Bubble Sort in TASM

[kornicameister]

recently I've started to learning assembler (using TASM syntax), my first task was to made bubble sort...
I made an implementation, I checked it in Turbo Debugger and elements are in fact replaced if the appropriate condition is met, elements are swapped

but the program runs into infinite loop and I can not notice why is that
here is the entire code

;program do metody babelkowej
 
.MODEL SMALL
; Algorytm sortowania babelkowego:
; 1. Start.
; 2. index = 0; zamiany = 0;
; 3. Odczytaj dwa sasiadujace elementy z tablicy o pozycjach index oraz
;    index + 1; jezeli pierwszy z odczytanych elementow jest wiekszy od
;    swojego nastepnika, to zamien elementy miejscami oraz zwieksz zamiany.
; 4. Zwieksz index o jeden.
; 5. Jezeli index < dlugosc_tablicy - 2 to skacz do 3.
; 6. Skacz do 2 jezeli zamiany rozne od zera.
; 7. Stop.
.DATA
 
	t DB 01h, 02h, 00h, 10h, 12h, 33h, 15h, 09h, 11h, 08h, 0Ah, 00h 
		; 1,   2,   0,  16,  18,  51,  21,   9,  17,   8,  10,   0 ;po sorcie2=>	00h,00h,01h,02h,08h,09h,0Ah,10h,11h,12h,15h,33h
	t_size	EQU 	11 			; t_lenght
	again_s	db 		0		; if again_s == 1, swap was made, again
.CODE
s:
	mov ax,@DATA
	mov ds,ax 
	;buble sort
prepare:
	mov bx, OFFSET t 		; load the t offset into bx
again:
	mov again_s,0			; XOR the flag
	xor si,si				; XOR used indexing register 
nextElement:
	mov ax,[bx+si]			; first compared element
	mov cx,[bx+si+1]                 ; second one
	cmp ax,cx				; comparing
	jbe	noSwap			; if first <= next jmp no swap
	xchg ax,cx				; else swap
	mov again_s,1			; set the flag		
noSwap:
	inc si					; increment index
	cmp si,t_size			; compare the index with t_size
	jb	nextElement		
	cmp again_s,1                     ; if last element, check the flag
	je	again
koniec:
	mov ax,4c00h
	int 21h
.STACK 100H
end s

;program do metody babelkowej

.MODEL SMALL
; Algorytm sortowania babelkowego:
; 1. Start.
; 2. index = 0; zamiany = 0;
; 3. Odczytaj dwa sasiadujace elementy z tablicy o pozycjach index oraz
;    index + 1; jezeli pierwszy z odczytanych elementow jest wiekszy od
;    swojego nastepnika, to zamien elementy miejscami oraz zwieksz zamiany.
; 4. Zwieksz index o jeden.
; 5. Jezeli index < dlugosc_tablicy - 2 to skacz do 3.
; 6. Skacz do 2 jezeli zamiany rozne od zera.
; 7. Stop.
.DATA

	t DB 01h, 02h, 00h, 10h, 12h, 33h, 15h, 09h, 11h, 08h, 0Ah, 00h 
		; 1,   2,   0,  16,  18,  51,  21,   9,  17,   8,  10,   0 ;po sorcie2=>	00h,00h,01h,02h,08h,09h,0Ah,10h,11h,12h,15h,33h
	t_size	EQU 	11 			; t_lenght
	again_s	db 		0		; if again_s == 1, swap was made, again
.CODE
s:
	mov ax,@DATA
	mov ds,ax 
	;buble sort
prepare:
	mov bx, OFFSET t 		; load the t offset into bx
again:
	mov again_s,0			; XOR the flag
	xor si,si				; XOR used indexing register 
nextElement:
	mov ax,[bx+si]			; first compared element
	mov cx,[bx+si+1]                 ; second one
	cmp ax,cx				; comparing
	jbe	noSwap			; if first <= next jmp no swap
	xchg ax,cx				; else swap
	mov again_s,1			; set the flag		
noSwap:
	inc si					; increment index
	cmp si,t_size			; compare the index with t_size
	jb	nextElement		
	cmp again_s,1                     ; if last element, check the flag
	je	again
koniec:
	mov ax,4c00h
	int 21h
.STACK 100H
end s

To copy to clipboard, switch view to plain text mode 

an important attention : I set t_size equal to 11, even though the real size equals 12. I made is so, because I tried to meet the condition when the index points at the last element (counting from 0)

[kornicameister]

fixed the issues
still do not know how to check if sorting is correct

I would be glad if someone just took a look on fixed code

;program do metody babelkowej
 
.MODEL SMALL
; Algorytm sortowania babelkowego:
; 1. Start.
; 2. index = 0; zamiany = 0;
; 3. Odczytaj dwa sasiadujace elementy z tablicy o pozycjach index oraz
;    index + 1; jezeli pierwszy z odczytanych elementow jest wiekszy od
;    swojego nastepnika, to zamien elementy miejscami oraz zwieksz zamiany.
; 4. Zwieksz index o jeden.
; 5. Jezeli index < dlugosc_tablicy - 2 to skacz do 3.
; 6. Skacz do 2 jezeli zamiany rozne od zera.
; 7. Stop.
.DATA
 
	t DB      01h, 02h, 00h, 10h, 12h, 33h,5h, 09h, 11h, 08h, 0Ah, 00h
	t_size	EQU 	3 			; dlugosc tablicy
	again_s	db 		0		; przyjmie 1 jesli byla zamiana
.CODE
s:
	mov ax,@DATA
	mov ds,ax	
	;buble sort
prepare:
	mov bx, OFFSET t 		
again:
	mov again_s,0			
	xor si,si				
nextElement:
	mov ah,[bx+si]			
	mov al,[bx+si+1]
	cmp ah,al				
	jbe	noSwap				
	mov [bx+si],al
	mov [bx+si+1],ah
	mov again_s,1				
noSwap:
	inc si					
	cmp si,t_size			
	jl  nextElement
 
	cmp again_s,1
	je	again=
 
	mov ax,4c00h
	int 21h
.STACK 100H
end s

;program do metody babelkowej

.MODEL SMALL
; Algorytm sortowania babelkowego:
; 1. Start.
; 2. index = 0; zamiany = 0;
; 3. Odczytaj dwa sasiadujace elementy z tablicy o pozycjach index oraz
;    index + 1; jezeli pierwszy z odczytanych elementow jest wiekszy od
;    swojego nastepnika, to zamien elementy miejscami oraz zwieksz zamiany.
; 4. Zwieksz index o jeden.
; 5. Jezeli index < dlugosc_tablicy - 2 to skacz do 3.
; 6. Skacz do 2 jezeli zamiany rozne od zera.
; 7. Stop.
.DATA

	t DB      01h, 02h, 00h, 10h, 12h, 33h,5h, 09h, 11h, 08h, 0Ah, 00h
	t_size	EQU 	3 			; dlugosc tablicy
	again_s	db 		0		; przyjmie 1 jesli byla zamiana
.CODE
s:
	mov ax,@DATA
	mov ds,ax	
	;buble sort
prepare:
	mov bx, OFFSET t 		
again:
	mov again_s,0			
	xor si,si				
nextElement:
	mov ah,[bx+si]			
	mov al,[bx+si+1]
	cmp ah,al				
	jbe	noSwap				
	mov [bx+si],al
	mov [bx+si+1],ah
	mov again_s,1				
noSwap:
	inc si					
	cmp si,t_size			
	jl  nextElement
	
	cmp again_s,1
	je	again=

	mov ax,4c00h
	int 21h
.STACK 100H
end s

To copy to clipboard, switch view to plain text mode 

[stampede]

In my free time I'm learning assembler as well - using HLA ( assembly language developed for academic purposes ), I've rewritten your program in this syntax and added sorting check:

program TestProg;
 
#include ("stdlib.hhf")
 
static
	t_size:	int32 := 12;
	again_s:	int8 := 0;
	t:              int8[12] := [1, 2, 0, 16, 44, 11,5, 9, 11, 8, 12, 3];
 
begin TestProg;	
 
again:
	mov (0,again_s);
	xor (esi,esi);
 
nextElement:
	mov( t[esi],ah);	;
	mov (t[esi+1],al);
	cmp (al,ah);				
	jbe	(noSwap);				
	mov (al,t[esi]);
	mov (ah,t[esi+1]);
	mov (1,again_s);
 
noSwap:
	inc (esi);			
	cmp (esi,t_size);
	jl  (nextElement);
	cmp (again_s,1);
	je	(again);
 
	xor(esi,esi);
print:
	stdout.put(t[esi], " ");
	inc(esi);
	cmp(esi,t_size);
	jl(print);
 
	stdout.put(nl);
	stdout.put("start check..." nl);
 
	xor(esi,esi);
check:
	mov(t[esi],ah);
	mov(t[esi+1],al);
	inc(esi);
	cmp(t_size,esi);
	jl(check_ok);
	cmp(al,ah);
	jbe(check);
 
check_failed:
	stdout.put("not ok" nl);
	jmp end_prog;
 
check_ok:
	stdout.put("check ok");
 
end_prog:
 
end TestProg;

program TestProg;

#include ("stdlib.hhf")

static
	t_size:	int32 := 12;
	again_s:	int8 := 0;
	t:              int8[12] := [1, 2, 0, 16, 44, 11,5, 9, 11, 8, 12, 3];

begin TestProg;	
	
again:
	mov (0,again_s);
	xor (esi,esi);

nextElement:
	mov( t[esi],ah);	;
	mov (t[esi+1],al);
	cmp (al,ah);				
	jbe	(noSwap);				
	mov (al,t[esi]);
	mov (ah,t[esi+1]);
	mov (1,again_s);

noSwap:
	inc (esi);			
	cmp (esi,t_size);
	jl  (nextElement);
	cmp (again_s,1);
	je	(again);
	
	xor(esi,esi);
print:
	stdout.put(t[esi], " ");
	inc(esi);
	cmp(esi,t_size);
	jl(print);

	stdout.put(nl);
	stdout.put("start check..." nl);
	
	xor(esi,esi);
check:
	mov(t[esi],ah);
	mov(t[esi+1],al);
	inc(esi);
	cmp(t_size,esi);
	jl(check_ok);
	cmp(al,ah);
	jbe(check);
	
check_failed:
	stdout.put("not ok" nl);
	jmp end_prog;
	
check_ok:
	stdout.put("check ok");
	
end_prog:

end TestProg;

To copy to clipboard, switch view to plain text mode 

Here you have more about HLA if you are interested: link.

[kornicameister]

I am literally forced to use tasm
but thank you for Your time