add discriminatorMapping when deriving a schema for co-products #1838
Conversation
|
Could you make a PR against |
|
|
||
| final case class Configuration(toEncodedName: String => String, discriminator: Option[String]) { | ||
| final case class Configuration(toEncodedName: String => String, discriminator: Option[String], toEncodedSubtypeName: TypeName => String) { |
There was a problem hiding this comment.
I think here we'd need to avoid depending on magnolia's class, as this is kind of an implementation detail. Maybe it would make sense to use SName => String here?
There was a problem hiding this comment.
Definitely! Also this is probably the reason CI failed here.
| @@ -244,7 +244,15 @@ class SchemaGenericAutoTest extends AsyncFlatSpec with Matchers { | |||
| ) | |||
| ) | |||
|
|
|||
| schemaType.asInstanceOf[SCoproduct[Entity]].discriminator shouldBe Some(SDiscriminator(FieldName("who_am_i"), Map.empty)) | |||
There was a problem hiding this comment.
would be great to add a test with non-identity subtype name transformations
Sure! (I was secretly hoping for this to make into 0.19.x :) but we can wait for 0.20.x) |
|
@yurique since this is binary incompatible, 0.20 is a better target :) |
Totally makes sense :) |
yurique
force-pushed
the
feature/discriminator-mapping-for-coproducts
branch
from
833f43f to
0a8962e
Compare
|
This now targets
|
|
Not sure why CI is not kicking in. |
|
Thanks! Two more things:
|
|
(no idea why CI didn't run) |
* rename toEncodedSubtypeName -> toDiscriminatorValue
I like it. Done.
Done. btw, what's the role of |
|
Thanks! :) |
|
SchemaMacros are version-specific extensions for |
Without
discriminatorMappingthe OpenAPI generator can't make sense of the schema and fails.I'm not sure if adding fields to
Configurationwas a right call – this breaks bin-compat, but I can re-work things here to re-usetoEncodedName(and deal with the$suffix on the call site).