this is pretty standard Jackson behaviour when

Question

< Tom Ellis> this is pretty standard Jackson behaviour when in doubt it does deserialise using Maps we see the same with generic lists It s nothing to do with http4k per se but about how you set up the Jackson mapper instance and annotate the classes for deserialisation with something like ` JsonTypeInfo` TBH I generally try to make DTO objects as dumb as possible as I hate all the annotations

Tom Ellis · Answer

hmm the thing is if i just do this it s fine ``` val jacksonResult = jacksonObjectMapper readValue lt Result lt MyClass gt gt response bodyString println jacksonResult body message ``` so i have a workaround it would just be nice to use the lens

dave · Answer

Interesting we re using convertValue instead of readValue Not sure of the difference Will have to investigate

Riku · Answer

I don t know if this is related at all but I also found cases when default mapper didn t work for mysterious reasons or reasons I never looked into but using our mapper always worked fine

Riku · Answer

```val mapper = KotlinModule asConfigurable withStandardMappings done registerModule JodaModule disableDefaultTyping configure DeserializationFeature FAIL ON UNKNOWN PROPERTIES false configure DeserializationFeature FAIL ON IGNORED PROPERTIES true configure DeserializationFeature USE BIG DECIMAL FOR FLOATS true configure DeserializationFeature USE BIG INTEGER FOR INTS true configure com fasterxml jackson databind SerializationFeature WRITE DATES AS TIMESTAMPS true ```

Riku · Answer

actually I now realize this is nothing to do with your case

dave · Answer

< Riku> what types didn t work for you with the default version

Riku · Answer

I can t remember off hand there were two cases

Riku · Answer

```data class EventInDto val name String val sponsor String val contact String val date Timestamp val details String val location String val status EventStatus = EventStatus PLANNING val maxParticipants Int val registrationOpens Timestamp val volume Int val sponsorLink String ````

Riku · Answer

that was one

dave · Answer

is that a SQL Timestamp

Riku · Answer

yeah that is my blurry memory

Riku · Answer

might run it tonight when at home just to check

dave · Answer

cool we can probably add that to the standard mappings

dave · Answer

let me know if we need to slightly smiling face

Riku · Answer

ah hey we are using readValue now too

Riku · Answer

because we validate the json we are doing ` Body string ContentType APPLICATION JSON map json gt EventInDto from json toLens `

Riku · Answer

and inside `from` readValue

Riku · Answer

but then else where we do use the mapper directly in the lens

Riku · Answer

and I think there was another spot it also fixed an issue

Riku · Answer

I guess that use of the lens is kind of hacky

dave · Answer

i ll fix it up to use readValue and also to support Timestamp then you can remove the hack

dave · Answer

and in future please let us know if you think you need to hack around stuff I am figuratively dying of professional embarrassment here wink

dave · Answer

< Riku> presumably timestamps are transmitted as longs over the wire or strings

dave · Answer

hmm timestamp as a number actually seems a little dodgy think we ll release with the other fix and look at timestamp later slightly smiling face

Riku · Answer

Ours are strings now i wonder if we even typed them correctly laughing Well if it aint broke