My concern is two-fold:
1. Should all long-lived branches that have multiple people collaborate together use branch-PRs?
2. Should branch-PRs have a different policy for acceptance, review and merge into a branch?
There have been a dozen or so branch-PRs that have been made against the openshift branch. A couple weeks ago, one of these PRs started to make some material changes to the nature of how workspace containers should be configured and set up. This PR made changes necessary for the openshift branch, but it's not clear that it would be appropriate for the master branch.
The branch maintainers wanted and opted to merge the branch early wanting to defer analysis for master later on. I raised some concerns - implying that the openshift branch, which now has a very large set of long-lived changes will have to have a complete review cycle before it is merged into master. Master maintainers are not able to use the history of branch-PR merges as a record of analysis that means the overall branchis suitable for merge into master.
So, I think it is good to discuss and determine the role of branch-PRs for long-lived branches, and then the merge + review + documentation requirements for merging branch-PRs.