Struct CompletedRecord 

pub struct CompletedRecord {
    pub task: TaskId,
    pub source: RunSource,
    pub state: RunState,
    pub exit_status: Option<ExitStatus>,
    pub stdout_hash: [u8; 32],
    pub stderr_hash: [u8; 32],
    pub materialised_outputs: Vec<CanonicalPath>,
}

Run-record fields produced by one run_task invocation: the observation of a task that the scheduler admitted into the lookup-then-spawn pipeline and that reached a run classification (EXEC-009).

Carries enough information for a scheduler layer to feed the per-TaskId predecessor_streams map that crate::cache_key::build_cache_key consumes: stdout_hash and stderr_hash are always present, whether the run was a cache hit (sourced from the manifest’s recorded hashes per CACHE-011) or a fresh run (computed from the captured byte buffers under the cache’s active HashAlgo).

A scheduler-level cascade decision produces RunOutcome::Skipped instead (EXEC-010 / EXEC-011); run_task itself never produces a Skipped because skipped tasks never enter the pipeline.

Fields

task: TaskId

The task this record describes.

source: RunSource

Whether the success came from a cache hit or a fresh run.

state: RunState

Terminal classification per EXEC-009.

exit_status: Option<ExitStatus>

Process exit status for a fresh run, None for a cache hit (no process ran).

stdout_hash: [u8; 32]

Hash of the run’s captured stdout under the cache’s HashAlgo. For RunSource::CacheHit this is read from the manifest; for RunSource::FreshRun it is computed from the captured byte buffer.

stderr_hash: [u8; 32]

Hash of the run’s captured stderr under the cache’s HashAlgo.

materialised_outputs: Vec<CanonicalPath>

Workspace-absolute paths the task materialised on disk.

Populated only for RunState::Succeeded; empty for RunState::Failed and RunState::Cancelled. The source is the executor’s output-resolution pass for a fresh run and haz_cache::Manifest::outputs for a cache hit.

Consumed by the scheduler’s runtime DAG validation pass (EXEC-019 runtime cycle detection and EXEC-020 runtime output-overlap detection) to derive runtime producer- matching edges and to populate the output-claim tracker.

Trait Implementations

impl Clone for CompletedRecord

fn clone(&self) -> CompletedRecord

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CompletedRecord

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

Formats the value using the given formatter.

impl PartialEq for CompletedRecord

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Eq for CompletedRecord

impl StructuralPartialEq for CompletedRecord