Files
crawos/source/kernel/features/strings.asm

339 lines
7.2 KiB
NASM
Executable File

; How string comparison works
; DI => scasb compares value stored in DI which is 's' with 's' stored in AX register
; and then increments DI if DF is 0
; v
; memory address 1: |s|n|a|t|0|
; memory address 2: |s|n|a|k|e|0|
; ^
; SI => lodsb loads value stored at SI to AX and then increments SI if DF is 0
; Additionaly, if the di string ends and the relevant character in si contains a space, it will still return
; the strings as being equal, this is to allow for command line arquments
os_compare_strings:
cld
.compare:
mov al, 0
scasb
je .di_ended
dec di
lodsb
scasb ; Compare di to si
jne .unequal ; If they are not equal, jump to .unequal
jmp .compare ; Finally, repeat
.unequal:
mov cl, 0 ; Change to 0 if unquality is proven
ret
.di_ended:
lodsb
cmp al, 20h ; 20h = space
je .equal
cmp al, 0
je .equal
jmp .unequal
.equal:
mov cl, 1
ret
; ------------------------------------------------------------------
; os_string_compare -- See if two strings match
; IN: SI = string one, DI = string two
; OUT: carry set if same, clear if different
; from MikeOS kernel
string_direct_compare:
pusha
.more:
mov al, [si] ; Retrieve string contents
mov bl, [di]
cmp al, bl ; Compare characters at current location
jne .not_same
cmp al, 0 ; End of first string? Must also be end of second
je .end
inc si
inc di
jmp .more
.not_same: ; If unequal lengths with same beginning, the byte
popa ; comparison fails at shortest string terminator
clc ; Clear carry flag
ret
.end: ; Both strings terminated at the same position
popa
stc ; Set carry flag
ret
; --------------------------------------------------------------------
; Copies a string from si to di
; IN:
; si: points to first character of source string
; OUT:
; di: points to first character of destination string
string_copy:
pusha
.loop:
; TODO could this cause issue when a character is > 1 byte?
lodsb ; Load si character into ax and increment si by 1
stosb ; Store ax in di location and increment di by 1
cmp ax, 0 ; if ax is a 0, quit
jne .loop
.done:
popa
ret
; --------------------------------------------------------------------
; Joins two strings together into a new string
; IN/OUT:
; ax: string1
; bx: string2
; cx: result string
string_join:
pusha
.copy_string1:
mov si, ax
mov di, cx
call string_copy
mov ax, cx ; TODO check this bit
call string_length
add cx, ax
.copy_string2:
mov si, bx
mov di, dx
call string_copy
.done:
popa
ret
; ------------------------------------------------------------------------
; TODO
string_cast_to_int:
ret
; ------------------------------------------------------------------------
; IN: AX = int
; OUT: AX = string location (output buffer)
string_cast_from_int: ; TODO I think this algorithm could be optimised
pusha
; Initialise base to decimal (base 10)
mov bx, 10
; clear the strigified int location
mov cx, 6
push ax
xor ax,ax
.clear_loop:
stosb
dec cx
cmp cx, 0
ja .clear_loop
; setup
pop ax
push ax
mov cx, -1
; loop over number to get the value of count (cx)
.count_loop:
xor dx,dx
div bx ; ax = ax // bx
inc cx
cmp ax, 0
ja .count_loop ; Jump if greater
pop ax
; loop over count downwards until count is 0
mov di, stringified_int
.stringify_loop:
mov cx,1 ;DEL
push cx
push bx
push ax
; Generate power of 10
mov ax, 1
mov cx,0 ;DEL
cmp cx, 1
jl .after_power_loop
.power_loop:
mul bx
dec cx
cmp cx, 0
ja .power_loop
.after_power_loop:
mov bx,ax
pop ax
pop cx
; ax = ax DIV bx
; dx = ax MOD bx
xor dx,dx
div bx
pop bx
add ax, 30h
stosb
mov ax, dx
dec cx
cmp cx, 0
ja .stringify_loop
.finish:
popa
;mov ax, stringified_int
ret
; Get the length of a string
; 'hello world', 0 is 11 characters long (excluding the terminator)
; input: si points to the string to be counted
; output: cl holds the length
string_length:
push si
xor cl,cl ; Clear the al register
.loop:
lodsb
cmp al, 0
je .finish
inc cl
jmp .loop
.finish:
pop si
ret
; convert a string to fat's filename format
; It will be capitalised and 11 characters long,
; 8 for the filename, 3 for the extension
; eg: 'file.txt' -> 'FILE TXT'
; input: si points to filename, di points to a free 11 bytes in memory
; output: di points to the fat formatted filename
os_format_fat_filename:
pusha
call string_upper_case
call string_length ; Stores the length of the string in cl
xor ch,ch ; Clear ch to reset it to 0
.character_loop:
lodsb
cmp al, 0
je .finish
cmp al, 2Eh ; 2Eh
je .add_spaces ; This will end up back at .character_loop
stosb
inc ch
jmp .character_loop
.add_spaces: ; Add the number of spaces as bl holds
mov al, ' ' ; 20h = space
; Work out the number of spaces in between
; the name and extension.
; 8 - name_length(ch)
xor bl, bl
sub bl, ch
add bl, 7
.spaces_loop:
stosb
cmp bl, 0
je .character_loop
dec bl
jmp .spaces_loop
.finish:
popa
ret
; Does the inverse of the previous
; Converts a fat filename back to human readable
; eg. 'KERNEL BIN' -> 'KERNEL.BIN'
; input: si points to fat filename (11 bytes)
; output:
; di points to the unformatted filename
; [file_name_length] stores the length of the filename
string_unformat_fat_filename:
pusha
xor ax,ax
xor dx,dx
mov cx, 11 ; Counter
.name_loop:
lodsb
stosb
dec cx
inc dx
cmp cx, 3
jne .name_loop
push si
mov si, di
.space_loop:
dec si
lodsb
dec si
dec dx
cmp al, 20h ; Space
je .space_loop
jmp .insert_stop
.insert_stop:
mov di, si
inc di
inc dx
pop si
mov al, 2Eh
stosb
mov cx, 3
.extension_loop:
lodsb
stosb
dec cx
inc dx
cmp cx, 0
jne .extension_loop
.finish:
inc dx
mov [file_name_length], dx
popa
ret
; Convert a string to all upper/lower case
; INPUT: si pointing to a string
; OUPUT: the same string in memory will now be capitalised/decapitalised
string_upper_case: ; to upper case
pusha
mov di, si
.loop:
lodsb ; Load the character into al
inc di
cmp al, 0
je .finish ; If it's null then the string is finished
cmp al, 7Ah ; 7Ah = 'z'
jns .loop ; Ignore if it's more than 'z'
cmp al, 61h ; 61h = 'a'
js .loop ; Ignore if it's less than 'a'
sub al, 20h ; Otherwise subtract 20h to capitalise it
dec di
stosb ; Store the new value
jmp .loop ; Next character
.finish:
popa
ret
os_lower_case: ; to lower case
pusha
mov di, si
.loop:
lodsb ; Load the character into al
inc di
cmp al, 0
je .finish ; If it's null then the string is finished
cmp al, 5Ah ; 5Ah = 'Z'
jns .loop ; Ignore if it's more than 'Z'
cmp al, 41h ; 41h = 'A'
js .loop ; Ignore if it's less than 'A'
add al, 20h ; Otherwise subtract 20h to capitalise it
dec di
stosb ; Store the new value
jmp .loop ; Next character
.finish:
popa
ret
string_lower_case:
ret
string_input:
ret
string_print_2hex:
ret