mirror of
https://github.com/multica-ai/multica.git
synced 2026-08-04 17:18:35 +02:00
28b6105edc
When an agent created a sub-issue while working on a human's behalf, that human received no notifications for it at all. issue_subscriber modelled ACTOR identity, so an agent-created, agent-assigned issue had a full subscriber list and zero members to deliver to. The platform already knew who the work was for (agent_task_queue.originator_user_id, MUL-4302); notification never asked. - attribution.DelegatedSubscriber: one shared rule over the same origin waterfall ClassifyDirect uses. agent_create subscribes the originator as 'delegated'; quick_create keeps the direct 'creator' tier; autopilot and degraded attribution subscribe nobody. - Delegated is a reduced delivery tier: in_review/done/cancelled/blocked plus failures and mentions. Routine churn is suppressed, and the parent bubble cannot re-deliver what the tier dropped. - Unsubscribe becomes stateful: an unsubscribed_at tombstone survives later rule passes, and opt_out_scope distinguishes "this issue" from "this subtree" so a narrow opt-out no longer silently suppresses future children. - Subtree unsubscribe is its own endpoint. A body flag cannot fail loudly against an older backend (Go drops unknown fields); an unknown route 404s, which the UI now surfaces with a distinct message. - Eligibility and the write share one statement under a (workspace, user) advisory lock that subtree unsubscribe and member revoke also take, closing the check-then-insert races. Revoke additionally clears the departing member's subscriptions in the same tx. - UI explains a delegated subscription and offers both unsubscribe scopes. Migrations 249/250 add the delegated reason, the opt-out tombstone, and the opt-out scope, using NOT VALID + VALIDATE CONSTRAINT so the widened CHECK does not scan issue_subscriber under an exclusive lock. Reviewed across eight rounds; an earlier write-time subtree roll-up was built and then removed in full once it proved unfixable without serializing every topology mutation. The parent's own status transition already carries that signal. Closes MUL-5483.
97 lines
3.6 KiB
TypeScript
97 lines
3.6 KiB
TypeScript
import { describe, expect, it } from "vitest";
|
|
import { InboxListSchema } from "./schemas";
|
|
|
|
/**
|
|
* Tests for mobile's CLIENT-SIDE parsing of GET /api/inbox.
|
|
*
|
|
* Scope, stated precisely because the name of this file used to overclaim:
|
|
* these are hand-written fixtures run against `InboxListSchema`. They pin how
|
|
* this client REACTS to a given payload. They cannot fail when the Go server
|
|
* starts sending something new — nothing here executes server code.
|
|
*
|
|
* The matching server-side guarantee is structural rather than a test: every
|
|
* `details` map in server/cmd/server/notification_listeners.go is typed
|
|
* `map[string]string`, so a non-string value is a compile error there.
|
|
*
|
|
* Why both halves exist: during MUL-5483 a new inbox type was added and the
|
|
* mobile label map was updated so `tsc` passed — but a NUMBER went into
|
|
* `details.child_count`, and `details` is `z.record(z.string(), z.string())`.
|
|
* Because the endpoint parses an ARRAY, one bad row fails the whole parse and
|
|
* `listInbox` falls back to `EMPTY_INBOX_LIST`: the entire mobile inbox
|
|
* renders empty, not just that row. The blast radius is what these tests
|
|
* document; the compile-time type is what prevents it.
|
|
*/
|
|
describe("inbox list schema", () => {
|
|
it("parses a row shaped like the documented server payload", () => {
|
|
const serverRow = {
|
|
id: "inbox-1",
|
|
workspace_id: "ws-1",
|
|
recipient_type: "member",
|
|
recipient_id: "user-1",
|
|
type: "status_changed",
|
|
severity: "info",
|
|
issue_id: "issue-1",
|
|
title: "P0: delegated subscription rule",
|
|
body: "",
|
|
actor_type: "agent",
|
|
actor_id: "agent-1",
|
|
read: false,
|
|
archived: false,
|
|
created_at: "2026-07-30T00:00:00Z",
|
|
// Every value is a string. A number here drops the whole list.
|
|
details: { from: "in_progress", to: "in_review" },
|
|
};
|
|
|
|
const parsed = InboxListSchema.safeParse([serverRow]);
|
|
expect(parsed.success).toBe(true);
|
|
expect(parsed.success && parsed.data[0]?.type).toBe("status_changed");
|
|
expect(parsed.success && parsed.data[0]?.details?.to).toBe("in_review");
|
|
});
|
|
|
|
it("rejects a numeric details value", () => {
|
|
const badRow = {
|
|
id: "inbox-2",
|
|
recipient_type: "member",
|
|
type: "status_changed",
|
|
details: { child_count: 3 },
|
|
};
|
|
|
|
expect(InboxListSchema.safeParse([badRow]).success).toBe(false);
|
|
});
|
|
|
|
it("keeps one malformed row from emptying the entire list observable", () => {
|
|
// Documents the blast radius that made this a P1 rather than a cosmetic bug:
|
|
// the schema is an array, so a single bad row invalidates every good one.
|
|
const good = {
|
|
id: "inbox-3",
|
|
recipient_type: "member",
|
|
type: "status_changed",
|
|
details: { from: "todo", to: "in_review" },
|
|
};
|
|
const bad = {
|
|
id: "inbox-4",
|
|
recipient_type: "member",
|
|
type: "status_changed",
|
|
details: { child_count: 3 },
|
|
};
|
|
|
|
expect(InboxListSchema.safeParse([good]).success).toBe(true);
|
|
expect(InboxListSchema.safeParse([good, bad]).success).toBe(false);
|
|
});
|
|
|
|
it("renders an unknown server type instead of dropping the row", () => {
|
|
// Mirrors the root CLAUDE.md API-compatibility rule and mobile's own
|
|
// "render every inbox type, never silently drop a category" parity rule: a
|
|
// type this build has never heard of must still parse.
|
|
const future = {
|
|
id: "inbox-5",
|
|
recipient_type: "member",
|
|
type: "some_future_type",
|
|
details: { anything: "still a string" },
|
|
};
|
|
|
|
const parsed = InboxListSchema.safeParse([future]);
|
|
expect(parsed.success).toBe(true);
|
|
});
|
|
});
|