.

8.11 The flavor Function

The flavor function, like the origin function, does not operate on the values of variables but rather it tells you something about a variable. Specifically, it tells you the flavor of a variable (see The Two Flavors of Variables).

The syntax of the flavor function is:

$(flavor variable)

Note that variable is the name of a variable to inquire about, not a reference to that variable. Therefore you would not normally use a ‘$’ or parentheses when writing it. (You can, however, use a variable reference in the name if you want the name not to be a constant.)

The result of this function is a string that identifies the flavor of the variable variable:

‘undefined’

if variable was never defined.

‘recursive’

if variable is a recursively expanded variable.

‘simple’

if variable is a simply expanded variable.

8.12 Functions That Control Make

These functions control the way make runs. Generally, they are used to provide information to the user of the makefile or to cause make to stop if some sort of environmental error is detected.

$(error text…)

Generates a fatal error where the message is text. Note that the error is generated whenever this function is evaluated. So, if you put it inside a recipe or on the right side of a recursive variable assignment, it won’t be evaluated until later. The text will be expanded before the error is generated.

For example,

ifdef ERROR1
$(error error is $(ERROR1))
endif

will generate a fatal error during the read of the makefile if the make variable ERROR1 is defined. Or,

ERR = $(error found an error!)

.PHONY: err
err: ; $(ERR)

will generate a fatal error while make is running, if the err target is invoked.

$(warning text…)

This function works similarly to the error function, above, except that make doesn’t exit. Instead, text is expanded and the resulting message is displayed, but processing of the makefile continues.

The result of the expansion of this function is the empty string.

$(info text…)

This function does nothing more than print its (expanded) argument(s) to standard output. No makefile name or line number is added. The result of the expansion of this function is the empty string.

8.13 The shell Function

The shell function is unlike any other function other than the wildcard function (see The Function wildcard) in that it communicates with the world outside of make.

The shell function performs the same function that backquotes (‘`’) perform in most shells: it does command expansion. This means that it takes as an argument a shell command and evaluates to the output of the command. The only processing make does on the result is to convert each newline (or carriage-return / newline pair) to a single space. If there is a trailing (carriage-return and) newline it will simply be removed.

The commands run by calls to the shell function are run when the function calls are expanded (see How make Reads a Makefile). Because this function involves spawning a new shell, you should carefully consider the performance implications of using the shell function within recursively expanded variables vs. simply expanded variables (see The Two Flavors of Variables).

Here are some examples of the use of the shell function:

contents := $(shell cat foo)

sets contents to the contents of the file foo, with a space (rather than a newline) separating each line.

files := $(shell echo *.c)

sets files to the expansion of ‘*.c’. Unless make is using a very strange shell, this has the same result as ‘$(wildcard *.c)’ (as long as at least one ‘.c’ file exists).

8.14 The guile Function