Compiling C++ Programs

Why Compile?

Recall:

Machine language instructions (ML) are the only instructions the processor/CPU can execute.
ML is just numbers. When the ML is put into memory, it can be fetched, decoded and executed by the processor. Different numbers will cause the processor to perform different actions.
Different type of processors usually have different ML's
Assembly language (AL) is a symbolic representation of ML.

A translation of the C++ program statements to machine language is necessary.

Stages in Compiling

A number of steps are involved in the compilation process.

                                                 other ML   ----+

         +--------------+   +----------+   +-----------+   +--------+

  C++ -->| PREPROCESSOR |-->| COMPILER |-->| ASSEMBLER |-->| LINKER |-->

Program  +--------------+   +----------+   +-----------+   +--------+

  Output:  Modified C++       Assembly      Executable     "Complete"

           program (copy)     language        code (ML)    Executable

Stages in Compiling - Explanation

PREPROCESSOR	The source program undergoes modification before it is compiled. A modified copy of the original program is created, however it is still valid C++. Things the preprocessor does includes: `#include`s are processed. In the copy of the original file, the `#include` is removed, and C++ statements are inserted at the point of the `#include` just as if the programmer had typed them in themself. Comments are removed.
COMPILER	The C++ statements are converted into AL statements.
ASSEMBLER	The AL statements are converted into ML called object code.
LINKER	Object code from multiple sources is combined together to form a complete executable program. For example the ML to perform input/output is almost always linked in since most programs do input or output.

Linking

Linking is an important phase of the compilation process.

All machine language can be compiled into a single executable program. This is called static linking. More common today is dynamic linking. With dynamic linking the "other" machine language remains separate and the linker arranges for your program to connect to the "other" machine language. Typing, at the prompt:

ldd executable_file_name

will show the dynamic libraries (files) the program requires. Try a.out for the executable_file_name.

When the program name is typed at the prompt the command interpreter, usually called the shell, causes the operating system to load the program into memory and begin executing it.

Note that the whole process is also called compiling and gets its name from the subpart, the conversion from C++ to AL.

`clang++` Compiling Options

`clang++ file.cpp`	Compile file file.cpp and create an executable named a.out
`clang++ file.cpp -o file2`	Compile file file.cpp, create an executable, and name the executable file2 instead of a.out
`clang++ -Wall file.cpp`	Output warnings Always compile with the warnings option
`clang++ -c file.cpp`	Run the first 3 stages, but not the linker. The machine language is put in file.o. This will be used when a program consists of more than one file, the linker will then be run on all the .o files.
`clang++ -E file.cpp`	Run the first stage of the compiler, the preprocessor, only. Output is sent to stdout (standard output, by default the terminal window).
`clang++ -H file.cpp`	Show header file inclusion only. Header files may include other header files. Including only iostream causes many other files to be included.
`clang++ --version`	Output the version of `clang++`