Firestore Data Mapping Does not work with Groovy Classes

[jjathman]

Describe the bug
It currently seems impossible to use a Groovy class as the object to save to Firestore due to the object mapping attempting to map some of the "hidden" fields that Groovy automatically adds to every class. I don't think there is any available configuration which allows for excluding these fields.

Sample

Data.groovy

@Document
class Data {
  @DocumentId
  String id
}

Attempting to save this document will generate this error:

reactor.core.Exceptions$ErrorCallbackNotImplemented: java.lang.IllegalArgumentException: Could not serialize object. Serializing Arrays is not supported, please use Lists instead (found in field 'metaClass.additionalMetaMethods')
Caused by: java.lang.IllegalArgumentException: Could not serialize object. Serializing Arrays is not supported, please use Lists instead (found in field 'metaClass.additionalMetaMethods')

I think there would be two relatively simple solutions that I would like to use. One would be to configure the mapper to always ignore certain classes MetaClass in this case, or allow the Exclude annotation to be applied to the class level and provide string properties to exclude. This would allow me to exclude the metaClass since I am unable to annotate it.

[elefeint]

Spring Data already has handling for excluding such autogenerated fields, so this is likely a matter of finding where the Firestore implementation missed plugging into the property exclusion process.

[jjathman]

That’s very interesting. Good find! Hopefully that means it’s not too much work to support it in the future.

[elefeint]

Leaving it open as a new feature.

[elefeint]

We've looked into this some more, and the object-to-document functionality lives entirely in the client library. I will close this issue; please watch googleapis/java-firestore#996 for updates.