I'm using the latest version of swagger-parser (2.1.13) and I've noticed an inconsistency in how a byte array schema's example field is read in and stored on the Schema (openapi 3.0.x) or JsonSchema (openapi 3.1.x).
Here's an example of a byte array property within an object schema:
Router:
description: A network router.
type: object
properties:
name:
description: The name of the router.
type: string
firmware:
description: The current firmware on the router.
type: string
format: byte
example: RGVmaW5ldGVseSBub3QgYSBmaXJtd2FyZS4gQ2hlY2sgaHR0cHM6Ly9vcGVud3J0Lm9yZyAhIQ==
Within an OpenAPI 3.0.x document, the firmware property's example field is stored as a byte[],
but if this same schema exists within an OpenAPI 3.1 document, the example field is stored as a String. It seems like these should be consistent. Pick one or the other and use it in each scenario in order to make it easier to write tooling code (i.e. SDK generator) that needs to deal with both versions of OpenAPI.
I think this inconsistency is rooted in the differences between the ByteArraySchema.cast() and JsonSchema.cast() methods.
I'm using the latest version of swagger-parser (2.1.13) and I've noticed an inconsistency in how a byte array schema's example field is read in and stored on the Schema (openapi 3.0.x) or JsonSchema (openapi 3.1.x).
Here's an example of a byte array property within an object schema:
Within an OpenAPI 3.0.x document, the
firmwareproperty's example field is stored as a byte[], but if this same schema exists within an OpenAPI 3.1 document, the example field is stored as a String. It seems like these should be consistent. Pick one or the other and use it in each scenario in order to make it easier to write tooling code (i.e. SDK generator) that needs to deal with both versions of OpenAPI.I think this inconsistency is rooted in the differences between the ByteArraySchema.cast() and JsonSchema.cast() methods.