Struct Overrides 

pub struct Overrides(/* private fields */);

Every severity override visible here, resolved the way --get resolves it.

ONE subprocess for the whole stage, and — more importantly — a VALUE. The dispatcher used to call severity_of inside the loop that classifies outcomes, which put a git spawn in the middle of a fold and made the fold impossible to test without a repository.

--get-regexp emits entries in precedence order, so folding with overwrite lands on the same answer --get gives. That equivalence is not assumed: the_batch_agrees_with_the_authority pins it against effective_override.

Implementations

impl Overrides

pub fn read() -> Overrides

Machine config AND the installed repo policy, folded on the specificity ladder: system < global < POLICY < local < worktree < command. --show-scope labels each config line; on a git too old to know the flag (exit 129 → None) this degrades, deliberately, to “ALL git config beats policy” — the fail-safe direction, because the alternative was an EMPTY override set silently discarding both.

pub fn from_scoped(scoped: &str, policy: &Policy) -> Overrides

Fold --show-scope output around the installed policy. Scope words system/global fold BELOW policy; everything else — local, worktree, command, and any word a future git invents — folds ABOVE, because misreading a local as below would let a file pulled from a remote silently override a person’s explicit setting on their own machine, and that is the worse direction to fail in.

pub fn from_plain_with_policy_below( plain: Option<String>, policy: &Policy, ) -> Overrides

The DEGRADED fold — policy below every config scope, i.e. all git config beats policy — for a git that cannot label scopes. pub because the fleet’s severity column must degrade the same way the dispatcher does, not invent a third opinion.

pub fn from_config(out: Option<String>) -> Overrides

Built from --get-regexp-shaped text (amont.severity.<check> <value> per line, in git’s own precedence order — system, then global, then local, then includes). pub so a reader that has already fetched those lines for its own reasons (the fleet dashboard needs the origin of each, which Overrides does not track) can still ask this — the one place that must not get precedence wrong — rather than re-deriving it from the lines by hand.

pub fn applied_to(&self, check: &str) -> Option<(&str, Severity)>

The configured key that applies to check, and what it says.

Several keys can name one check — pre-commit and clippy and pre-commit-clippy all reach pre-commit-clippy. The most specific wins, which is the rule anybody would guess and the only one that lets you downgrade a whole trigger and then exempt one check from it.

pub fn applied_with_source( &self, check: &str, ) -> Option<(&str, Severity, Source)>

As applied_to, keeping WHERE the winning key came from — the one reader (amont list) that reports provenance asks here rather than re-deriving the answer a second way.

pub fn of(&self, check: &dyn Check) -> Severity

The severity to apply to check, override or declared.

Trait Implementations

impl Clone for Overrides

fn clone(&self) -> Overrides

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Overrides

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

Formats the value using the given formatter.

impl Default for Overrides

fn default() -> Overrides

Returns the “default value” for a type.