Struct WaitObservation 

pub struct WaitObservation {

    pub checking_continuation: bool,
    pub background_work: Option<ApiBackgroundWork>,
    pub pending_elicitations: Vec<ElicitationRequest>,
    pub lifecycle: Option<ViewerLifecycleCategory>,
    pub resuming: bool,
    pub closing: bool,
    pub close_failure: Option<String>,
    pub cannot_take_prompt: bool,
    pub launch_failed: bool,
    pub launch_error: Option<String>,
    pub execution: MaterializedExecutionState,
    pub active_turn: Option<MaterializedTurn>,
    pub last_turn_outcome: Option<MaterializedTurnOutcome>,
    pub queued: usize,
    pub capacity_retry: Option<CapacityRetry>,
    pub retry_assessment_pending: bool,
    pub quota_recovery: Option<QuotaRecovery>,
    pub start_status: Option<StartStatus>,
    pub subagent: bool,
    pub report_pending_for: Option<String>,
    pub handback: Option<(String, String)>,
}

Everything one pass of the wait loop knows about a session.

Fields

checking_continuation: bool

background_work: Option<ApiBackgroundWork>

pending_elicitations: Vec<ElicitationRequest>

lifecycle: Option<ViewerLifecycleCategory>

resuming: bool

A resume operation owns this session now. Its durable record still says stopped — it stays stopped until the archive has been verified — so without this a wait would answer stopped for a session that is on its way up.

closing: bool

A close owns this session now. Like resuming, this ends nothing: the wait follows the close until it finishes.

close_failure: Option<String>

The reason a close recorded on a session that is alive again, which is what a close that failed leaves behind. It is published only until the next action or transition for the session succeeds, so it always refers to a close nobody has recovered from.

cannot_take_prompt: bool

The session cannot take a prompt yet: it is still provisioning, a lifecycle operation owns it, or its worker is not attached. A wait with no target turn keeps waiting, because “finished” would invite a prompt that is then refused.

launch_failed: bool

A recorded launch failure names this session.

launch_error: Option<String>

Why the launch failed, when a reason was recorded.

execution: MaterializedExecutionState

active_turn: Option<MaterializedTurn>

last_turn_outcome: Option<MaterializedTurnOutcome>

queued: usize

capacity_retry: Option<CapacityRetry>

retry_assessment_pending: bool

quota_recovery: Option<QuotaRecovery>

start_status: Option<StartStatus>

subagent: bool

This session is a Mjolnir sub-agent child, whose answer says where its report came from.

report_pending_for: Option<String>

The finished turn, by command id, whose child still owes its report. Mjolnir reminds the child to hand it back, so like an armed retry the turn is not an ending yet.

handback: Option<(String, String)>

The report a child handed back, and the turn it answers.

Implementations

impl WaitObservation

pub fn apply_subagent_report( &mut self, handback_tool: bool, report: &SubagentReport, now_ms: i64, )

Fold a child’s recorded report into this observation, judged against the turn this observation saw finish.

Trait Implementations

impl Clone for WaitObservation

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for WaitObservation

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

Formats the value using the given formatter.

impl Default for WaitObservation

fn default() -> Self

Returns the “default value” for a type.

impl PartialEq for WaitObservation

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for WaitObservation