The first time a crypto backend needs to get a fix backported into a Go minor release someone has to go to the backend repo and create a release branch from the commit that was used the create that release. This process is manual and error prone, and using the wrong commit is really disruptive.
I propose that the release process automatically creates a new release branch on every crypto backend at the correct commit, taken from the src/go.mod, the first time a minor release is created (that is, go1.x.0). While here, it could also push a tag in the form of go1.x, being x the Go toolchain minor release. That would not follow semantic version, but that's fine, our main use case is the Go toolchain, not external uses. Having the tag would make it easier to manually, or automatically, compare what changes in between two Go minor versions.
The first time a crypto backend needs to get a fix backported into a Go minor release someone has to go to the backend repo and create a release branch from the commit that was used the create that release. This process is manual and error prone, and using the wrong commit is really disruptive.
I propose that the release process automatically creates a new release branch on every crypto backend at the correct commit, taken from the
src/go.mod, the first time a minor release is created (that is,go1.x.0). While here, it could also push a tag in the form ofgo1.x, being x the Go toolchain minor release. That would not follow semantic version, but that's fine, our main use case is the Go toolchain, not external uses. Having the tag would make it easier to manually, or automatically, compare what changes in between two Go minor versions.