Support for protobuf base64 encoded descriptors.

[perj]

What is currently missing?

When using confluent-kafka-python as a Schema Registry client, it will upload a binary descriptor, you can see an example of a such here:
https://github.com/confluentinc/confluent-kafka-python/blob/e7e066120f05ca6f9c37b18d77f974f1a6cad222/examples/user_pb2.py#L22
That example will generate this POST request to /subjects/subject/versions:

{"schemaType": "PROTOBUF", "schema": "Cgp1c2VyLnByb3RvIkUKBFVzZXISDAoEbmFtZRgBIAEoCRIXCg9mYXZvcml0ZV9udW1iZXIYAiABKAMSFgoOZmF2b3JpdGVfY29sb3IYAyABKAliBnByb3RvMw=="}

This works when using CSR. When fetching back the subject, a textual representation is returned, a normalized form of https://github.com/confluentinc/confluent-kafka-python/blob/e7e066120f05ca6f9c37b18d77f974f1a6cad222/examples/user.proto

Karapace doesn't support this. It requires the textual represenation to be sent, making it incompatible with conluent-kafka-python.

How could this be improved?

GetDebugString() added here: protocolbuffers/protobuf#7498
returns this normalized form and should be applied server-side instead of having to do that before sending the request .

Is this a feature you would work on yourself?

[ ] I plan to open a pull request for this feature

[tvainika]

Implemented in #755