PAGE 60,132 TITLE FINDFILE.COM VER.2.0 14-AUG-83 19:15 comment * FINDFILE.COM VERSION 2.0 14-AUG-83 Written by Warren Craycroft 6236 Oakdale Ave. Oakland, CA 94605 (C) 1983 by Warren Craycroft. Permission is granted to copy and distribute this program, including source code, provided that no charge shall be made except for a reasonable charge for the media and handling, and that this notice shall remain intact in all copies. * comment * This program is a utility that can be used with DOS 2.0 to find the disk drive number that contains a certain file. A common use of FINDFILE is to help a batch file "orient" itself by discovering on which drive the Batch File itself is running, and on what drives other files (such as useful system utilities) can be found. A lot of this orientation can be done without bothering or confusing the user with a lot of questions or constraints in locating disks in certain drives. A key feature of this search (and where this method of orien- tation differs from the use of PATH or IF EXIST ) is that searching on the disk drives is done "benignly". By this we mean that FINDFILE doesn't come to a grinding halt if an empty disk drive is searched. Instead, it continues the search on the rest of the drives. This further relieves the user from having to locate certain disks on certain drives. The target file may be on the B: or C: drive with the A: drive empty, and FINDFILE will find it without the annoying "ignore, abort, retry" message. If the file is not found, then you may gently prod the user to place the required disk "in any drive" and try again. The syntax of the command line is the following: findfile fname where fname is a standard DOS path, filename, and extention. If you include a drive specification (like "a:" for example), the drive spec will be ignored by FINDFILE. FINDFILE benignly searches all disk drives that are claimed to present by the system board switches, in order A:, B:, C:, and D:. If the file is found, the disk number of the first disk on which it was found is returned in errorlevel and may be tested by a Batch File IF Subcommand. If fname found: errorlev = 0 if fname on A: errorlev = 1 if fname on B: errorlev = 2 if fname on C: errorlev = 3 if fname on D: If fname not found: errorlev = 255 If fname is missing: errorlev = 255 , and a CNTRL BREAK halts the Batch File * ; ; constant equates ; BEL_CHAR EQU 07 ;ascii BEL keycode CR EQU 0DH ;ascii carriage return LF EQU 0AH ;ascii line feed TAB_CHAR EQU 09 ;ascii tab BLANK_CHAR EQU ' ' ;ascii blank COLON_CHAR EQU ':' ;ascii colon NOT_FOUND_CODE EQU 255 ;errorlev returned for fname not ;found DISK_MASK EQU 00C0H ;mask for equipment word's disk ;count field ; ; ; declare a relocatable segment. Follow the .COM file requirements ; of entry point at 100H and making all seg register references relative ; to CS (no relocatable values MOV'ed into segment registers). ; ; COM_CODE SEGMENT ; ORG 80H ;PSP offset 80: user's command line PSP_CMD_LINE LABEL BYTE ;define a label for address refs ; ORG 100H ;for COM file ; ; ASSUME CS:COM_CODE,DS:COM_CODE ;tell assembler value of CS ASSUME ES:COM_CODE ; DS, and ES ; ; START PROC FAR ;FAR is meaningless; no RETS ; MOV SI,OFFSET PSP_CMD_LINE ;offset of user's cmd line SUB CX,CX ;clear CX MOV CL,[SI] ;byte count of line OR CL,CL ;zero? JZ NO_FNAME ;jump if yes, no fname param SUB BX,BX ;else zero index ; ; find first non-blank or non-tab character of command line ; STRIP_BLANKS: INC SI ;point to next char MOV AL,[SI] ;get next char from cmd line CMP AL,BLANK_CHAR ;is it blank? JE LOOP_BLANKS ;jump if yes, keep looking CMP AL,TAB_CHAR ;is it a tab? JNE GOT_NON_BLANK ;jump if not, first non-blank LOOP_BLANKS: LOOP STRIP_BLANKS ;else keep looking ; ; fname is missing from command line (all blanks or tabs, or else ; zero characters on line ; NO_FNAME: MOV DX,OFFSET ERROR_MESSAGE ;adr offset of error message MOV AH,9 ;DOS fn call, "print string" INT 21H ;call DOS INT 23H ;cause a control break, cause ;this is probably a bug in ;batch file creation. JMP NO_FILE_EXIT ;then exit with errorlevel = ; "NOT_FOUND_CODE" ; ; found first non-blank/tab character on command line ; at [SI][BX] . Check for disk designator " n: " and ; "strip" it from string by moving up pointer and decreasing ; remaining bytes count. ; GOT_NON_BLANK: CMP BYTE PTR [SI+1],COLON_CHAR ;2nd non-blank char ":"? JNE ADD_DRIVE ;if not, [SI] is start of strng INC SI ;else strip off the drive spec INC SI DEC CX ;and decr remaining byte count JZ NO_FNAME ;if string hits zero bytes left, DEC CX JZ NO_FNAME ; it's a "no param" error ; ; SI now points to start of ASCIIZ string that we want to use as ; the path/file name to search drives for ; ; Put a byte of zeros at the end of string, and an "A:" in front ; of string to initialize search. ; ADD_DRIVE: MOV BX,CX ;number of bytes remaining MOV BYTE PTR [SI][BX],0 ;points to 1 char past end ; of string MOV BYTE PTR [SI-2],'A' ;"A:" in front of string MOV BYTE PTR [SI-1],':' ; ; Replace the current critical error handler with our own ; ; First, get the address of the critical error handler assigned ; by "parent" process (probably system) to this process. Save its ; address, so that we can use it if our handler is handed a non-disk ; error ( "we don't do non-disk errors ..." ) ; MOV AH,35H ;DOS fn "get vector" MOV AL,24H ;"critical error" vector no. INT 21H ;call DOS MOV WORD PTR PARENT_HAND,BX ;save offset vector adr MOV WORD PTR PARENT_HAND+2,ES ;save seg vector adr ; ; Now, take control of critical error vector by assigning a handler ; to the vector address. ; ; DOS will restore the "parent"'s vector when this process returns to ; DOS. ; MOV DX,OFFSET CRIT_HAND ;address of handler MOV AL,24H ;the vector to be replaced MOV AH,25H ;DOS fn call "ret int vector" INT 21H ;call DOS and replace vector ; ; Get the number of drives that DOS knows about ; Dont disturb the current default disk when getting the number ; MOV AH,19H ;DOS fn "return current disk" INT 21H ;current disk returned in AL MOV DL,AL ;put current in DL for sel call MOV AH,0EH ;DOS fn "sel disk, return number INT 21H ;get number of drives SUB CH,CH ;clear CH MOV CL,AL ;number of drives INC CX ;is now (1,2,3, or 4) ; ; Now loop through all drives present, trying to find the file ; (using fn call 4EH) on each drive. If successful, return the ; disk number to caller in errorlevel. ; ; If not successful, return errorlevel = NOT_FOUND_CODE ; MOV DX,SI ;offset adr of drive spec and DEC DX ; filename DEC DX FIND_FILE: MOV AH,4EH ;DOS fn "find first file" PUSH CX ;save the loop count SUB CX,CX ;attribute says search ;"normal" files only PUSH [SI-2] ;save drive letter (DOS call ;destroys it if file found) INT 21H ;call DOS POP [SI-2] ;restore drive letter POP CX ;restore loop count JNC FOUND_FILE ;if carry not set, then search ;was successful - jump. INC BYTE PTR [SI-2] ;increment drive letter char LOOP FIND_FILE ;and look on next drive ; ; did not find file. Return NOT_FOUND_CODE in errorlevel ; NO_FILE_EXIT: MOV AL,NOT_FOUND_CODE ;file not found on any drive EXIT: MOV AH,4CH ;DOS fn call to exit w/ errlev INT 21H ;call DOS and exit ; ; found the file. Retrieve the drive letter used in this last call, ; map it to disk number (A = 0, B = 1, etc) and return it in errorlevel. ; FOUND_FILE: MOV AL,[SI-2] ;get the drive letter (upper ;case ascii code) SUB AL,'A' ;map it to 0,1,2,3 JMP EXIT ;and return it in errorlevel START ENDP page ; ; CRIT_HAND Critical Error Handler ; ; This handler will be called by DOS functions when any critical ; error (disk or non-disk) occurs during DOS INT 21H calls that ; are made during execution of this process. ; ; We only want to ignore disk errors, so we first test if the disk ; is the cause of error. ; ; If not a disk error, we jump to the parent's critical error handler ; (we JMP so that stack looks as if we were never there). ; ; If it is a disk error, we return an IGNORE code to the calling ; DOS function. The effect of all this is that when a DOS function ; tries to access a disk drive that has a door open, or no disk, or ; disk is unformatted, or disk is garbled, etc, the DOS function will ; simply give up and return to the function's caller "empty handed". ; There will be no "abort, retry, ignore" message from the critical ; error handler. ; ; on entry: Bit 7 of AH = 0 if disk error ; = 1 if not a disk error ; ; on exit: If disk error, we set AL = 0, "ignore error" ; ; registers affected: AH, AL ; CRIT_HAND PROC FAR ;RETS are typed FAR TEST AH,80H ;bit 7 on? JNZ NOT_DISK ;jump if not disk error, we dont ;treat it ; ; Treatment of disk errors is easy, just tell system to ignore them ; SUB AL,AL ;zero AL, "ignore" code IRET ;caller used INT so we IRET ; ; It was not a disk error, so let the "parent" critical error handler ; treat it by JMPing to its address that we saved. ; ; We JMP so we dont disturb the stack. Stack looks like we were ; never called. ; NOT_DISK: JMP DWORD PTR PARENT_HAND ;jump to parent crit err hand ; CRIT_HAND ENDP ; ; ; temporary storage of parent's critical handler address ; PARENT_HAND DD 1 DUP(?) ;double word for FAR address ; ; ; ; error message string ; ERROR_MESSAGE DB CR,LF,'***** No filename was given with ' DB 'FINDFILE *****',CR,LF,'$' ; ; COM_CODE ENDS END START