In 0.20.0, `NamedValueEncoder`/`Decoder` are marked
@InternalSerializationApi
. Will that remain the case in 1.0? If so, what should custom formats be using instead?
v
Vsevolod Tolstopyatov [JB]
03/06/2020, 8:48 AM
`AbstractDecoder`/`AbstractEncoder` will remain and can be used.
The problem with
Named*
(and
Tagged*
) is that they are way too specific and were never designed in a way that implies 3-rd party usages, inheritance compatibility and proper API.
Vsevolod Tolstopyatov [JB]
03/06/2020, 8:49 AM
It would be awesome to know why exactly you need
Named*
r
russhwolf
03/06/2020, 1:20 PM
Use-case is an integration I've been prototyping with Multiplatform Settings. Code at https://github.com/russhwolf/multiplatform-settings/pull/36 if you have suggestions. But in general named encoding seems like a pretty natural use-case to support here for common formats like json, xml, properties, etc.