Struct BaselineStaleness 

pub struct BaselineStaleness {
    pub entries: usize,
    pub matched: usize,
    pub current_findings: usize,
    pub change_scoped: bool,
}

One run’s view of a loaded baseline: everything needed to decide whether the baseline still describes the project.

Every command that accepts --baseline builds one of these and asks it the same two questions, so dead-code, dupes and health cannot answer “is this baseline stale enough to say something” three different ways.

Fields

entries: usize

Entries saved in the baseline file.

matched: usize

Baseline entries that matched something in this run.

current_findings: usize

Findings this run produced before the baseline filtered them. Zero means there was nothing to compare, either because the project is clean or the scope was empty, so staleness cannot be judged.

change_scoped: bool

True when this run analyzed only part of the project, so a whole-project baseline matches less of it for reasons that are not rot.

Implementations

impl BaselineStaleness

pub const fn stale_entries(&self) -> usize

Entries that matched no current finding on this run.

pub const fn warning(&self) -> BaselineStalenessWarning

The advisory warning this run prints on stderr by default.

Silent for a run narrowed to part of the project, because such a run legitimately sees only a slice of a whole-project baseline and re-saving from it would drop every entry the run never looked at. Silent as well when the run produced no findings: a cleaned project and a rotted baseline look identical from here, and the re-save advice is wrong when the right move is deleting the file.

pub const fn trips_gate(&self) -> bool

Whether the opt-in --fail-on-stale-baseline gate fires.

Deliberately stricter than Self::warning: any stale entry counts. The quarter threshold exists to keep an unasked-for line from training people to ignore it, and the empty-run silence exists because an advisory cannot tell rot from success; a repository that passes the flag has asked for both. Change-scope stays the one shared suppression, because a narrowed run still cannot judge a whole-project baseline.

pub fn to_envelope( &self, moved_entries: usize, scope_reasons: BaselineScopeReasons, unrecognised_format: bool, ) -> BaselineStaleness

This run’s machine-readable view of the baseline, for the JSON envelope.

Every member is derived here rather than in a consumer, so the advisory threshold and the gate rule have exactly one implementation. Only health can follow a file move; the commands that match entries by fingerprint pass 0.

scope_reasons and unrecognised_format come from the caller because only the command knows which channels it read and which format it read the file as; this struct carries the counts the analysis needs and nothing more. change_scoped and scope_reasons must agree, which is why every caller derives the boolean from the same reason set it passes here.

unrecognised_format widens gate_trips here, at the single site that builds it, so the opt-in gate, the exit code, every CI surface that reads the boolean and the MCP sentences move together: a file this command cannot read as its own suppresses nothing, which no count can express. Unlike the count rule it is not suppressed by change_scoped, because telling a foreign file from this command’s own needs no project-wide run.

Trait Implementations

impl Clone for BaselineStaleness

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for BaselineStaleness

impl Debug for BaselineStaleness

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for BaselineStaleness

impl PartialEq for BaselineStaleness

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for BaselineStaleness