> |This procedure gets a word from the input, actually it gets a token. |When a carriage return is found, the linecount is incremented | | | GetWord: call GetToken jc GetWordEnd movb al,InputWord call IsAlphaDotAL jnc InvalidStart clc GetWordEnd: ret InvalidStart: mov bx,#BadStartMessage call PanicRecover GetToken: push ax push bx push cx push dx push si push di call IgnoreSpaces jc GetTokenEnd mov di,#InputWord call GetChar jc GetTokenEnd call IsAlphaNumDotAL jnc NonAlphaNumeric call IsNumAL jnc MustBeWord |Only Numbers come here. |Numbers may be decimal,hexadecimal cmpb al,#'0' |leading digit is not zero - decimal jnz GetNumber stosb call GetChar cmpb al,#'X' |for hex jz GetHexNumber cmpb al,#'x' jz GetHexNumber call UnGetChar jmps GetNumber1 GetNumber: stosb GetNumber1: call GetChar jc NumberEnded1 call IsNumAL jc GetNumber NumberEnded: call UnGetChar NumberEnded1: clc GetTokenEnd: | pushf | mov bx,#InputWord | call DisplayOtherMessage | call DisplayMessage | db 'is the token got',0dh,0ah,0 | popf pushf xorb al,al stosb popf pop di pop si pop dx pop cx pop bx pop ax ret MustBeWord: stosb cmp di,#InputWord+WordSize| jnz StartWordSizeStillOK |If the identifier is too large then mov bx,#IdentLargeMessage |crib and come out call PanicRecover | StartWordSizeStillOK: | call GetChar jc WordEnded call IsAlphaNumDotAL |Succeeding can be als,nums,dots jc MustBeWord |If yes store them too call UnGetChar WordEnded: clc jmps GetTokenEnd GetHexNumber: orb al,#'a' - 'A' |downcase it, for later GetHexNumber1: stosb call GetChar jc NumberEnded1 call IsHexAL jc GetHexNumber1 jmps NumberEnded LostQuote: mov bx,#MissingQuoteMessage call PanicRecover NonAlphaNumeric: cmpb al,#''' jnz OnlyThisChar stosb call GetChar jc LostQuote stosb call GetChar jc LostQuote cmpb al,#''' jnz LostQuote stosb clc jmps GetTokenEnd OnlyThisChar: cmpb al,#'|' jz CommentStartIgnore cmpb al,#CR jz CommentStartIgnore cmpb al,#LF jz CommentStartIgnore stosb clc jmps GetTokenEnd CommentStartIgnore: call UnGetChar clc jmps GetTokenEnd IgnoreSpaces: call GetChar jc IgnoreSpacesEnd cmpb al,#' ' jz IgnoreSpaces cmpb al,#' ' jz IgnoreSpaces call UnGetChar clc IgnoreSpacesEnd: ret GetComma: cmpb InputWord,#',' jz GetCommaEnd push si call IgnoreSpaces jc GetCommaNo call GetChar cmpb al,#',' jz GetCommaYes GetCommaNo: mov bx,#CommaMessage call PutExpectedMessage GetCommaYes: pop si GetCommaEnd: ret | Assemble an instruction GetInstruction: mov si,#Instructions call MatchKeyword jc InstructionNoMatch mov ax,[si] inc si inc si call ax clc InstructionNoMatch: ret |IsAlphaAL |This function returns carry if A is alphabetic else it returns no carry IsAlphaAL: cmpb al,#'A' |First check whether it is in jc LessThanBigA |the capital letters of the alphabet cmpb al,#'[' |if >= A & <= Z then it is a valid jc WhatWeWant |character LessThanBigA: |Now checking for the lower case cmpb al,#'a' |if >= a jc LessThanSmallA |and <= z cmpb al,#'{' |then too it is a valid jc WhatWeWant |character LessThanSmallA: |If neither uppercase or lower cmpb al,#'_' |case then check for an underscore jnz NotWhatWeWant |The underscore is also treated stc |as an alphabetic character WhatWeWant: ret NotWhatWeWant: | clc ret IsNumAL: cmpb al,#'0' jc NotWhatWeWant cmpb al,#':' ret IsHexAlphaAL: cmpb al,#'A' jc NotWhatWeWant cmpb al,#'G' jc WhatWeWant cmpb al,#'a' jc NotWhatWeWant cmpb al,#'g' jnc NotWhatWeWant subb al,#'a' - 'A' stc ret |Surprise - converts to uppercase as well if a hex alphabet to take care. IsHexAL: call IsNumAL jc WhatWeWant call IsHexAlphaAL ret IsAlphaDotAL: call IsAlphaAL jc WhatWeWant cmpb al,#'.' jnz NotWhatWeWant stc ret IsAlphaNumDotAL: call IsAlphaDotAL jc WhatWeWant call IsNumAL ret |This function is called, when all the arguments of the current line |have been processed. i.e. when no more arguments are needed. This is |called before each statement is processed. It processes the comment |to the end of the line, if any, consumes the newline, skips to the |next line, If that line is empty (contains only white spaes) or |contains just a comment (with or without leasing spaces) then the |line is also consumed. This continues till a line, with a non- |comment, non whitespace start is found. The procedure also eats up |characters till the first non-white space character on that line. GetEndOfLine: | xor bp,bp |No line end found GetEndOfLine1: |In case of start of file call GetChar |Get the next input character jnc NotYetEofForEol |If EOF then abort inc bp |EOF also valid end of stmt. stc jmps GetEndOfLineEnd NotYetEofForEol: cmpb al,#' ' |If it is a space then consume jz GetEndOfLine1 |the space and continue cmpb al,#' ' |Do the same even if you jz GetEndOfLine1 |encounter a TAB character. cmpb al,#CR |Reached end of line on the input jz GetEndOfLine1 |Discard any carriage returns that cmpb al,#LF |were there in the input. So also jnz NoLine |the line feeds EolHaveLF: inc bp inc InputLineNumber jmps GetEndOfLine1 NoLine: cmpb al,#'|' |Is it the start of a comment? jz CommentStarted |If it isn't then we put back the call UnGetChar |extra character that we got. clc |Signal that a line was found GetEndOfLineEnd: |to the caller lahf or bp,bp |If a line end wasn't found jnz HadFoundEnd |then error mov bx,#ExtraCharsMessage |that extra characters were there call PanicRecover |at the end of line HadFoundEnd: | sahf ret |and get back. CommentStarted: |A UNIX pipe character was found call GetChar |Get the mext input character jc GetEndOfLineEnd |EOF found on the input cmpb al,#CR |If we reach a carraige return jz GetEndOfLine1 |it means that the comment ended cmpb al,#LF |So also for a linefeed (MINIX) jnz CommentStarted |So we start processing the next line jmps EolHaveLF |This function reads a character from a file, The input is buffered by the |InputBuffer. The function keeps track of where it is by keeping two pointers. |The InputBufferReadPtr (points to the character that has to be read) and |the InputBufferEndPtr (points beyond the last character that was read). If the |InputBufferReadPtr reaches the InputBufferEndPtr, then the characters in the |buffer have all been read. The buffer has to be loaded agein from the file. | |Destroys a lot of things | |Returns : No carry and character in al if successful | Carry if unsuccessful GetChar: mov ax,InputBufferReadPtr |If the read ptr is the same as the end cmp ax,InputBufferEndPtr |ptr, it means that all the characters jz InputBufferFinished |in the buffer have been read mov si,ax |If not then load the next character lodsb |and save the new read ptr mov InputBufferReadPtr,si | cmpb al,#26 |Not for MINIX. 1A means eof jz GetAtEndOfFile |for dos clc ret InputBufferFinished: movb ah,#ReadFunction |DOS function read from a file into the mov bx,InputFileHandle |input buffer. mov cx,#BufferSize mov dx,#InputBuffer int #DosInterrupt jnc BufferReadWasOK |If there was an error, then panic mov bx,#ErrorReadingInputMessage call Panic BufferReadWasOK: cmp ax,#0 |else if no chars were read, then jnz NotEndOfFile |it means that the end of file was GetAtEndOfFile: stc |reached. Return a carry to show that ret |couldn't return a value. NotEndOfFile: |If some characters were read, then mov si,#InputBuffer |the pointers to the beginning and add si,ax |the end of the buffer are made right mov InputBufferEndPtr,si |ax has the number of characters that sub si,ax |were read. Get the first character lodsb |into al, mov InputBufferReadPtr,si |Adjust the start buffer clc |Return success ret | UnGetChar: dec InputBufferReadPtr ret BackupInputWord: push si push di push cx mov si,#InputWord mov di,#BackupWord InputWordMoveEnd: mov cx,#WordSize rep movsb pop cx pop di pop si ret RestoreInputWord: push si push di push cx mov si,#BackupWord mov di,#InputWord jmp InputWordMoveEnd