would it be cleaner to use a json config, as we have a json-c dependency already?
JSON configs are rather difficult to edit (lots of quoting, no comments), which is why yaml has taken over for configs. I think either structured format is overkill for the flat "string attribute = string value" configuration supported by librdkafka.
JSON configs are rather difficult to edit (lots of quoting, no comments), which is why yaml has taken over for configs. I think either structured format is overkill for the flat "string attribute = string value" configuration supported by librdkafka.