REPE / REPZ

REPE and REPZ are two names for the F3h prefix, the same byte as REP. In front of CMPS or SCAS it adds a second exit: after each compare, CX is decreased and the loop also stops if ZF = 0. So it runs while CX is not 0 and the elements keep matching.

Decide with JE / JNE right after it. Testing CX alone is a classic bug: a difference in the last element also ends with CX = 0. And if CX can start at 0, check that first (JCXZ), since then no compare runs and ZF is left over from earlier code.

Finding where

After a difference SI and DI have already stepped past the differing elements. Going forward over bytes, the index of the first difference is SI - start - 1; the number of elements that matched is the same number. For words it is (SI - start - 2) / 2.

CLD
MOV     SI, OFFSET STR1
MOV     DI, OFFSET STR2
MOV     CX, 6
REPE    CMPSB
JE      SAME1    ; no difference at all
MOV     AX, SI
SUB     AX, OFFSET STR1
DEC     AX       ; index of the first difference: 3
ADD     AL, '0'  ; digit -> character
MOV     DL, AL
MOV     AH, 02h  ; DOS: print DL
INT     21h
JMP     DONE1
SAME1:
MOV     DL, '='
MOV     AH, 02h
INT     21h
DONE1:
 
STR1    DB      'hello!'
STR2    DB      'help!!'

With SCAS, REPE skips over elements equal to AL: MOV AL, ' ' / REPE SCASB steps over leading spaces, and the first other character is at DI - 1. In front of MOVS, STOS or LODS the prefix ignores ZF and acts as plain REP.