There’s enough cruft that is blocked on a major version bump that I think we should start discussing how we handle 2.0.
I think it’s important that we treat this as we would any other version. No breaking change we make should have a significant user impact. The only reason a change should happen in 2.0 instead of another version is because for whatever reason, it cannot have a deprecation cycle. The end result should be that updating to Diesel 2.0 should not be any harder than updating to any other version with #[deny(deprecated)].
However, Pascal has rightly pointed out that it’s only been a half year since 1.0 was released, and people will get the wrong impression.
so, I’m personally fine with the major version bump because I know what’s going on and it’s just a number to me
but other people will consider this a sign that the API changes often and that they were right not to use it because 1.0 was just half a year ago…
So I think we need to answer two major questions:
- How long is long enough before we can release 2.0 without folks worrying?
- How do we communicate that the major version bump is because of our commitment to semver, and does not contain substantial breaking changes?
