Question

I am tidying old code that used to use FastStrings and I’ve implemented an old routine of mine ‘PosAnyCase’ which should operate like ‘Pos’. (I was hoping that SearchBuf was better than calling UpperCase on both strings).

function PosAnyCase( const AFindStr, AStr : string ) : integer;
// Returns the position of this substring within a string ignoring case

I’m using SearchBuf as follows:

function PosAnyCase( const AFindStr, AStr : string ) : integer;
// Returns the position of this substring within a string ignoring case
var
  Start, ResultPos : PChar;
begin
  Start := PChar( AStr );

  ResultPos := SearchBuf(
    Start, ByteLength( AStr ),
    0, 0,
    AFindStr, [soDown] );

  if ResultPos = nil then
    Result := 0
   else
    Result := ResultPos-Start+1;
end;

When I call this routine from my unit tests, the following tests PASS:

  Check(
    PosAnyCase( '', '123' ) = 0 );
  Check(
    PosAnyCase( '2', '123' ) = 2 );
  Check(
    PosAnyCase( 'A', 'ABC' ) = 1 );
  Check(
    PosAnyCase( 'a', 'ABC' ) = 1 );
  Check(
    PosAnyCase( 'the', 'hellot there' ) = 8 );
  Check(
    PosAnyCase( 'THE', 'hellot there' ) = 8 );

But this test FAILS:

Check(
    PosAnyCase( 'nice', 'does not have n i c e' ) = 0 );

What am I doing wrong please? The documentation on SearchBuf is very limited….
Thanks

Answer 1

The call to ByteLength is incorrect. Although the documentation explicitly states that the parameter is the length in bytes, that is not the case. You should use Length instead because the function actually expects units of char rather than units of byte.