How to best build custom code based on Managed Beta package

Question

In the next release of our managed application we are going to deliver a big and disruptive data model change.

We need to give customers time to adjust their unmanaged extensions of our package in advance and think about the most flexible (for us and them) solution to do this.

Most flexible for us is to use a Managed Beta as we can change anything until our final release.

For our customers Managed Beta is a bad option as they have to delete all data AND code that references our metadata before they can use a new Beta or the final release.

We came up with another idea and ask for you feedback if that is going to work.

1. We create a Managed Beta 1
2. The customer created a Sandbox A of his Production org
3. The customer installs Managed Beta 1 in Sandbox A
4. The customer develops code and objects based in our package
5. We release Managed Beta 2
6. The customer installs it into Sandbox B (same prod org)
7. The customer moves his code and data extensions via Change set from Sandbox A to Sandbox

It boils down to the question: Is it possible to deploy code that is based on Managed Beta Version 1 (via Change Set) from a Sandbox A to a Sandbox B which has installed Managed Beta Version 2?

Answer 1

The answer depends on if anything from Managed beta package version 1, referenced in changeset wasn't removed from Managed beta package version 2.

Generally, to find the answer you have to try and see.

Also, if you won't remove anything from Managed beta package version 2, which was present in Managed beta package version 1, it should work.