I perhaps a little naïvely took them at their word:
Pull requests is the primary place of discussing code changes
Great discussion was had there ;-)
It’s fine. It’s their codebase and they have strong opinions about where they want to take it, etc. It’s just funny as for quite some time now I’ve been pondering “merging” policies via the management API, but it always felt too hacky to me. When operator policies came along I wished they had supported HA policies, but assumed it would be beyond me to add that, despite having dabbled in Erlang. But then recently this issue came up again so I decided to have a closer look at the code the end result being this PR which got shot down in flames faster than I ever imagined. So I probably will end up back at the API and doing it the “hacky” way.
I don’t agree with the reasoning though even if I do think it’s ok to say “no”… “it does not necessarily make sense to ‘pick the highest replication factor’” - well, hence the “opening for discussion” and “Initial work…”. Also “On top of that this does not support quorum queues” - again, hence the “opening for discussion” and “Initial work…”, but also, unless I’m grossly misunderstanding Quorum Queues, it’s not even applicable for Quorum Queues since they are replicated anyway.