Cherry-pick of 3419be8

@aruniverse Isn't this a breaking change that goes against our version policy?

@aruniverse Isn't this a breaking change that goes against our version policy?

yes while this has breaking changes, we can extract some of the changes (everything but removing the re-exports) and it should be ok

we can extract some of the changes (everything but removing the re-exports)

What's the reason for backporting this?

@aruniverse Isn't this a breaking change that goes against our version policy?

yes while this has breaking changes, we can extract some of the changes (everything but removing the re-exports) and it should be ok

If you aren't removing the re-exports, but just updating the other code not to rely on them, the title of the PR should be changed, and it should probably be explained in the description.

@aruniverse Isn't this a breaking change that goes against our version policy?

yes while this has breaking changes, we can extract some of the changes (everything but removing the re-exports) and it should be ok

If you aren't removing the re-exports, but just updating the other code not to rely on them, the title of the PR should be changed, and it should probably be explained in the description.

This PR is in draft, and not ready to be reviewed. Agreed, those changes will be done prior to publishing the PR.
No changes have made since the bp pr was automatically created by the mergify bot. You should also note, the conflict label.

we can extract some of the changes (everything but removing the re-exports)

What's the reason for backporting this?

Tried updating to a 5.0 pre-release in a test app, and ran into this breaking issue. Trying to understand the blast radius and how many shared pkgs would be impacted while providing an example for other consumers prior to 5.0