Issue a warning when final field is overwritten #414 #420
Conversation
|
Turns out we have more than 10 tests which do exactly that - overwrite final fields. What should we do - should we update them all? |
IMO yes we should - our code should set a good example |
|
+1. Yes, please fix any tests which overwrite final fields. |
alamar
force-pushed
the
Issue414
branch
2 times, most recently
from
db92898 to
828b5e4
Compare
De-finalize marshallable fields to remove warnings closes #414
|
Kudos, SonarCloud Quality Gate passed! |
| @@ -641,6 +650,14 @@ protected void readValue(Object o, Object defaults, ValueIn read, boolean overwr | |||
| } | |||
| } | |||
|
|
|||
| protected void triggerFinalWarning() { | |||
There was a problem hiding this comment.
Suggest assertNonFinal(). In this version, there will be a warning but in later, an Exception would be thrown.
| final List<Integer> numbers = new ArrayList<>(); | ||
| final Map<String, String> map1 = new LinkedHashMap<>(); | ||
| final Map<String, Double> map2 = new LinkedHashMap<>(); | ||
| Set<String> words = new LinkedHashSet<>(); |
There was a problem hiding this comment.
Can't mutable objects be final?
|
It's important to recognise how Java Serialization does this currently, as we can't change that behaviour. Java Serialization doesn't produce a warning when setting final fields, nor does it expect you to make a field that can be set non-final |
#414