mirror of
https://github.com/morluto/rea
synced 2026-10-26 06:16:17 -04:00
fix(environment): bound state reads and retain cleanup ownership (#6158)
* feat(environment): extract the initial state a capture's reads observed
Add the environment-state analysis view. From the parents the environment
manifest accepts, it extracts the data the capture's reads observed before
the first related write: entity types with an identity field (proved by a
path parameter or inferred from an identity name), entities with their
recorded values and provenance, documents for written non-entity paths, and
a render plan per read tool that is kept only when it reproduces the tool's
latest initial recordings from the state; every other read tool is a
recorded view with the reason. The Evidence is always inferred.
Response-body decoding moves to a domain owner shared with replay, exact
JSON parsing moves to the owner of the exact-number policy, the manifest and
state views share parent admission and record grouping, and
projectAnalysisView takes its accumulators as one object.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): bound state at redacted writes; infer identity only from items
A scenario write whose URL was redacted has no request line, so it was
dropped and reads after it counted as initial state. Such a successful
write now joins the timeline with an empty path, whose root relates to
every type, so it ends every type's state; a limitation counts them.
A list whose recordings were all empty passed the "every listed item
carries it" check vacuously and published an identity field nothing
showed; it now needs at least one item, else the view stays recorded
as no-identity.
A value that differed between two request lines of one item tool, such
as a query variant, was called volatile, and the gate then substituted
each recording's own value; only repeats of one request line now show
volatility, so differing lines go through fields-differ.
List membership is compared only while a list can still be the whole
collection, and merged entities are indexed by type and collection once,
so paged captures no longer cost pages times entities (8,000 pages of
40,000 entities: 5.7 s to 0.23 s).
Docs and the tool contract say scenario state keeps the capture's
redaction markers and how redacted writes bound it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): keep first sightings, one list pointer and linear field unions
Entities took the order of the first plan observed, and item plans run
before list plans, so a complete list read before its detail reads came
out reordered and lost "all" membership; each entity now keeps its
earliest recorded sighting.
A list plan reused the first recording's pointer for later recordings,
so a later response that was itself an entity wrapping a list was read
as that list; every recording must now locate the same list.
The field union scanned its ordered list for every key, quadratic in an
object's width; a set now answers membership.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): keep post-write values and partial lists out of the initial state
A value that changed between repeats of one request line was named
volatile whenever no write under the type's own root lay between, so a
write under another root (for example POST /api/dashboards/db before a
second GET /api/search) left its effect in the initial state. Any
recorded write between the two observations now leaves every tool of the
type recorded with reason fields-differ instead.
A list was published with membership "all" even when a total or count
envelope member exceeded the number it listed. Such a counter now makes
the membership "recorded".
Also drop a stale doc comment on ObservedEntity.order.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs(environment): scope item-shaped and deleted written paths out of documents
State documents are held only for list-shaped paths a POST, PUT or PATCH
tool writes. Reads of item-shaped written paths and of paths that are only
deleted stay recorded views; the guide now says so and the spec defers them
to a later increment of #4933.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): count writes in flight when a value changes between reads
The volatility check compared write start positions with the two reads,
so a write already in flight when the earlier read started was missed and
its effect was named volatile. A write now counts when its interval
overlaps the span from the earlier read's start to the later read's end.
The check also ran once per changed field and scanned every write. It now
runs once per pair of observations, answered by a binary search over the
writes in timeline order and the latest end among each prefix.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): catch cross-tool changes across writes; read only item totals
A value that changed across a write was caught only between repeats of
one tool's request line, so a list read before a write and an item read
after it put the post-write value into the initial state. Changes are
now classified per entity in timeline order after observation: across
any recorded write, a change between repeats of a line or between two
lines of the type rewrites the type (every tool recorded as
fields-differ); with no write between, a repeat's change is volatile.
The classification lives in environmentStateMerge.ts.
Any numeric member whose name contained "total" or "count" counted as
an item total, so total_pages, account_id or page_count made complete
lists partial. Item totals are now members whose name tokens include
total or count and none of page, pages or per; a string of digits counts
as a number; and a set has_more, next or next-page token member also
marks the list partial.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): keep post-write observations out of initial state
* fix(environment): retain field provenance and bound state inference work
* feat(environment): serve tool calls from the environment state
`environment-serve --state <state.json>` serves an environment from the
initial state an environment-state view extracted. Read tools whose views
are state-backed answer from a live state store: item views render the held
entity into the recorded envelope, list views render the collection or the
recorded members, and document views return the held value; a missing entity
or one lacking a required field is refused, and requests the state cannot
answer fall back to their recordings with the reason. The /agentenv data
plane implements data/reset, data/add (a rea-state/v1 document as a data
part, base64 bytes, a file:// path or an https URL, validated and replaced
atomically) and data/get, and the card advertises those operations and an
intake declaration. Stage 1 is read-only: write tools are not served in
state mode. Without --state nothing changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs(environment): record verification after merging the state view's review fixes
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(environment)!: serve the environment state by default
environment-serve derives the environment state at startup by running the
environment-state view in-process on the capture with the manifest's
template rules, so an environment comes with its state and a user supplies
nothing extra. The derived Evidence is the same one inspect-analysis-view
produces by hand. --state is removed; --replay-only serves exactly the
stateless replay of main. When the state cannot be derived, or its catalog
does not fit the discovery budget, the server starts in replay mode, logs
the reason and reports it in the stop summary's `state`. Exported bundles
serve the state by default too: export derives it the same way, so the
verifier's tool count and checked answer match what is served, and the
README describes the served mode.
BREAKING CHANGE: environment-serve answers read tools from the derived
state and does not serve write tools unless --replay-only is given; the
stop summary reports `state` instead of `state_evidence_id`.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): bound data-plane replies, keep batch order, cancel intake on stop
- One /agentenv response is held to 64 MiB: replies are measured as they
are produced and one past the budget becomes a -32000 error stating it.
- Every message of a batch enters the arrival-order queue on arrival, so
another request's reset never lands between them.
- Stopping the server, or a caller leaving, cancels the active download
and answers queued calls `cancelled` without running them.
- file:// URIs name only this host: another host or a network share is
refused before anything is opened.
- A loaded view value may not give an entity another identity.
- List answers refuse a member lacking a field every recording carried,
as item answers do.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): keep bodies too deep for the state's nesting out of the state
A recorded body at the JSON depth limit was accepted by extraction, then
nested a few levels deeper as a document or entity value, so sealing the
state Evidence threw and environment-serve and environment-export stopped
instead of falling back. stateBodyJson, the one rule extraction and
serving share, now treats a body nesting deeper than the limit minus 16
levels as unusable, so its tool stays a recorded view.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): measure data-plane replies before building them
83fcc739 bounded one /agentenv response to 64 MiB, but it still
serialized each reply whole before measuring it, so a batch of data/get
calls on a large state built every refused reply in full only to drop
it. Each reply is now measured in bounded JSON parts (the shared
jsonParts encoder), its parts kept only while it fits; a reply past the
budget is counted for its exact reply_bytes but never assembled. A test
checks that a refused read is never handed to JSON.stringify whole and
that its reply_bytes is exact.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): release batch replies as they are encoded; judge exact numbers by value
answerBatch kept every settled reply in its pending array until the
batch ended, so a batch alternating data/add and data/get held each
loaded state alive; each reply now leaves the queue as it is encoded.
The exact-number policy judged a number by its recorded spelling, so a
state loaded with 1e20 came back from data/get as
100000000000000000000, which data/add refused, breaking the promise
that data/get output loads back unchanged. A number is exact only when
both its recorded spelling and the one serialization writes are safe,
so integers from 2^53 below 1e21 are refused in any spelling; replay
keeps such bodies as recorded text. A domain test checks that every
accepted number is accepted again once written back.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): keep room for refusals and framing in the reply budget
The response budget counted accepted replies only: a budget error that
replaced a reply, and a batch's brackets and commas, were not charged,
so near 64 MiB later refusals pushed the body past the advertised
limit. Every reply still to come now keeps room for its own budget
error, and a batch keeps room for its framing, so a refusal always fits
and the body never passes the budget. Reply ids and which messages are
answered come from one place shared by the replies and the reserve.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): refuse a loaded state that data/get would write past the intake limit
data/add measured only the document it was given, while data/get
writes the state back with the members a document may omit (types, the
state Evidence id, view fields and record citations). A document just
under 32 MiB could therefore load and be reported larger than any intake
accepts, breaking the data/get then data/add round trip. replace now
measures the state as data/get writes it, part by part, and refuses it
past the limit, leaving the state unchanged. The limit moves to the
state document module, so the application store can read it; the
intake re-exports it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): keep merged export checks on the shared test graph
Main's export tests installed module mocks for the whole services
project, which reuses one module graph. A tarball read failure then
missed its importer, and a one-shot heap mock could be taken by another
file. Read those through bindings the tests replace for one case.
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(environment): keep a state file whose close failed owned until it closes
A file:// data/add that could not close its descriptor dropped the handle,
and an admission failure whose cleanup close failed dropped its owner, so
nothing could retry the close or report it.
The read now runs through withOwnedRegularFile, and a failed close is
retained in a server-lifetime ArtifactResourceScope. The next file read
retries it first and answers -32000 while it is still open; inline and
https documents still load. Closing the served environment retries it again
and rejects with the unfinished close.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(environment): attempt all state cleanup on close
* fix(environment): bound wide state JSON traversal
* fix(environment): scope cached state reply sizes to requests
* fix(domain): preserve JSON depth diagnostic guidance
---------
Co-authored-by: Edgar Arakelyan <edgar.arakelyan@scale.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
Co-authored-by: gensokyo-dev <gensokyo-dev@users.noreply.github.com>
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
This commit is contained in:
@@ -418,6 +418,30 @@ describe("EnvironmentStateStore", () => {
|
||||
});
|
||||
});
|
||||
|
||||
describe("caller-owned state replacement", () => {
|
||||
it("does not retain mutable data from a caller-owned replacement", () => {
|
||||
const store = new EnvironmentStateStore(state);
|
||||
const document = documentWith((value) => {
|
||||
const entity = value.rea_state.entities.find((item) => item.id === "2");
|
||||
if (!entity) throw new Error("Expected the user entity");
|
||||
entity.fields = { ...(entity.fields as object), login: "renamed" };
|
||||
});
|
||||
expect(store.replace(document).ok).toBe(true);
|
||||
const callerEntity = document.rea_state.entities.find(
|
||||
(entity) => entity.id === "2",
|
||||
);
|
||||
if (!callerEntity) throw new Error("Expected the caller's user entity");
|
||||
(callerEntity.fields as Record<string, unknown>).login = "later edit";
|
||||
|
||||
expect(store.snapshot().rea_state.entities).toContainEqual(
|
||||
expect.objectContaining({
|
||||
id: "2",
|
||||
fields: expect.objectContaining({ login: "renamed" }),
|
||||
}),
|
||||
);
|
||||
});
|
||||
});
|
||||
|
||||
describe("rea-state/v1 validation", () => {
|
||||
it.each([
|
||||
[
|
||||
@@ -570,3 +594,36 @@ describe("rea-state/v1 validation", () => {
|
||||
).toMatchObject({ path: "rea_state/version" });
|
||||
});
|
||||
});
|
||||
|
||||
describe("wide state value validation", () => {
|
||||
it("preserves a wide caller value through the state replacement boundary", () => {
|
||||
const wide = Array.from({ length: 200_000 }, (_, index) => index % 10);
|
||||
const wideObject = Object.fromEntries(
|
||||
Array.from({ length: 20_000 }, (_, index) => [
|
||||
`field_${String(index)}`,
|
||||
index,
|
||||
]),
|
||||
);
|
||||
const document = documentWith((value) => {
|
||||
const fields = value.rea_state.entities[0]!.fields as Record<
|
||||
string,
|
||||
unknown
|
||||
>;
|
||||
value.rea_state.entities[0]!.fields = {
|
||||
...fields,
|
||||
caller_wide: wide,
|
||||
caller_wide_object: wideObject,
|
||||
};
|
||||
});
|
||||
const store = new EnvironmentStateStore(state);
|
||||
|
||||
expect(store.replace(document).ok).toBe(true);
|
||||
const retained = store.snapshot().rea_state.entities[0]?.fields.caller_wide;
|
||||
if (!Array.isArray(retained)) throw new Error("Expected the wide value");
|
||||
expect(retained).toHaveLength(wide.length);
|
||||
expect(retained.at(-1)).toBe(wide.at(-1));
|
||||
const retainedObject =
|
||||
store.snapshot().rea_state.entities[0]?.fields.caller_wide_object;
|
||||
expect(retainedObject).toEqual(wideObject);
|
||||
});
|
||||
});
|
||||
|
||||
@@ -1,3 +1,4 @@
|
||||
import { types } from "node:util";
|
||||
import { z } from "zod";
|
||||
|
||||
import {
|
||||
@@ -16,7 +17,11 @@ import {
|
||||
} from "../../domain/analysisErrorCore.js";
|
||||
import { canonicalJson } from "../../domain/comparisonSemantics.js";
|
||||
import { projectInputIssues } from "../../domain/inputIssueProjection.js";
|
||||
import { jsonValueSchema } from "../../domain/jsonValue.js";
|
||||
import {
|
||||
jsonValueSchema,
|
||||
ownedJsonValueSchema,
|
||||
type JsonValue,
|
||||
} from "../../domain/jsonValue.js";
|
||||
import { err, ok, type Result } from "../../domain/result.js";
|
||||
import { compositeKey } from "../../domain/unicodeCodePointOrder.js";
|
||||
import type { EnvironmentState } from "./environmentState.js";
|
||||
@@ -37,37 +42,42 @@ export interface StateContent {
|
||||
readonly documents: readonly HeldDocument[];
|
||||
}
|
||||
|
||||
const fieldsSchema = z.record(z.string(), jsonValueSchema);
|
||||
const ordinalSchema = z.number().int().nonnegative();
|
||||
const stateDocumentSchemaFor = (jsonValue: z.ZodType<JsonValue, unknown>) => {
|
||||
const fieldsSchema = z.record(z.string(), jsonValue);
|
||||
const ordinalSchema = z.number().int().nonnegative();
|
||||
|
||||
const stateDocumentSchema = z.strictObject({
|
||||
rea_state: z.strictObject({
|
||||
version: z.literal(1),
|
||||
// `data/get` reports these; accepting them back lets its output be loaded unchanged.
|
||||
state_evidence_id: z.string().optional(),
|
||||
types: z.array(jsonValueSchema).optional(),
|
||||
entities: z.array(
|
||||
z.strictObject({
|
||||
type: z.string(),
|
||||
collection: z.string(),
|
||||
id: z.string(),
|
||||
fields: fieldsSchema,
|
||||
view_fields: z
|
||||
.array(z.strictObject({ tool: z.string(), fields: fieldsSchema }))
|
||||
.optional(),
|
||||
record_ordinals: z.array(ordinalSchema).optional(),
|
||||
}),
|
||||
),
|
||||
documents: z.array(
|
||||
z.strictObject({
|
||||
tool: z.string(),
|
||||
path: z.string(),
|
||||
value: jsonValueSchema,
|
||||
record_ordinal: ordinalSchema.optional(),
|
||||
}),
|
||||
),
|
||||
}),
|
||||
});
|
||||
return z.strictObject({
|
||||
rea_state: z.strictObject({
|
||||
version: z.literal(1),
|
||||
// `data/get` reports these; accepting them back lets its output be loaded unchanged.
|
||||
state_evidence_id: z.string().optional(),
|
||||
types: z.array(jsonValue).optional(),
|
||||
entities: z.array(
|
||||
z.strictObject({
|
||||
type: z.string(),
|
||||
collection: z.string(),
|
||||
id: z.string(),
|
||||
fields: fieldsSchema,
|
||||
view_fields: z
|
||||
.array(z.strictObject({ tool: z.string(), fields: fieldsSchema }))
|
||||
.optional(),
|
||||
record_ordinals: z.array(ordinalSchema).optional(),
|
||||
}),
|
||||
),
|
||||
documents: z.array(
|
||||
z.strictObject({
|
||||
tool: z.string(),
|
||||
path: z.string(),
|
||||
value: jsonValue,
|
||||
record_ordinal: ordinalSchema.optional(),
|
||||
}),
|
||||
),
|
||||
}),
|
||||
});
|
||||
};
|
||||
|
||||
const stateDocumentSchema = stateDocumentSchemaFor(jsonValueSchema);
|
||||
const ownedStateDocumentSchema = stateDocumentSchemaFor(ownedJsonValueSchema);
|
||||
|
||||
type ParsedDocument = z.output<typeof stateDocumentSchema>["rea_state"];
|
||||
type ParsedEntity = ParsedDocument["entities"][number];
|
||||
@@ -105,31 +115,82 @@ const memberProblem = (path: Path): AnalysisInputIssue => {
|
||||
|
||||
/** The path of the first `__proto__` member in document order; schema parsing would drop such a member rather than refuse it. */
|
||||
const prototypeMemberPath = (root: unknown): Path | undefined => {
|
||||
interface Node {
|
||||
interface Cursor {
|
||||
readonly value: unknown;
|
||||
readonly parent: Node | undefined;
|
||||
readonly parent: Cursor | undefined;
|
||||
readonly key: string | number;
|
||||
readonly entries?: Generator<string>;
|
||||
index: number;
|
||||
}
|
||||
const pathOf = (node: Node): (string | number)[] => {
|
||||
const pathOf = (node: Cursor): (string | number)[] => {
|
||||
const path: (string | number)[] = [];
|
||||
for (let at: Node = node; at.parent !== undefined; at = at.parent)
|
||||
for (let at: Cursor = node; at.parent !== undefined; at = at.parent)
|
||||
path.push(at.key);
|
||||
return path.reverse();
|
||||
};
|
||||
const pending: Node[] = [{ value: root, parent: undefined, key: "" }];
|
||||
for (let node = pending.pop(); node !== undefined; node = pending.pop()) {
|
||||
function* ownKeys(value: Record<string, unknown>): Generator<string> {
|
||||
// `for...in` is lazy. The ownership check keeps inherited enumerable
|
||||
// properties out, matching Object.entries without allocating all entries.
|
||||
for (const key in value) if (Object.hasOwn(value, key)) yield key;
|
||||
}
|
||||
const cursor = (
|
||||
value: unknown,
|
||||
parent: Cursor | undefined,
|
||||
key: string | number,
|
||||
): Cursor => ({
|
||||
value,
|
||||
parent,
|
||||
key,
|
||||
index: 0,
|
||||
...(!Array.isArray(value) && typeof value === "object" && value !== null
|
||||
? { entries: ownKeys(value as Record<string, unknown>) }
|
||||
: {}),
|
||||
});
|
||||
const pending: Cursor[] = [cursor(root, undefined, "")];
|
||||
while (pending.length > 0) {
|
||||
const node = pending[pending.length - 1]!;
|
||||
const { value } = node;
|
||||
if (typeof value !== "object" || value === null) continue;
|
||||
if (typeof value !== "object" || value === null) {
|
||||
pending.pop();
|
||||
continue;
|
||||
}
|
||||
// The canonical JSON validator reports proxies without invoking their traps.
|
||||
if (types.isProxy(value)) {
|
||||
pending.pop();
|
||||
continue;
|
||||
}
|
||||
const array = Array.isArray(value);
|
||||
if (!array && Object.hasOwn(value, "__proto__"))
|
||||
return [...pathOf(node), "__proto__"];
|
||||
const entries = Object.entries(value);
|
||||
for (let index = entries.length - 1; index >= 0; index -= 1)
|
||||
pending.push({
|
||||
value: entries[index]![1],
|
||||
parent: node,
|
||||
key: array ? index : entries[index]![0],
|
||||
});
|
||||
if (array) {
|
||||
if (node.index >= value.length) {
|
||||
pending.pop();
|
||||
continue;
|
||||
}
|
||||
const index = node.index++;
|
||||
const descriptor = Object.getOwnPropertyDescriptor(value, index);
|
||||
if (
|
||||
descriptor?.enumerable !== true ||
|
||||
!Object.hasOwn(descriptor, "value")
|
||||
)
|
||||
continue;
|
||||
const child: unknown = descriptor.value;
|
||||
if (typeof child === "object" && child !== null)
|
||||
pending.push(cursor(child, node, index));
|
||||
continue;
|
||||
}
|
||||
const next = node.entries?.next();
|
||||
if (next === undefined || next.done) {
|
||||
pending.pop();
|
||||
continue;
|
||||
}
|
||||
const key = next.value;
|
||||
const descriptor = Object.getOwnPropertyDescriptor(value, key);
|
||||
if (descriptor?.enumerable !== true || !Object.hasOwn(descriptor, "value"))
|
||||
continue;
|
||||
const child: unknown = descriptor.value;
|
||||
if (typeof child === "object" && child !== null)
|
||||
pending.push(cursor(child, node, key));
|
||||
}
|
||||
return undefined;
|
||||
};
|
||||
@@ -365,9 +426,12 @@ const heldContent = (
|
||||
export const parseStateDocument = (
|
||||
state: EnvironmentState,
|
||||
value: unknown,
|
||||
owned = false,
|
||||
): Result<StateContent, AnalysisInputError> => {
|
||||
const prototype = prototypeMemberPath(value);
|
||||
const parsed = stateDocumentSchema.safeParse(value);
|
||||
const parsed = (
|
||||
owned ? ownedStateDocumentSchema : stateDocumentSchema
|
||||
).safeParse(value);
|
||||
const initial = new Map(
|
||||
state.entities.map((entity) => [entityKey(entity), entity]),
|
||||
);
|
||||
|
||||
@@ -58,9 +58,21 @@ export class EnvironmentStateStore {
|
||||
this.#install(this.state);
|
||||
}
|
||||
|
||||
/** Validate a `rea-state/v1` document and, only when it is valid, make it the whole state. */
|
||||
/** Validate a caller-owned document and copy its JSON values before retaining them. */
|
||||
replace(document: unknown): Result<void, AnalysisInputError> {
|
||||
const content = parseStateDocument(this.state, document);
|
||||
return this.#replace(document, false);
|
||||
}
|
||||
|
||||
/** Validate and retain a JSON document created exclusively by the serving boundary. */
|
||||
replaceOwnedDocument(document: unknown): Result<void, AnalysisInputError> {
|
||||
return this.#replace(document, true);
|
||||
}
|
||||
|
||||
#replace(
|
||||
document: unknown,
|
||||
owned: boolean,
|
||||
): Result<void, AnalysisInputError> {
|
||||
const content = parseStateDocument(this.state, document, owned);
|
||||
if (!content.ok) return content;
|
||||
// data/get writes members a document may omit, so the state it reports must still be one data/add loads.
|
||||
if (jsonBytes(this.#snapshotOf(content.value)) > STATE_INTAKE_LIMIT_BYTES)
|
||||
|
||||
@@ -64,7 +64,9 @@ describe("jsonValueSchema depth bound", () => {
|
||||
const result = jsonValueSchema.safeParse(deepObject(MAX_JSON_DEPTH + 1));
|
||||
expect(result.success).toBe(false);
|
||||
if (result.success) return;
|
||||
expect(result.error.issues[0]?.message).toContain("maximum nesting depth");
|
||||
expect(result.error.issues[0]?.message).toBe(
|
||||
`JSON value exceeds maximum nesting depth of ${MAX_JSON_DEPTH}. This release has no runtime override.`,
|
||||
);
|
||||
});
|
||||
|
||||
it.each(["object", "array"])(
|
||||
|
||||
+116
-54
@@ -27,6 +27,10 @@ export const jsonValueValidationIssue = (root: unknown): string | undefined => {
|
||||
}
|
||||
};
|
||||
|
||||
function* ownEnumerableKeys(value: Record<string, unknown>): Generator<string> {
|
||||
for (const key in value) if (Object.hasOwn(value, key)) yield key;
|
||||
}
|
||||
|
||||
/** Validate JSON in bounded steps; cache only authenticated immutable roots. */
|
||||
export function* jsonValueValidationSteps(
|
||||
root: unknown,
|
||||
@@ -41,57 +45,104 @@ export function* jsonValueValidationSteps(
|
||||
if (immutable !== undefined) immutableValidation.set(immutable, issue);
|
||||
return issue;
|
||||
};
|
||||
const pending: { value: unknown; depth: number }[] = [
|
||||
{ value: root, depth: 0 },
|
||||
];
|
||||
type Frame =
|
||||
| {
|
||||
readonly kind: "array";
|
||||
readonly value: unknown[];
|
||||
readonly depth: number;
|
||||
index: number;
|
||||
}
|
||||
| {
|
||||
readonly kind: "object";
|
||||
readonly value: Record<string, unknown>;
|
||||
readonly depth: number;
|
||||
readonly keys: Generator<string>;
|
||||
};
|
||||
const pending: Frame[] = [];
|
||||
let value = root;
|
||||
let depth = 0;
|
||||
let inspect = true;
|
||||
let examined = 0;
|
||||
while (pending.length > 0) {
|
||||
if (examined++ === 4096) {
|
||||
examined = 0;
|
||||
yield;
|
||||
}
|
||||
const item = pending.pop();
|
||||
if (item === undefined) break;
|
||||
if (item.depth > MAX_JSON_DEPTH)
|
||||
return finish(
|
||||
`JSON value exceeds maximum nesting depth of ${MAX_JSON_DEPTH}. This release has no runtime override.`,
|
||||
);
|
||||
const value = item.value;
|
||||
if (typeof value === "object" && value !== null && types.isProxy(value))
|
||||
return finish("JSON values cannot contain proxies");
|
||||
if (
|
||||
value === null ||
|
||||
typeof value === "string" ||
|
||||
typeof value === "boolean"
|
||||
)
|
||||
for (;;) {
|
||||
if (inspect) {
|
||||
if (examined++ === 4096) {
|
||||
examined = 0;
|
||||
yield;
|
||||
}
|
||||
if (depth > MAX_JSON_DEPTH)
|
||||
return finish(
|
||||
`JSON value exceeds maximum nesting depth of ${MAX_JSON_DEPTH}. This release has no runtime override.`,
|
||||
);
|
||||
if (typeof value === "object" && value !== null && types.isProxy(value))
|
||||
return finish("JSON values cannot contain proxies");
|
||||
if (
|
||||
value === null ||
|
||||
typeof value === "string" ||
|
||||
typeof value === "boolean"
|
||||
) {
|
||||
inspect = false;
|
||||
} else if (typeof value === "number") {
|
||||
if (!Number.isFinite(value))
|
||||
return finish("JSON numbers must be finite");
|
||||
inspect = false;
|
||||
} else if (Array.isArray(value)) {
|
||||
pending.push({
|
||||
kind: "array",
|
||||
value,
|
||||
depth,
|
||||
index: value.length - 1,
|
||||
});
|
||||
inspect = false;
|
||||
} else if (isRecord(value)) {
|
||||
const prototype: unknown = Object.getPrototypeOf(value);
|
||||
if (prototype !== Object.prototype && prototype !== null)
|
||||
return finish("JSON objects require ordinary data properties");
|
||||
for (const key of Object.getOwnPropertySymbols(value))
|
||||
if (Object.getOwnPropertyDescriptor(value, key)?.enumerable === true)
|
||||
return finish("JSON object keys must be strings");
|
||||
pending.push({
|
||||
kind: "object",
|
||||
value,
|
||||
depth,
|
||||
keys: ownEnumerableKeys(value),
|
||||
});
|
||||
inspect = false;
|
||||
} else return finish("Invalid JSON value");
|
||||
continue;
|
||||
if (typeof value === "number") {
|
||||
if (!Number.isFinite(value)) return finish("JSON numbers must be finite");
|
||||
} else if (Array.isArray(value)) {
|
||||
for (let index = 0; index < value.length; index += 1) {
|
||||
const descriptor = Object.getOwnPropertyDescriptor(value, index);
|
||||
if (descriptor === undefined || !Object.hasOwn(descriptor, "value"))
|
||||
return finish(
|
||||
"JSON arrays require dense data properties without accessors",
|
||||
);
|
||||
pending.push({ value: descriptor.value, depth: item.depth + 1 });
|
||||
}
|
||||
|
||||
const frame = pending[pending.length - 1];
|
||||
if (frame === undefined) return finish(undefined);
|
||||
if (frame.kind === "array") {
|
||||
if (frame.index < 0) {
|
||||
pending.pop();
|
||||
continue;
|
||||
}
|
||||
} else if (isRecord(value)) {
|
||||
const prototype: unknown = Object.getPrototypeOf(value);
|
||||
if (prototype !== Object.prototype && prototype !== null)
|
||||
return finish("JSON objects require ordinary data properties");
|
||||
for (const key of Reflect.ownKeys(value)) {
|
||||
const descriptor = Object.getOwnPropertyDescriptor(value, key);
|
||||
if (descriptor?.enumerable !== true) continue;
|
||||
if (typeof key !== "string")
|
||||
return finish("JSON object keys must be strings");
|
||||
if (!Object.hasOwn(descriptor, "value"))
|
||||
return finish("JSON objects cannot contain accessors");
|
||||
pending.push({ value: descriptor.value, depth: item.depth + 1 });
|
||||
}
|
||||
} else return finish("Invalid JSON value");
|
||||
const index = frame.index--;
|
||||
const descriptor = Object.getOwnPropertyDescriptor(frame.value, index);
|
||||
if (descriptor === undefined || !Object.hasOwn(descriptor, "value"))
|
||||
return finish(
|
||||
"JSON arrays require dense data properties without accessors",
|
||||
);
|
||||
value = descriptor.value;
|
||||
depth = frame.depth + 1;
|
||||
inspect = true;
|
||||
continue;
|
||||
}
|
||||
|
||||
const next = frame.keys.next();
|
||||
if (next.done) {
|
||||
pending.pop();
|
||||
continue;
|
||||
}
|
||||
const descriptor = Object.getOwnPropertyDescriptor(frame.value, next.value);
|
||||
if (descriptor?.enumerable !== true) continue;
|
||||
if (!Object.hasOwn(descriptor, "value"))
|
||||
return finish("JSON objects cannot contain accessors");
|
||||
value = descriptor.value;
|
||||
depth = frame.depth + 1;
|
||||
inspect = true;
|
||||
}
|
||||
return finish(undefined);
|
||||
}
|
||||
|
||||
const preserveJsonValue = (value: JsonValue): JsonValue => {
|
||||
@@ -103,13 +154,17 @@ const preserveJsonValue = (value: JsonValue): JsonValue => {
|
||||
copy.push(preserveJsonValue(value[index]!));
|
||||
return copy;
|
||||
}
|
||||
if (value !== null && typeof value === "object")
|
||||
return Object.fromEntries<JsonValue>(
|
||||
Object.entries(value).map(([key, child]) => [
|
||||
key,
|
||||
preserveJsonValue(child),
|
||||
]),
|
||||
);
|
||||
if (value !== null && typeof value === "object") {
|
||||
const copy: Record<string, JsonValue> = {};
|
||||
for (const key of ownEnumerableKeys(value))
|
||||
Object.defineProperty(copy, key, {
|
||||
value: preserveJsonValue(value[key]!),
|
||||
enumerable: true,
|
||||
configurable: true,
|
||||
writable: true,
|
||||
});
|
||||
return copy;
|
||||
}
|
||||
return value;
|
||||
};
|
||||
|
||||
@@ -132,6 +187,13 @@ export const jsonValueSchema = z
|
||||
.transform(preserveJsonValue);
|
||||
jsonValueSchema._zod.toJSONSchema = () => ({});
|
||||
|
||||
/** Validate JSON that the current boundary exclusively owns, without a second full-tree copy. */
|
||||
export const ownedJsonValueSchema = z
|
||||
.unknown()
|
||||
.nonoptional()
|
||||
.pipe(validatedJsonValue);
|
||||
ownedJsonValueSchema._zod.toJSONSchema = () => ({});
|
||||
|
||||
const objectValidation = z.custom<Record<string, JsonValue>>(
|
||||
isRecord,
|
||||
"Expected JSON object",
|
||||
|
||||
@@ -60,7 +60,7 @@ const addState = async (
|
||||
if (!document.ok) return { error: document.error };
|
||||
// A file read is not interruptible, so a cancellation during it is honoured here, before the state changes.
|
||||
if (cancel.aborted) return { error: CANCELLED };
|
||||
const replaced = store.replace(document.value);
|
||||
const replaced = store.replaceOwnedDocument(document.value);
|
||||
return replaced.ok
|
||||
? { result: {} }
|
||||
: { error: documentError(replaced.error) };
|
||||
|
||||
@@ -193,6 +193,65 @@ const startRpc = (
|
||||
);
|
||||
};
|
||||
|
||||
interface CachedStateReplySize {
|
||||
readonly types: unknown;
|
||||
readonly documents: unknown;
|
||||
readonly bytes: number;
|
||||
}
|
||||
|
||||
// Repeated `data/get` snapshots have fresh wrappers but share the installed
|
||||
// state's member arrays. The request-scoped cache avoids retaining stale sizes.
|
||||
const cachedStateReplyResultBytes = (
|
||||
stateReplySizes: WeakMap<object, CachedStateReplySize>,
|
||||
reply: RpcReply,
|
||||
): number | undefined => {
|
||||
if (!("result" in reply) || !isRecord(reply.result)) return undefined;
|
||||
const parts = reply.result.parts;
|
||||
if (!Array.isArray(parts) || parts.length !== 1 || !isRecord(parts[0]))
|
||||
return undefined;
|
||||
const data = parts[0].data;
|
||||
if (!isRecord(data) || !isRecord(data.rea_state)) return undefined;
|
||||
const { types, entities, documents } = data.rea_state;
|
||||
if (!Array.isArray(entities) || !Array.isArray(documents)) return undefined;
|
||||
const cached = stateReplySizes.get(entities);
|
||||
return cached !== undefined &&
|
||||
cached.types === types &&
|
||||
cached.documents === documents
|
||||
? cached.bytes
|
||||
: undefined;
|
||||
};
|
||||
|
||||
const rememberStateReplyResultBytes = (
|
||||
stateReplySizes: WeakMap<object, CachedStateReplySize>,
|
||||
reply: RpcReply,
|
||||
replyBytes: number,
|
||||
): void => {
|
||||
if (!("result" in reply) || !isRecord(reply.result)) return;
|
||||
const parts = reply.result.parts;
|
||||
if (!Array.isArray(parts) || parts.length !== 1 || !isRecord(parts[0]))
|
||||
return;
|
||||
const data = parts[0].data;
|
||||
if (!isRecord(data) || !isRecord(data.rea_state)) return;
|
||||
const { types, entities, documents } = data.rea_state;
|
||||
if (!Array.isArray(entities) || !Array.isArray(documents)) return;
|
||||
// A one-character string marks the result slot. Removing its compact JSON
|
||||
// envelope yields the exact result size, independent of the RPC id.
|
||||
const markerBytes = Buffer.byteLength(
|
||||
JSON.stringify({ jsonrpc: "2.0", id: reply.id, result: "x" }),
|
||||
"utf8",
|
||||
);
|
||||
const bytes = replyBytes - markerBytes + 3;
|
||||
stateReplySizes.set(entities, { types, documents, bytes });
|
||||
};
|
||||
|
||||
const replyBytesFor = (reply: RpcReply, resultBytes: number): number =>
|
||||
resultBytes +
|
||||
Buffer.byteLength(
|
||||
JSON.stringify({ jsonrpc: "2.0", id: reply.id, result: "x" }),
|
||||
"utf8",
|
||||
) -
|
||||
3;
|
||||
|
||||
const overBudget = (replyBytes: number): RpcError => ({
|
||||
code: SERVER_ERROR,
|
||||
message: `the reply would take this response past its ${String(MAX_RPC_RESPONSE_BYTES)}-byte budget; the call ran, so repeat a read in another request`,
|
||||
@@ -224,11 +283,30 @@ const replyEncoder = (
|
||||
const refusals = messages.filter(expectsReply).map(refusalBytes);
|
||||
let reserved = refusals.reduce((sum, bytes) => sum + bytes, 0);
|
||||
let used = framed ? 2 + Math.max(0, refusals.length - 1) : 0;
|
||||
const stateReplySizes = new WeakMap<object, CachedStateReplySize>();
|
||||
let index = 0;
|
||||
return (reply) => {
|
||||
reserved -= refusals[index] ?? 0;
|
||||
index += 1;
|
||||
const room = MAX_RPC_RESPONSE_BYTES - used - reserved;
|
||||
const cachedResultBytes = cachedStateReplyResultBytes(
|
||||
stateReplySizes,
|
||||
reply,
|
||||
);
|
||||
const knownBytes =
|
||||
cachedResultBytes === undefined
|
||||
? undefined
|
||||
: replyBytesFor(reply, cachedResultBytes);
|
||||
if (knownBytes !== undefined && knownBytes > room) {
|
||||
const refused = JSON.stringify({
|
||||
jsonrpc: "2.0",
|
||||
id: reply.id,
|
||||
error: overBudget(knownBytes),
|
||||
});
|
||||
const refusedBytes = Buffer.byteLength(refused, "utf8");
|
||||
used += refusedBytes;
|
||||
return refused;
|
||||
}
|
||||
const parts: string[] = [];
|
||||
let bytes = 0;
|
||||
for (const part of bufferedJsonParts(jsonParts(reply))) {
|
||||
@@ -236,9 +314,13 @@ const replyEncoder = (
|
||||
if (bytes <= room) parts.push(part);
|
||||
}
|
||||
if (bytes <= room) {
|
||||
if (cachedResultBytes === undefined)
|
||||
rememberStateReplyResultBytes(stateReplySizes, reply, bytes);
|
||||
used += bytes;
|
||||
return parts.join("");
|
||||
}
|
||||
if (cachedResultBytes === undefined)
|
||||
rememberStateReplyResultBytes(stateReplySizes, reply, bytes);
|
||||
const refused = JSON.stringify({
|
||||
jsonrpc: "2.0",
|
||||
id: reply.id,
|
||||
@@ -418,8 +500,23 @@ export const serveEnvironment = async (
|
||||
},
|
||||
close: async () => {
|
||||
stop();
|
||||
await served.close();
|
||||
await stateFiles.close();
|
||||
const failures: unknown[] = [];
|
||||
try {
|
||||
await served.close();
|
||||
} catch (cause: unknown) {
|
||||
failures.push(cause);
|
||||
}
|
||||
try {
|
||||
await stateFiles.close();
|
||||
} catch (cause: unknown) {
|
||||
failures.push(cause);
|
||||
}
|
||||
if (failures.length === 1) throw failures[0];
|
||||
if (failures.length > 1)
|
||||
throw new AggregateError(
|
||||
failures,
|
||||
"Environment server and state-file cleanup both failed",
|
||||
);
|
||||
},
|
||||
});
|
||||
};
|
||||
|
||||
@@ -7,6 +7,32 @@ import { join } from "node:path";
|
||||
import { pathToFileURL } from "node:url";
|
||||
import { afterEach, describe, expect, it, onTestFinished, vi } from "vitest";
|
||||
|
||||
const closeFailureControl = vi.hoisted(() => ({
|
||||
failure: null as Error | null,
|
||||
}));
|
||||
|
||||
vi.mock("../http/serveStreamableHttp.js", async (importOriginal) => {
|
||||
const original =
|
||||
await importOriginal<typeof import("../http/serveStreamableHttp.js")>();
|
||||
return {
|
||||
...original,
|
||||
serveStreamableHttp: async (
|
||||
...args: Parameters<typeof original.serveStreamableHttp>
|
||||
) => {
|
||||
const served = await original.serveStreamableHttp(...args);
|
||||
return {
|
||||
...served,
|
||||
close: async () => {
|
||||
await served.close();
|
||||
const failure = closeFailureControl.failure;
|
||||
closeFailureControl.failure = null;
|
||||
if (failure !== null) throw failure;
|
||||
},
|
||||
};
|
||||
},
|
||||
};
|
||||
});
|
||||
|
||||
import { timedCaptureEvidence } from "../../../tests/fixtures/analysisView.js";
|
||||
import {
|
||||
environmentStateServeFixture,
|
||||
@@ -672,6 +698,7 @@ describe("state data plane response budget", () => {
|
||||
typeof value === "object" &&
|
||||
value !== null &&
|
||||
"result" in value &&
|
||||
typeof (value as { result?: unknown }).result === "object" &&
|
||||
(value as { id?: unknown }).id === 3,
|
||||
),
|
||||
).toBe(false);
|
||||
@@ -681,6 +708,56 @@ describe("state data plane response budget", () => {
|
||||
});
|
||||
});
|
||||
|
||||
describe("repeated state reads", () => {
|
||||
it("reuses the encoded size of a state for refused repeated data/get calls", async () => {
|
||||
const plane = dataPlane();
|
||||
const document = renamedDocument();
|
||||
const large = {
|
||||
rea_state: {
|
||||
...document.rea_state,
|
||||
entities: document.rea_state.entities.map((entity) =>
|
||||
entity.type === "/api/users" && entity.id === "2"
|
||||
? {
|
||||
...entity,
|
||||
fields: { ...entity.fields, large: "x".repeat(1024 * 1024) },
|
||||
}
|
||||
: entity,
|
||||
),
|
||||
},
|
||||
};
|
||||
await plane.add({ kind: "data", data: large });
|
||||
const fields = plane.store
|
||||
.snapshot()
|
||||
.rea_state.entities.find((entity) => entity.id === "2")?.fields;
|
||||
if (fields === undefined) throw new Error("Expected the loaded entity");
|
||||
|
||||
const originalKeys = Object.keys;
|
||||
let fieldEnumerations = 0;
|
||||
const keys = vi.spyOn(Object, "keys").mockImplementation((value) => {
|
||||
if (value === fields) fieldEnumerations += 1;
|
||||
return originalKeys(value);
|
||||
});
|
||||
try {
|
||||
const batch = await plane.post(
|
||||
Array.from({ length: 100 }, (_, index) => ({
|
||||
jsonrpc: "2.0",
|
||||
id: index + 1,
|
||||
method: "data/get",
|
||||
})),
|
||||
);
|
||||
expect(
|
||||
(batch.json as { error?: unknown }[]).filter((reply) => reply.error),
|
||||
).not.toHaveLength(0);
|
||||
expect(fieldEnumerations).toBeLessThan(65);
|
||||
expect(Buffer.byteLength(JSON.stringify(batch.json))).toBeLessThanOrEqual(
|
||||
MAX_RPC_RESPONSE_BYTES,
|
||||
);
|
||||
} finally {
|
||||
keys.mockRestore();
|
||||
}
|
||||
});
|
||||
});
|
||||
|
||||
describe("state data plane response budget reserve", () => {
|
||||
it("keeps room for every later refusal and the batch's framing, so the body never passes the budget", async () => {
|
||||
const document = renamedDocument();
|
||||
@@ -891,7 +968,47 @@ describe("serveEnvironment with a state", () => {
|
||||
await expect(environment.close()).resolves.toBeUndefined();
|
||||
expect(owners[0]?.handle.fd).toBe(-1);
|
||||
});
|
||||
});
|
||||
|
||||
describe("serveEnvironment cleanup retries", () => {
|
||||
it("attempts state-file cleanup after listener failure and keeps both failures retryable", async () => {
|
||||
const environment = await serveState({ source, state });
|
||||
const uri = await renamedStateUri();
|
||||
let failing = true;
|
||||
const owners = failFirstFileClose(() => failing);
|
||||
const response = await fetch(
|
||||
new URL(ENVIRONMENT_RPC_PATH, environment.url),
|
||||
{
|
||||
method: "POST",
|
||||
headers: { "content-type": "application/json" },
|
||||
body: JSON.stringify(addByUri(1, uri)),
|
||||
},
|
||||
);
|
||||
await expect(response.json()).resolves.toMatchObject({
|
||||
error: {
|
||||
code: -32000,
|
||||
message: expect.stringContaining("could not be closed"),
|
||||
},
|
||||
});
|
||||
|
||||
const listenerFailure = new Error("injected listener cleanup failure");
|
||||
closeFailureControl.failure = listenerFailure;
|
||||
await expect(environment.close()).rejects.toSatisfy((cause: unknown) => {
|
||||
if (!(cause instanceof AggregateError)) return false;
|
||||
return (
|
||||
cause.errors[0] === listenerFailure &&
|
||||
cause.errors[1] instanceof ArtifactResourceRetryFailure
|
||||
);
|
||||
});
|
||||
expect(owners[0]?.handle.fd).toBeGreaterThanOrEqual(0);
|
||||
|
||||
failing = false;
|
||||
await expect(environment.close()).resolves.toBeUndefined();
|
||||
expect(owners[0]?.handle.fd).toBe(-1);
|
||||
});
|
||||
});
|
||||
|
||||
describe("serveEnvironment with a state", () => {
|
||||
it("serves query POSTs from their recordings and leaves state-changing writes out", async () => {
|
||||
const environment = await serveState(
|
||||
environmentStateServeInputs(
|
||||
|
||||
Reference in New Issue
Block a user