Casts have higher precedence than binary expressions, so (int) 0.5f * 100
is
equivalent to ((int) 0.5f) * 100
= 0 * 100
= 0
, not (int) (0.5f * 100)
=
50
.
To avoid this common source of error, add explicit parentheses to make the precedence explicit. For example, instead of this:
long SIZE_IN_GB = (long) 1.5 * 1024 * 1024 * 1024; // this is 1GB, not 1.5GB!
// this is 0, not a long in the range [0, 1000000000]
long rand = (long) new Random().nextDouble() * 1000000000
Prefer:
long SIZE_IN_GB = (long) (1.5 * 1024 * 1024 * 1024);
long rand = (long) (new Random().nextDouble() * 1000000000);
Suppress false positives by adding the suppression annotation @SuppressWarnings("FloatCast")
to the enclosing element.