RepoConfig

Committed repository configuration.

Usage

Source

RepoConfig(
    version=CONFIG_VERSION,
    dataset_id=new_dataset_id(),
    snapshot_auto=False,
    verify_on_status=False,
    new_fork="lazy",
    new_record="fork",
    defaults=Policy(),
    vcs=dict(),
    backends=dict(),
    committed_import_query=None
)

Parameter Attributes

version: int = CONFIG_VERSION
dataset_id: str = new_dataset_id()

[dataset] id: 8 hex chars naming this dataset in every store it pins. Pins are tether.<id>.<hash>, working branches tether.ws.<id>.<ws>...; gc only touches refs in this namespace. Committed, so all clones share it.

snapshot_auto: bool = False

[snapshot] auto: whether the CLI’s status fingerprints every object on each run. Off by default: status shows the last snapshot and its age, status --snapshot (or tether snapshot) refreshes. commit and verify always fingerprint.

verify_on_status: bool = False
new_fork: str = "lazy"

[new] fork: lazy (default) creates a working branch on the first writable open; eager creates every branch during new.

new_record: str = "fork"

[new] record: fork (default) branches a pin = "record" object during new even when forks are lazy, since that branch is all that keeps its recorded state from expiring; defer leaves it for the first writable open like a pinned one (new --defer-record). Needs fork = "lazy".

defaults: Policy = Policy()
vcs: dict[str, Any] = dict()
backends: dict[str, dict[str, Any]] = dict()
committed_import_query: str | None = None
[import] query in the committed file: kept only so tether import can refuse it by name. A clone must not choose SQL that runs with your DSN; the default query is Secrets.import_query.

Attributes

Name Description
trunk [vcs] trunk (default main): the dataset bookmark that stands for

trunk

[vcs] trunk (default main): the dataset bookmark that stands for

trunk: str

every object’s upstream branch (locator.branch). Working on it writes to those branches directly; any other bookmark forks.