mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-06 05:56:19 +01:00
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - The board reads API data and keeps the current task on screen. > - A restart or temporary network failure can interrupt those reads. > - Separate page error handlers can hide cached data or expose raw server codes. > - The app needs one connection state, readable error copy and a consistent read policy. > - This change provides that foundation and moves affected read surfaces onto it. > - The dependent write layer in #17960 preserves edits and submissions across the same failures. ## Linked Issues or Issue Description **What happened?** Temporary API failures could replace loaded content with errors. Different surfaces handled the same outage differently. Some showed raw server codes. A successful query to an unrelated service could also falsely report that Paperclip had recovered. **Expected behavior** Keep cached content visible. Show one connection banner during sustained trouble. Confirm recovery through the Paperclip health probe or its WebSocket. Retry transient reads and show readable errors for permanent failures. **Steps to reproduce** 1. Open a populated task or workspace page. 2. Interrupt API requests with a 503 or a dropped connection. 3. Keep an unrelated external request in flight. 4. Confirm that cached content remains and that the external result cannot clear the outage. 5. Restore the API and check that reads refresh. The write layer first merged into this feature branch in #15769 now has its own dependent PR, #17960. This PR contains the shared foundation and read changes. Merge this PR first. ## What Changed - Classify API errors and provide readable error copy, including non-JSON proxy responses and Retry-After handling. - Use one connection store and health probe with bounded backoff. Confirm recovery from the server health probe or live-updates WebSocket. - Pause reads while disconnected, retry transient errors, and refresh active reads after recovery. - Keep cached data visible through shared query views and affected page, panel and plugin read surfaces. - Add one app-level connection banner and support the pending-write count consumed by the dependent write layer. - Add a query-error rendering guard. Call it directly from the existing CI step. Update development documentation. ## Verification - Historical-chat lookup now follows the same ready/stale roster state as rendering. Three regressions fail before the fix; the six new cases pass after it, including working Retry, fresh data, no-cache error, permanent 403 and company isolation. All six focused query/recovery suites pass 200 tests. Final-commit CI passes and Apex is 5/5 with no unresolved findings. Canary Dry Run initially refused a preview version that was already on npm. It passed after one rerun with fresh registry metadata; no check was disabled. - Integrated current master and retained both transcript interest registration and connectivity reporting on socket open. The seven focused recovery/query suites pass 288 tests, and UI typecheck and production build pass. All 9,451 UI tests across 799 files pass on the merge commit. Final-commit CI passes and Apex is 5/5 with no unresolved findings. - Current head `70d9335d732219352ef0f98414439766bbcb33df`: all hosted CI checks pass. Apex is **5/5** on that head with no unresolved findings. The branch has no merge conflict. GitHub still requires code-owner approval. - The foundation passes 9,430 UI tests in 804 files. After integrating current master page paths, 190 focused page tests pass. - Two unrelated-success regressions fail before the latest fix. With the fix, 62 focused UI tests, all six guard tests and UI TypeScript pass. - The composed foundation and write layer pass **9,505 UI tests in 808 files**, UI TypeScript, a production UI build, and query-error, token and map guards. - A real Chromium journey on the composed source injects a 503 during title save. The loaded task stays visible; one reconnect banner appears; the edit saves after recovery; another task stays unchanged; the saved title survives reload. - Full recursive local checks were not repeated because Runner compilation exhausted local disk during another repair. Hosted checks must pass before merge. No new staging deployment is claimed. - Latest review follow-up: a cached CaseDetail now survives a failed HTTP 429 refresh. Its page regression fails before the fix and passes after it. Eighty focused foundation tests, UI TypeScript, build and guards pass after that change. - Earlier CI evidence is retained: several foundation jobs received a runner shutdown signal together. The write-layer head also failed an unchanged Runner descriptor close test with `provider_transport_failed`. The same Runner source passed that suite in the foundation run. New CI runs cover the updated heads; no check or assertion was weakened. ## Risks - Shared retry defaults change the time before transient read errors appear. Permanent errors still fail promptly. - ApiError.message now contains readable copy. Callers that compare raw codes must use ApiError.code. - Cache success can come from another service or local data. It must not confirm Paperclip recovery. - Durable writes are reviewed separately in #17960. This PR does not change database or mutation contracts. ## Model Used Anthropic Claude Opus 5.5 (`claude-opus-5-5`) produced the original work with tool use. OpenAI Codex, GPT-6 family, assisted with restoration, integration, tests and review fixes. The Codex session does not expose its exact serving model ID or context window size. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: devinfoley <139239+devinfoley@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Paperclip <noreply@paperclip.ing> Co-authored-by: Amy (Opus) <claude@users.noreply.github.com>
322 lines
12 KiB
TypeScript
322 lines
12 KiB
TypeScript
import { getPageVisibility, getVisibilityHeaderValue } from "@/lib/page-visibility";
|
|
import { tenantSessionRecovery } from "@/lib/tenant-session-recovery";
|
|
import { apiErrorMessage, parseRetryAfter } from "./errors";
|
|
import { readApiJson } from "./response";
|
|
|
|
const BASE = "/api";
|
|
|
|
export interface ApiErrorOptions {
|
|
/** Raw server error code; defaults to `body.error` when that is a string. */
|
|
code?: string | null;
|
|
/** Server-requested delay from the `Retry-After` header, in milliseconds. */
|
|
retryAfterMs?: number | null;
|
|
}
|
|
|
|
export class ApiError extends Error {
|
|
status: number;
|
|
body: unknown;
|
|
/**
|
|
* The raw `body.error` string (for example `tenant_app_unavailable`). Match
|
|
* on this, never on `message`: `message` is readable copy for people and may
|
|
* differ from the code (see `apiErrorMessage` in `api/errors.ts`).
|
|
*/
|
|
code: string | null;
|
|
retryAfterMs: number | null;
|
|
|
|
constructor(message: string, status: number, body: unknown, options: ApiErrorOptions = {}) {
|
|
super(message);
|
|
this.name = "ApiError";
|
|
this.status = status;
|
|
this.body = body;
|
|
const rawError = body && typeof body === "object" ? (body as { error?: unknown }).error : undefined;
|
|
this.code = options.code !== undefined ? options.code : typeof rawError === "string" ? rawError : null;
|
|
this.retryAfterMs = options.retryAfterMs ?? null;
|
|
}
|
|
}
|
|
|
|
/** Build the `ApiError` for a non-OK response whose body has been read. */
|
|
export function apiErrorFromResponse(res: Response, body: unknown): ApiError {
|
|
return new ApiError(apiErrorMessage(res.status, body), res.status, body, {
|
|
retryAfterMs: parseRetryAfter(res.headers?.get("Retry-After") ?? null),
|
|
});
|
|
}
|
|
|
|
// --- Dev/QA outage simulator ----------------------------------------------------
|
|
//
|
|
// Reproduces a deploy blip without a deploy. Dev and QA builds only:
|
|
//
|
|
// localStorage.paperclipSimulateOutage = "503:tenant_app_unavailable" // until removed
|
|
// localStorage.paperclipSimulateOutage = "503:tenant_app_unavailable@20" // for 20 seconds
|
|
// localStorage.paperclipSimulateOutage = "502" // non-JSON proxy page
|
|
// localStorage.paperclipSimulateOutage = "network" // fetch rejects (connection drop)
|
|
//
|
|
// Every same-origin API request made through `apiFetch` (the client, health,
|
|
// session) fails with the chosen shape. Remove the key to end the outage.
|
|
|
|
export const OUTAGE_SIMULATOR_STORAGE_KEY = "paperclipSimulateOutage";
|
|
|
|
export type SimulatedOutage =
|
|
| { kind: "network"; durationMs: number | null }
|
|
| { kind: "status"; status: number; code: string | null; durationMs: number | null };
|
|
|
|
export function parseSimulatedOutage(raw: string | null | undefined): SimulatedOutage | null {
|
|
const match = /^\s*(network|(\d{3})(?::([A-Za-z0-9_.-]+))?)(?:@(\d+(?:\.\d+)?))?\s*$/.exec(raw ?? "");
|
|
if (!match) return null;
|
|
const durationMs = match[4] ? Math.round(Number(match[4]) * 1000) : null;
|
|
if (match[1] === "network") return { kind: "network", durationMs };
|
|
const status = Number(match[2]);
|
|
if (status < 400 || status > 599) return null;
|
|
return { kind: "status", status, code: match[3] ?? null, durationMs };
|
|
}
|
|
|
|
function outageSimulatorEnabled(): boolean {
|
|
return import.meta.env.DEV || import.meta.env.MODE === "qa";
|
|
}
|
|
|
|
let simulatedOutageWindow: { raw: string; endsAt: number | null } | null = null;
|
|
|
|
/** The outage to simulate for the next request, or null. Never throws. */
|
|
export function activeSimulatedOutage(now = Date.now()): SimulatedOutage | null {
|
|
if (!outageSimulatorEnabled() || typeof window === "undefined") return null;
|
|
let raw: string | null = null;
|
|
try {
|
|
raw = window.localStorage?.getItem(OUTAGE_SIMULATOR_STORAGE_KEY) ?? null;
|
|
} catch {
|
|
return null;
|
|
}
|
|
const outage = parseSimulatedOutage(raw);
|
|
if (!raw || !outage) {
|
|
simulatedOutageWindow = null;
|
|
return null;
|
|
}
|
|
if (simulatedOutageWindow?.raw !== raw) {
|
|
simulatedOutageWindow = { raw, endsAt: outage.durationMs === null ? null : now + outage.durationMs };
|
|
}
|
|
if (simulatedOutageWindow.endsAt !== null && now >= simulatedOutageWindow.endsAt) {
|
|
simulatedOutageWindow = null;
|
|
try {
|
|
window.localStorage.removeItem(OUTAGE_SIMULATOR_STORAGE_KEY);
|
|
} catch {
|
|
// Storage can be unavailable; the outage still ends for this tab.
|
|
}
|
|
return null;
|
|
}
|
|
return outage;
|
|
}
|
|
|
|
function simulatedOutageResponse(outage: Extract<SimulatedOutage, { kind: "status" }>): Response {
|
|
if (!outage.code) {
|
|
// What a proxy serves while the app is down: an HTML page, not JSON.
|
|
return new Response("<!doctype html><title>Unavailable</title>", {
|
|
status: outage.status,
|
|
headers: { "Content-Type": "text/html" },
|
|
});
|
|
}
|
|
return new Response(JSON.stringify({ error: outage.code }), {
|
|
status: outage.status,
|
|
headers: { "Content-Type": "application/json" },
|
|
});
|
|
}
|
|
|
|
/**
|
|
* `fetch` for same-origin API calls. Identical to `fetch` except that the
|
|
* dev/QA outage simulator can intercept it.
|
|
*/
|
|
export async function apiFetch(input: string, init?: RequestInit): Promise<Response> {
|
|
const outage = activeSimulatedOutage();
|
|
if (outage) {
|
|
if (init?.signal?.aborted) throw abortError();
|
|
if (outage.kind === "network") throw new TypeError("Failed to fetch");
|
|
return simulatedOutageResponse(outage);
|
|
}
|
|
return fetch(input, init);
|
|
}
|
|
|
|
export interface RequestOptions {
|
|
/** Abort signal wired through to `fetch` and coalescing (per-caller). */
|
|
signal?: AbortSignal;
|
|
/** Extra request headers (e.g. the async-import opt-in). Mutations only. */
|
|
headers?: Record<string, string>;
|
|
/** The `fetch` cache mode. Use `"no-store"` for a response that must never
|
|
* come from the browser's HTTP cache. */
|
|
cache?: RequestCache;
|
|
}
|
|
|
|
function abortError(): DOMException {
|
|
return new DOMException("The operation was aborted.", "AbortError");
|
|
}
|
|
|
|
/**
|
|
* Non-authoritative observability hints (PAP-12556 / Phase 1). The server treats
|
|
* these as scheduling/telemetry only and never as security signals.
|
|
*/
|
|
function applyObservabilityHeaders(headers: Headers) {
|
|
if (headers.has("X-Paperclip-Tab-Visible")) return; // caller override wins
|
|
const visibility = getPageVisibility();
|
|
headers.set("X-Paperclip-Tab-Visible", getVisibilityHeaderValue(visibility));
|
|
if (typeof window !== "undefined" && window.location) {
|
|
headers.set("X-Paperclip-Route", window.location.pathname);
|
|
}
|
|
}
|
|
|
|
export async function requestResponse(path: string, init?: RequestInit): Promise<Response> {
|
|
const headers = new Headers(init?.headers ?? undefined);
|
|
const body = init?.body;
|
|
if (!(body instanceof FormData) && !headers.has("Content-Type")) {
|
|
headers.set("Content-Type", "application/json");
|
|
}
|
|
applyObservabilityHeaders(headers);
|
|
|
|
const res = await apiFetch(`${BASE}${path}`, {
|
|
credentials: "include",
|
|
...init,
|
|
headers,
|
|
});
|
|
if (!res.ok) {
|
|
const errorBody = await readApiJson(res);
|
|
const recovery = tenantSessionRecovery.recoverIfNeeded(res.status, errorBody);
|
|
if (recovery) return recovery;
|
|
throw apiErrorFromResponse(res, errorBody);
|
|
}
|
|
return res;
|
|
}
|
|
|
|
async function request<T>(path: string, init?: RequestInit): Promise<T> {
|
|
const res = await requestResponse(path, init);
|
|
if (res.status === 204) return undefined as T;
|
|
return readApiJson<T>(res);
|
|
}
|
|
|
|
// --- In-tab request coalescing for identical safe GETs -----------------------
|
|
//
|
|
// Multiple callers issuing the same GET while one is in flight share a single
|
|
// underlying fetch. Each caller keeps its own abort semantics: aborting one
|
|
// caller only cancels the shared fetch when *every* caller has aborted.
|
|
// Mutations are never coalesced.
|
|
|
|
interface InflightGet {
|
|
promise: Promise<unknown>;
|
|
controller: AbortController;
|
|
refs: Set<symbol>;
|
|
}
|
|
|
|
const inflightGets = new Map<string, InflightGet>();
|
|
|
|
function coalescedGet<T>(path: string, options?: RequestOptions): Promise<T> {
|
|
const signal = options?.signal;
|
|
if (signal?.aborted) return Promise.reject(abortError());
|
|
|
|
let entry = inflightGets.get(path);
|
|
if (!entry) {
|
|
const controller = new AbortController();
|
|
const promise = request<T>(path, {
|
|
method: "GET",
|
|
signal: controller.signal,
|
|
...(options?.cache ? { cache: options.cache } : {}),
|
|
});
|
|
const created: InflightGet = { promise, controller, refs: new Set() };
|
|
// Clear the shared entry once settled so later calls issue a fresh request.
|
|
promise.then(
|
|
() => {
|
|
if (inflightGets.get(path) === created) inflightGets.delete(path);
|
|
},
|
|
() => {
|
|
if (inflightGets.get(path) === created) inflightGets.delete(path);
|
|
},
|
|
);
|
|
inflightGets.set(path, created);
|
|
entry = created;
|
|
}
|
|
|
|
const activeEntry = entry;
|
|
const ref = Symbol("caller");
|
|
activeEntry.refs.add(ref);
|
|
|
|
const releaseRef = () => {
|
|
if (!activeEntry.refs.delete(ref)) return;
|
|
// Last caller gone before the fetch settled → abort the shared request.
|
|
if (activeEntry.refs.size === 0 && inflightGets.get(path) === activeEntry) {
|
|
inflightGets.delete(path);
|
|
activeEntry.controller.abort();
|
|
}
|
|
};
|
|
|
|
return new Promise<T>((resolve, reject) => {
|
|
const onAbort = () => {
|
|
signal?.removeEventListener("abort", onAbort);
|
|
releaseRef();
|
|
reject(abortError());
|
|
};
|
|
if (signal) signal.addEventListener("abort", onAbort);
|
|
|
|
activeEntry.promise.then(
|
|
(value) => {
|
|
signal?.removeEventListener("abort", onAbort);
|
|
activeEntry.refs.delete(ref);
|
|
resolve(value as T);
|
|
},
|
|
(err) => {
|
|
signal?.removeEventListener("abort", onAbort);
|
|
activeEntry.refs.delete(ref);
|
|
reject(err);
|
|
},
|
|
);
|
|
});
|
|
}
|
|
|
|
/**
|
|
* Stop later callers from joining the in-flight GET for `path`.
|
|
*
|
|
* Coalescing keys on the path alone, so a GET issued under one account's session
|
|
* can be joined by a caller that runs after the account changed — and handed the
|
|
* previous account's response. Detaching leaves that request to settle for the
|
|
* callers that asked for it, and makes the next call issue a fresh one. It does
|
|
* not abort, because those callers still want what they asked for.
|
|
*/
|
|
export function detachInflightGet(path: string): void {
|
|
inflightGets.delete(path);
|
|
}
|
|
|
|
function isRequestOptions(value: unknown): value is RequestOptions {
|
|
return typeof value === "object" && value !== null && "signal" in value;
|
|
}
|
|
|
|
export const api = {
|
|
get: <T>(path: string, options?: RequestOptions) => coalescedGet<T>(path, options),
|
|
post: <T>(path: string, body: unknown, options?: RequestOptions) =>
|
|
request<T>(path, {
|
|
method: "POST",
|
|
body: JSON.stringify(body),
|
|
signal: options?.signal,
|
|
...(options?.headers ? { headers: options.headers } : {}),
|
|
}),
|
|
postForm: <T>(path: string, body: FormData, options?: RequestOptions) =>
|
|
request<T>(path, {
|
|
method: "POST",
|
|
body,
|
|
signal: options?.signal,
|
|
// Never set Content-Type here — the browser sets multipart/form-data with
|
|
// the boundary. Extra headers (e.g. an async opt-in) may still ride along.
|
|
...(options?.headers ? { headers: options.headers } : {}),
|
|
}),
|
|
put: <T>(path: string, body: unknown, options?: RequestOptions) =>
|
|
request<T>(path, { method: "PUT", body: JSON.stringify(body), signal: options?.signal }),
|
|
/** Raw binary upload (e.g. one chunked import-transfer part); the body travels as-is. */
|
|
putRaw: <T>(path: string, body: Blob, options?: RequestOptions) =>
|
|
request<T>(path, {
|
|
method: "PUT",
|
|
body,
|
|
signal: options?.signal,
|
|
headers: { "Content-Type": "application/octet-stream", ...(options?.headers ?? {}) },
|
|
}),
|
|
patch: <T>(path: string, body: unknown, options?: RequestOptions) =>
|
|
request<T>(path, { method: "PATCH", body: JSON.stringify(body), signal: options?.signal }),
|
|
delete: <T>(path: string, bodyOrOptions?: unknown, options?: RequestOptions) => {
|
|
const requestOptions = isRequestOptions(bodyOrOptions) ? bodyOrOptions : options;
|
|
const body = bodyOrOptions === undefined || isRequestOptions(bodyOrOptions) ? undefined : JSON.stringify(bodyOrOptions);
|
|
return request<T>(path, { method: "DELETE", ...(body === undefined ? {} : { body }), signal: requestOptions?.signal });
|
|
},
|
|
deleteWithBody: <T>(path: string, body: unknown, options?: RequestOptions) =>
|
|
request<T>(path, { method: "DELETE", body: JSON.stringify(body), signal: options?.signal }),
|
|
};
|