Struct PlatformExecutionProvider 

pub struct PlatformExecutionProvider { /* private fields */ }

Production provider set. Native and isolated allocations are routed by the immutable execution kind in the policy/journal; an unavailable isolated backend can never fall through to native process launch.

Implementations

impl PlatformExecutionProvider

pub fn new(host_id: HostId) -> Self

pub fn with_isolation_settings(self, settings: IsolationSettings) -> Self

Applies the operator’s isolation settings to the OCI provider.

pub fn with_runner_env_file(self, path: PathBuf) -> Self

Applies the host’s runner.env at path to every native runner. Isolated runners do not get it: their environment is the image’s.

pub fn with_dependency_caches(self, caches: DependencyCaches) -> Self

Gives every native runner persistent dependency caches. Isolated runners do not get them: their providers mount no host directory.

pub fn with_docker_cleanup(self, cleanup: DockerCleanup) -> Self

pub fn with_runner_pool(self, pool: Result<Option<RunnerPool>, String>) -> Self

Starts every native runner as an account of pool of its own, which owns its attempt directory; the pool shares the caches. None keeps runners in the daemon’s account.

Trait Implementations

impl Debug for PlatformExecutionProvider

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

Formats the value using the given formatter.

impl ExecutionProvider for PlatformExecutionProvider

fn probe(&self, policy: &ScalePolicy) -> ProviderCapability

fn resolve( &self, policy: &ScalePolicy, ) -> Result<Option<ResolvedEnvironment>, FailureReason>

Resolve Auto and image identity before allocation or GitHub JIT. Native has no image.

fn prepare( &self, attempt: &RunnerAttempt, policy: &ScalePolicy, ) -> Result<PreparedEnvironment, FailureReason>

fn start( &self, prepared: PreparedEnvironment, attempt: &RunnerAttempt, handoff: OneTimeJitHandoff<'_>, ) -> Result<EnvironmentIdentity, ProcessStartFailure>

fn inspect( &self, attempt: &RunnerAttempt, ) -> Result<EnvironmentState, FailureReason>

fn stop(&self, attempt: &RunnerAttempt) -> Result<ProviderStop, FailureReason>

fn destroy( &self, attempt: &RunnerAttempt, ) -> Result<ProviderDestroy, FailureReason>

fn recover( &self, attempt: &RunnerAttempt, ) -> Result<EnvironmentState, FailureReason>

fn enumerate_owned(&self, host_id: HostId) -> Vec<EnvironmentIdentity>

fn diagnostics(&self, attempt: &RunnerAttempt) -> Vec<ProviderDiagnostic>

fn owns(&self, attempt: &RunnerAttempt) -> Result<bool, FailureReason>

Independent resource metadata must match the journal before destructive calls.

fn spawn( &self, attempt: &RunnerAttempt, config: &EncodedJitConfig, ) -> Result<u32, ProcessStartFailure>

fn is_alive(&self, attempt: &RunnerAttempt) -> Result<bool, FailureReason>

fn recovered_pid( &self, attempt: &RunnerAttempt, ) -> Result<Option<u32>, FailureReason>

Durable identity observed for a process that spawned before the starting journal write survived.

fn completed_successfully(&self, attempt: &RunnerAttempt) -> bool

True only for a child this invocation owned and reaped with a successful exit status. A recovered process that is merely gone answers false.

fn record_terminate_intent( &self, attempt: &RunnerAttempt, ) -> Result<(), FailureReason>

fn has_terminate_intent(&self, attempt: &RunnerAttempt) -> bool

fn terminate(&self, attempt: &RunnerAttempt) -> Result<(), FailureReason>