freepascal.compiler/rtl/inc/astrings.inc

{
    This file is part of the Free Pascal run time library.
    Copyright (c) 1999-2000 by Michael Van Canneyt,
    member of the Free Pascal development team.

    This file implements AnsiStrings for FPC

    See the file COPYING.FPC, included in this distribution,
    for details about the copyright.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

 **********************************************************************}

{ This will release some functions for special shortstring support }
{ define EXTRAANSISHORT}

{
  This file contains the implementation of the AnsiString type,
  and all things that are needed for it.
  AnsiString is defined as a 'silent' pchar :
  a pchar that points to :

  @-8  : SizeInt for reference count;
  @-4  : SizeInt for size;
  @    : String + Terminating #0;
  Pchar(Ansistring) is a valid typecast.
  So AS[i] is converted to the address @AS+i-1.

  Constants should be assigned a reference count of -1
  Meaning that they can't be disposed of.
}

Type
  PAnsiRec = ^TAnsiRec;
  TAnsiRec = Packed Record
    CodePage    : TSystemCodePage;
    ElementSize : Word;
{$ifdef CPU64}	
    { align fields  }
	Dummy       : DWord;
{$endif CPU64}
    Ref         : SizeInt;
    Len         : SizeInt;
    First       : Char;
  end;

Const
  AnsiRecLen = SizeOf(TAnsiRec);
  AnsiFirstOff = SizeOf(TAnsiRec)-1;

{****************************************************************************
                    Internal functions, not in interface.
****************************************************************************}

Function NewAnsiString(Len : SizeInt) : Pointer;
{
  Allocate a new AnsiString on the heap.
  initialize it to zero length and reference count 1.
}
Var
  P : Pointer;
begin
  { request a multiple of 16 because the heap manager alloctes anyways chunks of 16 bytes }
  GetMem(P,Len+AnsiRecLen);
  If P<>Nil then
   begin
     PAnsiRec(P)^.Ref:=1;         { Set reference count }
     PAnsiRec(P)^.Len:=0;         { Initial length }
     PAnsiRec(P)^.CodePage:=DefaultSystemCodePage;
     PAnsiRec(P)^.ElementSize:=SizeOf(AnsiChar);
     PAnsiRec(P)^.First:=#0;      { Terminating #0 }
     inc(p,AnsiFirstOff);         { Points to string now }
   end;
  NewAnsiString:=P;
end;


{$ifndef FPC_SYSTEM_HAS_ANSISTR_DECR_REF}
Procedure fpc_ansistr_decr_ref (Var S : Pointer); [Public,Alias:'FPC_ANSISTR_DECR_REF'];  compilerproc;
{
  Decreases the ReferenceCount of a non constant ansistring;
  If the reference count is zero, deallocate the string;
}
Var
  p: PAnsiRec;
Begin
  { Zero string }
  If S=Nil then 
    exit;
  { check for constant strings ...}
  p:=PAnsiRec(S-AnsiFirstOff);
  If p^.ref<0 then exit;
  { declocked does a MT safe dec and returns true, if the counter is 0 }
  If declocked(p^.ref) then
    begin
      FreeMem(p);
      s:=nil;
    end;
end;

{$endif FPC_SYSTEM_HAS_ANSISTR_DECR_REF}

{ also define alias for internal use in the system unit }
Procedure fpc_ansistr_decr_ref (Var S : Pointer); [external name 'FPC_ANSISTR_DECR_REF'];

Procedure fpc_AnsiStr_Incr_Ref (S : Pointer); [Public,Alias:'FPC_ANSISTR_INCR_REF'];  compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
Begin
  If S=Nil then
    exit;
  { Let's be paranoid : Constant string ??}
  If PAnsiRec(S-AnsiFirstOff)^.Ref<0 then exit;
  inclocked(PAnsiRec(S-AnsiFirstOff)^.Ref);
end;


{ also define alias which can be used inside the system unit }
Procedure fpc_AnsiStr_Incr_Ref (S : Pointer); [external name 'FPC_ANSISTR_INCR_REF'];

Procedure fpc_AnsiStr_Assign (Var DestS : Pointer;S2 : Pointer);[Public,Alias:'FPC_ANSISTR_ASSIGN'];  compilerproc;
{
  Assigns S2 to S1 (S1:=S2), taking in account reference counts.
}
begin
  if DestS=S2 then
    exit;
  If S2<>nil then
    If PAnsiRec(S2-AnsiFirstOff)^.Ref>0 then
      inclocked(PAnsiRec(S2-AnsiFirstOff)^.Ref);
  { Decrease the reference count on the old S1 }
  fpc_ansistr_decr_ref (DestS);
  { And finally, have DestS pointing to S2 (or its copy) }
  DestS:=S2;
end;

{ alias for internal use }
Procedure fpc_AnsiStr_Assign (Var S1 : Pointer;S2 : Pointer);[external name 'FPC_ANSISTR_ASSIGN'];

{$ifndef STR_CONCAT_PROCS}

function fpc_AnsiStr_Concat (const S1,S2 : AnsiString): ansistring; compilerproc;
Var
  Size,Location : SizeInt;
  pc : pchar;
begin
  { only assign if s1 or s2 is empty }
  if (S1='') then
    begin
      result:=s2;
      exit;
    end;
  if (S2='') then
    begin
      result:=s1;
      exit;
    end;
  Location:=Length(S1);
  Size:=length(S2);
  SetLength(result,Size+Location);
  pc:=pchar(result);
  Move(S1[1],pc^,Location);
  inc(pc,location);
  Move(S2[1],pc^,Size+1);
end;


function fpc_AnsiStr_Concat_multi (const sarr:array of Ansistring): ansistring; compilerproc;
Var
  i  : Longint;
  p  : pointer;
  pc : pchar;
  Size,NewLen : SizeInt;
begin
  { First calculate size of the result so we can do
    a single call to SetLength() }
  NewLen:=0;
  for i:=low(sarr) to high(sarr) do
    inc(NewLen,length(sarr[i]));
  SetLength(result,NewLen);
  pc:=pchar(result);
  for i:=low(sarr) to high(sarr) do
    begin
      p:=pointer(sarr[i]);
      if assigned(p) then
        begin
          Size:=length(ansistring(p));
          Move(pchar(p)^,pc^,Size+1);
          inc(pc,size);
        end;
    end;
end;

{$else STR_CONCAT_PROCS}

procedure fpc_AnsiStr_Concat (var DestS:RawByteString;const S1,S2 : RawByteString{$ifdef FPC_HAS_CPSTRING};cp : TSystemCodePage{$endif FPC_HAS_CPSTRING}); compilerproc;
Var
  Size,Location : SizeInt;
  same : boolean;
  S1CP, S2CP, DestCP: TSystemCodePage;
  U: UnicodeString;
begin
  { if codepages are differ then concat using unicodestring }
  S1CP:=StringCodePage(S1);
  if (S1CP=$ffff) or (S1CP=0) then
    S1CP:=DefaultSystemCodePage;
  S2CP:=StringCodePage(S2);
  if (S2CP=$ffff) or (S2CP=0) then
    S2CP:=DefaultSystemCodePage;
{$ifdef FPC_HAS_CPSTRING}
  if (Pointer(DestS)=nil) then
    DestCP:=cp
  else
    DestCP:=StringCodePage(DestS);
{$else FPC_HAS_CPSTRING}
  DestCP:=StringCodePage(DestS);
{$endif FPC_HAS_CPSTRING}
  if (DestCP=$ffff) or (DestCP=0) then
    DestCP:=DefaultSystemCodePage;
  if (S1CP<>DestCP) or (S2CP<>DestCP) then
    begin
      U:=UnicodeString(S1)+UnicodeString(S2);
      DestS:='';
      widestringmanager.Unicode2AnsiMoveProc(PUnicodeChar(Pointer(U)),DestS,DestCP,Length(U));
      exit;
    end;
  { only assign if s1 or s2 is empty }
  if (S1='') then
    begin
      DestS:=s2;
      exit;
    end;
  if (S2='') then
    begin
      DestS:=s1;
      exit;
    end;
  Location:=Length(S1);
  Size:=length(S2);
  { Use Pointer() typecasts to prevent extra conversion code }
  if Pointer(DestS)=Pointer(S1) then
    begin
      same:=Pointer(S1)=Pointer(S2);
      SetLength(DestS,Size+Location);
      if same then
        Move(Pointer(DestS)^,(Pointer(DestS)+Location)^,Size)
      else
        Move(Pointer(S2)^,(Pointer(DestS)+Location)^,Size+1);
    end
  else if Pointer(DestS)=Pointer(S2) then
    begin
      SetLength(DestS,Size+Location);
      Move(Pointer(DestS)^,(Pointer(DestS)+Location)^,Size+1);
      Move(Pointer(S1)^,Pointer(DestS)^,Location);
    end
  else
    begin
      DestS:='';
      SetLength(DestS,Size+Location);
      SetCodePage(DestS,DestCP,false);
      Move(Pointer(S1)^,Pointer(DestS)^,Location);
      Move(Pointer(S2)^,(Pointer(DestS)+Location)^,Size+1);
    end;
end;


procedure fpc_AnsiStr_Concat_multi (var DestS:RawByteString;const sarr:array of RawByteString{$ifdef FPC_HAS_CPSTRING};cp : TSystemCodePage{$endif FPC_HAS_CPSTRING}); compilerproc;
Var
  lowstart,i  : Longint;
  p,pc        : pointer;
  Size,NewLen,
  OldDestLen  : SizeInt;
  destcopy    : pointer;  
  DestCP      : TSystemCodePage;
  U           : UnicodeString;
  sameCP      : Boolean;
  tmpStr      : RawByteString;
  tmpCP       : TSystemCodePage;
begin
  if high(sarr)=0 then
    begin
      DestS:='';
      exit;
    end;
{$ifdef FPC_HAS_CPSTRING}
  if (Pointer(DestS)=nil) then
    DestCP:=cp
  else
    DestCP:=StringCodePage(DestS);
{$else FPC_HAS_CPSTRING}
  DestCP:=StringCodePage(DestS);
{$endif FPC_HAS_CPSTRING}
  if (DestCP=$ffff) or (DestCP=0) then
    DestCP:=DefaultSystemCodePage;
  sameCP:=true;
  lowstart:=low(sarr);
  for i:=lowstart to high(sarr) do
    begin
      if (DestCP<>StringCodePage(sarr[i])) then
        begin
          sameCP:=false;
          break;
        end;
    end;
  if not sameCP then
    begin
      U:='';
      for i:=lowstart to high(sarr) do begin
        tmpCP:=StringCodePage(sarr[i]);
        if (tmpCP=$ffff) or (tmpCP=0) then
          begin
            tmpStr:=sarr[i];
            SetCodePage(tmpStr,DefaultSystemCodePage,False);
            U:=U+UnicodeString(tmpStr);
          end
        else
          U:=U+UnicodeString(sarr[i]);
      end;

      DestS:='';
      widestringmanager.Unicode2AnsiMoveProc(PUnicodeChar(Pointer(U)),DestS,DestCP,Length(U));
      exit;            
    end;

  destcopy:=nil;
  lowstart:=low(sarr);
  if Pointer(DestS)=Pointer(sarr[lowstart]) then
    inc(lowstart);
  { Check for another reuse, then we can't use
    the append optimization }
  for i:=lowstart to high(sarr) do
    begin
      if Pointer(DestS)=Pointer(sarr[i]) then
        begin
          { if DestS is used somewhere in the middle of the expression,
            we need to make sure the original string still exists after
            we empty/modify DestS                                       }
          destcopy:=pointer(dests);
          fpc_AnsiStr_Incr_Ref(destcopy);
          lowstart:=low(sarr);
          break;
        end;
    end;
  { Start with empty DestS if we start with concatting
    the first array element }
  if lowstart=low(sarr) then
    DestS:='';
  OldDestLen:=length(DestS);
  { Calculate size of the result so we can do
    a single call to SetLength() }
  NewLen:=0;
  for i:=low(sarr) to high(sarr) do
    inc(NewLen,length(sarr[i]));
  SetLength(DestS,NewLen);
  if (StringCodePage(DestS) <> DestCP) then
    SetCodePage(DestS,DestCP,False);
  { Concat all strings, except the string we already
    copied in DestS }
  pc:=Pointer(DestS)+OldDestLen;
  for i:=lowstart to high(sarr) do
    begin
      p:=pointer(sarr[i]);
      if assigned(p) then
        begin
          Size:=length(ansistring(p));
          Move(p^,pc^,Size+1);
          inc(pc,size);
        end;
    end;
  fpc_AnsiStr_Decr_Ref(destcopy);
end;
{$endif STR_CONCAT_PROCS}


{$ifdef EXTRAANSISHORT}
Procedure AnsiStr_ShortStr_Concat (Var S1: AnsiString; Var S2 : ShortString);
{
  Concatenates a Ansi with a short string; : S2 + S2
}
Var
  Size,Location : SizeInt;
begin
  Size:=Length(S2);
  Location:=Length(S1);
  If Size=0 then
    exit;
  { Setlength takes case of uniqueness
    and alllocated memory. We need to use length,
    to take into account possibility of S1=Nil }
  SetLength (S1,Size+Length(S1));
  Move (S2[1],Pointer(Pointer(S1)+Location)^,Size);
  PByte( Pointer(S1)+length(S1) )^:=0; { Terminating Zero }
end;
{$endif EXTRAANSISHORT}

{$ifdef FPC_HAS_CPSTRING}
Function fpc_AnsiStr_To_AnsiStr (const S : RawByteString;cp : TSystemCodePage): RawByteString; [Public, alias: 'FPC_ANSISTR_TO_ANSISTR']; compilerproc;
{
  Converts an AnsiString to an AnsiString taking code pages into care
}
Var
  Size : SizeInt;
  temp : UnicodeString;
begin
  result:='';
  Size:=Length(S);
  if Size>0 then
    begin
      if (cp=0) then
        cp:=DefaultSystemCodePage;
      if (StringCodePage(S)=cp) then
        begin
          SetLength(result,Size);
          Move(S[1],result[1],Size);
          PAnsiRec(Pointer(result)-AnsiFirstOff)^.CodePage:=cp;
        end
      else
        begin
          temp:=S;
          Size:=Length(temp);
          widestringmanager.Unicode2AnsiMoveProc(PUnicodeChar(Pointer(temp)),result,cp,Size);
        end;
    end;
end;

Function fpc_AnsiStr_To_AnsiStr (const S : RawByteString;cp : TSystemCodePage): RawByteString; [external name 'FPC_ANSISTR_TO_ANSISTR'];
{$endif FPC_HAS_CPSTRING}

{$ifndef FPC_STRTOSHORTSTRINGPROC}

{ the following declaration has exactly the same effect as                   }
{ procedure fpc_AnsiStr_To_ShortStr (Var S1 : ShortString;S2 : Pointer);     }
{ which is what the old helper was, so we don't need an extra implementation }
{ of the old helper (JM)                                                     }
function fpc_AnsiStr_To_ShortStr (high_of_res: SizeInt;const S2 : Ansistring): shortstring; [Public, alias: 'FPC_ANSISTR_TO_SHORTSTR']; compilerproc;
{
  Converts a AnsiString to a ShortString;
}
Var
  Size : SizeInt;
begin
  if S2='' then
   fpc_AnsiStr_To_ShortStr:=''
  else
   begin
     Size:=Length(S2);
     If Size>high_of_res then
      Size:=high_of_res;
     Move (S2[1],fpc_AnsiStr_To_ShortStr[1],Size);
     byte(fpc_AnsiStr_To_ShortStr[0]):=byte(Size);
   end;
end;

{$else FPC_STRTOSHORTSTRINGPROC}

procedure fpc_AnsiStr_To_ShortStr (out res: shortstring; const S2 : RawByteString);[Public, alias: 'FPC_ANSISTR_TO_SHORTSTR'];  compilerproc;
{
  Converts a AnsiString to a ShortString;
}
Var
  Size : SizeInt;
begin
  if S2='' then
   res:=''
  else
   begin
     Size:=Length(S2);
     If Size>high(res) then
      Size:=high(res);
     Move (S2[1],res[1],Size);
     byte(res[0]):=byte(Size);
   end;
end;

{$endif FPC_STRTOSHORTSTRINGPROC}


Function fpc_ShortStr_To_AnsiStr (Const S2 : ShortString{$ifdef FPC_HAS_CPSTRING};cp : TSystemCodePage{$endif FPC_HAS_CPSTRING}): RawByteString; compilerproc;
{
  Converts a ShortString to a AnsiString;
}
Var
  Size : SizeInt;
{$ifndef FPC_HAS_CPSTRING}
  cp : TSystemCodePage;
{$endif FPC_HAS_CPSTRING}
begin
{$ifdef FPC_HAS_CPSTRING}
  if (cp=0) then
    cp:=DefaultSystemCodePage;
{$else FPC_HAS_CPSTRING}
  cp:=DefaultSystemCodePage;
{$endif FPC_HAS_CPSTRING}
  Size:=Length(S2);
  Setlength (fpc_ShortStr_To_AnsiStr,Size);
  if Size>0 then
    begin
      Move(S2[1],Pointer(fpc_ShortStr_To_AnsiStr)^,Size);
      SetCodePage(fpc_ShortStr_To_AnsiStr,cp,False);
    end
end;

Function fpc_Char_To_AnsiStr(const c : Char{$ifdef FPC_HAS_CPSTRING};cp : TSystemCodePage{$endif FPC_HAS_CPSTRING}): RawByteString; compilerproc;
{
  Converts a Char to a AnsiString;
}
{$ifndef FPC_HAS_CPSTRING}
var
  cp : TSystemCodePage;
{$endif FPC_HAS_CPSTRING}
begin
{$ifdef FPC_HAS_CPSTRING}
  if (cp=0) then
    cp:=DefaultSystemCodePage;
{$else FPC_HAS_CPSTRING}
  cp:=DefaultSystemCodePage;
{$endif FPC_HAS_CPSTRING}
  Setlength (fpc_Char_To_AnsiStr,1);
  PByte(Pointer(fpc_Char_To_AnsiStr))^:=byte(c);
  { Terminating Zero }
  PByte(Pointer(fpc_Char_To_AnsiStr)+1)^:=0;
  SetCodePage(fpc_Char_To_AnsiStr,cp,False);
end;


Function fpc_PChar_To_AnsiStr(const p : pchar{$ifdef FPC_HAS_CPSTRING};cp : TSystemCodePage{$endif FPC_HAS_CPSTRING}): RawByteString; compilerproc;
Var
  L : SizeInt;
{$ifndef FPC_HAS_CPSTRING}
  cp : TSystemCodePage;
{$endif FPC_HAS_CPSTRING}
begin
  if (not assigned(p)) or (p[0]=#0) Then
    L := 0
  else
    l:=IndexChar(p^,-1,#0);
  SetLength(fpc_PChar_To_AnsiStr,L);
  if L > 0 then
    begin
{$ifdef FPC_HAS_CPSTRING}
      if (cp=0) then
        cp:=DefaultSystemCodePage;
{$else FPC_HAS_CPSTRING}
      cp:=DefaultSystemCodePage;
{$endif FPC_HAS_CPSTRING}
      Move (P[0],Pointer(fpc_PChar_To_AnsiStr)^,L);
      SetCodePage(fpc_PChar_To_AnsiStr,cp,False);
    end;
end;


Function fpc_CharArray_To_AnsiStr(const arr: array of ansichar; {$ifdef FPC_HAS_CPSTRING}cp : TSystemCodePage;{$endif FPC_HAS_CPSTRING}zerobased: boolean = true): RawByteString; compilerproc;
var
  i  : SizeInt;
{$ifndef FPC_HAS_CPSTRING}
  cp : TSystemCodePage;
{$endif FPC_HAS_CPSTRING}
begin
  if (zerobased) then
    begin
      if (arr[0]=#0) Then
        i := 0
      else
      begin
        i:=IndexChar(arr,high(arr)+1,#0);
        if i = -1 then
          i := high(arr)+1;
      end;
    end
  else
    i := high(arr)+1;
  SetLength(fpc_CharArray_To_AnsiStr,i);
  if i > 0 then
    begin
{$ifdef FPC_HAS_CPSTRING}
      if (cp=0) then
        cp:=DefaultSystemCodePage;
{$else FPC_HAS_CPSTRING}
      cp:=DefaultSystemCodePage;
{$endif FPC_HAS_CPSTRING}
      Move (arr[0],Pointer(fpc_CharArray_To_AnsiStr)^,i);
      SetCodePage(fpc_CharArray_To_AnsiStr,cp,False);
    end;
end;

{$ifndef FPC_STRTOCHARARRAYPROC}

{ note: inside the compiler, the resulttype is modified to be the length }
{ of the actual chararray to which we convert (JM)                       }
function fpc_ansistr_to_chararray(arraysize: SizeInt; const src: ansistring): fpc_big_chararray; [public, alias: 'FPC_ANSISTR_TO_CHARARRAY']; compilerproc;
var
  len: SizeInt;
begin
  len := length(src);
  if len > arraysize then
    len := arraysize;
{$push}{$r-}
  { make sure we don't try to access element 1 of the ansistring if it's nil }
  if len > 0 then
    move(src[1],fpc_ansistr_to_chararray[0],len);
  { fpc_big_chararray is defined as array[0..0], see compproc.inc why }
  fillchar(fpc_ansistr_to_chararray[len],arraysize-len,0);
{$pop}
end;

{$else ndef FPC_STRTOCHARARRAYPROC}

procedure  fpc_ansistr_to_chararray(out res: array of char; const src: RawByteString); compilerproc;
var
  len: SizeInt;
begin
  len := length(src);
  if len > length(res) then
    len := length(res);
{$push}{$r-}
  { make sure we don't try to access element 1 of the ansistring if it's nil }
  if len > 0 then
    move(src[1],res[0],len);
  { fpc_big_chararray is defined as array[0..0], see compproc.inc why }
  fillchar(res[len],length(res)-len,0);
{$pop}
end;

{$endif ndef FPC_STRTOCHARARRAYPROC}

Function fpc_AnsiStr_Compare(const S1,S2 : RawByteString): SizeInt;[Public,Alias : 'FPC_ANSISTR_COMPARE'];  compilerproc;
{
  Compares 2 AnsiStrings;
  The result is
   <0 if S1<S2
   0 if S1=S2
   >0 if S1>S2
}
Var
  MaxI,Temp : SizeInt;
  cp1,cp2 : TSystemCodePage;
  r1,r2 : RawByteString;
  u1,u2 : UTF8String;
begin
  if pointer(S1)=pointer(S2) then
    begin
      result:=0;
      exit;
    end;
  if (pointer(S1)=nil) then
    begin
      result:=-Length(S2);
      exit;
    end;
  if (pointer(S2)=nil) then
    begin
      result:=Length(S1);
      exit;
    end;
  cp1:=StringCodePage(S1);
  cp2:=StringCodePage(S2);
  if (cp1=cp2) then
    begin
      Maxi:=Length(S1);
      temp:=Length(S2);
      If MaxI>Temp then
        MaxI:=Temp;
      if MaxI>0 then
        begin
          result:=CompareByte(S1[1],S2[1],MaxI);
          if result=0 then
            result:=Length(S1)-Length(S2);
        end
      else
        result:=Length(S1)-Length(S2);
    end
  else
    begin
      r1:=S1;
      if (cp1=$ffff) or (cp1=0) then
        SetCodePage(r1,DefaultSystemCodePage,false);
      r2:=S2;
      if (cp2=$ffff) or (cp2=0) then
        SetCodePage(r2,DefaultSystemCodePage,false);
      //convert them to utf8 then compare
      SetCodePage(r1,65001);
      SetCodePage(r2,65001);
      Result := fpc_AnsiStr_Compare(r1,r2);
    end;
end;

Function fpc_AnsiStr_Compare_equal(const S1,S2 : RawByteString): SizeInt;[Public,Alias : 'FPC_ANSISTR_COMPARE_EQUAL'];  compilerproc;
{
  Compares 2 AnsiStrings for equality/inequality only;
  The result is
   0 if S1=S2
   <>0 if S1<>S2
}
Var
  MaxI,Temp : SizeInt;
  cp1,cp2 : TSystemCodePage;
  r1,r2 : RawByteString;
begin
  if pointer(S1)=pointer(S2) then
    begin
      result:=0;
      exit;
    end;
  cp1:=StringCodePage(S1);
  cp2:=StringCodePage(S2);
  if (cp1=cp2) then
    begin
      Maxi:=Length(S1);
      temp:=Length(S2);
      Result := Maxi - temp;
      if Result = 0 then
        if MaxI>0 then
          result:=CompareByte(S1[1],S2[1],MaxI);
    end
  else
    begin
      r1:=S1;
      if (cp1=$ffff) or (cp1=0) then
        SetCodePage(r1,DefaultSystemCodePage,false);
      r2:=S2;
      if (cp2=$ffff) or (cp2=0) then
        SetCodePage(r2,DefaultSystemCodePage,false);
      //convert them to utf8 then compare
      SetCodePage(r1,65001);
      SetCodePage(r2,65001);
      Maxi:=Length(r1);
      temp:=Length(r2);
      Result := Maxi - temp;
      if Result = 0 then
        if MaxI>0 then
          result:=CompareByte(r1[1],r2[1],MaxI);
    end;
end;

{$ifdef VER2_4}
// obsolete but needed for boostrapping with 2.4
Procedure fpc_AnsiStr_CheckZero(p : pointer);[Public,Alias : 'FPC_ANSISTR_CHECKZERO'];  compilerproc;
begin
  if p=nil then
    HandleErrorFrame(201,get_frame);
end;

Procedure fpc_AnsiStr_CheckRange(len,index : SizeInt);[Public,Alias : 'FPC_ANSISTR_RANGECHECK'];  compilerproc;
begin
  if (index>len) or (Index<1) then
    HandleErrorFrame(201,get_frame);
end;

{$else VER2_4}
Procedure fpc_AnsiStr_CheckRange(p: Pointer; index: SizeInt);[Public,Alias : 'FPC_ANSISTR_RANGECHECK'];  compilerproc;
begin
  if (p=nil) or (index>PAnsiRec(p-AnsiFirstOff)^.Len) or (Index<1) then
    HandleErrorFrame(201,get_frame);
end;
{$endif VER2_4}

Procedure fpc_AnsiStr_SetLength (Var S : RawByteString; l : SizeInt{$ifdef FPC_HAS_CPSTRING};cp : TSystemCodePage{$endif FPC_HAS_CPSTRING});[Public,Alias : 'FPC_ANSISTR_SETLENGTH'];  compilerproc;
{
  Sets The length of string S to L.
  Makes sure S is unique, and contains enough room.
}
Var
  Temp : Pointer;
  lens, lena,
  movelen : SizeInt;
begin
  if (l>0) then
    begin
      if Pointer(S)=nil then
       begin
         GetMem(Pointer(S),AnsiRecLen+L);
         PAnsiRec(S)^.Ref:=1;
{$ifdef FPC_HAS_CPSTRING}
         if (cp=0) then
           cp:=DefaultSystemCodePage;
         PAnsiRec(S)^.CodePage:=cp;
{$else}
         PAnsiRec(S)^.CodePage:=DefaultSystemCodePage;
{$endif FPC_HAS_CPSTRING}
         PAnsiRec(S)^.ElementSize:=1;
         inc(Pointer(S),AnsiFirstOff);
       end
      else if PAnsiRec(Pointer(S)-AnsiFirstOff)^.Ref=1 then
        begin
          Dec(Pointer(S),AnsiFirstOff);
          lens:=MemSize(Pointer(s));
          lena:=AnsiRecLen+L;
          { allow shrinking string if that saves at least half of current size }
          if (lena>lens) or ((lens>32) and (lena<=(lens div 2))) then
            reallocmem(pointer(S),AnsiRecLen+L);
          Inc(Pointer(S),AnsiFirstOff);
        end
      else
        begin
          { Reallocation is needed... }
          Temp:=Pointer(NewAnsiString(L));
{$ifdef FPC_HAS_CPSTRING}
          PAnsiRec(Pointer(Temp)-AnsiFirstOff)^.CodePage:=cp;
{$endif FPC_HAS_CPSTRING}

          { also move terminating null }
          lens:=succ(length(s));
          if l<lens then
            movelen:=l
          else
            movelen:=lens;
          Move(Pointer(S)^,Temp^,movelen);
          { ref count dropped to zero in the mean time? }
          If (PAnsiRec(Pointer(S)-AnsiFirstOff)^.Ref>0) and
            declocked(PAnsiRec(Pointer(S)-AnsiFirstOff)^.Ref) then
            freemem(PAnsiRec(Pointer(s)-AnsiFirstOff));
          Pointer(S):=Temp;
       end;
      { Force nil termination in case it gets shorter }
      PByte(Pointer(S)+l)^:=0;
      PAnsiRec(Pointer(S)-AnsiFirstOff)^.Len:=l;
    end
  else
    begin
      { Length=0 }
      if Pointer(S)<>nil then
        fpc_ansistr_decr_ref (Pointer(S));
      Pointer(S):=Nil;
    end;
end;

{$ifdef EXTRAANSISHORT}
Function fpc_AnsiStr_ShortStr_Compare (Var S1 : Pointer; Var S2 : ShortString): SizeInt;  compilerproc;
{
  Compares a AnsiString with a ShortString;
  The result is
   <0 if S1<S2
   0 if S1=S2
   >0 if S1>S2
}
Var
  i,MaxI,Temp : SizeInt;
begin
  Temp:=0;
  i:=0;
  MaxI:=Length(AnsiString(S1));
  if MaxI>byte(S2[0]) then
    MaxI:=Byte(S2[0]);
  While (i<MaxI) and (Temp=0) do
   begin
     Temp:= PByte(S1+I)^ - Byte(S2[i+1]);
     inc(i);
   end;
  AnsiStr_ShortStr_Compare:=Temp;
end;
{$endif EXTRAANSISHORT}


{*****************************************************************************
                     Public functions, In interface.
*****************************************************************************}

function fpc_truely_ansistr_unique(Var S : Pointer): Pointer;
Var
  SNew : Pointer;
  L    : SizeInt;
begin
  L:=PAnsiRec(Pointer(S)-AnsiFirstOff)^.len;
  SNew:=NewAnsiString (L);
  Move (Pointer(S)^,SNew^,L+1);
  PAnsiRec(SNew-AnsiFirstOff)^.len:=L;
  PAnsiRec(SNew-AnsiFirstOff)^.CodePage:=PAnsiRec(Pointer(S)-AnsiFirstOff)^.CodePage;
  fpc_ansistr_decr_ref (Pointer(S));  { Thread safe }
  pointer(S):=SNew;
  pointer(result):=SNew;
end;


{$ifndef FPC_SYSTEM_HAS_ANSISTR_UNIQUE}
// MV: inline the basic checks for case that S is already unique.
// Rest is too complex to inline, so factor that out as a call.
Function fpc_ansistr_Unique(Var S : Pointer): Pointer; [Public,Alias : 'FPC_ANSISTR_UNIQUE']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
{
  Make sure reference count of S is 1,
  using copy-on-write semantics.
}
begin
  pointer(result) := pointer(s);
  If Pointer(S)=Nil then
    exit;
  if PAnsiRec(Pointer(S)-AnsiFirstOff)^.Ref<>1 then
    result:=fpc_truely_ansistr_unique(s);
end;
{$endif FPC_SYSTEM_HAS_ANSISTR_UNIQUE}


Procedure fpc_ansistr_append_char(Var S : AnsiString;c : char); [Public,Alias : 'FPC_ANSISTR_APPEND_CHAR']; compilerproc;
begin
  SetLength(S,length(S)+1);
  // avoid unique call
  PChar(Pointer(S)+length(S)-1)^:=c;
  PByte(Pointer(S)+length(S))^:=0; { Terminating Zero }
end;

Procedure fpc_ansistr_append_shortstring(Var S : AnsiString;const Str : ShortString); [Public,Alias : 'FPC_ANSISTR_APPEND_SHORTSTRING']; compilerproc;
var
   ofs : SizeInt;
begin
   if Str='' then
     exit;
   ofs:=Length(S);
   SetLength(S,ofs+length(Str));
   { the pbyte cast avoids an unique call which isn't necessary because SetLength was just called }
   move(Str[1],(pointer(S)+ofs)^,length(Str));
   PByte(Pointer(S)+length(S))^:=0; { Terminating Zero }
end;

Procedure fpc_ansistr_append_ansistring(Var S : AnsiString;const Str : AnsiString); [Public,Alias : 'FPC_ANSISTR_APPEND_ANSISTRING']; compilerproc;
var
   ofs, strlength: SizeInt;
   samestring: boolean;
begin
   if Str='' then
     exit;
   samestring := pointer(s) = pointer(str);
   { needed in case s and str are the same string }
   strlength := length(str);
   ofs:=Length(S);
   SetLength(S,ofs+strlength);
   { the pbyte cast avoids an unique call which isn't necessary because SetLength was just called }
   if not(samestring) then
     move(Str[1],(pointer(S)+ofs)^,strlength+1)
   else
     { the setlength may have relocated the string, so str may no longer be valid }
     move(S[1],(pointer(S)+ofs)^,strlength+1)
end;

Function Fpc_Ansistr_Copy (Const S : RawByteString; Index,Size : SizeInt) : AnsiString;compilerproc;
var
  ResultAddress : Pointer;
begin
  ResultAddress:=Nil;
  dec(index);
  if Index < 0 then
    Index := 0;
  { Check Size. Accounts for Zero-length S, the double check is needed because
    Size can be maxint and will get <0 when adding index }
  if (Size>Length(S)) or
     (Index+Size>Length(S)) then
   Size:=Length(S)-Index;
  If Size>0 then
   begin
     If Index<0 Then
      Index:=0;
     ResultAddress:=Pointer(NewAnsiString (Size));
     if ResultAddress<>Nil then
      begin
        Move (Pointer(Pointer(S)+index)^,ResultAddress^,Size);
        PByte(ResultAddress+Size)^:=0;
        PAnsiRec(ResultAddress-AnsiFirstOff)^.Len:=Size;
        PAnsiRec(ResultAddress-AnsiFirstOff)^.CodePage:=PAnsiRec(Pointer(S)-AnsiFirstOff)^.CodePage;
      end;
   end;
   fpc_ansistr_decr_ref(Pointer(fpc_ansistr_copy));
  Pointer(fpc_ansistr_Copy):=ResultAddress;
end;

Function Pos (Const Substr : ShortString; Const Source : RawByteString) : SizeInt;

var
  i,MaxLen : SizeInt;
  pc : pchar;
begin
  Pos:=0;
  if Length(SubStr)>0 then
   begin
     MaxLen:=Length(source)-Length(SubStr);
     i:=0;
     pc:=@source[1];
     while (i<=MaxLen) do
      begin
        inc(i);
        if (SubStr[1]=pc^) and
           (CompareByte(Substr[1],pc^,Length(SubStr))=0) then
         begin
           Pos:=i;
           exit;
         end;
        inc(pc);
      end;
   end;
end;


Function Pos (Const Substr : RawByteString; Const Source : RawByteString) : SizeInt;
var
  i,MaxLen : SizeInt;
  pc : pchar;
begin
  Pos:=0;
  if Length(SubStr)>0 then
   begin
     MaxLen:=Length(source)-Length(SubStr);
     i:=0;
     pc:=@source[1];
     while (i<=MaxLen) do
      begin
        inc(i);
        if (SubStr[1]=pc^) and
           (CompareByte(Substr[1],pc^,Length(SubStr))=0) then
         begin
           Pos:=i;
           exit;
         end;
        inc(pc);
      end;
   end;
end;


{ Faster version for a char alone. Must be implemented because   }
{ pos(c: char; const s: shortstring) also exists, so otherwise   }
{ using pos(char,pchar) will always call the shortstring version }
{ (exact match for first argument), also with $h+ (JM)           }
Function Pos (c : Char; Const s : RawByteString) : SizeInt;
var
  i: SizeInt;
  pc : pchar;
begin
  pc:=@s[1];
  for i:=1 to length(s) do
   begin
     if pc^=c then
      begin
        pos:=i;
        exit;
      end;
     inc(pc);
   end;
  pos:=0;
end;


{$ifndef FPUNONE}
Function fpc_Val_Real_AnsiStr(Const S : AnsiString; out Code : ValSInt): ValReal; [public, alias:'FPC_VAL_REAL_ANSISTR']; compilerproc;
Var
  SS : String;
begin
  fpc_Val_Real_AnsiStr := 0;
  if length(S) > 255 then
    code := 256
  else
    begin
      SS := S;
      Val(SS,fpc_Val_Real_AnsiStr,code);
    end;
end;
{$endif}


Function fpc_Val_Currency_AnsiStr(Const S : AnsiString; out Code : ValSInt): Currency; [public, alias:'FPC_VAL_CURRENCY_ANSISTR']; compilerproc;
Var
  SS : String;
begin
  if length(S) > 255 then
    begin
      fpc_Val_Currency_AnsiStr := 0;
      code := 256;
    end
  else
    begin
      SS := S;
      Val(SS,fpc_Val_Currency_AnsiStr,code);
    end;
end;


Function fpc_Val_UInt_AnsiStr (Const S : AnsiString; out Code : ValSInt): ValUInt; [public, alias:'FPC_VAL_UINT_ANSISTR']; compilerproc;
Var
  SS : ShortString;
begin
  fpc_Val_UInt_AnsiStr := 0;
  if length(S) > 255 then
    code := 256
  else
    begin
      SS := S;
      Val(SS,fpc_Val_UInt_AnsiStr,code);
    end;
end;


Function fpc_Val_SInt_AnsiStr (DestSize: SizeInt; Const S : AnsiString; out Code : ValSInt): ValSInt; [public, alias:'FPC_VAL_SINT_ANSISTR']; compilerproc;
Var
  SS : ShortString;
begin
  fpc_Val_SInt_AnsiStr:=0;
  if length(S)>255 then
    code:=256
  else
    begin
       SS := S;
       fpc_Val_SInt_AnsiStr := int_Val_SInt_ShortStr(DestSize,SS,Code);
    end;
end;


{$ifndef CPU64}

Function fpc_Val_qword_AnsiStr (Const S : AnsiString; out Code : ValSInt): qword; [public, alias:'FPC_VAL_QWORD_ANSISTR']; compilerproc;
Var
  SS : ShortString;
begin
  fpc_Val_qword_AnsiStr:=0;
  if length(S)>255 then
    code:=256
  else
    begin
       SS := S;
       Val(SS,fpc_Val_qword_AnsiStr,Code);
    end;
end;


Function fpc_Val_int64_AnsiStr (Const S : AnsiString; out Code : ValSInt): Int64; [public, alias:'FPC_VAL_INT64_ANSISTR']; compilerproc;
Var
  SS : ShortString;
begin
  fpc_Val_int64_AnsiStr:=0;
  if length(S)>255 then
    code:=256
  else
    begin
       SS := s;
       Val(SS,fpc_Val_int64_AnsiStr,Code);
    end;
end;

{$endif CPU64}


{$ifndef FPUNONE}
procedure fpc_AnsiStr_Float(d : ValReal;len,fr,rt : SizeInt;out s : ansistring);[public,alias:'FPC_ANSISTR_FLOAT']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
var
  ss: ShortString;
begin
  str_real(len,fr,d,treal_type(rt),ss);
  s:=ss;
end;
{$endif}

procedure fpc_ansistr_enum(ordinal,len:sizeint;typinfo,ord2strindex:pointer;out s:ansistring);[public,alias:'FPC_ANSISTR_ENUM'];compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}

var ss:shortstring;

begin
  fpc_shortstr_enum(ordinal,len,typinfo,ord2strindex,ss);
  s:=ss;
end;


procedure fpc_ansistr_bool(b : boolean;len:sizeint;out s:ansistring);[public,alias:'FPC_ANSISTR_BOOL'];compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
var
  ss:shortstring;
begin
  fpc_shortstr_bool(b,len,ss);
  s:=ss;
end;


function fpc_val_enum_ansistr(str2ordindex:pointer;const s:ansistring;out code:valsint):longint; [public, alias:'FPC_VAL_ENUM_ANSISTR']; compilerproc;

begin
  fpc_val_enum_ansistr:=fpc_val_enum_shortstr(str2ordindex,s,code);
end;


{$ifdef FPC_HAS_STR_CURRENCY}
procedure fpc_AnsiStr_Currency(c : currency;len,fr : SizeInt;out s : ansistring);[public,alias:'FPC_ANSISTR_CURRENCY']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
var
  ss: ShortString;
begin
  str(c:len:fr,ss);
  s:=ss;
end;
{$endif FPC_HAS_STR_CURRENCY}

Procedure fpc_AnsiStr_UInt(v : ValUInt;Len : SizeInt; out S : AnsiString);[Public,Alias : 'FPC_ANSISTR_VALUINT']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
Var
  SS : ShortString;
begin
  str(v:Len,SS);
  S:=SS;
end;



Procedure fpc_AnsiStr_SInt(v : ValSInt;Len : SizeInt; out S : AnsiString);[Public,Alias : 'FPC_ANSISTR_VALSINT']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
Var
  SS : ShortString;
begin
  str (v:Len,SS);
  S:=SS;
end;


{$ifndef CPU64}

Procedure fpc_AnsiStr_QWord(v : QWord;Len : SizeInt; out S : AnsiString);[Public,Alias : 'FPC_ANSISTR_QWORD']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
Var
  SS : ShortString;
begin
  str(v:Len,SS);
  S:=SS;
end;

Procedure fpc_AnsiStr_Int64(v : Int64; Len : SizeInt; out S : AnsiString);[Public,Alias : 'FPC_ANSISTR_INT64']; compilerproc; {$IFNDEF VER2_0} Inline; {$ENDIF}
Var
  SS : ShortString;
begin
  str (v:Len,SS);
  S:=SS;
end;

{$endif CPU64}

Procedure Delete (Var S : RawByteString; Index,Size: SizeInt);
Var
  LS : SizeInt;
begin
  ls:=Length(S);
  If (Index>LS) or (Index<=0) or (Size<=0) then
    exit;
  UniqueString (S);
  If (Size>LS-Index) then   // Size+Index gives overflow ??
     Size:=LS-Index+1;
  If (Size<=LS-Index) then
    begin
      Dec(Index);
      Move(PByte(Pointer(S))[Index+Size],PByte(Pointer(S))[Index],LS-Index-Size+1);
    end;
  Setlength(S,LS-Size);
end;


Procedure Insert (Const Source : RawByteString; Var S : RawByteString; Index : SizeInt);
var
  Temp : RawByteString;
  LS : SizeInt;
begin
  If Length(Source)=0 then
   exit;
  if index <= 0 then
   index := 1;
  Ls:=Length(S);
  if index > LS then
   index := LS+1;
  Dec(Index);
  Pointer(Temp) := NewAnsiString(Length(Source)+LS);
  SetLength(Temp,Length(Source)+LS);
  If Index>0 then
    move (Pointer(S)^,Pointer(Temp)^,Index);
  Move (Pointer(Source)^,PByte(Temp)[Index],Length(Source));
  If (LS-Index)>0 then
    Move(PByte(Pointer(S))[Index],PByte(temp)[Length(Source)+index],LS-Index);
  S:=Temp;
end;


Function StringOfChar(c : char;l : SizeInt) : AnsiString;
begin
  SetLength(StringOfChar,l);
  FillChar(Pointer(StringOfChar)^,Length(StringOfChar),c);
end;

Procedure SetString (Out S : AnsiString; Buf : PChar; Len : SizeInt); {$IFNDEF VER2_0} Inline; {$ENDIF}
begin
  SetLength(S,Len);
  If (Buf<>Nil) then
    Move (Buf^,Pointer(S)^,Len);
end;

Procedure SetString (Out S : AnsiString; Buf : PWideChar; Len : SizeInt);
begin
  if (Buf<>nil) and (Len>0) then
    widestringmanager.Wide2AnsiMoveProc(Buf,S,DefaultSystemCodePage,Len)
  else
    SetLength(S, Len);
end;

function upcase(const s : ansistring) : ansistring;
var
  i : SizeInt;
begin
  Setlength(result,length(s));
  for i := 1 to length (s) do
    result[i] := upcase(s[i]);
end;


function lowercase(const s : ansistring) : ansistring;
var
  i : SizeInt;
begin
  Setlength(result,length(s));
  for i := 1 to length (s) do
    result[i] := lowercase(s[i]);
end;


function StringCodePage(const S: RawByteString): TSystemCodePage; overload;
  begin
{$ifdef FPC_HAS_CPSTRING}
    if assigned(Pointer(S)) then
      Result:=PAnsiRec(pointer(S)-AnsiFirstOff)^.CodePage
    else
{$endif FPC_HAS_CPSTRING}
      Result:=DefaultSystemCodePage;
  end;


function StringElementSize(const S: RawByteString): Word; overload;
  begin
    if assigned(Pointer(S)) then
      Result:=PAnsiRec(pointer(S)-AnsiFirstOff)^.ElementSize
    else
      Result:=SizeOf(AnsiChar);
  end;


function StringRefCount(const S: RawByteString): SizeInt; overload;
  begin
    if assigned(Pointer(S)) then
      Result:=PAnsiRec(pointer(S)-AnsiFirstOff)^.Ref
    else
      Result:=0;
  end;


procedure SetCodePage(var s : RawByteString; CodePage : TSystemCodePage; Convert : Boolean = True);
  begin
    if (S='') or (StringCodePage(S)=CodePage) then
      exit
    else if Convert then
      begin
{$ifdef FPC_HAS_CPSTRING}
        s:=fpc_AnsiStr_To_AnsiStr(s,CodePage);
{$else FPC_HAS_CPSTRING}
        UniqueString(s);
        PAnsiRec(pointer(s)-AnsiFirstOff)^.CodePage:=CodePage;
{$endif FPC_HAS_CPSTRING}
      end
    else
      begin
        UniqueString(s);
        PAnsiRec(pointer(s)-AnsiFirstOff)^.CodePage:=CodePage;
      end;
  end;

procedure SetMultiByteConversionCodePage(CodePage: TSystemCodePage);
  begin
    DefaultSystemCodePage:=CodePage;
  end;