> > > Hi all, this PR #15216 `about modifying the version recorded to v0.13.0 > > > on v0.13.0 branch` **is blocked by CI**, but this won't block our > > > `cherry-picking soft deadline` and `cherry-picking hard deadline`. > > > cc @driazati > > > > > > Can anybody help to see the CI issue? It seems these thing block the > > process of release schedule several times. > > Hi all, this Jenkins CI issue is about our Jenkins config for those [non-main > branchs will merge with latest main when CI > running](https://github.com/apache/tvm/blob/main/ci/jenkins/generated/arm_jenkinsfile.groovy#L148-L154), > this has no relation with Github branch protection. > > Besides, this CI issue exposes our release step about version modification > having some latent problem. We should modify version to 0.13.0 on main branch > first and then do v0.13.0 branch cutting from main.: > > * First, that can avoid conflict between main and non-main branch with CI, > especially for one file in two branch existing divergence; > * Second, because downstream companies codes are always from release branch, > PR about version number is introduced on main but not release branch. > Downstream companies can avoid conflicts about version number when git merge > this release branch codes based on last release branch codes.
I think branch protection is no problem. Because The git merge operation is actually a protection mechanism for our publishing branch. Fundamentally, ensure that the PR introduced by the release branch is in the form of Cherry Pick on the main branch. And in theory, as long as the PR of cherry pick is in order, there will be no conflicts on the publishing branch. The behavior of version modification on release branch is not in the form of Cherry Pick, and after branch cutting, we made changes on the main branch, which caused divergence thus conflicts. **I think we can first modify the version number on the main branch next time, because two reason as above** -- Reply to this email directly or view it on GitHub: https://github.com/apache/tvm/issues/15134#issuecomment-1627682176 You are receiving this because you are subscribed to this thread. Message ID: <apache/tvm/issues/15134/1627682...@github.com>
