kolibrios-fun/kernel/trunk/fs/parse_fn.inc
pathoswithin 5435e675b8 UTF-8 based disk system, UTF-16 path input
git-svn-id: svn://kolibrios.org@6471 a494cfbc-eb01-0410-851d-a64ba20cac60
2016-08-15 16:55:03 +00:00

494 lines
10 KiB
PHP

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
;; ;;
;; Copyright (C) KolibriOS team 2004-2016. All rights reserved. ;;
;; Distributed under terms of the GNU General Public License. ;;
;; ;;
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
$Revision$
iglobal
; pointer to memory for path replace table,
; size of one record is 128 bytes: 64 bytes for search pattern + 64 bytes for replace string
; start with one entry: sys -> <sysdir>
full_file_name_table dd sysdir_name
.size dd 1
tmp_file_name_size dd 1
endg
uglobal
; Parser_params will initialize: sysdir_name = "sys", sysdir_path = <sysdir>
sysdir_name rb 64
sysdir_path rb 64
sysdir_name1 rb 64
sysdir_path1 rb 64
; for example:
;dir_name1 db 'KolibriOS',0
; rb 64-8
;dir_path1 db 'HD0/1',0
; rb 64-6
endg
uglobal
tmp_file_name_table dd ?
endg
; use bx_from_load and init system directory /sys
proc Parser_params
locals
buff db 4 dup(?) ; for test cd
endl
mov eax, [OS_BASE+0x10000+bx_from_load]
mov ecx, sysdir_path
mov [ecx-64], dword 'sys'
cmp al, 'r'; if ram disk
jnz @f
mov [ecx], dword 'RD/?'
mov [ecx+3], byte ah
mov [ecx+4], byte 0
ret
@@:
cmp al, 'm'; if ram disk
jnz @f
mov [ecx], dword 'CD?/'; if cd disk {m}
mov [ecx+4], byte '1'
mov [ecx+5], dword '/KOL'
mov [ecx+9], dword 'IBRI'
mov [ecx+13], byte 0
.next_cd:
mov [ecx+2], byte ah
inc ah
cmp ah, '5'
je .not_found_cd
lea edx, [buff]
pushad
stdcall read_file, read_firstapp, edx, 0, 4
popad
cmp [edx], dword 'MENU'
jne .next_cd
jmp .ok
@@:
sub al, 49
mov [ecx], dword 'HD?/'; if hard disk
mov [ecx+2], byte al
mov [ecx+4], byte ah
mov [ecx+5], dword '/KOL'
mov [ecx+9], dword 'IBRI'
mov [ecx+13], byte 0
.ok:
.not_found_cd:
ret
endp
proc load_file_parse_table
stdcall kernel_alloc, 0x1000
mov [tmp_file_name_table], eax
mov edi, eax
mov esi, sysdir_name
mov ecx, 128/4
rep movsd
invoke ini.enum_keys, conf_fname, conf_path_sect, get_every_key
mov eax, [tmp_file_name_table]
mov [full_file_name_table], eax
mov eax, [tmp_file_name_size]
mov [full_file_name_table.size], eax
ret
endp
uglobal
def_val_1 db 0
endg
proc get_every_key stdcall, f_name, sec_name, key_name
mov esi, [key_name]
mov ecx, esi
cmp byte [esi], '/'
jnz @f
inc esi
@@:
mov edi, [tmp_file_name_size]
shl edi, 7
cmp edi, 0x1000
jae .stop_parse
add edi, [tmp_file_name_table]
lea ebx, [edi+64]
@@:
cmp edi, ebx
jae .skip_this_key
lodsb
test al, al
jz @f
or al, 20h
stosb
jmp @b
@@:
stosb
invoke ini.get_str, [f_name], [sec_name], ecx, ebx, 64, def_val_1
cmp byte [ebx], '/'
jnz @f
lea esi, [ebx+1]
mov edi, ebx
mov ecx, 63
rep movsb
@@:
push ebp
mov ebp, [tmp_file_name_table]
mov ecx, [tmp_file_name_size]
jecxz .noreplace
mov eax, ecx
dec eax
shl eax, 7
add ebp, eax
.replace_loop:
mov edi, ebx
mov esi, ebp
@@:
lodsb
test al, al
jz .doreplace
mov dl, [edi]
inc edi
test dl, dl
jz .replace_loop_cont
or dl, 20h
cmp al, dl
jz @b
jmp .replace_loop_cont
.doreplace:
cmp byte [edi], 0
jz @f
cmp byte [edi], '/'
jnz .replace_loop_cont
@@:
lea esi, [ebp+64]
call .replace
jc .skip_this_key2
.replace_loop_cont:
sub ebp, 128
loop .replace_loop
.noreplace:
pop ebp
inc [tmp_file_name_size]
.skip_this_key:
xor eax, eax
inc eax
ret
.skip_this_key2:
pop ebp
jmp .skip_this_key
.stop_parse:
xor eax, eax
ret
endp
proc get_every_key.replace
; in: ebx->destination, esi->first part of name, edi->second part of name
; maximum length is 64 bytes
; out: CF=1 <=> overflow
; 1) allocate temporary buffer in stack
sub esp, 64
; 2) save second part of name to temporary buffer
push esi
lea esi, [esp+4] ; esi->tmp buffer
xchg esi, edi ; edi->tmp buffer, esi->source
@@:
lodsb
stosb
test al, al
jnz @b
; 3) copy first part of name to destination
pop esi
mov edi, ebx
@@:
lodsb
test al, al
jz @f
stosb
jmp @b
@@:
; 4) restore second part of name from temporary buffer to destination
; (may cause overflow)
lea edx, [ebx+64] ; limit of destination
mov esi, esp
@@:
cmp edi, edx
jae .overflow
lodsb
stosb
test al, al
jnz @b
; all is OK
add esp, 64 ; CF is cleared
ret
.overflow:
; name is too long
add esp, 64
stc
ret
endp
cp866toUpper:
; convert cp866 character in al to uppercase
cmp al, 'a'
jb .ret
cmp al, 'z'
jbe @f
cmp al, 0xA0
jb .ret
cmp al, 0xB0
jb @f
cmp al, 0xE0
jb .ret
cmp al, 0xF0
jb .rus
cmp al, 0xF7
ja .ret
and eax, -2
.ret:
ret
@@:
sub eax, 32
ret
.rus:
sub eax, 0xE0-0x90
ret
utf16toUpper:
; convert UTF-16 character in ax to uppercase
cmp ax, 'a'
jb .ret
cmp ax, 'z'
jbe @f
cmp ax, 430h
jb .ret
cmp ax, 450h
jb @f
cmp ax, 460h
jnc .ret
sub eax, 80
.ret:
ret
@@:
sub eax, 32
ret
uni2ansi_char:
; convert UNICODE character in ax to ANSI character in al using cp866 encoding
cmp ax, 0x80
jb .ret
cmp ax, 0xB6
jz .B6
cmp ax, 0x400
jb .unk
cmp ax, 0x410
jb @f
cmp ax, 0x440
jb .rus1
cmp ax, 0x450
jb .rus2
cmp ax, 0x460
jb @f
.unk:
mov al, '_'
.ret:
ret
.B6:
mov al, 20
ret
.rus1: ; 0x410-0x43F -> 0x80-0xAF
add al, 0x70
ret
.rus2: ; 0x440-0x44F -> 0xE0-0xEF
add al, 0xA0
ret
@@:
push ecx edi
mov ecx, 8
mov edi, .table
repnz scasb
mov ah, cl
pop edi ecx
jnz .unk
mov al, 0xF7
sub al, ah
ret
.table db 1, 51h, 4, 54h, 7, 57h, 0Eh, 5Eh
ansi2uni_char:
; convert ANSI character in al to UNICODE character in ax, using cp866 encoding
movzx eax, al
cmp al, 0x80
jb @f ; 0x00-0x7F - trivial map
cmp al, 0xB0
jb .rus ; 0x80-0xAF -> 0x410-0x43F
cmp al, 0xE0
jb .unk
cmp al, 0xF0
jb .rus2 ; 0xE0-0xEF -> 0x440-0x44F
cmp al, 0xF8
jnc .unk
mov al, [eax+uni2ansi_char.table-0xF0]
add ax, 400h
ret
@@:
cmp al, 20
jnz .ret
mov al, 0xB6
.ret:
ret
.rus:
add ax, 0x410-0x80
ret
.rus2:
add ax, 0x440-0xE0
ret
.unk:
mov al, '_'
ret
cp866toUTF8_string:
; in:
; esi -> cp866 string (could be zero terminated)
; edi -> buffer for UTF-8 string
; ecx = buffer size (signed)
lodsb
call ansi2uni_char
push eax
call UTF16to8
pop eax
js @f
test eax, eax
jnz cp866toUTF8_string
@@:
ret
; SF=1 -> counter
; ZF=1 -> zero char
UTF16to8_string:
; in:
; esi -> UTF-16 string (could be zero terminated)
; edi -> buffer for UTF-8 string
; ecx = buffer size (signed)
xor eax, eax
@@:
lodsw
push eax
call UTF16to8
pop eax
js @f
test eax, eax
jnz @b
@@:
ret
UTF16to8:
; in:
; eax = UTF-16 char
; edi -> buffer for UTF-8 char (increasing)
; ecx = byte counter (decreasing)
dec ecx
js .ret
cmp eax, 80h
jnc @f
stosb
test eax, eax ; SF=0
.ret:
ret
@@:
dec ecx
js .ret
cmp eax, 800h
jnc @f
shl eax, 2
shr al, 2
or eax, 1100000010000000b
xchg al, ah
stosw
ret
@@:
dec ecx
js .ret
shl eax, 4
shr ax, 2
shr al, 2
or eax, 111000001000000010000000b
bswap eax
shr eax, 8
stosb
shr eax, 8
stosw
ret
utf8to16:
; in: esi -> UTF-8 char (increasing)
; out: ax = UTF-16 char
lodsb
test al, al
jns .got
shl al, 2
jnc utf8to16
@@:
shl ax, 8
lodsb
test al, al
jns .got
shl al, 2
jc @b
shr ah, 2
shl ax, 3
jnc @f
shl eax, 3
lodsb
test al, al
jns .got
shl al, 2
jc @b
shr eax, 2
ret
@@:
shr ax, 5
ret
.got:
xor ah, ah
ret
strlen:
; in: esi -> source
; out: ecx = length
push edi eax
or ecx, -1
mov edi, esi
xor eax, eax
repnz scasb
inc ecx
not ecx
pop eax edi
ret