Enhance KotlinSerializer with value codecs for widening primitive conversion. #1301
Conversation
…version. JAVA-5303
vbabanin
requested review from
a team and
rozza
and removed request for
a team
|
Kotlinx Json has a configuration to allow for lenient parsing: https://github.com/Kotlin/kotlinx.serialization/blob/master/docs/json.md#lenient-parsing Should this be configurable? |
|
@rozza I think in this case it's better to be consistent with other codecs and also with the way numeric values are treated in MongoDB is basically interchangeable and compatible with each other, so long as there is no loss of precision or magnitude. |
|
@vbabanin can this come out of Draft now? |
JAVA-5303
| <!-- MongoDB status: "False Positive", SpotBugs rank: 17 --> | ||
| <Class name="org.bson.codecs.kotlinx.KotlinSerializerCodec$Companion"/> | ||
| <Method name="~.*validateAnnotations.*"/> | ||
| <Bug pattern="BC_BAD_CAST_TO_ABSTRACT_COLLECTION"/> |
There was a problem hiding this comment.
The warning is triggered by an implicit cast generated during Kotlin to Java bytecode compilation, which is not directly controllable in the source Kotlin code. As this casting is safely handled by the Kotlin compiler and does not reflect an actual error in our Kotlin source, we can suppress this warning.
| @@ -34,8 +34,18 @@ import org.bson.BsonReader | |||
| import org.bson.BsonReaderMark | |||
| import org.bson.BsonType | |||
| import org.bson.BsonValue | |||
| import org.bson.codecs.BooleanCodec | |||
There was a problem hiding this comment.
I thought we discussed not using codecs (so to not confuse the two concepts of codecs and kserializers. Instead use the NumberCodecHelper directly (which is what the numeric codecs are doing in decode anyway).
JAVA-5303