Proposal: rebase on primary upstream

[blaggacao]

An optomized workflow would be in my opinion represented as follows:

...
origin: <url>
upstream: 
  url: <url>
  ref: <ref>
  rebase: <default: true>  # target will be rebased on upstream, first
remotes:
  - name: <name>
    url: <url>
    merge:
       - <ref>
       - <ref>
target: <branchname on local & origin>
...

[blaggacao]

@dataclass
class TargetDirectory(DataClassYAMLMixin):
    name: GitSubmoduleSpec


@dataclass
class GitSubmoduleSpec(DataClassYAMLMixin):
    origin: str
    upstream: Upstream
    remotes: List[Remote] = field(default_factory=list)
    targets: List[Target] = field(default_factory=list)
    postprocess: List[str] = field(default_factory=list)

@dataclass
class Target(DataClassYAMLMixin):
    name: str
    patches: List[Patch] = field(default_factory=list)

@dataclass
class Patch(DataClassYAMLMixin):
    remote: str
    refs: List[str] = field(default_factory=list)


@dataclass
class Upstream(DataClassYAMLMixin):
    url: str
    ref: str
    rebase: bool = True


@dataclass
class Remote(DataClassYAMLMixin):
    name: str
    url: str

[blaggacao]

Why targets? -> maintaining a set of anxilliary pull requests against upstream in origin (lower the cost of contribution...)

First target, hence would be the locally checked out one...