Floating Point Conversion

By your friendly neighborhood Bilal.

isinteger() (validates a string of numbers if its an integer) �atol() (converts a string of numbers into a long integer)

Now we all (well most of us) completed assignment 2 and have used both of these.

​

• We use isinteger() to verify that our input was an integer.

​

• We use atol() (or atolong) to convert the ascii characters of the string into a long integer.

​

Here is how it Worked.

​

• String a = “355”;
• String b = “123fdgh”;

​

• Bool a_valid = isinteger(a); - returns true because string a is an int.
• Bool b_valid = isinteger(b); - returns false because string b is not an int.

Once a string is validated as an int we can call the atol function to convert it into a long integer!

​

• long int num = atol(a);

cout << num << endl;

output: 355 as int not as string.

Now what if the string was a floating point number instead of an integer?���**Surprise it is!!

isfloat() (validates a string of numbers if its a floating point number) �atof() (converts a string of numbers into a floating point number)

​

• We use isfloat() to verify our input was a floating point .

​

• We use atof() to convert the ascii characters of the string into a floating point number.

​

We Use Different Functions For Floats.

So What Is The Problem?

These exist as library functions:

​

• Isinteger() by Professor Holliday

isdigit() library function.

​

• Atol()

​

• Atof()

This does not exist as a library function:

​

Isfloat()

​

Solution

• Take Professor Holliday’s isinteger() function (or create your own) and include the only difference between integers and floats.

​

​

A DECIMAL!

​

• Therefore instead of validating the input of just a digit ( ‘0’, ’1’, ’2’, ’3’, ’5’, ’6’ ,’7’ , ’8’, ’9’), as in isdigit() and isinteger(), include the input of a decimal ( ‘.’ )!

From isinteger() to isfloat()

bool isinteger(char w[])

{

bool result = true; 

  int start = 0;

  if (w[0] == '-' || w[0] == '+') start = 1;

  unsigned long int k = start;

  while( !(w[k]=='\0') && result )

   {

result = result && isdigit(w[k]);

       k++;

   }

  return result;

}

From isinteger() to isfloat()

bool isinteger(char w[])

{

bool result = true; 

  int start = 0;

  if (w[0] == '-' || w[0] == '+') start = 1;

  unsigned long int k = start;

  while( !(w[k]=='\0') && result )

   {

result = result && isdigit(w[k])

|| result && (w[k] == ‘.’);

       k++;

   }

  return result;

}

From isinteger() to isfloat()

bool isfloat(char w[])

{

bool result = true;

bool onedecimal = false; 

  int start = 0;

  if (w[0] == '-' || w[0] == '+') start = 1;

  unsigned long int k = start;

  while( !(w[k]=='\0') && result )

   {

if (w[k] == ‘.’ && !onedecimal) onedecimal = true;

else {

result = result && isdigit(w[k])

|| result && (w[k] == ‘.’); }

       k++;

   }

  return result && onedecimal;}

Atof()

• Exactly the same way you call atol()
• mov rax, 0

mov rdi, rsp

call atol   

​

• Only difference is passing a 1 to rax instead of 0 to indicate the passing of a float number

and no need to copy rsp to rdi because atof will return its value into xmm0 register

• mov rax, 1

call atof

​

​

I Believe In You All!

But nonetheless any questions?