Continuing from our previous example, The linker can now takes over, it searches for the actual object code implementation containing  scanf and printf, (along with any other dependencies), and combines it with the original main.o code  from the example above.

In combining the object files, the references are resolved so that scanf and printf implementation can be accessed by the main.o machine code  in the same way that find_sum can be accessed from the original object code file. The result is a stand alone executable file which is essentially our program.

In summary, Linkers combine the generated object code from the translation process to other dependencies or references needed to execute the program, it converts reference placeholders to actual addresses, thus linking the main program to its dependencies to create fully functioning executable code.

© 2024  Vedesh Kungebeharry. All rights reserved. 

The optimized TAC is now used to generate assembly code.

The assembly code is converted into object code in a process  known as assembly.

This object code contains machine instructions corresponding to the source code, but may not be fully functional for execution as yet.

Consider the following example in source code in a file “main.c” :

#include <stdio.h>
//function prototype omitted for simplicity
int find_sum(int num1, int num2) {
    return num1 + num2;
}

int main() {
    int num1, num2;
    
    printf("Enter the first number: ");
    scanf("%d", &num1);

    printf("Enter the second number: ");
    scanf("%d", &num2);

    int result = find_sum(num1, num2);
    printf("The sum of %d and %d is %d\n", num1, num2, result);

    return 0;
}

The machine code for the main and find_sum function is generated, and stored in the object file, main.o . Since scanf and printf is used but exist in other libraries, the compiler creates code know as references (placeholders) to call scanf and printf, but doesn’t include the implementation of of scanf and printf in in the main.o object file. The creation of main.o was the final step of the code generation.

Now we move on to the last phase to use a linkers combine  various bits of object code to create one executable file.

© 2024  Vedesh Kungebeharry. All rights reserved. 

Three address code (TAC) has the advantage of being easily modified  to optimize the code. Code optimization is where the newly generated intermediate code is processed to remove redundancy in order to produce efficient TAC that will eventually be used to generate object code.

Our previous example is not sufficient to illustrate code optimization; we will use the following source code:

#include <stdio.h>
int main()
{
    int a=1;
    int b=2;
    int c=3;
    
    a= a+b;
    a= a+c; //a is now a+b+c
    c= a-b; //in the end,c=a+c
    return 0;
} 

In TAC, here’s a comparison between the newly generated intermediate code and the optimized  code :

int main() {
    int a = 1;
    int b = 2;
    int c = 3;

    t1 = a + b; // t1 = 1 + 2
    a = t1;     // a = t1 (a is now a+b)

    t2 = a + c; // t2 = (a+b) + 3
    a = t2;     // a = t2 (a is now a+b+c)

    t3 = a - b; // t3 = (a+b+c) - 2
    c = t3;     // c = t3 (c is now a+b+c)

    return 0;
}

int main() {
    int a = 6; // Optimized: a = 6 (1 + 2 + 3)
               // (Constant Folding)
    int c = a; // Optimized: c = a (c is now a+b+c)
               // (Algebraic simplification)
    return 0;
}

From here, we move onto the final stage, Code Generation, which produces object code from our optimized TAC

© 2024  Vedesh Kungebeharry. All rights reserved. 

After all structures are generated without any errors, a simplified intermediate code is generated. In c compilation, the intermediate code is three address code (TAC).

TAC :

• consists of simplified instructions, and operations only take 3 operands , a destination , 2 source operands and an operation.  
  
  
• TAC uses multiple intermediate variables to arrive at the value for complex expressions.

For example, the following operation,

Could be represented as:

t1 = b * b      // Compute the square of b and store 
                //it in temporary variable t1
t2 = 4 * a      // Compute 4 times the value of a and 
                //store it in temporary variable t2
t3 = t2 * c     // Multiply the result of t2 by c 
                //and store it in temporary variable t3
t4 = t1 - t3    // Subtract the value of t3 from t1 
                //and store it in temporary variable t4
D = t4          // Assign the value in t4 to variable D

Note that the above TAC code is not exact, it has been simplified for the purposes of demonstration.

The actual source code and corresponding TAC code is shown here:

#include <stdio.h>
int main()
{   
    float D,b,a,c;
    D= b*b - 4*a*c; 
    return 0;
}

main ()
{
  int D.3251;

  {
    float D;
    float b;
    float a;
    float c;

    _1 = b * b;
    _2 = a * 4.0e+0;
    _3 = c * _2;
    D = _1 - _3;
    D.3251 = 0;
    return D.3251;
  }
  D.3251 = 0;
  return D.3251;
}

The TAC was obtained by running the command “gcc -fdump-tree-all -o program.exe source.c”

See all files here: https://drive.google.com/drive/folders/1dDYCaNddQFpAF2oDknWrEFkPpj692xQ7?usp=sharing

© 2024  Vedesh Kungebeharry. All rights reserved. 

Lexical Errors

Syntax Errors

Semantic Errors

© 2024  Vedesh Kungebeharry. All rights reserved. 

Checks in semantic analysis include:

• Type checking :
  • Type mismatch checking. E.g,

  int x = “Hello”;  

• Type cast checking. E.g,

int radius = 3;
float Pi = 3.14;
float diameter = 2*radius*Pi; 
/*radius will be typecast to float
during the calculation, i.e 3.00 
(This result is reflected in the next step, 
intermediate code generation
 as  TAC)
*/

the TAC code generated from the above example:

t1 = 2.0  // Represents the float constant 2.0
t2 = (float)radius  // Type casting 'radius' to float
t3 = t1 * t2  // Multiplying 2.0 by the type-casted 'radius'
t4 = t3 * Pi  // Multiplying the result by 'Pi'

• Variable Declaration checks :
  
  
  • Ensuring all variables are declared. Eg,

int x = 10;
int y = 10;
sum = x+y; //sum was not declared

• Reserved identifier / keyword use. E.g,

int char = '1';//char is a reserved identifier
int char_2;

int if; //if is a reserved keyword
int if_2; 

• Variable scope check
  
  A unique variable declared within a function cannot be used by any other structures or functions outside of the original function block. Eg,

void demoFunction() 
{
    int uniqueX = 10;
}

int main() 
{
    printf("%d\n", uniqueX); // Error: 'uniqueX' is out of scope
    return 0;
}

• Function Declaration and Invocation checks:
  • Function declaration check – missed function declaration. e.g

int main() {
    // Calling a function before it is declared or defined
    hello(); // Error: 'hello' is called before declaration/definition
    
    return 0;
}
/*the declaration was omitted, 
this function declaration
is required: 

void hello();

*/
void hello() {
    // Function definition for 'hello'
    printf("Hello, world!\n");
}

•  Function declaration check – missing function implementation when a function prototype (function declaration) was declared. E.g,

// Function prototype (declaration)
int multiply(int a, int b);

int main() {
    int result = multiply(3, 7); /* Error: Function 
                                'multiply' 
                                is declared but not 
                                defined/implemented*/    
    return 0;
}

• Function invocation checks.

(NB, function invocation simply means to call a function correctly)

int add(int a, int b);

int add(int a, int b) {
    return a + b;
}

int main() {
    int sum1 = add(5, 10, 15); /* Error: Too many arguments
                                in the function call*/
                                
    int sum2 = add(5);/*Error: Missing arguments 
                        in the function call*/
                        
    int sum3 = add(1,2) /*proper invocation of add,
                        no error here*/
    return 0;
}

• Control Flow Checks -Verify the correct structure and use of control structures: if, while, for etc
  
  

Other checks exist , we have only examined some of the main semantic checks.

Recall that once all semantic checks are passed, Intermediate code generation occurs next.

Examples of Semantic errors in C

© 2024  Vedesh Kungebeharry. All rights reserved.