Module 7: Build Automation with Make

• First read this page then start the module with the GitHub classroom link below.
• Github Classroom Link: https://classroom.github.com/a/uwjQefOh

Compilation

When you compile a C++ program, the compiler often does the following steps:

1. Preprocessing: Handles macros, #include directives, and other preprocessor commands.
2. Compilation: Translates C++ code into assembly language. This is where a .s file can be generated if desired.
3. Assembly: Converts the assembly code into machine code, producing an object file (usually with a .o or .obj extension).
4. Linking: Combines multiple object files and resolves references to produce the final executable or shared library.

Exercise 1

Go to the exercise1 directory of your module 8 GitHub repository. Put your answers to the questions below in the README.md in your module 8 GitHub repository.

1. Run the following commands to inspect the file type of each file.

   $ file calculate.cpp
   $ file functions.cpp
   $ file functions.h
   

2. Read the contents of the calculate.cpp, functions.cpp, and functions.h file.
3. What is the purpose of including the functions.h file in calculate.cpp?
4. Run the following commands to run the preprocessing step.

   $ g++ -E -P calculate.cpp > calculate.i
   $ g++ -E -P functions.cpp > functions.i
   

5. Read the contents of calculate.i and functions.i.
6. What has changed compared to the .cpp versions of these files?
7. Now run the following commands to run the compilation step.

   $ g++ -S calculate.i
   $ g++ -S functions.i
   

8. Run the following commands to inspect the file type of each file.

   $ file calculate.s
   $ file functions.s
   

9. Read the contents of calculate.s and functions.s.
10. What has changed compared to the .i versions of these files?
11. Now run the following commands to run the assembly step.

    $ g++ -c calculate.s
    $ g++ -c functions.s
    

12. Run the following commands to inspect the file type of each file.

    $ file calculate.o
    $ file functions.o
    

13. Read the contents of the calculate.o and functions.o by running.

    $ xxd -bits calculate.o
    $ xxd -bits functions.o
    

14. What has changed compared to the .s versions of these files?
15. Now run the following command to run the linking step.

    $ g++ calculate.o functions.o -o calculate
    

16. Why did we have to provide both .o files at the same time when linking?
17. Run the following commands to inspect the file type of the executable.

    $ file calculate
    

18. How is the calculate executable file different from the .o files?
19. Run the executable.

    $ ./calculate
    

20. In Linux (and Unix-like systems), the return value (or exit status) of the last command executed is stored in a special shell variable ?. You can access this value using $?.

    $ echo $?
    

21. Is the return value what you expected from running calculate?

Build Automation Tools: Make

Makefile: The make utility reads a special file known as a makefile.

• The makefile (or Makefile) describes the files involved in the project and the dependencies among them.
• Each group of lines in a makefile has the following form.

  target: dependencies
  <Tab> commands
  

• In this context,
  • target is a file to be created
  • dependencies is a list of files on which the target depends
  • commands is a list of commands used to (re)create the target.
• If any of the files in a dependency list is modified, make will recompile and/or relink the target.
• Important: a Tab character must precede the commands — not spaces.

Exercise 2

Go to the exercise2 directory of your module 8 GitHub repository. Put your answers to the questions below in the README.md in your module 8 GitHub repository.

1. Review the Makefile in the exercise2 directory.
2. Run the make command. Copy and paste the result of issuing the make command.
3. Run the make command again. Copy and paste the result of issuing the make command.
4. Run the commands:

   $ touch functions.h
   $ make
   

   Copy and paste the result of issuing both commands.

5. What did the touch command do in this case (man touch)?

Exercise 3

Go to the exercise3 directory of your module 8 GitHub repository. Put your answers to the questions below in the README.md in your module 8 GitHub repository.

• One way to avoid having all files be recompiled each time is to make use of object files.
• To create an object file for the hello.cpp source file, you could issue the command g++ -c hello.cpp (but do not do so here).
• The result would be a file hello.o. This file is not executable because no linking has taken place, only compilation — the result of using the -c flag.
• Create a rule in your Makefile for creating hello.o similar to the following (remember tab, not spaces!):

  hello.o: hello.cpp functions.h
    g++ -g -Wall -c hello.cpp
  

• Notice that the dependencies for hello.o are hello.cpp and functions.h.
• The hello.cpp source file as a dependency should be obvious. The function.h is also a dependency because it is included in the file hello.cpp.
• Now create similar rules for creating main.o and factorial.o. (What dependencies do these two need? Make sure to modify the corresponding commands in the current Makefile as well.)

1. Run make. List all .o files created by the make command.
2. Run make. Explain the result.
3. Run make main.o. Copy and paste the result of issuing this command.
4. Run

   $ make factorial.o 
   $ ls -al
   

   Is there an executable present for running main?

5. Create another rule as the last rule in Makefile that will create the target hello based on the dependencies hello.o, main.o, and factorial.o. Hint: g++ hello.o main.o factorial.o -o hello
6. Run:

   $ make hello
   $ ./hello
   

   Copy and paste the output from issuing both commands.

7. Run:

   $ rm *.o hello
   $ make
   $ ./hello
   

   Copy and paste the output from issuing the commands.

8. Explain the previous output.
9. Make the necessary modification to have hello be the default target created by make.
10. Run:

    $ touch functions.h
    $ make
    

    Copy and paste the output from issuing the commands.

11. Explain the previous output.
12. Run:

    $ touch hello.cpp
    $ make
    

    Copy and paste the output from issuing the commands.

13. Explain the previous output.
14. Run: rm *.o hello
15. Could we get make to handle cleanup work like this for us? Yes — the target does not have to be a file-to-be-created. In the examples we have seen thus far, the command has been a compile command which naturally creates a file. We can issue other command-prompt commands that do not result in a created file. The typical way to have make handle the cleanup is to create a new rule at the end of your makefile. The rule should have a target called clean (no dependencies) with the command /bin/rm -f *.o hello. Add this rule to the end Makefile.
16. Run:

    $ make
    $ make clean
    

17. Copy and paste the output from issuing the commands.
18. Should the clean target be first in the makefile? Explain.