Open aflinton opened 2 years ago
Obviously I can copy the various "original" jars into the zip but it would possibly be easier to include the various schematypes in a single large shaded jar as then a user knows all the dependencies will be there.
Thanks for creating this issue. It could certainly be useful to create a combined shaded jar. We would have to change the module hierarchy or do some maven magic here. Can I also understand more on why using two shaded jars together does not work for you ?
Just size really. Each shaded jar is about 50MB & contains almost exactly the same thing with a few classes missing at the specific converter level. If I include the "original" jars from the other converters I get what I am looking for but I simply think that it would be helpful to have an all in one. The basic use case is that we can start with JSON with a schema until we've got what we need then empty / rebuild the queue and switch to avro for speed etc.
Please upvote if this feature is important. We are also open to accepting PR if many users feel this feature is important.
At the moment large shaded jars are created for a given schema registry type (e.g. avro (schema-registry-kafkaconnect-converter) or jsonschema (jsonschema-kafkaconnect-converter)). Is it possible/ could it be possible just to have all the options in one shaded jar? Otherwise I am having unpack zips for upload to S3 for use with MSK, change the avro jar for the jsonschema jar etc.