question about assignments...

[Andres Gonzalez]

class A {
  transient float e = 1;         // 1
  volatile float f  = 1;         // 2
  final float g = 1;             // 4
  private final float k = 1;     // 5
  private transient float l = 1; // 6
}

the above code compiles fine... but I've got a question.. Floating point literals are defined as double, so my first guess was that this code didn't compile because it tried to fit a double value into a float.
Now, if modify line 1:

  transient float e = 1.0;         // 1

it raises the compiler error I was expecting. so, is line 1 of the initial code treated as an int? not only for line 1 but for all other lines?
[ June 23, 2003: Message edited by: Andres Gonzalez ]

[Valentin Crettaz]

Floating point literals are defined as double
This means that 1.0 will always be a double value and not a float value. In order to specify a float, you must append 'f' or 'F' after the value, as in 1.0f or 1.0F
Actually the int value is widened to a float value as per JLS 5.1.2 Widening Primitive Conversion

[Miki Muzsi]

"1" is a valid IntegerLiteral (see eventually JLS-2.0 chapters: 5.18.2 and 3.10.1 for more info). So the value corresponding top the "1" literal is an int type (having value), which is implicitly converted to a float.
So:
float f = 1; // valid, "1" is an IntegerLiteral
float f = 1.0; // INVALID, "1.0" is a FloatingPointLiteral, that does not end with 'f' nor 'F'