jobq: one insertion entry point, and make it atomic

Three findings from the operator's review, all correct.

1. Two insert_job's. Graph::insert_job had no caller outside hive-jobq's
   own tests -- production only ever went through Scheduler::insert_job.
   It existed because the graph-level one got written first. Deleted; the
   tests moved onto a Scheduler, which is where insertion belongs anyway.

2. insert_job was not atomic, and the previous commit made that worse: a
   forward edge or forward parent surfaced mid-loop, leaving the nodes
   before it in the graph, and resolve_wanted ran after every insert, so
   an unknown handle failed once the whole job was already committed.
   The module documented this under "Partial insertion" instead of fixing
   it -- prose describing a hole is not a design.

   All three are decidable from what the builder holds, so
   check_declaration_order now runs before the first insert and the loop
   indexes ids directly. A malformed job leaves the graph untouched.
   What remains mid-insert is the graph's own rejection (out-of-group
   dep, empty DepWhen); closing that needs a dry-run validate on Graph,
   which is a separate change.

3. DagSpec no longer boxes its recipe: it is generic over the closure,
   which travels from the template that built it straight into submit.
   The box bought type inference, and paying for it costs annotations --
   `|b: &Job|` at each declaration site (the field needs an HRTB, and an
   unannotated closure binds one lifetime) and `+ use<>` on each
   returning signature (or the opaque type captures the caller's borrows).
   Erasure is still needed where several recipe shapes share one type:
   the boxed Declare stays for the executor's append_subgraph, and a test
   table uses an erase() helper.
This commit is contained in:
atlas 2026-08-02 14:39:37 +02:00 committed by mara
commit f035b63b9a
8 changed files with 211 additions and 176 deletions

View file

@ -179,7 +179,7 @@ impl Default for JobQueue {
/// Propagates a crate graph-insert error (malformed dep/parent / dep-scope).
fn insert_group(
inner: &mut QueueInner,
declare: Declare,
declare: impl FnOnce(&Job),
group_parent: Option<NodeId>,
) -> anyhow::Result<()> {
inner
@ -216,15 +216,19 @@ impl JobQueue {
self.inner.lock().expect("job_queue mutex poisoned")
}
/// Submit a DAG. Validates the spec, inserts a [`NodeKind::Dag`] **container
/// node** carrying the group's metadata, then inserts the template's nodes as
/// its subtree (their roots re-parented to the container). Returns the
/// container's id as the DAG id — its rolled-up state is the DAG state.
/// Submit a DAG: insert a [`NodeKind::Dag`] **container node** carrying the
/// group's metadata, then insert the template's nodes as its subtree (their
/// roots re-parented to the container). Returns the container's id as the
/// DAG id — its rolled-up state is the DAG state.
///
/// Takes the spec's recipe by generic, not as a boxed [`Declare`]: a spec
/// travels from the template that built it directly into this call, so
/// there is nothing to allocate for.
///
/// # Errors
/// Propagates the spec-validation error (empty / cyclic / bad parent) or a
/// graph-insert error (dependencies that aren't dependency-topological).
pub fn submit(&self, spec: DagSpec) -> anyhow::Result<u64> {
/// Propagates a graph-insert error (dependencies that aren't
/// dependency-topological).
pub fn submit<F: FnOnce(&Job)>(&self, spec: DagSpec<F>) -> anyhow::Result<u64> {
let mut inner = self.lock();
let container = inner
.sched

View file

@ -475,16 +475,23 @@ impl NodeKind {
/// moment it inserts. A shape that has been declared is therefore always
/// insertable — a dangling edge or a cycle cannot be expressed, so there is
/// nothing left for a submit-time validation pass to reject.
pub struct DagSpec {
///
/// Generic over the recipe rather than boxing it: a spec goes from the template
/// that returns it straight to the `submit` that consumes it, so the closure's
/// concrete type is known the whole way and needs neither an allocation nor a
/// `Send` bound. (The executor's `append_subgraph` is the case that *does* need
/// a boxed [`super::Declare`] — its recipes are collected into a `Vec` and
/// applied later, across a task boundary.)
pub struct DagSpec<F> {
pub source: Source,
/// Free-form "why".
pub reason: String,
/// Declares the DAG's nodes — their edges, grouping and resources — onto
/// the builder the queue hands it.
pub declare: super::Declare,
pub declare: F,
}
impl std::fmt::Debug for DagSpec {
impl<F> std::fmt::Debug for DagSpec<F> {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
// The recipe is a closure; there is nothing to show of it, and its
// nodes do not exist until the queue runs it.

View file

@ -26,11 +26,11 @@ use std::sync::Arc;
use super::model::{DagSpec, NodeKind};
use super::templates::{RebuildOpts, node, rebuild_nodes};
use super::{Declare, Job, Source, templates};
use super::{Job, Source, templates};
use crate::coordinator::Coordinator;
use crate::lifecycle;
fn submit_and_emit(coord: &Arc<Coordinator>, spec: super::DagSpec) -> u64 {
fn submit_and_emit<F: FnOnce(&Job)>(coord: &Arc<Coordinator>, spec: super::DagSpec<F>) -> u64 {
let id = coord
.job_queue
.submit(spec)
@ -169,7 +169,7 @@ fn restart_chain(b: &Job, agent: &str, graceful: bool, running: bool) {
/// and each keeps its own root, so the per-agent subgraphs are independent and
/// run concurrently, each on its own lease. Rebasing one subgraph's indices
/// onto another's used to be a function.
fn power_dag(source: Source, reason: String, declare: Declare) -> DagSpec {
fn power_dag<F: FnOnce(&Job)>(source: Source, reason: String, declare: F) -> DagSpec<F> {
DagSpec {
source,
reason,
@ -188,12 +188,12 @@ pub(crate) fn stop_spec(
graceful: bool,
source: Source,
reason: String,
) -> DagSpec {
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let targets = targets.to_vec();
power_dag(
source,
reason,
Box::new(move |b| {
Box::new(move |b: &Job| {
for (agent, running) in targets {
stop_chain(b, &agent, graceful, running);
}
@ -211,12 +211,12 @@ pub(crate) fn start_spec(
targets: &[(String, bool, bool)],
source: Source,
reason: String,
) -> DagSpec {
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let targets = targets.to_vec();
power_dag(
source,
reason,
Box::new(move |b| {
Box::new(move |b: &Job| {
for (agent, running, stale) in targets {
start_chain(b, &agent, running, stale);
}
@ -230,12 +230,12 @@ pub(crate) fn restart_spec(
graceful: bool,
source: Source,
reason: String,
) -> DagSpec {
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let targets = targets.to_vec();
power_dag(
source,
reason,
Box::new(move |b| {
Box::new(move |b: &Job| {
for (agent, running) in targets {
restart_chain(b, &agent, graceful, running);
}

View file

@ -241,7 +241,7 @@ pub(crate) fn rebuild_nodes<'a>(
/// already holding it rather than deadlocking against it.
pub(crate) fn deploy_rebuild_nodes(agent: &str, approval_id: i64) -> Declare {
let agent = agent.to_owned();
Box::new(move |b| {
Box::new(move |b: &Job| {
let roots = rebuild_nodes(
b,
&agent,
@ -274,12 +274,17 @@ pub(crate) fn deploy_rebuild_nodes(agent: &str, approval_id: i64) -> Declare {
/// whole `StopForUpdate`→`Swap`→`PostSwap` subtree, so those three cover every
/// node. Edging `Reconcile` alone would not do: it is `AfterAny` `Prebuild`, so
/// it reaches `Done` even after a failed swap and the tail would report success.
pub fn rebuild(agent: &str, source: Source, reason: String, relock: bool) -> DagSpec {
pub fn rebuild(
agent: &str,
source: Source,
reason: String,
relock: bool,
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let agent = agent.to_owned();
DagSpec {
source,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let roots = rebuild_nodes(
b,
&agent,
@ -319,12 +324,16 @@ pub fn rebuild(agent: &str, source: Source, reason: String, relock: bool) -> Dag
///
/// The window still spans the container build, as it must: `prepare_deploy`
/// leaves `flake.lock` staged-uncommitted for the build's whole duration.
pub fn approval_deploy(agent: &str, approval_id: i64, reason: String) -> DagSpec {
pub fn approval_deploy(
agent: &str,
approval_id: i64,
reason: String,
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let agent = agent.to_owned();
DagSpec {
source: Source::Approval,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let a = || agent.clone();
let window = node(
b,
@ -371,12 +380,16 @@ pub fn approval_deploy(agent: &str, approval_id: i64, reason: String) -> DagSpec
/// single-node lifecycle DAGs that exercise per-agent lease serialization
/// in the queue tests); production paths no longer emit a bare reconcile.
#[cfg(test)]
pub fn reconcile_only(agent: &str, source: Source, reason: String) -> DagSpec {
pub fn reconcile_only(
agent: &str,
source: Source,
reason: String,
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let agent = agent.to_owned();
DagSpec {
source,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let _reconcile = node(b, NodeKind::Reconcile { agent });
}),
}
@ -393,12 +406,12 @@ pub fn reconcile_only(agent: &str, source: Source, reason: String) -> DagSpec {
/// container was never created). Closed by a `ResolveApproval` tail root edged
/// `AfterAny` onto `Provision` — the DAG's only other group-root, so its roll-up
/// already carries the whole cascade.
pub fn spawn(agent: &str, approval_id: i64, reason: String) -> DagSpec {
pub fn spawn(agent: &str, approval_id: i64, reason: String) -> DagSpec<impl FnOnce(&Job) + use<>> {
let agent = agent.to_owned();
DagSpec {
source: Source::Approval,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let a = || agent.clone();
let provision = node(b, NodeKind::Provision { agent: a() });
let create = node(b, NodeKind::Create { agent: a() }).part_of(provision);
@ -417,12 +430,17 @@ pub fn spawn(agent: &str, approval_id: i64, reason: String) -> DagSpec {
/// effect in the container. Group-roots are `WritePermFile` plus the rebuild
/// subgraph's `MetaSync` / `Prebuild` / `Reconcile`, so the `EmitRebuilt` tail
/// edges all four.
pub fn perm_change(agent: &str, source: Source, reason: String, payload: PermPayload) -> DagSpec {
pub fn perm_change(
agent: &str,
source: Source,
reason: String,
payload: PermPayload,
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let agent = agent.to_owned();
DagSpec {
source,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let write = node(
b,
NodeKind::WritePermFile {
@ -463,11 +481,11 @@ pub fn meta_update(
source: Source,
reason: String,
approval_id: Option<i64>,
) -> DagSpec {
) -> DagSpec<impl FnOnce(&Job) + use<>> {
DagSpec {
source,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let lock = node(
b,
NodeKind::MetaLock {
@ -501,11 +519,11 @@ pub fn reparent(
moves: Vec<(hive_types::Ident, Option<hive_types::Ident>)>,
source: Source,
reason: String,
) -> DagSpec {
) -> DagSpec<impl FnOnce(&Job) + use<>> {
DagSpec {
source,
reason,
declare: Box::new(move |b| {
declare: Box::new(move |b: &Job| {
let _reparent = node(b, NodeKind::Reparent { moves });
}),
}

View file

@ -9,15 +9,29 @@
use super::model::NodeKind;
use super::*;
fn submit(q: &JobQueue, spec: DagSpec) -> u64 {
fn submit<F: FnOnce(&Job)>(q: &JobQueue, spec: DagSpec<F>) -> u64 {
q.submit(spec).expect("valid spec")
}
/// Erase a spec's recipe to the boxed [`Declare`] so specs of *different*
/// shapes can share one type — e.g. a table of `(name, spec)` cases.
///
/// Production never needs this: each submit path builds one spec and hands it
/// straight to `submit`, so the concrete closure type is known end to end. A
/// test table is the case where several shapes must be one type.
fn erase<F: FnOnce(&Job) + Send + 'static>(spec: DagSpec<F>) -> DagSpec<Declare> {
DagSpec {
source: spec.source,
reason: spec.reason,
declare: Box::new(spec.declare),
}
}
fn ident(s: &str) -> hive_types::Ident {
hive_types::Ident::parse(s).expect("valid test ident")
}
fn rebuild(agent: &str, reason: &str) -> DagSpec {
fn rebuild(agent: &str, reason: &str) -> DagSpec<impl FnOnce(&Job) + use<>> {
templates::rebuild(agent, Source::Manual, reason.to_owned(), true)
}
@ -25,14 +39,22 @@ fn rebuild(agent: &str, reason: &str) -> DagSpec {
/// shape (`[Signal→Drain→] StopForUpdate → Reconcile`, no `SetWanted` head)
/// most queue-mechanics tests assume. Mirrors the pre-dynamic
/// `templates::restart` (which is now the state-aware `submit::restart_spec`).
fn restart_online(agents: &[&str], graceful: bool, reason: &str) -> DagSpec {
fn restart_online(
agents: &[&str],
graceful: bool,
reason: &str,
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let targets: Vec<(String, bool)> = agents.iter().map(|a| ((*a).to_owned(), true)).collect();
submit::restart_spec(&targets, graceful, Source::Manual, reason.to_owned())
}
/// Stop DAG spec with every agent treated as **running** — the online shape
/// (`SetWanted → [Signal→Drain→](graceful) Reconcile`).
fn stop_online(agents: &[&str], graceful: bool, reason: &str) -> DagSpec {
fn stop_online(
agents: &[&str],
graceful: bool,
reason: &str,
) -> DagSpec<impl FnOnce(&Job) + use<>> {
let targets: Vec<(String, bool)> = agents.iter().map(|a| ((*a).to_owned(), true)).collect();
submit::stop_spec(&targets, graceful, Source::Manual, reason.to_owned())
}
@ -197,7 +219,7 @@ fn graceful_rebuild_chain_drains_before_stopping() {
DagSpec {
source: Source::AutoUpdate,
reason: "sweep".to_owned(),
declare: Box::new(|b| {
declare: Box::new(|b: &Job| {
templates::rebuild_nodes(
b,
"agent-a",
@ -246,7 +268,7 @@ fn non_graceful_rebuild_has_no_signal_or_drain() {
DagSpec {
source: Source::Manual,
reason: "manual".to_owned(),
declare: Box::new(|b| {
declare: Box::new(|b: &Job| {
templates::rebuild_nodes(
b,
"agent-a",
@ -695,7 +717,7 @@ fn append_subgraph_roots_on_emitter_and_rebases_local_deps() {
let spec = DagSpec {
source: Source::AutoUpdate,
reason: "sweep".to_owned(),
declare: Box::new(|b| {
declare: Box::new(|b: &Job| {
let _lock = templates::node(
b,
NodeKind::MetaLock {
@ -715,7 +737,7 @@ fn append_subgraph_roots_on_emitter_and_rebases_local_deps() {
// match the sweep arm of `run_meta_lock` or this stops tracking production.
let subgraph = |agent: &str| -> Declare {
let agent = agent.to_owned();
Box::new(move |b| {
Box::new(move |b: &Job| {
templates::rebuild_nodes(
b,
&agent,
@ -821,7 +843,7 @@ fn meta_update_grows_cascade_in_dag() {
// Simulate the executor growing the cascade in-DAG (`relock = false` — a
// cascade child must not re-lock and revert the parent's bump).
for agent in ["alice", "bob"] {
let declare: Declare = Box::new(move |b| {
let declare: Declare = Box::new(move |b: &Job| {
templates::rebuild_nodes(
b,
agent,
@ -1073,25 +1095,37 @@ fn cancelled_power_op_runs_no_compensating_node() {
for graceful in [false, true] {
for running in [false, true] {
let targets = vec![("agent-a".to_owned(), running)];
// Erased to `DagSpec<Declare>`: three different recipe types have to
// sit in one array.
let cases = [
(
"restart",
false,
submit::restart_spec(&targets, graceful, Source::Manual, "bounce".to_owned()),
erase(submit::restart_spec(
&targets,
graceful,
Source::Manual,
"bounce".to_owned(),
)),
),
(
"stop",
true,
submit::stop_spec(&targets, graceful, Source::Manual, "stop".to_owned()),
erase(submit::stop_spec(
&targets,
graceful,
Source::Manual,
"stop".to_owned(),
)),
),
(
"start",
true,
submit::start_spec(
erase(submit::start_spec(
&[("agent-a".to_owned(), running, false)],
Source::Manual,
"start".to_owned(),
),
)),
),
];
for (name, writes_intent, spec) in cases {