There are a few other dependencies that are probably no longer used
and can be removed. I'm not talking about the netty-related
dependencies, because they seem to be used transitively and required
to be in the classpath, but the others are good candidates, I think.
For example, org.caffinitas.ohc:
>
> I propose we remove commons-codec on trunk.
> The only usage I found was from CASSANDRA-12790 - Support InfluxDb
metrics reporter configuration, which relied on commons-codec and
metrics-reporter-config, which will be removed as part of CASSANDRA-18743.
> The only question is whether we can rem
I would remove it all in 5.0 but that's just me. I do not think that the
deprecation is a must and it is just unnecessary exercise and we are just red
taping here.
Major releases are good for dropping the "baggage" like this. Do we really want
to wait until 6.0 is out to cut off the dead weight
If we're going to do bulk dependency pruning, we should minimize the number of
deprecation plans that users need to prepare for. There will likely be a few
more dependencies we clean up around 5.0, so sticking with 5.0 deprecation and
6.0 removal for all of them would likely make our users' live
