Integer() function depends on SetFloatDelta()

[cpyrgas]

As seen below, the result of Integer() function depends on the SetFloatDelta() setting :

FUNCTION Start() AS VOID
	SetFloatDelta(0.0)
	? "Float delta 0:"
	? Integer(0.1)   // 0, OK
	? Integer(0.001) // 0, OK

	SetFloatDelta(0.01)
	
	? "Float delta 0.01:"
	? Integer(0.1)   // 0, OK
	? Integer(0.001) // 1, wrong, 0 in VO

Problem is in this code of Integer() function:

FUNCTION Integer(nValue AS USUAL) AS USUAL
...
        IF nValue > 0.0 // this comparison depends on SetFloatDelta()
            result := Floor(nValue)
        ELSE
            result := Ceil(nValue)
        ENDIF

This can be fixed by casting the number to a REAL8 for the comparison, but we need to search for other similar occasions in the runtime.

[RobertvanderHulst]

When nValue is of type LONG or INT64 you do not want to compare with a fractional value at all.
I think the function should:

• have a check for the type: LONG and INT64 should simply return the value
• float should cast the value to a float and retrieve the Value part of the float to compare to the real8 value 0.0
• Currency should compare with $0.0 and use a Floor() and Ceil() for Currency values