Question

I’m trying to compare 2 strings to see if they’re equal in assembly. Case sensitive. They’re passed in from a cpp wrapper. Here is what I have so far:

I stored one string in esi, the other in edi:

LOOP:
    mov al, [esi + edx*4]
    mov bl, [edi + edx*4]
    sub al, bl
    INC edx
    je LOOP
    jmp END_LOOP  ; if it's not equal, do some stuff at the end

END_LOOP:

What am I doing wrong? How does it know when I’m at the end of a char array?

Answer 1

Since these strings are C++ strings, and I’m guessing they are made of chars and not wide characters, then each character is one byte in size so you should not be multiplying the index register edx by 4. Also, you cannot increment edx before immediately before your je instruction, because that only jumps is edx has been bumped to zero.

Finally, the strings end with a 0 byte. So you are looking for either al or bl to be zero to know when to stop your loop.

You can modify your code along these lines:

    xor edx, edx         ; make sure edx is 0 to start with
LOOP:
    mov al, [esi + edx]
    mov bl, [edi + edx]
    inc edx              ; prepare for next char
    cmp al, bl           ; compare two current characters
    jne DIFFERENT        ; not equal, get out, you are DONE!
    cmp al, 0            ; equal so far, are you at the end?
    je SAME              ; got to end of both strings, you're good, get out
    jmp LOOP             ; okay well they agree so far, go to next char
DIFFERENT:
    ; Do what you need to do for the strings being different
    ;
    ;
    jmp DONE
SAME:
    ; Do what you need to do for the strings being the same
    ;
    ;
DONE:

However I suggest just doing a search for string comparison in x86. There is a cmps instruction. You can even arrange to call the strncmp function if you like. There are several ways to go about this.