Struct RepositoryEvidence 

pub struct RepositoryEvidence<'a> { /* private fields */ }

The evidence of a real repository: its ledger, its claim matrix joined to that ledger, its index and its tree, judged at the working tree by the evidence module’s own derivation — evidence::report for a claim, evidence::freshness() over the same evidence::compare for a test — so a criterion is current exactly when the rest of Majordomus would call its evidence current. The claim join and the comparisons run git, so each is computed once, and only when a criterion or a governance entry asks.

use majordomus_cli::index::Index;
use majordomus_cli::intent::{EvidenceLookup, RepositoryEvidence};
let evidence = RepositoryEvidence::load(index).expect("the ledger is readable");
// a file the repository does not hold is answered without running git
assert!(!evidence.file("no/such/file.md"));

Implementations

impl<'a> RepositoryEvidence<'a>

pub fn load(index: &'a Index) -> Result<RepositoryEvidence<'a>, Error>

The evidence of the repository an index was built from. A ledger this executable cannot read is an error, never an empty ledger: a misread one could report a pass that was never recorded.

use majordomus_cli::index::Index;
use majordomus_cli::intent::RepositoryEvidence;
// a ledger that cannot be read is an error here, never an empty ledger
let evidence = RepositoryEvidence::load(index).expect("the ledger is readable");
let _ = &evidence;

Trait Implementations

impl EvidenceLookup for RepositoryEvidence<'_>

fn test(&self, id: &TestId, scope: &[String]) -> TestStanding

What the ledger holds for one test, by its identity: whether its source is in the tree and what its latest run came to. scope is the code the plan declares for the criterion — the scope of every issue that serves it — which the judgement counts among the test’s inputs beside what the claims that test proves name.

fn claim(&self, id: &str) -> Option<ClaimStanding>

What a claim’s evidence says, judged as IntentEvidenceState::judged judges it, with the claim report’s verdict behind it; None when no such claim is declared — which is a refusal, not an unproven claim.

fn deployment(&self, id: &str) -> bool

Whether a deployment object with this id exists. A deployment criterion resolves through this and is still never counted as met.

fn object(&self, kind: &str, identity: &str) -> bool

Whether an object of kind answers to identity (rule matches with or without its @version). This is what makes a governance: entry resolve or fail.

fn file(&self, path: &str) -> bool

Whether a repository-relative file exists, for a file: governance entry and for the test sources a criterion names.