regarding float and long

[Harivenkatesh Polnati]

float has 4 bytes and long has 8 bytes but i faced a problem
float f = 10L; is valid..
why it is possible?

[Campbell Ritchie]

Have you read the Java Language Specification section which covers this question? It is not easy to read, I am afraid.

[Henry Wong]

Harivenkatesh Polnati wrote:float has 4 bytes and long has 8 bytes but i faced a problem
float f = 10L; is valid..
why it is possible?

To summarize it, basically, when it comes to implicit casting, the JLS is only concerned about range (and not about precision). In the case of float, even though it is smaller in footprint than a long, it has a larger range than long (even though the specification doesn't count infinity and negative infinity).

Of course, with much lower precision, there are lots of numbers that can be represented by the long, that is only "rounded off" by the float.

Henry

[Harivenkatesh Polnati]

thank you for your answers

[Campbell Ritchie]

You’re welcome
That’s what it meant in the JLS link about losing information

float f1 = 10L;
System.out.println(f1); // 10 approx
float f2 = 123456789101112L;
System.out.println(f2); //1.23457e14 approx. Loses 9 digits precision