Converting String to Integer

[Cmaranec]

Hello,
how can i convert integer to (for example) char dd[128] ? I tried to create function like IntToStr (from Pascal), but it didn't work...

[Bughunter]

1. Divide the value by the base of the numbering system (for decimal, the base is 10)
2. Convert the remainder value of the division to ASCII (for decimal, this is usually done by adding 0x30)
3. Check if the quotient of the division is 0, if it is _not_ then go back to step 1.
4. When the loop ends, reverse the string buffer (for example '54321' becomes '12345')

[Cmaranec]

Ok, i will try it, thanks

[Cmaranec]

Yes, it works fine, thanks.

Question 2: How can i convert string to integer? (with error code) i have char firstc[64]; and char secndc[64]; and int first; and int secnd; and i want to "copy" all from 'firstc' to 'first' and from 'secndc' to 'secnd'.

[Brendan]

Yes, it works fine, thanks.

Question 2: How can i convert string to integer? (with error code) i have char firstc[64]; and char secndc[64]; and int first; and int secnd; and i want to "copy" all from 'firstc' to 'first' and from 'secndc' to 'secnd'.

Something like this perhaps?

Code: Select all

char *convertStringToInt(int *number, char *string) {
    if(strncasecmp(string, "0x", 2) == 0) {
        // Hexadecimal
        return convertStringToIntWithBase(number, &string[2], 16);
    }
    if(*string == '0') {
        // Octal???
        return convertStringToIntWithBase(number, &string[1], 8);
    }
    // Decimal
    return convertStringToIntWithBase(number, string, 10);
}


char *convertStringToIntWithBase(int *number, char *string, int base) {
    int digit;

    *number = 0;

    for(;;) {
        if( (*string >= '0') && (*string <= '9') )  {
            digit = *string - '0';
        } else if( (*string >= 'A') && (*string <= 'Z') )  {
            digit = *string - 'A' + 10;
        } else if( (*string >= 'a') && (*string <= 'z') )  {
            digit = *string - 'a' + 10;
        } else {
            break;
        }
        if(digit >= base) {
            break;
        }
        if(*number >= INT_MAX / base) {
            return NULL; // Will overflow...
        }
        *number = *number * base;
        if(*number >= INT_MAX - digit) {
            return NULL; // Will overflow...
        }
        *number = *number + digit;
        string++;
    }
    return string;
}

Of course you can do more. For e.g. supporting binary, handling leading '-' and '+' characters, supporting arbitrary number bases (e.g. "0r6r54321" could be a number in base 6), handling exponents (e.g. "-1E2" could be -100 in decimal), etc. You could also combine all of that, so that "-0r4r32100E-0r16r1" is a base 4 number with a negative base 16 exponent (in this case, -57 in decimal).


Cheers,

Brendan

[Cmaranec]

Thank you, i am doing simple calculator

[Bughunter]

Yes, it works fine, thanks.

Great job, congrats!

[Cmaranec]

Sorry, have you got other code for convert string to integer?

[Bughunter]

Basically, you make a loop and pick a character each loop, convert the character back to a raw value, then multiply that by 10 each loop iteration. Oh yes, start at the end of the string and decrease your counter every iteration.

I'll see if I have an example laying around.

[Bughunter]

Here is an example:

Code: Select all

int atodw(const char *str) {
    int value = 0;

    for(int counter = strlen(str)-1, multiplier = 1; !(counter < 0); --counter, multiplier *= 10) {
        value += (str[counter] - 0x30) * multiplier;
    }
   
    return value;
}

[XCHG]

Here is a function that I have coded for my kernel that converts an unsigned string to a 32-bit Unsigned DWORD value. This one is optimized and it wouldn't use the MUL instruction at all for multiplication:

Code: Select all

; --------------------------------------------------
  __StrToDWORD:
    ; DWORD __StrToDWORD (const char* InStr)
    PUSH    EBX
    PUSH    EDX
    PUSH    EBP
    MOV     EBP , ESP
    MOV     EBX , DWORD PTR [EBP + 0x10]
    XOR     EAX , EAX
    .Loop:
      MOV     EDX , DWORD PTR [EBX]
      AND     EDX , 0x000000FF
      JE      .EP
      AND     EDX , 0x0000000F
      ADD     EAX , EAX
      LEA     EAX , [EAX + 0x04*EAX]
      ADD     EAX , EDX
      INC     EBX
      JMP     .Loop
    .EP:
      POP     EBP
      POP     EDX
      POP     EBX
    RET     0x04
; --------------------------------------------------

[Bughunter]

I have written a similar function ASM some time ago. Cmaranec wanted a C function so why post the ASM one? Since Cmaranec also asked how to convert a string back to an integer, I don't think he understands much of ASM since he's just exploring calculators

Imho, you shouldn't overthrow someone with 'complicated' stuff that they will probably not (yet) understand.

[Cmaranec]

OK, thanks to all, i will try bughunter's example

[Cmaranec]

Wow, thanks ! All the functions works well

[Bughunter]

Wow, thanks ! All the functions works well

That's good to hear