Inconsistent behavior of JSON decoders for Schema.CaseClass* and Schema.GenericRecord

[plokhotnyuk]

Here is a table of differences or missing feature for both types of schemas, where + and - are used to mark support or missing support accordingly:

Issue	Schema.CaseClass*	Schema.GenericRecord	Fix
Using of default values for "transient" (skip on write) fields	+	-	#764
Detection of duplicated field/alias names	+	-	#765
Use field name in an error for duplicated fields	-	-	#768
Detection of missing fields	+	-
Use field/alias name in an optional error for extra fields	-	+
Using empty collection values for parsed JSON null	+	-
Using None values for parsed JSON null	+	-
Handling of different discriminator types	+	-