iosswiftrangedoublecgfloat

Is CGFloat more accurate than Double?

Double & CGFloat both print the same value to the console (#2), but only the Double is issued the warning (#1):

(#1) - The warning

the warning

(#2) - Printing each to the console

DOUBLE  = 9.223372036854776e+18
CGFLOAT = 9.223372036854776e+18

Many articles mention that a CGFloat is a Float on 32-bit platforms & a Double on a 64-bit platform, but doesn't that only refer to the backing storage? Is there more going on behind the scenes which causes the CGFloat to be more accurate than the Double?

Solution

  • No, CGFloat does not represent a BigDecimal-type number, with infinite precision. CGFloat is defined as a struct with either wraps a Double or a Float, depending on whether you are on a 32-bit or 64-bit platform; i.e., on a 64-bit platform, it is equivalent to a Double, with no more or less precision or range.

    The difference in the warning you see is that Double is a type native to Swift, which the compiler knows about, and has enough information about in order to be able to provide such a warning. CGFloat, however, is defined by Apple frameworks, and the compiler does not know enough about it to produce such a warning.