Any class definition may be annotated with @EqualsAndHashCode
to let lombok generate implementations of the
equals(Object other)
and hashCode()
methods. By default, it'll use all non-static, non-transient
fields, but you can exclude more fields by naming them in the optional exclude
parameter to the annotation.
By setting callSuper
to true, you can include the equals
and hashCode
methods of your superclass in the generated methods.
For hashCode
, the result of super.hashCode()
is included in the hash algorithm, and for equals
, the generated method will return
false if the super implementation thinks it is not equal to the passed in object. Be aware that not all equals
implementations handle this situation properly.
However, lombok-generated equals
implementations do handle this situation properly, so you can safely call your superclass equals if it, too,
has a lombok-generated equals
method.
Setting callSuper
to true when you don't extend anything (you extend java.lang.Object
) is a compile-time error, because it would turn
the generated equals()
and hashCode()
implementations into having the same behaviour as simply inheriting these methods from java.lang.Object
:
only the same object will be equal to each other and will have the same hashCode. Obviously, inheriting java.lang.Object
is the right strategy if you want this behaviour.
Not setting callSuper
to true when you extend another class generates a warning, because unless the superclass has no (equality-important) fields, lombok
cannot generate an implementation for you that takes into account the fields declared by your superclasses. You'll need to write your own implementations, or rely on the
callSuper
chaining facility.
Arrays are 'deep' compared/hashCoded, which means that arrays that contain themselves will result in StackOverflowError
s. However,
this behaviour is no different from e.g. ArrayList
.
You may safely presume that the hashCode implementation used will not change between versions of lombok, however this guarantee is not set in stone; if there's a significant performance improvement to be gained from using an alternate hash algorithm, that will be substituted in a future version.
For the purposes of equality, 2 NaN
(not a number) values for floats and doubles are considered equal, eventhough 'NaN == NaN' would
return false. This is analogous to java.lang.Double
's equals method, and is in fact required to ensure that comparing an object
to an exact copy of itself returns true
for equality.
If there is any method named either hashCode
or equals
, regardless of parameters or return type,
that method will not be generated, and a warning is emitted instead. hashCode
and equals
need to be in sync with
each other, which lombok cannot guarantee if it is only generating one of the two methods, hence you always get a warning if one or both
of the methods already exist.
Attempting to exclude fields that don't exist or would have been excluded anyway (because they are static) results in warnings on the named fields. You therefore don't have to worry about typos.