From: "Carlo Hogeveen" 
This is a multi-part message in MIME format.
------=_NextPart_000_004A_01BD7D41.1C024920
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Hi Skip and others
Version 2 of the Findws macro adressed Skip's
request for a find _again_ function, and here is
version 3 that solves a bug.
If "Great" is at the end of a line,
and "Small Britain" is on the next line,
then Findws would erroneously find "Great Britain".
The attached Findws.s solves this problem.
I'm thinking about expanding Findws, so it can
replace the standard Find(), RepeatFind() and Replace()
in the .UI file.
One benefit of this is, that it would no longer take up
extra keys and it would work transparent for the
usual Find() etc.
Another benefit would be, that the user can once again
choose his and her own options.
The user would then have to add an extra option S to
indicate that "Space matches whitespace", and could
add an extra option M to mark the found text as a block.
I would take the opportunity to let Find(), Replace() and
"Replace_by" all use the same find_history, and to let
the option box list all options
I suppose that in case of a found text that spans lines,
Replace() should add the Replace_by text to the end
of the first found line, though this might make the line
longer than the screen.
Perhaps the "S" option or something like it  would be
a nice one for TSE's wishlist.
I'm thinking out loud here, because what I'm going to build
won't necessarily have to reflect just my own whishes.
Any ideas or comments?
Carlo
------=_NextPart_000_004A_01BD7D41.1C024920
Content-Type: application/octet-stream;
	name="findws.s"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment;
	filename="findws.s"
/*
   Macro.         Findws.
   Author.        Carlo Hogeveen (hyphen@xs4all.nl)
   Date-written.  12 may 1998.
   Version        3.
   This macro finds a sequence of words separated by whitespace.
   You enter the words separated by a single space, which the macro =
matches
   with whitespace.
   Whitespace is any combination of spaces, horizontal_tabs and =
end_of_lines.
   The words may not contain spaces and are searched using =
case-insensitive
   regular expressions starting right from the cursor position.
   Findws can be started from the menu with Macro Execute, or from =
another
   macro or using a key definition with ExecMacro().
   It can be started with or without parameters.
   By searching "_again_" the previous find is repeated.
   Examples:
      begfile()
      execmacro("findws")
      execmacro("findws great britain")
      execmacro("findws _again_")
*/
#define horizontal_tab 9
#define space 32
integer find_history =3D 0
string previous_findws[255] =3D ""
proc main()
   string word[255] =3D ""
   string searching[255] =3D query(macrocmdline)
   integer found =3D false
   integer found_line =3D 0
   integer found_column =3D 0
   integer word_line =3D 0
   integer word_column =3D 0
   integer token_counter =3D 0
// set(break, on)
   if searching =3D=3D ""
      ask("Enter words separated by spaces", searching, find_history)
   else
      if lower(searching) =3D=3D "_again_"
         searching =3D previous_findws
      endif
   endif
   if searching  ""
      previous_findws =3D searching
   endif
   pushposition()
   repeat
      token_counter =3D 0
      repeat
         token_counter =3D token_counter + 1
         word =3D gettoken(searching, " ", token_counter)
         if word  ""
            if token_counter =3D=3D 1
               if lfind(word, "ix+")
                  found =3D true
                  found_line =3D currline()
                  found_column =3D currcol()
               else
                  found =3D false
                  word  =3D ""
               endif
            else
               // Cursor is on previous found word: goto next =
whitespace.
               while currchar()  space
               and   currchar()  horizontal_tab
               and   currchar()  _at_eol_
                  nextchar()
               endwhile
               // Cursor is on first whitespace: goto next =
non-whitespace.
               while currchar() =3D=3D space
               or    currchar() =3D=3D horizontal_tab
               or    currchar() =3D=3D _at_eol_
                  nextchar()
               endwhile
               // Check if the word under the cursor is the next search =
word.
               word_line =3D currline()
               word_column =3D currcol()
               if  lfind(word, "cix")
               and currline() =3D=3D word_line
               and currcol() =3D=3D word_column
                  // Yes, the next search word was and is under the =
cursor.
               else
                  found =3D false
               endif
            endif
         endif
      until found =3D=3D false
      or    word  =3D=3D ""
   until word =3D=3D ""
   if found
      killposition()
      gotoline(found_line)
      gotocolumn(found_column)
   else
      alarm()
      popposition()
      message(searching, " not found.")
   endif
end
proc whenloaded()
    find_history =3D getfreehistory("findws:find")
end
------=_NextPart_000_004A_01BD7D41.1C024920--
---
---------------