Struct SignalPolicy 

pub struct SignalPolicy {
    pub sigterm_grace_seconds: u32,
    pub sigkill_force: bool,
    pub sighup_strategy: SighupStrategy,
    pub start_suspended: bool,
}

Signal policy — how the Process responds to signals.

Fields

sigterm_grace_seconds: u32

Grace before escalating SIGTERM → SIGKILL.

sigkill_force: bool

Permit force-reap via SIGKILL (default: allow).

sighup_strategy: SighupStrategy

How SIGHUP is handled.

start_suspended: bool

Start suspended — requires SIGCONT to transition past Forking.

Implementations

impl SignalPolicy

pub fn has_sighup_strategy(&self, kind: SighupStrategy) -> bool

Closed-set-driven presence probe — does this SignalPolicy carry the given SighupStrategy discriminator on its Self::sighup_strategy slot? The ONE substrate primitive that owns the (SignalPolicy, SighupStrategy) -> bool scalar-carrier walk shape.

Third representation kind on the presence-probe axis

The workspace-wide closed-set-driven presence-probe algebra spans three underlying representation kinds; every downstream consumer composes through the SAME has(kind: K) -> bool shape regardless of the field’s Rust type:

• Option-slot (populated-slot semantics) — crate::intent::Intent::has and crate::lifetime::Lifetime::has, both bodies kind.select(self).is_some() over a tagged-union parent whose discriminator’s select projects onto an Option<&V>.
• Slice (walk-a-Vec semantics) — crate::boundary::ConditionSliceExt::has_kind on &[Condition] and crate::spec::DependsOnSliceExt::has_must_reach on &[DependsOn], both bodies self.iter().any(|c| c.kind == kind) over a per-entry discriminator field.
• Scalar (variant-equality semantics) — THIS primitive on spec.signals.sighup_strategy, body self.sighup_strategy == kind over a non-Option, non-Vec closed-set-discriminator field. SighupStrategy is a #[derive(DeriveClosedSet)] implementor with a Default impl, so the field is ALWAYS one of the ALL variants — there is no absent state to detect, and the probe answers “does the carrier’s variant equal this discriminator” rather than “is this slot populated”.

Semantics — VARIANT match, not POPULATED slot

has_sighup_strategy(kind) returns true iff self.sighup_strategy == kind. On a SignalPolicy::default (sighup_strategy: SighupStrategy::default() = Reconverge) the probe returns true for SighupStrategy::Reconverge and false for every other variant — distinct from the Option-slot axis where a default carrier returns false for EVERY kind. The SighupStrategy::default arm’s answer is legitimate operator signal: a Process that left :signals :sighupStrategy at the substrate default IS configured for Reconverge, and a :requires (sighup-Reconverge) should pass; only an operator who deliberately overrode the strategy to Restart or Noop fails the tag on this axis.

Sibling to the presence-probe algebra

Same shape (has(kind)), same axis (closed-set discriminator on a ProcessSpec field), same operator-facing answer (does this spec carry this discriminator on this axis). A future unification (a trait for closed-set-driven presence probes that admits all three representation kinds — Option-slot, slice, scalar) lands as ONE peer trait with every current implementor picking up the trait default in lockstep.

Compounding

A future closed-set-discriminator scalar field on ProcessSpec (or any of its nested structs) that wants a <prefix>-<kind> require-tag family — a verification-phase-<kind> on spec.compliance.<binding>.phase, a routing-form-<kind> on spec.routing.as_ref().map(|r| r.form), a future intent-kind scalar discriminator on any scalar-enum spec field — lands as ONE peer inherent method with the same one-line self.<field> == kind body and routes through the same strip_and_classify_prefixed_kind::<K, _> shape in tatara-check.

Theory anchor: THEORY.md §II.1 invariant 5 — composition preserves proofs; the scalar-carrier presence-probe body lives at ONE substrate site so every downstream (sighup-<kind> require-tag family in tatara-check, closed-set audit dispatchers, future variant additions on SighupStrategy) binds through the SAME shape rather than restating the signals.sighup_strategy == kind closure body at each callsite. THEORY.md §VI.1 — generation over composition; a future SighupStrategy variant (a Suspend that maps SIGHUP onto crate::phase::ProcessPhase::Zombie via SighupStrategy::sighup_target) lands at ONE ALL entry + ONE as_str arm on the closed set and the probe picks it up mechanically without further per-consumer edits.

pub fn has_sighup_target(&self, phase: ProcessPhase) -> bool

Derived-Option-child presence probe on the SIGHUP-target projection — true iff this policy’s stored Self::sighup_strategy transitions this Process INTO phase on SIGHUP reception (as read through SighupStrategy::has_target, which composes SighupStrategy::sighup_target with an == Some(phase) gate).

The one-line collapse of the signals.sighup_strategy.sighup_target() == Some(phase) closure body lifted to ONE substrate owner past the ★★ PRIME-DIRECTIVE ≥ 2 duplication threshold — the sighup-target-<phase> require-tag prefix family in [tatara-check]’s point-domain require-tag classifier is the first workspace-wide consumer. Composes through SighupStrategy::has_target so a regression in the SighupStrategy::sighup_target projection surfaces at ONE substrate site and every downstream (this method, the require-tag classifier, closed-set audit dispatchers walking crate::phase::ProcessPhase::ALL against SIGHUP transition semantics) inherits the shift.

Semantics — DERIVED sighup_target, not raw strategy equality

Sibling to Self::has_sighup_strategy on the SAME Self::sighup_strategy field: has_sighup_strategy asks “does this policy CARRY the queried strategy literal” (raw stored-scalar equality); has_sighup_target asks “would this policy TRANSITION the Process into the queried phase on SIGHUP” (compound sighup_target projection). The two probes coexist because a future SighupStrategy variant that reuses an existing target (a hypothetical Refresh mapping to Reconverging) would satisfy has_sighup_target(Reconverging) alongside the existing Reconverge-carrying policies — distinct stored variants, same derived target — while has_sighup_strategy(Reconverge) stays keyed strictly on the stored variant.

Sibling scalar-carrier probes

• Self::has_sighup_strategy — required parent × defaulted scalar child (stored). SAME parent, SAME field, RAW discriminator equality.
• crate::routing::RoutingSpec::has_form — Option-parent × derived-scalar-child. Peer on the “derived child” axis, but the derived projection returns raw [RoutingForm] rather than Option<RoutingForm> (there is no None-projecting variant).
• THIS — required-parent × derived-Option-child. First occupant on the (required-parent × derived-Option-child) corner of the presence-probe algebra: the child projection returns Option<K> because one variant (SighupStrategy::Noop) transitions into no phase at all, and the None arm short-circuits every query to false.

Compounding

The point-domain require-tag surface in tatara-reconciler::bin::tatara-check composes this primitive with the closed-set crate::phase::ProcessPhase’s autoderived FromStr through the strip_and_classify_prefixed_kind substrate to publish the TWENTY-SIXTH closed-set-driven prefix family in the point- domain classifier’s dispatch table (sighup-target-<phase>), byte-for-byte symmetrical with the sibling sighup-<kind> family that composes through Self::has_sighup_strategy.

A future SighupStrategy variant reaches this probe through ONE ALL entry + one as_str arm + one sighup_target arm alone — no per-caller edit at this composition, no per-consumer restatement of the signals.sighup_strategy.sighup_target() == Some(phase) closure body. A future crate::phase::ProcessPhase variant that a hypothetical fresh strategy targets is reached by pairing a new sighup_target arm with a new ALL entry — again ONE substrate edit and every downstream inherits the shift.

Theory anchor: THEORY.md §II.1 invariant 5 — composition preserves proofs; the derived-Option-child presence-probe composition (self.sighup_strategy.has_target(phase)) lives at ONE substrate site so every downstream (sighup-target-<phase> require-tag family in tatara-check, closed-set audit dispatchers, future variant additions on either SighupStrategy or crate::phase::ProcessPhase) binds through the SAME has_sighup_target(phase) shape rather than restating the signals.sighup_strategy.sighup_target() == Some(phase) chain at each callsite. THEORY.md §VI.1 — generation over composition; the SighupStrategy::sighup_target projection is the ONE per-variant edit site every future strategy variant reaches through, and this method inherits the shift automatically.

Trait Implementations

impl Clone for SignalPolicy

fn clone(&self) -> SignalPolicy

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for SignalPolicy

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

Formats the value using the given formatter.

impl Default for SignalPolicy

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for SignalPolicy

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl JsonSchema for SignalPolicy

fn schema_name() -> String

The name of the generated JSON Schema.

fn schema_id() -> Cow<'static, str>

Returns a string that uniquely identifies the schema produced by this type.

fn json_schema(generator: &mut SchemaGenerator) -> Schema

Generates a JSON Schema for this type.

fn is_referenceable() -> bool

Whether JSON Schemas generated for this type should be re-used where possible using the $ref keyword.

impl Serialize for SignalPolicy

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.