@bastani/atomic 0.9.44 → 0.9.45

package/CHANGELOG.md

@@ -2,6 +2,40 @@
 
 ## [Unreleased]
 
+## [0.9.45] - 2026-10-25
+
+Custom storage adapters now plug into workflow journals as first-class citizens, and large local worker pools start reliably on busy machines.
+
+- **Journal adapters that build on Atomic's own journal:** a storage adapter loaded by path can import `@bastani/atomic/workflows` and get the running Atomic's module, so `SqlRunJournal` subclasses and journal errors such as `StaleLeaseError` work instead of failing with `Storage unavailable` ([#3963](https://github.com/bastani-inc/atomic/issues/3963), [#3969](https://github.com/bastani-inc/atomic/pull/3969))
+- **Fast status, resume and receipts for custom journals:** a run journal that implements the optional `changeToken()`, `runVersions()`, `snapshotTail()` and `appendMany()` members gets the same faster `/workflow status`, resume and batched delivery receipts as the built-in SQL journal ([#3964](https://github.com/bastani-inc/atomic/issues/3964), [#3968](https://github.com/bastani-inc/atomic/pull/3968))
+- **Worker pools that start on small or busy machines:** without an explicit `workflows.localWorkers.startupTimeoutMs`, the startup deadline scales with pool size and machine load instead of failing after 30 s ([#3962](https://github.com/bastani-inc/atomic/issues/3962), [#3965](https://github.com/bastani-inc/atomic/pull/3965))
+
+### Added
+
+- `StorageProvider` journal factories accept `hostModules`, and `loadStorageAdapter()` takes an optional third `hostModules` argument, to give adapter modules extra live host modules by specifier ([#3963](https://github.com/bastani-inc/atomic/issues/3963))
+- The `RunJournal` interface in `@bastani/atomic/storage` has optional `changeToken()`, `runVersions()`, `snapshotTail()` and `appendMany()` members, with the `RunVersion`, `RunSnapshotTail` and `RunJournalBatchAppend` types. A run journal from a storage adapter, or any custom `RunJournal`, that implements them gets the same faster `/workflow status`, resume and batched delivery receipts as the built-in SQL journal, and `registerRunJournalConformance` checks them when present ([#3964](https://github.com/bastani-inc/atomic/issues/3964))
+
+### Fixed
+
+- A journal adapter loaded by path (`ATOMIC_WORKFLOWS_JOURNAL`, `workflows.journal`, or a run's `journal: { adapter }`) can now import `@bastani/atomic/workflows`. It gets the workflows module of the running Atomic that opened the journal, so an `openRunJournal` hook can build on `SqlRunJournal` and its errors, such as `StaleLeaseError`, pass Atomic's `instanceof` checks. Before, the import resolved to a path inside `dist/index.js` and the journal failed with `Storage unavailable` ([#3963](https://github.com/bastani-inc/atomic/issues/3963))
+
+## [0.9.45-alpha.1] - 2026-10-25
+
+Custom storage adapters now plug into workflow journals as first-class citizens, and large local worker pools start reliably on busy machines.
+
+- **Journal adapters that build on Atomic's own journal:** a storage adapter loaded by path can import `@bastani/atomic/workflows` and get the running Atomic's module, so `SqlRunJournal` subclasses and journal errors such as `StaleLeaseError` work instead of failing with `Storage unavailable` ([#3963](https://github.com/bastani-inc/atomic/issues/3963), [#3969](https://github.com/bastani-inc/atomic/pull/3969))
+- **Fast status, resume and receipts for custom journals:** a run journal that implements the optional `changeToken()`, `runVersions()`, `snapshotTail()` and `appendMany()` members gets the same faster `/workflow status`, resume and batched delivery receipts as the built-in SQL journal ([#3964](https://github.com/bastani-inc/atomic/issues/3964), [#3968](https://github.com/bastani-inc/atomic/pull/3968))
+- **Worker pools that start on small or busy machines:** without an explicit `workflows.localWorkers.startupTimeoutMs`, the startup deadline scales with pool size and machine load instead of failing after 30 s ([#3962](https://github.com/bastani-inc/atomic/issues/3962), [#3965](https://github.com/bastani-inc/atomic/pull/3965))
+
+### Added
+
+- `StorageProvider` journal factories accept `hostModules`, and `loadStorageAdapter()` takes an optional third `hostModules` argument, to give adapter modules extra live host modules by specifier ([#3963](https://github.com/bastani-inc/atomic/issues/3963))
+- The `RunJournal` interface in `@bastani/atomic/storage` has optional `changeToken()`, `runVersions()`, `snapshotTail()` and `appendMany()` members, with the `RunVersion`, `RunSnapshotTail` and `RunJournalBatchAppend` types. A storage adapter's run journal that implements them gets the same faster `/workflow status`, resume and batched delivery receipts as the built-in SQL journal ([#3964](https://github.com/bastani-inc/atomic/issues/3964))
+
+### Fixed
+
+- A storage adapter loaded by path to open a workflow journal can now import `@bastani/atomic/workflows`. It gets the running Atomic's workflows module, so an `openRunJournal` hook can build on `SqlRunJournal` and its errors, such as `StaleLeaseError`, pass Atomic's `instanceof` checks. Before, the import resolved to a path inside `dist/index.js` and the journal failed with `Storage unavailable` ([#3963](https://github.com/bastani-inc/atomic/issues/3963))
+
 ## [0.9.44] - 2026-10-25
 
 Large agent swarms and pooled workflow fan-outs now run for hours without freezing the TUI, dropping messages or slowing down, and recover cleanly from a crash.

package/dist/builtin/intercom/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bastani/intercom",
-  "version": "0.9.44",
+  "version": "0.9.45",
   "private": true,
   "description": "Atomic extension providing a private coordination channel between parent and child agent sessions. Fork of: https://github.com/nicobailon/pi-intercom",
   "contributors": [

package/dist/builtin/mcp/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bastani/mcp",
-  "version": "0.9.44",
+  "version": "0.9.45",
   "private": true,
   "description": "Atomic's pi-compatible MCP integration for server connections, OAuth, tools, resources, and terminal management.",
   "contributors": [

package/dist/builtin/subagents/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bastani/subagents",
-  "version": "0.9.44",
+  "version": "0.9.45",
   "private": true,
   "description": "Atomic extension for delegating tasks to subagents with parallel execution. Fork of: https://github.com/nicobailon/pi-subagents",
   "contributors": [

package/dist/builtin/web-access/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bastani/web-access",
-  "version": "0.9.44",
+  "version": "0.9.45",
   "private": true,
   "description": "Atomic extension for web search, URL fetching, GitHub repo cloning, PDF/video extraction. Fork of: https://github.com/nicobailon/pi-web-access",
   "contributors": [

package/dist/builtin/workflows/CHANGELOG.md

@@ -6,6 +6,23 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 
 ## [Unreleased]
 
+## [0.9.45] - 2026-10-25
+
+### Fixed
+
+- A large local worker pool no longer fails at startup with `Local worker startup timed out after 30000 ms` on a small or busy machine. When `localWorkers.startupTimeoutMs` is not set, the readiness deadline now scales with the number of workers and the 1-minute load average per CPU core, from 30000 ms up to 600000 ms; on an idle machine with spare cores it stays 30000 ms. If workers still miss the deadline, the error now says they were too slow to start and names `workflows.localWorkers.startupTimeoutMs` ([#3962](https://github.com/bastani-inc/atomic/issues/3962))
+
+## [0.9.45-alpha.1] - 2026-10-25
+
+### Added
+
+- A run journal from a storage adapter, or any custom `RunJournal`, can now opt into the faster `/workflow status`, resume and batched delivery receipts that the built-in SQL journal uses, by implementing the optional `changeToken()`, `runVersions()`, `snapshotTail()` and `appendMany()` members. `registerRunJournalConformance` checks them when present ([#3964](https://github.com/bastani-inc/atomic/issues/3964))
+
+### Fixed
+
+- Journal adapters loaded by path (`ATOMIC_WORKFLOWS_JOURNAL`, `workflows.journal`, or a run's `journal: { adapter }`) can import `@bastani/atomic/workflows` and receive the workflows module that opened the journal, including its `SqlRunJournal` and journal error classes ([#3963](https://github.com/bastani-inc/atomic/issues/3963))
+- A large local worker pool no longer fails at startup with `Local worker startup timed out after 30000 ms` on a small or busy machine. When `localWorkers.startupTimeoutMs` is not set, the readiness deadline now scales with the number of workers and the 1-minute load average per CPU core, from 30000 ms up to 600000 ms; on an idle machine with spare cores it stays 30000 ms. If workers still miss the deadline, the error now says they were too slow to start and names `workflows.localWorkers.startupTimeoutMs` ([#3962](https://github.com/bastani-inc/atomic/issues/3962))
+
 ## [0.9.44] - 2026-10-25
 
 ### Fixed

package/dist/builtin/workflows/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bastani/workflows",
-  "version": "0.9.44",
+  "version": "0.9.45",
   "private": true,
   "description": "Atomic extension for multi-stage workflow authoring and execution.",
   "contributors": [

package/dist/builtin/workflows/src/durable/journal/internal.d.ts

@@ -1,50 +1,19 @@
 import type { SqliteExecutor } from "@bastani/atomic/storage";
-import { type AppendOutcome, type JournalEntry, type Lease, type RunEvent, type RunJournal } from "./types.js";
+import { type AppendOutcome, type Lease, type RunJournal, type RunJournalBatchAppend, type RunSnapshotTail, type RunVersion } from "./types.js";
+export type { RunJournalBatchAppend as JournalBatchAppend, RunSnapshotTail, RunVersion } from "./types.js";
 export declare const JOURNAL_READ_PAGE_SIZE = 256;
 /** Event types the durable backend folds without replacing the run's handle; every other event carries it. */
 export declare const NON_SNAPSHOT_EVENT_TYPES: readonly ["CheckpointRecorded", "MessageSettled", "MessageDelivered"];
-/** A run's latest event. A run whose history was deleted and recorded again ends with a different event. */
-export interface RunVersion {
-    readonly seq: number;
-    readonly at: number;
-}
-/**
- * A run's latest snapshot event (one carrying the run's handle) and the checkpoints recorded after it: enough to
- * list the run in a catalog without reading its history.
- */
-export interface RunSnapshotTail {
-    /** The run's latest event of any type. */
-    readonly version: RunVersion;
-    readonly snapshot: JournalEntry;
-    readonly laterCheckpoints: number;
-    /** `completedAt` of the last of `laterCheckpoints`. */
-    readonly lastCheckpointAt?: number;
-}
-/** Lets a durable backend re-read only the runs that changed; kept off the public journal shape. */
+/** Lets a durable backend re-read only the runs that changed. */
 export interface RunJournalVersions {
-    /** Moves whenever any connection, in this process or another, commits a change to the journal's store. */
     changeToken(): Promise<string>;
-    /** Every listed run with its latest event; a run with no events has version `{ seq: 0, at: 0 }`. */
     runVersions(): Promise<ReadonlyMap<string, RunVersion>>;
-    /** Undefined for a run with no snapshot event. */
     snapshotTail(runId: string): Promise<RunSnapshotTail | undefined>;
 }
-export declare function registerRunJournalVersions(journal: RunJournal, versions: RunJournalVersions): void;
-/** Change tracking for a journal that registered it; other journals are re-read in full. */
+/** Change tracking for a journal that implements all three members; other journals are re-read in full. */
 export declare function runJournalVersions(journal: RunJournal): RunJournalVersions | undefined;
-/**
- * One append of a batch. `context` runs a function in the async context the append was issued from; that context,
- * not the batch's, decides the append's pool append guard.
- */
-export interface JournalBatchAppend {
-    readonly event: RunEvent;
-    readonly key?: string;
-    readonly context: <R>(action: () => R) => R;
-}
-/** Appends to one run in one transaction; each append is settled on its own. Kept off the public journal shape. */
-export type RunJournalAppendMany = (lease: Lease, appends: readonly JournalBatchAppend[]) => Promise<readonly PromiseSettledResult<AppendOutcome>[]>;
-export declare function registerRunJournalAppendMany(journal: RunJournal, appendMany: RunJournalAppendMany): void;
-/** Batched appends for a journal that registered them; other journals append one event at a time. */
+export type RunJournalAppendMany = (lease: Lease, appends: readonly RunJournalBatchAppend[]) => Promise<readonly PromiseSettledResult<AppendOutcome>[]>;
+/** Batched appends for a journal that implements them; other journals append one event at a time. */
 export declare function runJournalAppendMany(journal: RunJournal): RunJournalAppendMany | undefined;
 export type JournalAppendGuard = (tx: SqliteExecutor, runId: string) => Promise<void>;
 export interface SqlPoolStorage {

package/dist/builtin/workflows/src/durable/journal/resolution.d.ts

@@ -9,6 +9,8 @@ export interface RunJournalSettings {
     };
     storage?: StorageSettingsSelection;
 }
+/** Sets the live modules a journal adapter imports by specifier; the workflows entry modules provide themselves. */
+export declare function provideRunJournalHostModules(modules: Readonly<Record<string, object>>): void;
 export interface RunJournalResolutionOptions {
     journal?: RunJournalOption;
     settings?: RunJournalSettings;

package/dist/builtin/workflows/src/durable/journal/sql-run-journal.d.ts

@@ -1,4 +1,5 @@
 import { type SqliteDatabase, type StorageDescriptor } from "@bastani/atomic/storage";
+import { type JournalBatchAppend, type RunSnapshotTail, type RunVersion } from "./internal.js";
 import { type AppendOutcome, type JournalEntry, type Lease, type LeaseOutcome, type LeaseState, type RunEvent, type RunHeader, type RunJournal, type RunSummary } from "./types.js";
 export { JOURNAL_READ_PAGE_SIZE } from "./internal.js";
 export declare class SqlRunJournal implements RunJournal {
@@ -27,13 +28,13 @@ export declare class SqlRunJournal implements RunJournal {
      * `data_version` moves when another connection commits and `total_changes()` when this one writes; the
      * table aggregates also catch run changes if the adapter does not serve every statement from one connection.
      */
-    private changeToken;
-    private runVersions;
+    changeToken(): Promise<string>;
+    runVersions(): Promise<ReadonlyMap<string, RunVersion>>;
     /**
      * The latest snapshot event is found by walking the run's `(run_id, seq)` key back from its last event, so
      * only the events after it are visited and only its payload is read.
      */
-    private snapshotTail;
+    snapshotTail(runId: string): Promise<RunSnapshotTail | undefined>;
     private requireRun;
     /** Whether this run exists only in the legacy tables, whose events are then read in place. */
     private isUnconvertedLegacyRun;
@@ -58,7 +59,7 @@ export declare class SqlRunJournal implements RunJournal {
      * one append's guard or payload failure is reported for it alone and the rest still commit. A lease or storage
      * failure rejects the whole batch, with nothing written.
      */
-    private appendMany;
+    appendMany(lease: Lease, appends: readonly JournalBatchAppend[]): Promise<readonly PromiseSettledResult<AppendOutcome>[]>;
     /** Validate an append outside any transaction; its payload error is raised where `append` always raised it. */
     private prepare;
     private insert;

package/dist/builtin/workflows/src/durable/journal/types.d.ts

@@ -37,7 +37,7 @@ export declare const JOURNAL_LEASE: {
     readonly absentEpoch: 0;
 };
 export type JournalLeaseSemantics = typeof JOURNAL_LEASE;
-export type { AppendOutcome, JournalEntry, Lease, LeaseOutcome, LeaseState, RunEvent, RunEventType, RunHeader, RunJournal, RunSummary, } from "@bastani/atomic/storage";
+export type { AppendOutcome, JournalEntry, Lease, LeaseOutcome, LeaseState, RunEvent, RunEventType, RunHeader, RunJournal, RunJournalBatchAppend, RunSnapshotTail, RunSummary, RunVersion, } from "@bastani/atomic/storage";
 export { RUN_EVENT_TYPES } from "@bastani/atomic/storage";
 export interface JournalLeaseRow {
     [JOURNAL_COLUMNS.holderId]: string | null;