Struct Scanner 

pub struct Scanner { /* private fields */ }

Embedded consumers identify candidates independently of redacted display labels. Embedded consumers identify candidates independently of redacted display labels. Embedded consumers identify candidates independently of redacted display labels. One loaded ruleset reused across candidate snapshots.

Embedding hosts must use panic=unwind and configure payload-free panic output before spawning workers. The library catches load/matcher unwinds but deliberately does not replace process-wide hooks; Rust invokes the host’s hook before an unwind can be caught.

Implementations

impl Scanner

Load once and scan exact snapshots without reading candidate files from the live worktree.

pub fn load( runtime_rules_path: &Path, builtin_rules: bool, explicit: bool, ) -> Result<Scanner>

What: Load runtime rules and the optional shipped baseline through the incumbent cache/validation path. Why: The embedding adapter does not create a second compiler, cache or builtin-rules loader.

In TS you’d write (pseudocode):

function loadScanner(runtimeRulesPath, builtinRules, explicitlySelectedPath): Scanner;

pub fn cache_warnings(&self) -> &[CacheWarning]

Borrow fixed-token cache warnings without recovering their fields from serialized terminal text.

pub fn scan( &self, identity: usize, logical_path: &Path, bytes: &[u8], ) -> CandidateScan

What: Scan a candidate’s logical pathname and exact immutable bytes. Why: Historical/index candidates must never substitute the current worktree file or its temporary materialization name.

In TS you’d write (pseudocode):

function scan(identity, logicalPath, exactBytes): CandidateScan;