Forum Discussion

leonk's avatar
leonk
New Contributor
15 years ago

Testing inequality



I have the following code to test if a reading is within +/-25% of the expected value.



function testWithinLimits()

{

   var expectedVal = 200;

   var ErrorInPercent = 25;

   var UpperLimit = expectedVal/1000 * ((100 + ErrorInPercent)/100);

   var LowerLimit = expectedVal/1000 * ((100 - ErrorInPercent)/100);



   var readVal = "0.15";

   var reading = StrToFloat(readVal);

   bResult1 = (reading <= UpperLimit);

   bResult2 = (reading >= LowerLimit);



   strMsg = FloatToStr(reading) + " <= " + FloatToStr(UpperLimit) + " is : " + bResult1;

   strMsg += "\r\n";

   strMsg += FloatToStr(reading) + " >= " + FloatToStr(LowerLimit) + " is : " + bResult2;  

   

   ShowMessage(strMsg);



}



The expected value is 0.2 so a range from 0.15 to 0.25 (inclusive) is allowed. My reading is 0.15 but the function shows a false result.



Have I missed a type conversion or made a serious error?



I'm using TestComplete v6.52.419.7.



Thanks, Leon.
  • leonk's avatar
    leonk
    New Contributor
    15 years ago


    Stripping some of the extra code away gives me this...



    function theSame()

    {



       var expectedVal = 200;

       var ErrorInPercent = 25;

       var x = expectedVal/1000 *  (100-ErrorInPercent) / 100;

       var y = expectedVal/1000 * ((100-ErrorInPercent)/100);



       ShowMessage((x == y)?"True":"False");   



    }



    I am getting a False result though believe x and y should be equal if I have my operator precedence correct.
  • YMinaev's avatar
    YMinaev
    Staff
    15 years ago
    Hi Leon,



    The same behavior is observed in external JScript scripts (launched from JS files), so it is not related to TC. This can be a problem in the MS JScript engine or just a floating-point computation inaccuracy.