Turn a string with ( multiple) \n embedded into a Pascal string declaration

couple of questions...
1) how do you know when you have reached the end of the string?  eg. " hey ho"
2) would it have to handle "commented code"
eg. ignore "bob" in the following example

fred <<< a frog said hi
// bob <<< a cow replaied hi yourself
bill <<< a chicken ignored them both...
 
my c skills are pretty rusty so I don't remember all the syntax :-)

loki .. as you assume, no comments and a full string supplied.

Thanks for the quick response, and I will start testing your code in a few minutes.

Thanks again

yes, it works beautifully, except that if there is a lot of text with no \n anywhere the Delphi string generated
will be too long (the compiler does not  like strings that go past col 127 or something)

About the only solution I can see s to add another stringlist in there, and check for a space inside a quoted string near some preferred column boundary

I am having a go at that now.

Doesn't StringReplace do the job? (I think that function was implemented on D7)

ShowMessage(StringReplace('My multiline\ntestmessage\n\nby luis','\n',#13+#10,[rfReplaceAll]))

Also I beleive you can turn back the parameters to convert #13#10 to \n

> will be too long (the compiler does not  like strings that go past col 127 or something)

ah, personally, I'd do a final loop through the stringlist, and split lines > 127 into 2 lines.
I'll take a look at it tomorrow and see if I can make it look nice

well, this is my approach to the long lines problem .. it is not very pretty, but it seems to work Loki, unless you think you have a much nicer way of doing that reformatting, I don't wan you to spend any more time on it .. I am happy to give you the points, thanks for getting me most of the way.


function TForm1.ConvertCStringToDelphiString(InCString: string): string;

var
  sl3          : Tstringlist;

  function SplitLongStringAtPos(const s: string; const StartFrom: integer; const Prefer: integer): Boolean;
  var
    numspaces, len: integer;
    spacesAt   : array of Boolean;
    PlusesAt   : array of Boolean;
    InsideQuotes: array of Boolean;
    inQuote    : Boolean;
    s1, s2     : string;
    k, nearest, nearestpos: integer;
  begin
    result := false;
    len := length(s);
    setlength(spacesAt, len + 1);
    setlength(PlusesAt, len + 1);
    setLength(InsideQuotes, len + 1);
    numspaces := 0;
    inquote := false;
    for k := StartFrom to len do begin
      spacesAt[k] := false;
      PlusesAt[k] := false;
      insideQuotes[k] := false;
      if inquote then insideQuotes[k] := true;
      if s[k] = '+' then PlusesAt[k] := true;
      if s[k] = '''' then inquote := not (inquote);
      if s[k] = ' ' then begin
        spacesAt[k] := true;
        inc(numspaces);
      end;
    end;                                // end for

    if numspaces = 0 then exit;         // cannot do it

    // ------------------------------------------------------------------------------
   // now look for a space near the preferred boundary,  outside the quotes
    nearest := maxInt; nearestPos := -1;
    for k := 1 to len do begin
      if not spacesAt[k] then continue;
      if insideQuotes[k] then continue;
      if abs(k - Prefer) < nearest then begin
        nearestpos := k;
        nearest := abs(k - Prefer);
      end;
    end;                                // end for

    // check if within tolerance (say 10 characters)
    if nearest < 10 then begin
      s1 := copy(s, 1, nearestpos);
      sl3.add(s1);
      s2 := copy(s, nearestpos + 1, len);
      if length(s2) < Prefer then begin
        sl3.add(s2);
        result := true;
        exit;
      end;
      // second bit of string is still too long .. recurse down
      result := SplitLongStringAtPos(s2, 1, prefer);
      exit;
    end;


    // ------------------------------------------------------------------------------
   // now look for a  plus near the preferred boundary, outside the quotes
    nearest := maxInt; nearestPos := -1;
    for k := 1 to len do begin
      if not PlusesAt[k] then continue;
      if insideQuotes[k] then continue;
      if abs(k - Prefer) < nearest then begin
        nearestpos := k;
        nearest := abs(k - Prefer);
      end;
    end;                                // end for

    // check if within tolerance (say 10 characters)
    if nearest < 10 then begin
      s1 := copy(s, 1, nearestpos);
      sl3.add(s1);
      s2 := copy(s, nearestpos + 1, len);
      if length(s2) < Prefer then begin
        sl3.add(s2);
        result := true;
        exit;
      end;
      // second bit of string is still too long .. recurse down
      result := SplitLongStringAtPos(s2, 1, prefer);
      exit;
    end;


    // ------------------------------------------------------------------------------
   // now look for a space near the preferred boundary inside a quoted string

    nearest := maxInt; nearestPos := -1;
    for k := 1 to len do begin
      if not spacesAt[k] then continue;
      if insideQuotes[k] = false then continue;
      if abs(k - Prefer) < nearest then begin
        nearestpos := k;
        nearest := abs(k - Prefer);
      end;
    end;                                // end for

    // check if within tolerance, again
    if nearest < 10 then begin
      s1 := copy(s, 1, nearestpos - 1) + '''';
      sl3.add(s1);
      s2 := '+''' + copy(s, nearestpos, len);
      if length(s2) < Prefer then begin
        sl3.add(s2);
        result := true;
        exit;
      end;
      // second bit of string is still too long .. recurse down
      result := SplitLongStringAtPos(s2, 1, prefer);
      exit;
    end;




    // ------------------------------------------------------------------------------
   //  we could not find any space near the preferred boundary (within tolerance)
   //  so we do a brute force split at position Prefer

    inquote := false;
    for k := StartFrom to Prefer do begin
      if s[k] = '''' then inquote := not (inquote);
    end;                                // end for
    nearestpos := Prefer;
    case inquote of
      true: begin
          s1 := copy(s, 1, nearestpos - 1) + '''';
          sl3.add(s1);
          s2 := '+''' + copy(s, nearestpos, len);
        end;
      false: begin
          s1 := copy(s, 1, nearestpos);
          sl3.add(s1);
          s2 := copy(s, nearestpos + 1, len);
        end;
    end;                                // end case
    if length(s2) < Prefer then begin
      sl3.add(s2);
      result := true;
      exit;
    end;
    // second bit of string is still too long .. recurse down
    result := SplitLongStringAtPos(s2, 1, prefer);
  end;

const
  vardeclare   = '<<<';
  newline      = '\n';
  replaceCRLF  = #39 + ' + #13#10 +' + #13#10 + #39;
  emptyCRLFline = #39#39 + ' + #13#10 +';
var
  i, crit      : integer;
  s            : string;
  FullString   : string;
  VarPart      : string;
  indent       : string;                // for nice formatting only
  OutStrings   : TStrings;
  sl           : Tstringlist;
  need2Split   : Boolean;

begin
  result := '';
  OutStrings := TStringList.Create;
  try
    // firstly, remove any white space
    OutStrings.Text := InCString;
    for i := pred(OutStrings.Count) downto 0 do begin
      s := TRIM(OutStrings[i]);
      if (s = '') then OutStrings.Delete(i) else OutStrings[i] := s;
    end;
    // now concatenate the lines as 1 line with actual cr/lfs replaced as single spaces
    FullString := StringReplace(OutStrings.Text, #13#10, ' ', [rfReplaceAll]);
    // gives us
    // mystring <<< this is a test \nof some stuff and this is some \n\n\nmore testing of the same \nstuff ok

    // Determine the variable
    i := pos(vardeclare, FullString);
    if i = 0 then raise Exception.Create('no variable declaration');

    VarPart := Trim(copy(FullString, 1, i - 1));
    delete(FullString, 1, i + length(vardeclare) - 1);
    FullString := #39 + trim(FullString) + #39; // remove spaces between variable and text
    // gives us 'this is a test of \n some stuff and this is some \n\nmore testing of the same \nstuff ok'
    FullString := StringReplace(FullString, newline, replaceCRLF, [rfReplaceAll, rfIgnoreCase]);
    { gives us
      'this is a test of ' + #13#10 +
      'some stuff and this is some ' + #13#10 +
      '' + #13#10 +
      '' + #13#10 +
      'more testing of the same ' + #13#10 +
      'stuff ok'
      }

    // concatenate multiple CRLFs
    OutStrings.Text := FullString;
    i := 2;                             // start at the 3rd line.. no point starting earlier
    while (i < OutStrings.Count) do begin
      if (OutStrings[i] = emptyCRLFline) then begin
        s := OutStrings[i - 1];
        delete(s, length(s) - 1, 2);    // remove the " +" portion
        s := s + '#13#10 +';
        OutStrings[i - 1] := s;
        Outstrings.Delete(i);
      end
      else
        inc(i);
    end;

    // nicely indent each line
    Varpart := VarPart + ' := ';
    SetLength(indent, length(VarPart));
    FillChar(indent[1], length(VarPart), ord(' '));

    crit := 70;
    need2split := false;
    for i := 0 to Outstrings.count - 1 do begin
      if length(outstrings[i]) > crit then need2Split := true;
    end;




    if need2Split = false then begin
      for i := pred(OutStrings.Count) downto 0 do begin
        if i <> 0 then OutStrings[i] := indent + OutStrings[i];
      end;
      result := VarPart + OutStrings.Text;
      // remove final CRLF if applicable
      if copy(result, length(result) - 1, 2) = #13#10 then
        delete(result, length(result) - 1, 2);
      result := result + ';';
    end else begin


      // -- now split long lines, as needed,  at some notional boundary

      sl := Tstringlist.create;
      for i := 0 to outstrings.count - 1 do begin
        if length(outstrings[i]) <= crit then begin
          sl.add(outstrings[i]);
        end else begin
          sl3 := Tstringlist.create;
          if SplitLongStringAtPos(outstrings[i], 1, crit) then begin
            sl.addstrings(sl3);
          end
          else begin
            sl.add('// warning .. could not split this string');
            sl.add(outstrings[i]);
          end;
          sl3.free;
        end;
      end;
      for i := pred(sl.Count) downto 0 do begin
        if i <> 0 then sl[i] := indent + sl[i];
      end;

      result := VarPart + sl.Text;
      // remove final CRLF if applicable
      if copy(result, length(result) - 1, 2) = #13#10 then
        delete(result, length(result) - 1, 2);
      result := result + ';';
      sl.free;
    end;

  finally
    OutStrings.Clear;
    OutStrings.Free;
  end;
end;




procedure TForm1.btnProcessHeredocClick(Sender: TObject);
begin
  dsmemo2.lines.text := ConvertCStringToDelphiString(dsmemo1.lines.Text);
end;

------------------------------ test string input------------------------------------------------
s <<< this
is
ggggggggggggggggggggggggggggggggggggggggg************************************************************************************************************************************gggggg    dduh
    gggggggggggggggggggggggggggggggggggggggggg       yyyyyyyyyyyyyyyyyyyyyyyy

  rrrrrrrrrrrrr
rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr  555555555555555555555             66666666666666666           999999999999   a        test\n
of some stuff
       and this is some \n\n\nmore
 testing of the same \nstuff
   ok

---------------------------------  output ---------------------------------------

s := 'this is ggggggggggggggggggggggggggggggggggggggggg*******************'
     +'*******************************************************************'
     +'**********************************************gggggg    dduh'
     +' gggggggggggggggggggggggggggggggggggggggggg       yyyyyyyyyyyyyyyyyyyyyyyy'
     +' rrrrrrrrrrrrr rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr'
     +'  555555555555555555555             66666666666666666          '
     +' 999999999999   a        test' + #13#10 +
     ' of some stuff and this is some ' + #13#10#13#10#13#10 +
     'more testing of the same ' + #13#10 +
     'stuff ok';

your code seems to work pretty well, I like the "try to find the best place to split" way you are doing it :-)