{ This file is part of the Free Pascal run time library. Copyright (c) 1999-2005 by Florian Klaempfl, member of the Free Pascal development team. This file implements support routines for WideStrings with 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 file contains the implementation of the WideString type, and all things that are needed for it. WideString is defined as a 'silent' pwidechar : a pwidechar that points to : @-8 : SizeInt for reference count; @-4 : SizeInt for size; size=number of bytes, not the number of chars. Divide or multiply with sizeof(WideChar) to convert. This is needed to be compatible with Delphi and Windows COM BSTR. @ : String + Terminating #0; Pwidechar(Widestring) is a valid typecast. So WS[i] is converted to the address @WS+i-1. Constants should be assigned a reference count of -1 Meaning that they can't be disposed of. } Type PWideRec = ^TWideRec; TWideRec = Packed Record Len : DWord; First : WideChar; end; Const WideRecLen = SizeOf(TWideRec); WideFirstOff = SizeOf(TWideRec)-sizeof(WideChar); { Default WideChar <-> Char conversion is to only convert the lower 127 chars, all others are translated to spaces. These routines can be overwritten for the Current Locale } procedure DefaultWide2AnsiMove(source:pwidechar;var dest:ansistring;len:SizeInt); var i : SizeInt; begin setlength(dest,len); for i:=1 to len do begin if word(source^)<256 then dest[i]:=char(word(source^)) else dest[i]:='?'; inc(source); end; end; procedure DefaultAnsi2WideMove(source:pchar;var dest:widestring;len:SizeInt); var i : SizeInt; begin setlength(dest,len); for i:=1 to len do begin dest[i]:=widechar(byte(source^)); inc(source); end; end; (* Procedure UniqueWideString(Var S : WideString); [Public,Alias : 'FPC_WIDESTR_UNIQUE']; { Make sure reference count of S is 1, using copy-on-write semantics. } begin end; *) {**************************************************************************** Internal functions, not in interface. ****************************************************************************} procedure WideStringError; begin HandleErrorFrame(204,get_frame); end; {$ifdef WideStrDebug} Procedure DumpWideRec(S : Pointer); begin If S=Nil then Writeln ('String is nil') Else Begin With PWideRec(S-WideFirstOff)^ do begin Write ('(Len:',len); Writeln (' Ref: ',ref,')'); end; end; end; {$endif} Function NewWideString(Len : SizeInt) : Pointer; { Allocate a new WideString on the heap. initialize it to zero length and reference count 1. } Var P : Pointer; begin {$ifdef MSWINDOWS} if winwidestringalloc then begin P:=SysAllocStringLen(nil,Len); if P=nil then WideStringError; end else {$endif MSWINDOWS} begin GetMem(P,Len*sizeof(WideChar)+WideRecLen); If P<>Nil then begin PWideRec(P)^.Len:=Len*2; { Initial length } PWideRec(P)^.First:=#0; { Terminating #0 } inc(p,WideFirstOff); { Points to string now } end else WideStringError; end; NewWideString:=P; end; Procedure DisposeWideString(Var S : Pointer); { Deallocates a WideString From the heap. } begin If S=Nil then exit; {$ifndef MSWINDOWS} Dec (S,WideFirstOff); Freemem(S); {$else MSWINDOWS} if winwidestringalloc then SysFreeString(S) else begin Dec (S,WideFirstOff); Freemem(S); end; {$endif MSWINDOWS} S:=Nil; end; var __data_start: byte; external name '__data_start__'; __data_end: byte; external name '__data_end__'; function IsWideStringConstant(S: pointer): boolean;{$ifdef SYSTEMINLINE}inline;{$endif} { Returns True if widestring is constant (located in .data section); } begin Result:=(S>=@__data_start) and (S<@__data_end); end; Procedure fpc_WideStr_Decr_Ref (Var S : Pointer);[Public,Alias:'FPC_WIDESTR_DECR_REF']; compilerproc; { Decreases the ReferenceCount of a non constant widestring; If the reference count is zero, deallocate the string; } Type pSizeInt = ^SizeInt; Begin { Zero string } if S=Nil then exit; if not IsWideStringConstant(S) then DisposeWideString(S); end; { alias for internal use } Procedure fpc_WideStr_Decr_Ref (Var S : Pointer);[external name 'FPC_WIDESTR_DECR_REF']; Procedure fpc_WideStr_Incr_Ref(Var S : Pointer);[Public,Alias:'FPC_WIDESTR_INCR_REF']; compilerproc; var p : pointer; Begin If S=Nil then exit; p:=NewWidestring(length(WideString(S))); move(s^,p^,(length(WideString(s))+1)*sizeof(widechar)); // double #0 too s:=p; end; { alias for internal use } Procedure fpc_WideStr_Incr_Ref (Var S : Pointer);[external name 'FPC_WIDESTR_INCR_REF']; {$ifndef FPC_STRTOSHORTSTRINGPROC} function fpc_WideStr_To_ShortStr (high_of_res: SizeInt;const S2 : WideString): shortstring;[Public, alias: 'FPC_WIDESTR_TO_SHORTSTR']; compilerproc; { Converts a WideString to a ShortString; } Var Size : SizeInt; temp : ansistring; begin result:=''; Size:=Length(S2); if Size>0 then begin If Size>high_of_res then Size:=high_of_res; widestringmanager.Wide2AnsiMoveProc(PWideChar(S2),temp,Size); result:=temp; end; end; {$else FPC_STRTOSHORTSTRINGPROC} procedure fpc_WideStr_To_ShortStr (out res: ShortString;const S2 : WideString); [Public, alias: 'FPC_WIDESTR_TO_SHORTSTR'];compilerproc; { Converts a WideString to a ShortString; } Var Size : SizeInt; temp : ansistring; begin res:=''; Size:=Length(S2); if Size>0 then begin If Size>high(res) then Size:=high(res); widestringmanager.Wide2AnsiMoveProc(PWideChar(S2),temp,Size); res:=temp; end; end; {$endif FPC_STRTOSHORTSTRINGPROC} Function fpc_ShortStr_To_WideStr (Const S2 : ShortString): WideString;compilerproc; { Converts a ShortString to a WideString; } Var Size : SizeInt; begin result:=''; Size:=Length(S2); if Size>0 then begin widestringmanager.Ansi2WideMoveProc(PChar(@S2[1]),result,Size); { Terminating Zero } PWideChar(Pointer(fpc_ShortStr_To_WideStr)+Size*sizeof(WideChar))^:=#0; end; end; Function fpc_WideStr_To_AnsiStr (const S2 : WideString): AnsiString; compilerproc; { Converts a WideString to an AnsiString } Var Size : SizeInt; begin result:=''; Size:=Length(S2); if Size>0 then widestringmanager.Wide2AnsiMoveProc(PWideChar(Pointer(S2)),result,Size); end; Function fpc_AnsiStr_To_WideStr (Const S2 : AnsiString): WideString; compilerproc; { Converts an AnsiString to a WideString; } Var Size : SizeInt; begin result:=''; Size:=Length(S2); if Size>0 then widestringmanager.Ansi2WideMoveProc(PChar(S2),result,Size); end; Function fpc_PWideChar_To_WideStr(const p : pwidechar): widestring; compilerproc; var Size : SizeInt; begin result:=''; if p=nil then exit; Size := IndexWord(p^, -1, 0); Setlength(result,Size); if Size>0 then begin Move(p^,PWideChar(Pointer(result))^,Size*sizeof(WideChar)); { Terminating Zero } PWideChar(Pointer(result)+Size*sizeof(WideChar))^:=#0; end; end; { checked against the ansistring routine, 2001-05-27 (FK) } Procedure fpc_WideStr_Assign (Var S1 : Pointer;S2 : Pointer);[Public,Alias:'FPC_WIDESTR_ASSIGN']; compilerproc; { Assigns S2 to S1 (S1:=S2), taking in account reference counts. } begin if S1=S2 then exit; if S2<>nil then begin if IsWideStringConstant(S1) then begin S1:=NewWidestring(length(WideString(S2))); move(s2^,s1^,(length(WideString(s1))+1)*sizeof(widechar)); end else {$ifdef MSWINDOWS} if winwidestringalloc then begin if SysReAllocStringLen(S1, S2, Length(WideString(S2))) = 0 then WideStringError; end else {$endif MSWINDOWS} begin SetLength(WideString(S1),length(WideString(S2))); move(s2^,s1^,(length(WideString(s1))+1)*sizeof(widechar)); end; end else begin { Free S1 } fpc_widestr_decr_ref (S1); S1:=nil; end; end; { alias for internal use } Procedure fpc_WideStr_Assign (Var S1 : Pointer;S2 : Pointer);[external name 'FPC_WIDESTR_ASSIGN']; {$ifndef STR_CONCAT_PROCS} function fpc_WideStr_Concat (const S1,S2 : WideString): WideString; compilerproc; Var Size,Location : SizeInt; pc : pwidechar; begin { only assign if s1 or s2 is empty } if (S1='') then begin result:=s2; exit; end; if (S2='') then begin result:=s1; exit; end; Location:=Length(S1); Size:=length(S2); SetLength(result,Size+Location); pc:=pwidechar(result); Move(S1[1],pc^,Location*sizeof(WideChar)); inc(pc,location); Move(S2[1],pc^,(Size+1)*sizeof(WideChar)); end; function fpc_WideStr_Concat_multi (const sarr:array of Widestring): widestring; compilerproc; Var i : Longint; p : pointer; pc : pwidechar; Size,NewSize : SizeInt; begin { First calculate size of the result so we can do a single call to SetLength() } NewSize:=0; for i:=low(sarr) to high(sarr) do inc(Newsize,length(sarr[i])); SetLength(result,NewSize); pc:=pwidechar(result); for i:=low(sarr) to high(sarr) do begin p:=pointer(sarr[i]); if assigned(p) then begin Size:=length(widestring(p)); Move(pwidechar(p)^,pc^,(Size+1)*sizeof(WideChar)); inc(pc,size); end; end; end; {$else STR_CONCAT_PROCS} procedure fpc_WideStr_Concat (var DestS:Widestring;const S1,S2 : WideString); compilerproc; Var Size,Location : SizeInt; same : boolean; begin { 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*sizeof(WideChar))^,(Size)*sizeof(WideChar)) else Move(Pointer(S2)^,(Pointer(DestS)+Location*sizeof(WideChar))^,(Size+1)*sizeof(WideChar)); end else if Pointer(DestS)=Pointer(S2) then begin SetLength(DestS,Size+Location); Move(Pointer(DestS)^,(Pointer(DestS)+Location*sizeof(WideChar))^,(Size+1)*sizeof(WideChar)); Move(Pointer(S1)^,Pointer(DestS)^,Location*sizeof(WideChar)); end else begin DestS:=''; SetLength(DestS,Size+Location); Move(Pointer(S1)^,Pointer(DestS)^,Location*sizeof(WideChar)); Move(Pointer(S2)^,(Pointer(DestS)+Location*sizeof(WideChar))^,(Size+1)*sizeof(WideChar)); end; end; procedure fpc_WideStr_Concat_multi (var DestS:Widestring;const sarr:array of Widestring); compilerproc; Var i : Longint; p,pc : pointer; Size,NewLen : SizeInt; DestTmp : Widestring; begin if high(sarr)=0 then begin DestS:=''; exit; end; { 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(DestTmp,NewLen); pc:=pwidechar(DestTmp); for i:=low(sarr) to high(sarr) do begin p:=pointer(sarr[i]); if assigned(p) then begin Size:=length(widestring(p)); Move(p^,pc^,(Size+1)*sizeof(WideChar)); inc(pc,size*sizeof(WideChar)); end; end; DestS:=DestTmp; end; {$endif STR_CONCAT_PROCS} Function fpc_Char_To_WideStr(const c : Char): WideString; compilerproc; { Converts a Char to a WideString; } begin Setlength(fpc_Char_To_WideStr,1); fpc_Char_To_WideStr[1]:=c; { Terminating Zero } PWideChar(Pointer(fpc_Char_To_WideStr)+sizeof(WideChar))^:=#0; end; Function fpc_WChar_To_WideStr(const c : WideChar): WideString; compilerproc; { Converts a WideChar to a WideString; } begin Setlength (fpc_WChar_To_WideStr,1); fpc_WChar_To_WideStr[1]:= c; end; Function fpc_WChar_To_AnsiStr(const c : WideChar): AnsiString; compilerproc; { Converts a WideChar to a AnsiString; } begin widestringmanager.Wide2AnsiMoveProc(@c, fpc_WChar_To_AnsiStr, 1); end; Function fpc_UChar_To_WideStr(const c : WideChar): WideString; compilerproc; { Converts a WideChar to a WideString; } begin Setlength (fpc_UChar_To_WideStr,1); fpc_UChar_To_WideStr[1]:= c; end; Function fpc_PChar_To_WideStr(const p : pchar): WideString; compilerproc; Var L : SizeInt; begin if (not assigned(p)) or (p[0]=#0) Then begin fpc_pchar_to_widestr := ''; exit; end; l:=IndexChar(p^,-1,#0); widestringmanager.Ansi2WideMoveProc(P,fpc_PChar_To_WideStr,l); end; Function fpc_CharArray_To_WideStr(const arr: array of char; zerobased: boolean = true): WideString; compilerproc; var i : SizeInt; begin if (zerobased) then begin if (arr[0]=#0) Then begin fpc_chararray_to_widestr := ''; exit; end; i:=IndexChar(arr,high(arr)+1,#0); if i = -1 then i := high(arr)+1; end else i := high(arr)+1; SetLength(fpc_CharArray_To_WideStr,i); widestringmanager.Ansi2WideMoveProc (pchar(@arr),fpc_CharArray_To_WideStr,i); end; {$ifndef FPC_STRTOCHARARRAYPROC} { inside the compiler, the resulttype is modified to that of the actual } { chararray we're converting to (JM) } function fpc_widestr_to_chararray(arraysize: SizeInt; const src: WideString): fpc_big_chararray;[public,alias: 'FPC_WIDESTR_TO_CHARARRAY']; compilerproc; var len: SizeInt; temp: ansistring; begin len := length(src); { make sure we don't dereference src if it can be nil (JM) } if len > 0 then widestringmanager.wide2ansimoveproc(pwidechar(@src[1]),temp,len); len := length(temp); if len > arraysize then len := arraysize; {$r-} move(temp[1],fpc_widestr_to_chararray[0],len); fillchar(fpc_widestr_to_chararray[len],arraysize-len,0); {$ifdef RangeCheckWasOn} {$r+} {$endif} end; { inside the compiler, the resulttype is modified to that of the actual } { widechararray we're converting to (JM) } function fpc_widestr_to_widechararray(arraysize: SizeInt; const src: WideString): fpc_big_widechararray;[public,alias: 'FPC_WIDESTR_TO_WIDECHARARRAY']; compilerproc; var len: SizeInt; begin len := length(src); if len > arraysize then len := arraysize; {$r-} { make sure we don't try to access element 1 of the ansistring if it's nil } if len > 0 then move(src[1],fpc_widestr_to_widechararray[0],len*SizeOf(WideChar)); fillchar(fpc_widestr_to_widechararray[len],(arraysize-len)*SizeOf(WideChar),0); {$ifdef RangeCheckWasOn} {$r+} {$endif} end; { inside the compiler, the resulttype is modified to that of the actual } { chararray we're converting to (JM) } function fpc_ansistr_to_widechararray(arraysize: SizeInt; const src: AnsiString): fpc_big_widechararray;[public,alias: 'FPC_ANSISTR_TO_WIDECHARARRAY']; compilerproc; var len: SizeInt; temp: widestring; begin len := length(src); { make sure we don't dereference src if it can be nil (JM) } if len > 0 then widestringmanager.ansi2widemoveproc(pchar(@src[1]),temp,len); len := length(temp); if len > arraysize then len := arraysize; {$r-} move(temp[1],fpc_ansistr_to_widechararray[0],len*sizeof(widechar)); fillchar(fpc_ansistr_to_widechararray[len],(arraysize-len)*SizeOf(WideChar),0); {$ifdef RangeCheckWasOn} {$r+} {$endif} end; function fpc_shortstr_to_widechararray(arraysize: SizeInt; const src: ShortString): fpc_big_widechararray;[public,alias: 'FPC_SHORTSTR_TO_WIDECHARARRAY']; compilerproc; var len: longint; temp : widestring; begin len := length(src); { make sure we don't access char 1 if length is 0 (JM) } if len > 0 then widestringmanager.ansi2widemoveproc(pchar(@src[1]),temp,len); len := length(temp); if len > arraysize then len := arraysize; {$r-} move(temp[1],fpc_shortstr_to_widechararray[0],len*sizeof(widechar)); fillchar(fpc_shortstr_to_widechararray[len],(arraysize-len)*SizeOf(WideChar),0); {$ifdef RangeCheckWasOn} {$r+} {$endif} end; {$else ndef FPC_STRTOCHARARRAYPROC} procedure fpc_widestr_to_chararray(out res: array of char; const src: WideString); compilerproc; var len: SizeInt; temp: ansistring; begin len := length(src); { make sure we don't dereference src if it can be nil (JM) } if len > 0 then widestringmanager.wide2ansimoveproc(pwidechar(@src[1]),temp,len); len := length(temp); if len > length(res) then len := length(res); {$r-} move(temp[1],res[0],len); fillchar(res[len],length(res)-len,0); {$ifdef RangeCheckWasOn} {$r+} {$endif} end; procedure fpc_widestr_to_widechararray(out res: array of widechar; const src: WideString); compilerproc; var len: SizeInt; begin len := length(src); if len > length(res) then len := length(res); {$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*SizeOf(WideChar)); fillchar(res[len],(length(res)-len)*SizeOf(WideChar),0); {$ifdef RangeCheckWasOn} {$r+} {$endif} end; {$endif ndef FPC_STRTOCHARARRAYPROC} Function fpc_WideStr_Compare(const S1,S2 : WideString): SizeInt;[Public,Alias : 'FPC_WIDESTR_COMPARE']; compilerproc; { Compares 2 WideStrings; The result is <0 if S10 if S1>S2 } Var MaxI,Temp : SizeInt; begin if pointer(S1)=pointer(S2) then begin fpc_WideStr_Compare:=0; exit; end; Maxi:=Length(S1); temp:=Length(S2); If MaxI>Temp then MaxI:=Temp; Temp:=CompareWord(S1[1],S2[1],MaxI); if temp=0 then temp:=Length(S1)-Length(S2); fpc_WideStr_Compare:=Temp; end; Function fpc_WideStr_Compare_Equal(const S1,S2 : WideString): SizeInt;[Public,Alias : 'FPC_WIDESTR_COMPARE_EQUAL']; compilerproc; { Compares 2 WideStrings for equality only; The result is 0 if S1=S2 <>0 if S1<>S2 } Var MaxI : SizeInt; begin if pointer(S1)=pointer(S2) then exit(0); Maxi:=Length(S1); If MaxI<>Length(S2) then exit(-1) else exit(CompareWord(S1[1],S2[1],MaxI)); end; Procedure fpc_WideStr_CheckZero(p : pointer);[Public,Alias : 'FPC_WIDESTR_CHECKZERO']; compilerproc; begin if p=nil then HandleErrorFrame(201,get_frame); end; Procedure fpc_WideStr_CheckRange(len,index : SizeInt);[Public,Alias : 'FPC_WIDESTR_RANGECHECK']; compilerproc; begin if (index>len div 2) or (Index<1) then HandleErrorFrame(201,get_frame); end; Procedure fpc_WideStr_SetLength(Var S : WideString; l : SizeInt);[Public,Alias : 'FPC_WIDESTR_SETLENGTH']; compilerproc; { Sets The length of string S to L. Makes sure S is unique, and contains enough room. } Var Temp : Pointer; movelen: SizeInt; begin if (l>0) then begin if Pointer(S)=nil then begin { Need a complete new string...} Pointer(s):=NewWideString(l); end { windows doesn't support reallocing widestrings, this code is anyways subject to be removed because widestrings shouldn't be ref. counted anymore (FK) } else if {$ifdef MSWINDOWS} not winwidestringalloc and {$endif MSWINDOWS} not IsWideStringConstant(pointer(S)) then begin Dec(Pointer(S),WideFirstOff); if SizeUInt(L*sizeof(WideChar)+WideRecLen)>MemSize(Pointer(S)) then reallocmem(pointer(S), L*sizeof(WideChar)+WideRecLen); Inc(Pointer(S), WideFirstOff); end else begin { Reallocation is needed... } Temp:=Pointer(NewWideString(L)); if Length(S)>0 then begin if l < succ(length(s)) then movelen := l { also move terminating null } else movelen := succ(length(s)); Move(Pointer(S)^,Temp^,movelen * Sizeof(WideChar)); end; fpc_widestr_decr_ref(Pointer(S)); Pointer(S):=Temp; end; { Force nil termination in case it gets shorter } PWord(Pointer(S)+l*sizeof(WideChar))^:=0; {$ifdef MSWINDOWS} if not winwidestringalloc then {$endif MSWINDOWS} PWideRec(Pointer(S)-WideFirstOff)^.Len:=l*sizeof(WideChar); end else begin { Length=0 } if Pointer(S)<>nil then fpc_widestr_decr_ref (Pointer(S)); Pointer(S):=Nil; end; end; {***************************************************************************** Public functions, In interface. *****************************************************************************} Function fpc_widestr_Unique(Var S : Pointer): Pointer; [Public,Alias : 'FPC_WIDESTR_UNIQUE']; compilerproc; begin pointer(result) := pointer(s); end; Function Fpc_WideStr_Copy (Const S : WideString; Index,Size : SizeInt) : WideString;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(NewWideString (Size)); if ResultAddress<>Nil then begin Move (PWideChar(S)[Index],ResultAddress^,Size*sizeof(WideChar)); PWideRec(ResultAddress-WideFirstOff)^.Len:=Size*sizeof(WideChar); PWideChar(ResultAddress+Size*sizeof(WideChar))^:=#0; end; end; fpc_widestr_decr_ref(Pointer(fpc_widestr_copy)); Pointer(fpc_widestr_Copy):=ResultAddress; end; Function Pos (Const Substr : WideString; Const Source : WideString) : SizeInt; var i,MaxLen : SizeInt; pc : pwidechar; 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 (CompareWord(Substr[1],pc^,Length(SubStr))=0) then begin Pos:=i; exit; end; inc(pc); end; end; end; { Faster version for a widechar alone } Function Pos (c : WideChar; Const s : WideString) : SizeInt; var i: SizeInt; pc : pwidechar; 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; Function Pos (c : WideChar; Const s : AnsiString) : SizeInt; begin result:=Pos(c,WideString(s)); end; Function Pos (c : AnsiString; Const s : WideString) : SizeInt;{$ifdef SYSTEMINLINE}inline;{$endif} begin result:=Pos(WideString(c),s); end; Function Pos (c : ShortString; Const s : WideString) : SizeInt;{$ifdef SYSTEMINLINE}inline;{$endif} begin result:=Pos(WideString(c),s); end; Function Pos (c : WideString; Const s : AnsiString) : SizeInt;{$ifdef SYSTEMINLINE}inline;{$endif} begin result:=Pos(c,WideString(s)); 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 $h+ (JM) } Function Pos (c : Char; Const s : WideString) : SizeInt; var i: SizeInt; wc : widechar; pc : pwidechar; begin wc:=c; pc:=@s[1]; for i:=1 to length(s) do begin if pc^=wc then begin pos:=i; exit; end; inc(pc); end; pos:=0; end; Procedure Delete (Var S : WideString; Index,Size: SizeInt); Var LS : SizeInt; begin If Length(S)=0 then exit; if index<=0 then exit; LS:=PWideRec(Pointer(S)-WideFirstOff)^.Len div sizeof(WideChar); if (Index<=LS) and (Size>0) then begin UniqueString (S); if Size+Index>LS then Size:=LS-Index+1; if Index+Size<=LS then begin Dec(Index); Move(PWideChar(S)[Index+Size],PWideChar(S)[Index],(LS-Index-Size+1)*sizeof(WideChar)); end; Setlength(s,LS-Size); end; end; Procedure Insert (Const Source : WideString; Var S : WideString; Index : SizeInt); var Temp : WideString; LS : SizeInt; begin If Length(Source)=0 then exit; if index <= 0 then index := 1; Ls:=Length(S); if index > LS then index := LS+1; Dec(Index); Pointer(Temp) := NewWideString(Length(Source)+LS); SetLength(Temp,Length(Source)+LS); If Index>0 then move (PWideChar(S)^,PWideChar(Temp)^,Index*sizeof(WideChar)); Move (PWideChar(Source)^,PWideChar(Temp)[Index],Length(Source)*sizeof(WideChar)); If (LS-Index)>0 then Move(PWideChar(S)[Index],PWideChar(temp)[Length(Source)+index],(LS-Index)*sizeof(WideChar)); S:=Temp; end; function UpCase(const s : WideString) : WideString; begin result:=widestringmanager.UpperWideStringProc(s); end; Procedure SetString (Out S : WideString; Buf : PWideChar; Len : SizeInt); begin SetLength(S,Len); If (Buf<>Nil) and (Len>0) then Move (Buf[0],S[1],Len*sizeof(WideChar)); end; Procedure SetString (Out S : WideString; Buf : PChar; Len : SizeInt); begin SetLength(S,Len); If (Buf<>Nil) and (Len>0) then widestringmanager.Ansi2WideMoveProc(Buf,S,Len); end; {$ifndef FPUNONE} Function fpc_Val_Real_WideStr(Const S : WideString; out Code : ValSInt): ValReal; [public, alias:'FPC_VAL_REAL_WIDESTR']; compilerproc; Var SS : String; begin fpc_Val_Real_WideStr := 0; if length(S) > 255 then code := 256 else begin SS := S; Val(SS,fpc_Val_Real_WideStr,code); end; end; {$endif} function fpc_val_enum_widestr(str2ordindex:pointer;const s:widestring;out code:valsint):longint;compilerproc; var ss:shortstring; begin if length(s)>255 then code:=256 else begin ss:=s; val(ss,fpc_val_enum_widestr,code); end; end; Function fpc_Val_Currency_WideStr(Const S : WideString; out Code : ValSInt): Currency; [public, alias:'FPC_VAL_CURRENCY_WIDESTR']; compilerproc; Var SS : String; begin if length(S) > 255 then begin fpc_Val_Currency_WideStr:=0; code := 256; end else begin SS := S; Val(SS,fpc_Val_Currency_WideStr,code); end; end; Function fpc_Val_UInt_WideStr (Const S : WideString; out Code : ValSInt): ValUInt; [public, alias:'FPC_VAL_UINT_WIDESTR']; compilerproc; Var SS : ShortString; begin fpc_Val_UInt_WideStr := 0; if length(S) > 255 then code := 256 else begin SS := S; Val(SS,fpc_Val_UInt_WideStr,code); end; end; Function fpc_Val_SInt_WideStr (DestSize: SizeInt; Const S : WideString; out Code : ValSInt): ValSInt; [public, alias:'FPC_VAL_SINT_WIDESTR']; compilerproc; Var SS : ShortString; begin fpc_Val_SInt_WideStr:=0; if length(S)>255 then code:=256 else begin SS := S; fpc_Val_SInt_WideStr := int_Val_SInt_ShortStr(DestSize,SS,Code); end; end; {$ifndef CPU64} Function fpc_Val_qword_WideStr (Const S : WideString; out Code : ValSInt): qword; [public, alias:'FPC_VAL_QWORD_WIDESTR']; compilerproc; Var SS : ShortString; begin fpc_Val_qword_WideStr:=0; if length(S)>255 then code:=256 else begin SS := S; Val(SS,fpc_Val_qword_WideStr,Code); end; end; Function fpc_Val_int64_WideStr (Const S : WideString; out Code : ValSInt): Int64; [public, alias:'FPC_VAL_INT64_WIDESTR']; compilerproc; Var SS : ShortString; begin fpc_Val_int64_WideStr:=0; if length(S)>255 then code:=256 else begin SS := S; Val(SS,fpc_Val_int64_WideStr,Code); end; end; {$endif CPU64} {$ifndef FPUNONE} procedure fpc_WideStr_Float(d : ValReal;len,fr,rt : SizeInt;out s : WideString);compilerproc; var ss : shortstring; begin str_real(len,fr,d,treal_type(rt),ss); s:=ss; end; {$endif} procedure fpc_widestr_enum(ordinal,len:sizeint;typinfo,ord2strindex:pointer;out s:widestring);compilerproc; var ss:shortstring; begin fpc_shortstr_enum(ordinal,len,typinfo,ord2strindex,ss); s:=ss; end; {$ifdef FPC_HAS_STR_CURRENCY} procedure fpc_WideStr_Currency(c : Currency;len,fr : SizeInt;out s : WideString);compilerproc; var ss : shortstring; begin str(c:len:fr,ss); s:=ss; end; {$endif FPC_HAS_STR_CURRENCY} Procedure fpc_WideStr_SInt(v : ValSint; Len : SizeInt; out S : WideString);compilerproc; Var SS : ShortString; begin Str (v:Len,SS); S:=SS; end; Procedure fpc_WideStr_UInt(v : ValUInt;Len : SizeInt; out S : WideString);compilerproc; Var SS : ShortString; begin str(v:Len,SS); S:=SS; end; {$ifndef CPU64} Procedure fpc_WideStr_Int64(v : Int64; Len : SizeInt; out S : WideString);compilerproc; Var SS : ShortString; begin Str (v:Len,SS); S:=SS; end; Procedure fpc_WideStr_Qword(v : Qword;Len : SizeInt; out S : WideString);compilerproc; Var SS : ShortString; begin str(v:Len,SS); S:=SS; end; {$endif CPU64} { converts an utf-16 code point or surrogate pair to utf-32 } function utf16toutf32(const S: WideString; const index: SizeInt; out len: longint): UCS4Char; [public, alias: 'FPC_WIDETOUTF32']; var w: widechar; begin { UTF-16 points in the range #$0-#$D7FF and #$E000-#$FFFF } { are the same in UTF-32 } w:=s[index]; if (w<=#$d7ff) or (w>=#$e000) then begin result:=UCS4Char(w); len:=1; end { valid surrogate pair? } else if (w<=#$dbff) and { w>=#$d7ff check not needed, checked above } (index=#$dc00) and (s[index+1]<=#$dfff) then { convert the surrogate pair to UTF-32 } begin result:=(UCS4Char(w)-$d800) shl 10 + (UCS4Char(s[index+1])-$dc00) + $10000; len:=2; end else { invalid surrogate -> do nothing } begin result:=UCS4Char(w); len:=1; end; end; function UnicodeToUtf8(Dest: PChar; Source: PWideChar; MaxBytes: SizeInt): SizeInt;{$ifdef SYSTEMINLINE}inline;{$endif} begin if assigned(Source) then Result:=UnicodeToUtf8(Dest,MaxBytes,Source,IndexWord(Source^,-1,0)) else Result:=0; end; function UnicodeToUtf8(Dest: PChar; MaxDestBytes: SizeUInt; Source: PWideChar; SourceChars: SizeUInt): SizeUInt; var i,j : SizeUInt; w : word; lw : longword; len : longint; begin result:=0; if source=nil then exit; i:=0; j:=0; if assigned(Dest) then begin while (i=MaxDestBytes then break; Dest[j]:=char($c0 or (w shr 6)); Dest[j+1]:=char($80 or (w and $3f)); inc(j,2); end; $800..$d7ff,$e000..$ffff: begin if j+2>=MaxDestBytes then break; Dest[j]:=char($e0 or (w shr 12)); Dest[j+1]:=char($80 or ((w shr 6) and $3f)); Dest[j+2]:=char($80 or (w and $3f)); inc(j,3); end; $d800..$dbff: {High Surrogates} begin if j+3>=MaxDestBytes then break; if (i= $dc00) and (word(Source[i+1]) <= $dfff) then begin lw:=longword(utf16toutf32(Source[i] + Source[i+1], 1, len)); Dest[j]:=char($f0 or (lw shr 18)); Dest[j+1]:=char($80 or ((lw shr 12) and $3f)); Dest[j+2]:=char($80 or ((lw shr 6) and $3f)); Dest[j+3]:=char($80 or (lw and $3f)); inc(j,4); inc(i); end; end; end; inc(i); end; if j>SizeUInt(MaxDestBytes-1) then j:=MaxDestBytes-1; Dest[j]:=#0; end else begin while i= $dc00) and (word(Source[i+1]) <= $dfff) then begin inc(j,4); inc(i); end; end; end; inc(i); end; end; result:=j+1; end; function Utf8ToUnicode(Dest: PWideChar; Source: PChar; MaxChars: SizeInt): SizeInt;{$ifdef SYSTEMINLINE}inline;{$endif} begin if assigned(Source) then Result:=Utf8ToUnicode(Dest,MaxChars,Source,strlen(Source)) else Result:=0; end; function UTF8ToUnicode(Dest: PWideChar; MaxDestChars: SizeUInt; Source: PChar; SourceBytes: SizeUInt): SizeUInt; const UNICODE_INVALID=63; var InputUTF8: SizeUInt; IBYTE: BYTE; OutputUnicode: SizeUInt; PRECHAR: SizeUInt; TempBYTE: BYTE; CharLen: SizeUint; LookAhead: SizeUInt; UC: SizeUInt; begin if not assigned(Source) then begin result:=0; exit; end; result:=SizeUInt(-1); InputUTF8:=0; OutputUnicode:=0; PreChar:=0; if Assigned(Dest) Then begin while (OutputUnicode13) and FALSE then begin //Expand to crlf, conform UTF-8. //This procedure will break the memory alocation by //FPC for the widestring, so never use it. Condition never true due the "and FALSE". if OutputUnicode+10 do begin TempBYTE:=(TempBYTE shl 1) and $FE; inc(CharLen); end; //Test for the "CharLen" conforms UTF-8 string //This means the 10xxxxxx pattern. if SizeUInt(InputUTF8+CharLen-1)>SourceBytes then begin //Insuficient chars in string to decode //UTF-8 array. Fallback to single char. CharLen:= 1; end; for LookAhead := 1 to CharLen-1 do begin if ((byte(Source[InputUTF8+LookAhead]) and $80)<>$80) or ((byte(Source[InputUTF8+LookAhead]) and $40)<>$00) then begin //Invalid UTF-8 sequence, fallback. CharLen:= LookAhead; break; end; end; UC:=$FFFF; case CharLen of 1: begin //Not valid UTF-8 sequence UC:=UNICODE_INVALID; end; 2: begin //Two bytes UTF, convert it UC:=(byte(Source[InputUTF8]) and $1F) shl 6; UC:=UC or (byte(Source[InputUTF8+1]) and $3F); if UC <= $7F then begin //Invalid UTF sequence. UC:=UNICODE_INVALID; end; end; 3: begin //Three bytes, convert it to unicode UC:= (byte(Source[InputUTF8]) and $0F) shl 12; UC:= UC or ((byte(Source[InputUTF8+1]) and $3F) shl 6); UC:= UC or ((byte(Source[InputUTF8+2]) and $3F)); if (UC <= $7FF) or (UC >= $FFFE) or ((UC >= $D800) and (UC <= $DFFF)) then begin //Invalid UTF-8 sequence UC:= UNICODE_INVALID; End; end; 4: begin //Four bytes, convert it to two unicode characters UC:= (byte(Source[InputUTF8]) and $07) shl 18; UC:= UC or ((byte(Source[InputUTF8+1]) and $3F) shl 12); UC:= UC or ((byte(Source[InputUTF8+2]) and $3F) shl 6); UC:= UC or ((byte(Source[InputUTF8+3]) and $3F)); if (UC < $10000) or (UC > $10FFFF) then begin UC:= UNICODE_INVALID; end else begin { only store pair if room } dec(UC,$10000); if (OutputUnicode 0 then begin PreChar:=UC; Dest[OutputUnicode]:=WideChar(UC); inc(OutputUnicode); end; InputUTF8:= InputUTF8 + CharLen; end; end; Result:=OutputUnicode+1; end else begin while (InputUTF813) and FALSE then begin //Expand to crlf, conform UTF-8. //This procedure will break the memory alocation by //FPC for the widestring, so never use it. Condition never true due the "and FALSE". inc(OutputUnicode,2); PreChar:=10; end else begin inc(OutputUnicode); PreChar:=IBYTE; end; end else begin inc(OutputUnicode); PreChar:=IBYTE; end; inc(InputUTF8); end else begin TempByte:=IBYTE; CharLen:=0; while (TempBYTE and $80)<>0 do begin TempBYTE:=(TempBYTE shl 1) and $FE; inc(CharLen); end; //Test for the "CharLen" conforms UTF-8 string //This means the 10xxxxxx pattern. if SizeUInt(InputUTF8+CharLen-1)>SourceBytes then begin //Insuficient chars in string to decode //UTF-8 array. Fallback to single char. CharLen:= 1; end; for LookAhead := 1 to CharLen-1 do begin if ((byte(Source[InputUTF8+LookAhead]) and $80)<>$80) or ((byte(Source[InputUTF8+LookAhead]) and $40)<>$00) then begin //Invalid UTF-8 sequence, fallback. CharLen:= LookAhead; break; end; end; UC:=$FFFF; case CharLen of 1: begin //Not valid UTF-8 sequence UC:=UNICODE_INVALID; end; 2: begin //Two bytes UTF, convert it UC:=(byte(Source[InputUTF8]) and $1F) shl 6; UC:=UC or (byte(Source[InputUTF8+1]) and $3F); if UC <= $7F then begin //Invalid UTF sequence. UC:=UNICODE_INVALID; end; end; 3: begin //Three bytes, convert it to unicode UC:= (byte(Source[InputUTF8]) and $0F) shl 12; UC:= UC or ((byte(Source[InputUTF8+1]) and $3F) shl 6); UC:= UC or ((byte(Source[InputUTF8+2]) and $3F)); If (UC <= $7FF) or (UC >= $FFFE) or ((UC >= $D800) and (UC <= $DFFF)) then begin //Invalid UTF-8 sequence UC:= UNICODE_INVALID; end; end; 4: begin //Four bytes, convert it to two unicode characters UC:= (byte(Source[InputUTF8]) and $07) shl 18; UC:= UC or ((byte(Source[InputUTF8+1]) and $3F) shl 12); UC:= UC or ((byte(Source[InputUTF8+2]) and $3F) shl 6); UC:= UC or ((byte(Source[InputUTF8+3]) and $3F)); if (UC < $10000) or (UC > $10FFFF) then UC:= UNICODE_INVALID else { extra character character } inc(OutputUnicode); end; 5,6,7: begin //Invalid UTF8 to unicode conversion, //mask it as invalid UNICODE too. UC:=UNICODE_INVALID; end; end; if CharLen > 0 then begin PreChar:=UC; inc(OutputUnicode); end; InputUTF8:= InputUTF8 + CharLen; end; end; Result:=OutputUnicode+1; end; end; function UTF8Encode(const s : WideString) : UTF8String; var i : SizeInt; hs : UTF8String; begin result:=''; if s='' then exit; SetLength(hs,length(s)*3); i:=UnicodeToUtf8(pchar(hs),length(hs)+1,PWideChar(s),length(s)); if i>0 then begin SetLength(hs,i-1); result:=hs; end; end; const SNoWidestrings = 'This binary has no widestrings support compiled in.'; SRecompileWithWidestrings = 'Recompile the application with a widestrings-manager in the program uses clause.'; procedure unimplementedwidestring; begin {$ifdef FPC_HAS_FEATURE_CONSOLEIO} If IsConsole then begin Writeln(StdErr,SNoWidestrings); Writeln(StdErr,SRecompileWithWidestrings); end; {$endif FPC_HAS_FEATURE_CONSOLEIO} HandleErrorFrame(233,get_frame); end; {$warnings off} function GenericWideCase(const s : WideString) : WideString; begin unimplementedwidestring; end; function CompareWideString(const s1, s2 : WideString) : PtrInt; begin unimplementedwidestring; end; function CompareTextWideString(const s1, s2 : WideString): PtrInt; begin unimplementedwidestring; end; {$warnings on} function CharLengthPChar(const Str: PChar): PtrInt;forward; procedure initwidestringmanager; begin fillchar(widestringmanager,sizeof(widestringmanager),0); {$ifndef HAS_WIDESTRINGMANAGER} widestringmanager.Wide2AnsiMoveProc:=@defaultWide2AnsiMove; widestringmanager.Ansi2WideMoveProc:=@defaultAnsi2WideMove; widestringmanager.UpperWideStringProc:=@GenericWideCase; widestringmanager.LowerWideStringProc:=@GenericWideCase; {$endif HAS_WIDESTRINGMANAGER} widestringmanager.CompareWideStringProc:=@CompareWideString; widestringmanager.CompareTextWideStringProc:=@CompareTextWideString; widestringmanager.CharLengthPCharProc:=@CharLengthPChar; end;