Commit Graph

6 Commits

Author SHA1 Message Date
MechaCat02
95d6b95272 feat(executor): expose ctx.trigger_depth and pin the invoke depth bump
`invoke_callee_sees_incremented_depth` asserted NOTHING — it ended `let _ = resp`
with a comment that `ctx.trigger_depth` wasn't exposed, so it would have passed
even if `invoke` forwarded the depth unchanged, i.e. never incrementing the
chain-depth counter that bounds runaway trigger loops.

`build_ctx_map` now surfaces `ctx.trigger_depth` (read-only: 0 for direct ingress,
+1 per synchronous `invoke` re-entry or dispatched handler) — the value the
author intended to read and a legitimate diagnostic for a script. The test now
pins the increment: the caller is depth 0, the invoke callee must see 1.

Mutation-verified: forwarding the depth unchanged makes the callee see 0 and the
test fails.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-15 19:38:31 +02:00
MechaCat02
1844bef132 feat(executor): thread the defining node into ExecRequest (Phase 4b C2)
Carry the resolved script's owner (its defining node) from dispatch into
the executor so `import` resolution can be lexical (§5.5). The executing
app (`app_id`) stays the SDK isolation boundary; `script_owner` is a
separate axis — the lexical origin for imports.

- `ExecRequest.script_owner: Option<ScriptOwner>` (serde default; `None`
  falls back to `App(app_id)` in the engine for old payloads / cluster wire).
- `ScriptOwner` gains `Serialize`/`Deserialize` for the wire.
- The engine computes `default_origin` and hands it to the resolver
  (used fully in C3; the resolve() lookup already uses it).
- Every dispatch site sets it from the resolved `Script.owner()`:
  the 4 dispatcher arms (queue/trigger/http/invoke_async, via a new
  `ResolvedTrigger.script_owner`), the orchestrator id-bypass, and the
  `invoke()` SDK path (new `ResolvedScript.owner`, so a group script
  invoked by an app resolves imports from the group).

Behaviour-preserving: app scripts still resolve `App(app_id)`; group
scripts can't yet carry imports (lifted in C4).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-26 07:17:33 +02:00
MechaCat02
343f6d3b4d feat(vars): VarsService + vars:: SDK + config capabilities
Scripts can now read group-inherited config via vars::get(key) / vars::all().

- shared: VarsService trait (read-only get/all) + NoopVarsService; added as
  the 14th field on the Services bundle.
- manager-core: VarsServiceImpl resolves the calling app's config via
  config_resolver (derives app_id from cx.app_id; AppVarsRead-gated for
  authed principals, script-as-gate for anon).
- executor-core: vars:: Rhai bridge (get/all), registered in the SDK.
- authz: AppVarsRead/Write, GroupVarsRead/Write, GroupSecretsRead/Write
  capabilities (group caps resolve via the Phase-2 ancestor walk; the
  GroupSecretsRead human-read gate is group_admin-only, distinct from an
  app's runtime injection).
- wired VarsServiceImpl into the picloud binary; updated the executor-core
  test Services::new call sites.

386 manager-core lib tests green; clippy clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-24 20:59:23 +02:00
MechaCat02
ae2134e62d feat(executor): expose ctx.request.method to scripts (F4)
The request map exposed path/headers/body/params/query/rest but not the
HTTP method, forcing one script per verb (the E2E app needed 7 scripts
where ~3 would do). Add a `method` field to ExecRequest (populated from
the HTTP method on the sync-execute bypass and the HTTP outbox payload;
empty for non-HTTP triggers) and surface it as `ctx.request.method`,
uppercased. A single script can now branch GET/POST on one route.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-12 20:43:10 +02:00
MechaCat02
c3baa87415 chore(v1.1.9): clippy clean — workspace -D warnings
Surfaced + fixed during the F3 attestation:

- executor-core/sdk/queue.rs: drop redundant .map(|()| ()) call sites;
  enqueue_blocking discards QueueMessageId via .map(|_id| ()) (intentional)
- executor-core/sdk/retry.rs: hoist use std::collections::hash_map::DefaultHasher
  + use std::hash::Hasher to the top of the file (clippy::items_after_statements);
  replace `as i64` casts with i64::try_from + clear comment about jitter
  bound (clippy::cast_possible_wrap)
- executor-core/sdk/invoke.rs: move _LIMITS_IS_COPY const before #[cfg(test)]
  mod tests (clippy::items_after_test_module)
- manager-core/dispatcher.rs: dispatch_one_queue gains #[allow(too_many_lines)]
  (it's the queue tick's whole logic — split makes it less readable than
  the lint); .map(...).unwrap_or(default) → .map_or(default, ...)
- picloud/lib.rs: Limits { trigger_depth_max, ..Limits::default() } via
  struct-update instead of let-mut-assign (clippy::field_reassign_with_default)
- tests: r#"..."# → r"..." where there are no `"` inside (clippy::needless_raw_string_hashes);
  combine InvokeTarget::Name | InvokeTarget::Path arms with identical bodies
  in sdk_invoke.rs (clippy::match_same_arms)

cargo fmt --all -- --check 2>&1 | tail -3: no output (exit 0)
cargo clippy --workspace --all-targets --all-features -- -D warnings:
  Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.41s
Workspace lib smoke: 432 passing (306 manager-core + 74 executor-core + 34 shared + 18 orchestrator-core).
DB-gated E2E (against docker compose postgres on localhost:15432):
  queue_e2e: 4 ok / invoke_e2e: 4 ok / retry_e2e: 3 ok / migration_queue_messages: 4 ok / schema_snapshot: 1 ok

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-07 11:02:22 +02:00
MechaCat02
302c1df577 feat(v1.1.9): invoke:: Rhai SDK module — sync re-entry + invoke_async
executor-core/sdk/invoke.rs adds `invoke()` and `invoke_async()` as
top-level Rhai helpers (no `::` namespace, mirroring the brief):

  invoke(target, args)        -> Dynamic  (sync, returns callee's value)
  invoke_async(target, args)  -> String   (fire-and-forget; returns execution_id)

Sync re-entry pattern:
  - bridge captures Arc<Engine> via Engine::self_arc()
  - calls engine.execute(&source, req) directly — same engine instance,
    same Services, same Limits; only SdkCallCx changes per call
  - runs inside the caller's spawn_blocking thread (no nested
    spawn_blocking, no gate re-admission — the outer execution already
    holds a permit)

Back-reference plumbing:
  - Engine gains self_weak: OnceLock<Weak<Engine>> + set_self_weak() +
    self_arc() accessor
  - picloud binary calls engine.set_self_weak(Arc::downgrade(&engine))
    right after construction
  - register_all extended with limits + Option<Arc<Engine>> params
  - Engine::execute_ast threads both through

Limits.trigger_depth_max added — mirrors TriggerConfig::max_trigger_depth.
Default 8; the picloud binary syncs from TriggerConfig::from_env() so
PICLOUD_MAX_TRIGGER_DEPTH governs both the dispatcher's fan-out cap and
the invoke depth bound.

Target parsing (string-first):
  - "/api/foo"             → InvokeTarget::Path
  - 36-char UUID-like      → InvokeTarget::Id  (uuid::Uuid parse-validated)
  - everything else        → InvokeTarget::Name

Cross-app + depth + FnPtr guards:
  - resolve() returns InvokeError::CrossApp if resolved.app_id != cx.app_id
  - bridge throws "invoke: depth limit exceeded (max N)" when
    cx.trigger_depth + 1 > limits.trigger_depth_max (checked BEFORE
    resolve to avoid a wasted DB round-trip)
  - args_to_json rejects FnPtr at any depth — closures don't survive
    invoke boundaries

invoke_async path:
  - calls services.invoke.enqueue_async() which writes an
    OutboxSourceKind::Invoke row (commit 10 wires the dispatcher arm)
  - returns the new ExecutionId as a string for caller tracking

Integration tests (sdk_invoke.rs, 6 binaries):
  - sync invoke returns callee's value (script returns body.x + 1)
  - cross-app invoke rejected (target in other app)
  - depth limit exceeded (recursive script that calls itself; throws
    before stack overflow)
  - callee receives incremented depth in cx (smoke — strict assertion
    needs ctx.trigger_depth surface, deferred to v1.2)
  - args_to_json rejects FnPtr in payload
  - invoke_async returns valid UUID string + queues args payload

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-06-06 19:56:05 +02:00