ProfitPress Mega CDROM2 …eeware (MSDOS)(1992)(Eng)/ProfitPress-MegaCDROM2.B6I/PROG/PASCAL/CONV_P18.ZIP/STRINGS.DOC

Text File | 1989-11-21 | 17.4 KB | 409 lines

[This document, and STRINGS.TPU, is extracted from SIMTEL20's
 PD2:<MSDOS.TURBOPAS>TPSTRING.ARC.
 STRINGS.TPU is actually the author's STR50.TPU.
]

             Turbo Pascal Rexx STRINGS Unit
        Version 1.2 for Turbo Pascal 4.0, 5.0 & 5.5

      STRINGS.TPU is a Turbo Pascal unit containing 29 string-related
      functions implemented in assembler.  These routines are highly
|     optimized (check out the benchmark program) and make extensive use
      of the 80x86 string oriented opcodes.  IBM mainframe hackers will
|     notice that this package is substantially equivalent to the string
      routines available in IBM's Rexx language.

|     Three versions are included:  STR40.TPU, STR50.TPU and STR55.TPU
|     for TPas versions 4.0, 5.0 and 5.5, respectively.  Be sure to
|     rename the appropriate file to STRINGS.TPU before usage.

            Function Descriptions
            ---------------------

function left(str:string; width:byte; pad:char):string;
     Returns STR left justified in a field of width WIDTH, padded out
     with PAD characters.
     Ex: left('hello',10,'=') returns 'hello====='
         left('hello there',10,'=') returns 'hello ther'

function right(str:string; width:byte; pad:char):string;
     Returns STR right justified in a field of width WIDTH, padded out
     with PAD characters.
     Ex: right('hello',10,'>') returns '>>>>>hello'
         right('hello there',10,'>') returns 'ello there'

function center(str:string; width:byte; pad:char):string;
     Returns STR centered in a string of length WIDTH, padded out
     with PAD chars.
     Ex: center('ABC',8,' ') returns '  ABC   '
         center(' ABC ',8,'-') returns '- ABC --'
         center('ABCDE',3,'-') returns 'ABC'

function strip(str:string; opt,ch:char):string;
     Strips leading and/or trailing CH characters from STR.
     Setting OPT to L, T or B causes leading, trailing, or both
     leading and trailing characters to be stripped.
     Ex: strip('   abcdef  ','L',' ') returns 'abcdef  '
         strip('   abcdef  ','t',' ') returns '   abcdef'
         strip('   abcdef  ','b',' ') returns 'abcdef'
         strip('++ abcdef +','B','+') returns ' abcdef '

function lastpos(findstr,instr:string; start:byte):byte;
     Returns the position of the last occurrance of FINDSTR in INSTR,
     searching backwards from the character position START.  If START
     is 0, the search begins at the end of INSTR.  Returns 0 if the
     string is not found.
     Ex: lastpos('he','he was the best',15) returns 9.
         lastpos('he','he was the best',6) returns 1.
         lastpos('he','he was the best',0) returns 9.
         lastpos('he','he was the best',1) returns 0.

function firstpos(findstr,instr:string; start:byte):byte;
     This function was included for completeness.  It works exactly
     the same way as Turbo's built in POS function, except for the
     presence of the START option.  It is equivalent to:
     start-1+pos(findstr,copy(instr,start,length(instr)-start+1));
     except for being more efficient.
     Ex: firstpos('he','he was the best',15) returns 0.
         firstpos('he','he was the best',6) returns 9.
         firstpos('he','he was the best',0) returns 1.
         firstpos('he','he was the best',1) returns 1.
|     The timing benchmark in BENCHMRK.PAS indicates this runs
|     over NINE times faster than Tpas's POS function, on my NEC
|     V20 CPU (a Japanese clone of an 8088).
|     (this will vary with the specific strings involved)

function copies(str:string; count:byte):string;
     Returns COUNT copies of STR concatenated together.
     If the length of n(<=count) copies of STR would exceed 255,
     n-1 copies are returned.
     Ex:  copies('----+',4) returns '----+----+----+----+'

function overlay(new,str:string; pos:byte; pad:char):string;
     Returns the string STR, overlayed by the string NEW, starting
     at character position POS, padding out STR with PAD characters
     if necessary.
     Ex: overlay('aygu','Ronald Reagan',9,'+') returns 'Ronald Raygun'
         overlay('abc','xyz',6,'+') returns 'xyz++abc'

|function instr(new,str:string; pos:byte; pad:char):string;
|     Returns the string STR after insertion of the string NEW,
|     starting at character position POS, padding out STR with PAD
|     characters if necessary.
|     Ex: instr('abcdef','123456',4,'+') returns '123abcdef456'
|         instr('abcdef','123456',10,'+') returns '123456+++abcdef'
|     While having similar function to TPas's INSERT procedure, the
|     timing benchmark in BENCHMRK.PAS indicates instr is about 60%
|     faster.

|function delstr(str:string; pos,len:byte):string;
|     Returns the string STR after deletion of LEN characters,
|     starting at character position POS.
|     Ex: delstr('abcdefgh',4,2) returns 'abcfgh'
|         delstr('abcdefgh',4,20) returns 'abc'
|     While having similar function to TPas's DELETE procedure, the
|     timing benchmark in BENCHMRK.PAS indicates delstr is about 50%
|     faster.

|function substr(str:string; pos,len:byte):string;
|     Returns a substring from STR, starting at position POS
|     and continuing for LEN characters, or until the end of STR.
|     Ex: substr('1234567890',4,2) returns '45'
|         substr('1234567890',4,20) returns '4567890'
|     This function is identical to TPas's COPY function, and it
|     executes at the same speed.  So why did I include it?  Who
|     knows?  I guess I like typing substr better than typing copy...

function uppercase(str:string):string;
     Folds the argument STR to uppercase.
     Ex: uppercase('abcdef123') returns 'ABCDEF123'

function lowercase(str:string):string;
     Folds the argument STR to lowercase.
     Ex: lowercase('ABCDEF123') returns 'abcdef123'

function words(str:string):byte;
     Returns the number of (blank delimited) words in the string STR.
     Ex: words('two four six      eight') returns 4.

function werd(str:string; n:byte):string;
     Returns the N'th (blank delimited) word from the string STR.
     The strange spelling is to avoid conflict with Tpas's WORD type.
     Ex: werd('two     air is humin',3) returns 'is'
         werd('two     air is humin',5) returns ''

function subword(str:string; n,count:byte):string;
     Returns COUNT words from STR, starting at the N'th word.
     Embedded blanks are preserved.
     Ex: subword('one two three  four',2,3) returns 'two three  four'
         subword('one two three  four',2,0) returns ''
         subword('one two  three ',2,3) returns 'two  three'
         subword('one two  three ',4,2) returns ''

function delword(str:string; n,count:byte):string;
     Returns STR with COUNT words deleted, starting at word N.
     Preceeding blanks are preserved.
     Ex: delword('here   we go again',2,2) returns 'here   again'

function pos2word(str:string; pos:byte):byte;
     Returns the number of the word in STR pointed to by POS.  If
     POS points to a blank, the number of the following word is
     returned.  If POS points after the last word or end of STR,
     or POS is 0, then 0 is returned.
     Ex: pos2word('abc def ghi ',4) returns 2.
         pos2word('abc def ghi ',6) returns 2.
         pos2word('abc def ghi ',11) returns 3.
         pos2word('abc def ghi ',12) returns 0.
         pos2word('abc def ghi ',0) returns 0.

function word2pos(str:string; wrd:byte):byte;
     Returns the position in STR of the first character in the WRD'th
     word.  If WRD>WORDS(STR) or WRD=0, returns 0.
     Ex: word2pos('  abcd e  fghi jk',2) returns 8.
     Note that this function is equivalent to Rexx's WORDINDEX
     function.

function space(str:string; spc:byte):string;
     Returns STR with each word separated by SPC blanks.
     Ex: space('  here  we   go again  ',0) returns 'herewegoagain'
         space('  here  we   go again  ',1) returns 'here we go again'
|     If the length of the output string would exceed 255, the string
|     is truncated at the end of the last whole word which fits.

function justify(str:string; len:byte):string;
     Distributes blanks between words in STR so that length(STR)=LEN.
     Ex: justify(' a b cd ef ',10)='a b  cd ef'
|     The length of STR should be <= than LEN.
|     This runs about eight times faster than the previous version,
|     which was written in pascal.
|     See usage notes below for important information on usage.

function translate(str,intable,outable:string):string;
     Returns STR after translation via the map INTABLE->OUTABLE.
     In other words, each occurrance in STR of the i'th character
     in INTABLE is replaced by the i'th character in OUTABLE.
     Ex: translate('ABC BDE',' BCF','XYZ ') returns 'AYZXYDE'
     INTABLE and OUTABLE should be of the same length.

function verify(str,ref:string; opt:char; start:byte):byte;
     Returns the position of the first character in STR (after START)
     which matches/doesn't match a character in REF.  Setting OPT to
     'M' or 'N' returns matching or non-matching character positions,
     respectively.
     Ex: verify('abcd1ef','0123456789','M',0) returns 5.
         verify('123a125','0123456789','n',0) returns 4.

function compare(s1,s2:string):byte;
     Compares S1 to S2 and returns the position of the first
     characters which don't match, or 0 if all characters match.
     Ex: compare('hello','hello there') returns 6.
         compare('hello','hexlo') returns 3.
         compare('hello','hello') returns 0.
         compare('','') returns 0.

function xrange(c1,c2:char):string;
     Returns a string containing all characters from C1 to C2
|     inclusive.  Note the ordering of C1 & C2 can now be reversed.
     Ex: xrange('a','h') returns 'abcdefgh'
|         xrange('h','a') returns 'hgfedcba'

function reverse(str:string):string;
     Returns contents of STR in reverse order.
     Ex: reverse('hello there') returns 'ereht olleh'

function abbrev(str,abbr:string; len:byte):boolean;
     Returns true if ABBR is an 'acceptable' abbreviation for STR.
     The criterion is:
        length(ABBR)>=LEN and ABBR=left(STR,length(ABBR),' ')
     LEN should be set <= length(STR).
     Ex: abbrev('DELETE','DEL',3)=true
         abbrev('DELETE','DELY',3)=false
         abbrev('DELETE','DELET',3)=true
         abbrev('DELETE','DELETEX',3)=false

function d2x(i:word):xstr;
     (XSTR is defined in the TPU as STRING[4])
     Returns a four byte string equal to the hex representation of I.
     Ex: d2x(255) returns '00FF'

function x2d(x:xstr):word;
     Returns the numeric value represented by the xstr X.  Upper
     and lower case A-F are valid on input.  No checking is done for
     the validity of the characters in X, so garbage input gives
     garbage output.  If the validity of X is in doubt, use the
     VERIFY function first:
     validx:=(verify(x,'0123456789ABCDEFabcdef','N')=0);
     Ex: x2d('7F') returns 127.

                 Usage Notes
                 -----------

      Note that Rexx's FIND and WORDLENGTH functions can be readily
      synthesized using functions in this package:

      {find returns the number of the word in str1 where str2 starts}
      find(str1,str2) ::= pos2word(str1,firstpos(str2,str1,1))

      {wordlength returns the length of a word in str}
      wordlength(str,n) ::= length(word(str,n))

      The previous version of justify did an implicit space(str,1) upon
      entry.  However, it turns out that any reasonable text formatting
      algorithm will require that a space() be done BEFORE the call to
      justify (see TXTFMT.PAS for an example) so the implicit call is
      actually redundant, so I got rid of it.  This also allows the
      added flexibility of inserting an extra space after periods (at
      the end of sentences) before calling justify, if you're picky
      about such things.  Justify DOES do an implicit strip(str,'B',' ')
      upon entry.

Why the lack of VAR string parameters? (in case you were wondering)

      I initially wrote this unit to use VAR string formal parameters in
      the interest of speed, but I've since discovered that when calling
      external assembler routines with value string formal parameters,
      when the actual parameter is a string variable (as opposed to a
      string expression), TP passes a pointer to the ACTUAL variable,
      not a copy of the variable.  While this behavior isn't consistent
      with the pascal standard, assembler isn't pascal, and it does give
      the programmer the best of both worlds:  fast execution and low
      stack overhead when using string variables as parameters, and the
      flexibility of value parameters.  At any rate, string variable
      function parameters are NEVER modified.

Differences between releases:

Release 1.05:

      Converted ABBREV and DELWORD to asm.  Included turbo compiler
      directives {$N-,E-,D-,L-,B-,I-,R-,S+,V-} for optimization.  Added
      PAD parameter to OVERLAY.  Added CHGSTR.  Fixed bug in FIRSTPOS.
      Changed behavior of SUBWORD when COUNT=0, to match Rexx's SUBWORD
      function.

Release 1.1:

      Compiled for Tpas 5.0.  Modified to cooperate with Tpas's dead
      code elimination feature.  Added X2D & D2X.  Changed JUSTIFY to
      use integer arithmetic for added speed.  Changed name of function
      WORD to WERD to avoid conflict with Tpas's new WORD type.  Added
      type XSTR for use with X2D & D2X.

Release 1.11:

      Tweaked LEFT, RIGHT, SPACE and COMPARE for optimization.

|Release 1.2:
|
|     Added SUBSTR, INSTR & DELSTR.  Converted JUSTIFY to asm.  Got rid
|     of CHGSTR (it didn't seem as useful as I'd first thought, it's not
|     part of standard Rexx, and it was out of place as a pascal routine
|     among asm routines).  Changed truncation behavior in SPACE when
|     output string exceeds 255 chars.  Changed XRANGE to allow reverse
|     ordering of C1 and C2, to produce reversed output.  Included
|     benchmark and test suite programs.  Included TPU versions for TP
|     4.0, 5.0 and 5.5.

Examples:

      Capitalize the first char in each word in a string S:
      for i:=1 to words(S);
      j:=word2pos(S,i);
      S:=overlay(upcase(S[j]),S,j);
      end;

      Find the lowest-ordered alphabetic character in a string STR
      of uppercase characters:
      lochar:=chr(ord('A')+verify(xrange('A','Z'),STR,'M',0)-1);

      Find the highest-ordered alphabetic character in a string STR
      of uppercase characters:
      hichar:=
     chr(ord('Z')-verify(xrange('Z','A'),STR,'M',0)+1);

      Replace non-alphabetic chars with blanks in a string S:
      S:=translate(S,xrange('!','9')+xrange(':','@')+
         xrange('[','`')+xrange('{',''),left('',33,' '));
      Or:
      S:=translate(S,translate(xrange('!',''),xrange('A','Z')+xrange('a','z'),
         left('',52,' ')),left('',ord('')-ord('!')+1,' '));

      Generate a sorted string S consisting of chars between '0' and 'z',
      none of which occur in an alphanumeric string STR:
      S:=space(translate(xrange('0','z'),STR,left('',length(STR),' ')),0);

      Generate a sorted string S, each of who's characters occurs at least
      once in an alphanumeric string STR:
      S:=space(translate(xrange('0','z'),translate(xrange('0','z'),
         STR,left('',length(STR),' ')),left('',75,' ')),0);

      Permute the characters in a 'MM/DD/YY' date string to allow for
      easy date comparison:
      translate('78312645','12345678','12/25/88') returns '88/12/25'.

      A simple text formatting example is in TXTFMT.PAS.


                 The Fine Print
                 --------------

      STRINGS.TPU is copyright 1989 by Richard Winkel.  The ASM & OBJ
      files are available from me for $10 if you send me a formatted
      360K disk in a stamped self addressed mailer.  If you want to
      avoid the hassle, add $3 and I'll buy the floppy, mailer and
      postage.

Note: Purchase of the ASM & OBJ files confers an unlimited license to
      incorporate these routines into your own programs for any purpose,
      commercial or otherwise.

Send cash, check or money order to:

   Richard Winkel
   Route 1, box 193
   Harrisburg, MO.   65256

Note: You don't need an assembler to recompile the TPU, as long as
      you have the OBJ files and a Turbo Pascal compiler.

Internet address:
[e-mail],
[e-mail] or
[e-mail]

------------------------------- cut here ---------------------------

          Calling Syntax Guide to STRINGS.TPU

type xstr=string[4];

function abbrev(str,abbr:string; len:byte):boolean;
function center(s:string; width:byte; pad:char):string;
function compare(s1,s2:string):byte;
function copies(str:string; count:byte):string;
function d2x(i:word):xstr;
function delstr(str:string; pos,len:byte):string;
function delword(str:string; n,len:byte):string;
function firstpos(findstr,instr:string; start:byte):byte;
function instr(new,str:string; pos:byte; pad:char):string;
function justify(str:string; len:byte):string;
function lastpos(findstr,instr:string; start:byte):byte;
function left(str:string; width:byte; pad:char):string;
function lowercase(s:string):string;
function overlay(new,str:string; pos:byte; pad:char):string;
function pos2word(s:string; pos:byte):byte;
function reverse(str:string):string;
function right(str:string; width:byte; pad:char):string;
function space(str:string; spc:byte):string;
function strip(s:string; opt,ch:char):string; {opt='L', 'T' or 'B'}
function substr(str:string; pos,len:byte):string;
function subword(str:string; num,count:byte):string;
function translate(str,intable,outable:string):string;
function uppercase(s:string):string;
function verify(str,ref:string; opt:char; start:byte):byte; {opt='M' or 'N'}
function werd(s:string;c:byte):string;
function word2pos(s:string; wrd:byte):byte;
function words(s:string):byte;
function x2d(str:xstr):word;
function xrange(c1,c2:char):string;