Files
翅膀喵andRhenCloud 978c53eeb3 fix(security): harden webhook delivery pipeline and group storage (#11)
* fix(log): preserve structured error details

Error objects lose their useful details when serialized as plain JSON.
Keep the message, stack and cause in structured log output.

* fix(runtime): bound shared in-flight waits

A canceled request can leave its shared promise pending forever.
Bound joins and coordinate a single replacement so later requests
make progress without multiplying upstream work.

* fix(bot): reject oversized callback bodies

Reject declared oversized bot callbacks before reading their bodies.
This keeps unauthenticated requests from wasting worker resources.

* test(d1): exercise real sqlite transactions

Exercise real foreign keys, transactions and conditional statements
instead of mocks that cannot reproduce lost updates or cascades.

* fix(storage): preserve claims across kv fallback

Only missing tables justify KV fallback; other failures must surface.
Remember fallback writes so migration to D1 cannot forget claims
and replay completed work. Share marker checks without extra writes.

* fix(filter): use isolated linear-time patterns

Backtracking patterns can exhaust worker CPU on untrusted payloads.
Use linear-time matching and separate bounded cache namespaces so
glob and regex entries cannot corrupt each other.

* fix(filter): read each event branch correctly

Deployment and check branches live in event-specific nested fields.
Read those fields and remove only the actual leading ref prefix.

* fix(filter): read only own payload properties

Inherited object properties are not webhook payload data.
Read own properties only so JSONPath cannot traverse prototypes.

* fix(filter): validate executable condition trees

Malformed trees and unsupported patterns must not silently turn an
exclude rule into a pass. Validate the shape and only the patterns
the evaluator actually executes.

* fix(i18n): bound and cache translation loads

Repeated misses and simultaneous override loads spend KV reads for
the same language. Cache confirmed misses and share bounded loads
without caching read failures as missing translations.

* i18n(bot): explain command and account failures

Explain account linking, command failures and skipped delivery
outcomes consistently in both supported languages.

* fix(github): bound and reuse app token fetches

Repeated token exchanges waste requests and JWT work for every event.
Reuse valid installation tokens, share refreshes and back off briefly
after failures while keeping every shared wait bounded.

* fix(github): restrict enrichment request urls

Webhook-provided URLs must never receive an installation token on
another host. Restrict enrichment to GitHub repository API paths,
reject redirects and bound optional requests.

* fix(formatter): preserve invalid code points

Out-of-range numeric entities can throw while formatting a delivery.
Keep invalid code points verbatim instead of aborting the message.

* fix(formatter): preserve bounded markdown text

Code spans, fences, indentation and autolinks are Markdown content,
not HTML tags to strip. Preserve them while bounding cleanup input
and reusing results across routes for the same event.

* fix(formatter): strip only leading ref prefixes

A short branch or tag may itself contain refs/heads or refs/tags.
Strip one leading full-ref prefix at the caller and keep short names
unchanged when generating labels and links.

* fix(gitea): link push authors to their forge

A Gitea author is not the same-named GitHub account.
Build author profile links from the originating forge.

* fix(formatter): render deployment state values

The deployment status object stringifies as [object Object].
Render its actual state in the status field.

* fix(formatter): read security severity levels

The scanner severity is not the security severity of an alert.
Prefer the security level so critical and high alerts get the right
label and color.

* fix(formatter): hide actions on other forges

Action buttons call the GitHub API as the linked user.
Hide them for other forges to avoid operating on a same-named
GitHub repository.

* fix(check): isolate updates by check suite

Same-named checks from different suites must not edit one message.
Include the suite or app identity and retain the former key for
controlled adoption of existing tracked messages.

* fix(logs): scope queries before limiting results

Filter authorized groups in SQL before LIMIT and aggregate only
visible rows. Preserve current group attribution for late writes,
surface read failures and combine related totals in one scan.

* perf(d1): index pruning and failed send queries

Pruning and failure views should not scan all historical rows.
Index the high-value access paths without adding a write-heavy
index for infrequent delivery-detail lookups.

* fix(discord): bound individual api requests

An unbounded API request can outlive an update-chain lock.
Bound each request so another holder cannot race a late edit.

* fix(discord): defer long rate-limit waits

Long rate-limit sleeps block delivery progress inside a worker.
Return a retryable 429 for the queue when the wait exceeds the bound.

* fix(discord): encode channel path segments

Stored channel and message ids are path segments, not URL paths.
Encode them so aliases cannot address another Discord resource.

* fix(discord): reuse nonces after lost responses

A successful create may lose its response. Reuse a stable nonce
with enforce_nonce so retries recover that message instead of
posting duplicates or accepting an unreadable success.

* fix(discord): constrain embeds and role mentions

Respect combined embed and component limits as well as field caps.
Restrict allowed mentions to validated route roles so payload text
cannot notify everyone.

* fix(discord): resume bounded command syncs

Application changes, command rollbacks and large guild lists can
leave commands stale indefinitely. Track the current hash per app
and resume bounded paginated work without needless KV writes.

* fix(github): verify installation management

An installation id is user-controlled, and repository access is not
installation management. Verify the owning account or an active
organization admin using the user token and read:org scope.

* fix(github): resolve immutable member ids

A login can be renamed and reclaimed before its first sign-in.
Resolve the current immutable user id when membership is granted
so a future holder of the login cannot inherit access.

* fix(auth): revoke stale sessions and tokens

Revoked or rotated credentials must stop working despite stale KV
replicas. Track the live token and revocations in D1 and preserve
immutable identity bindings rather than trusting a recycled login.

* fix(auth): bind sessions and tokens to identity

Automatic retries can repeat non-idempotent user actions.
Disable those retries and expose a write hook so bot handlers can
record completion immediately before the GitHub write.

* fix(discord): defer and deduplicate interactions

Acknowledge slow actions before Discord times out and deduplicate
interaction writes. Preserve comment text and reject edits whose
source changed after the modal opened.

* fix(telegram): always verify webhook secrets

Webhook verification must not silently disable itself when the
optional secret is absent. Derive a stable secret from the bot
token and reject callbacks when verification is unavailable.

* fix(telegram): bound individual api requests

Bound individual platform calls so they cannot outlive the lock
protecting a message update or stall a whole delivery invocation.

* fix(telegram): honor nested rate-limit delays

Telegram places retry_after inside parameters, not at the top level.
Read that value and defer long waits to the delivery queue.

* fix(telegram): report unreadable send results

An unreadable successful response may already have sent the message.
Report an explicit uncertain result instead of retrying a create
without an idempotency key or pretending it succeeded.

* fix(telegram): keep editable messages as text

Guessing photo versus text from message content makes later edits
use the wrong API and a smaller caption budget. Always send text
and use the avatar as a link preview.

* fix(telegram): render hard breaks as newlines

Telegram HTML does not support br tags. Render Markdown hard breaks
as plain newlines so otherwise valid messages are accepted.

* fix(telegram): clamp visible html characters

Telegram counts visible characters, not HTML source bytes.
Truncate without splitting tags, entities or surrogate pairs.

* fix(telegram): honor the author emoji toggle

The author marker must obey the group emoji preference just like
formatter-added decoration.

* fix(telegram): keep completed commands claimed

Redelivery after a completed write must never post another comment.
Keep completion claims despite reply failures, scope updates per bot
and only issue account-login links in private chats.

* fix(feishu): bound individual api requests

Bound individual platform calls so they cannot outlive the lock
protecting a message update or stall a whole delivery invocation.

* fix(feishu): classify retryable api failures

Business error codes can be transient even in HTTP 200 responses.
Classify token and rate-limit errors for retry and defer long waits.

* fix(feishu): retry sends with stable uuids

An unreadable response does not prove a send failed.
Reuse the same uuid for retries so recovering the result cannot
create a second Feishu message.

* fix(feishu): refresh and evict tokens safely

Separate app token caches, stop reusing refused tokens and share
refreshes. Cache failures briefly and keep valid tokens in memory
when a best-effort KV cache write fails.

* fix(feishu): use the card update endpoint

Delayed card updates use the card update endpoint; the config
endpoint cannot update an interaction response.

* fix(feishu): neutralize untrusted mentions

Forge text is untrusted and can contain at-all tags.
Neutralize mentions in card bodies and footers before rendering.

* fix(feishu): authenticate and dedupe callbacks

Authenticate signed or token-verified callbacks before dispatching
commands. Decode encrypted envelopes and handle card actions using
the correct fields without claiming irrelevant chat events.

* fix(target): normalize actual destinations

Different target spellings can name the same actual destination.
Normalize channel, thread and topic identities for ownership checks
and per-route duplicate suppression.

* fix(config): transact per-group configuration

Whole-table writes and cached snapshots can overwrite other groups
or resurrect removed configuration. Use per-group guarded writes,
one-time identity migration and explicit unreadable-row failures.

* fix(group): preserve identity through lifecycle

Group names are mutable and reusable, so they cannot own durable
state. Preserve immutable identities across lifecycle operations and
commit invite use with membership under transaction-time guards.

* fix(oauth): confirm account links and installs

A login link or installation id is not proof of account ownership.
Verify installation management, bind browser state and require an
explicit account-link confirmation before granting chat access.

* fix(queue): fence leases and record target outcomes

Retries must remember completed targets and distinguish active
lease holders. Fence state updates, preserve permanent outcomes and
reject missing parked payloads instead of dispatching empty data.

* fix(dispatch): isolate and resume tenant sends

Check tenant sources before filters and never dispatch orphan routes.
Record each target outcome, isolate update chains by forge and group
identity, and keep stale events from overwriting newer messages.

* fix(queue): bound batches and resume deliveries

A queue invocation has shared query and subrequest limits.
Use a small batch and shared send budget, renew fenced leases and
resume only unfinished work without exhausting unstarted messages.

* fix(admin): guard revisions and tenant targets

Read-time authorization alone cannot protect concurrent edits.
Guard revisions, current roles and destination ownership in the
write transaction, and use per-group endpoints for isolated changes.

* fix(webhook): confirm only accepted deliveries

Acknowledging before queue acceptance permanently loses deliveries.
Confirm only accepted work, release failed claims, scope replay
identities and strip tenant-supplied installation authority.

* fix(storage): prune abandoned parked payloads

Oversized payloads can remain after interrupted enqueue attempts.
Prune objects beyond their retry lifetime without deleting payloads
that a current delivery may still need.

* fix(console): expose api conflict statuses

Callers need the HTTP status to distinguish stale edits from other
failures and reload current data after a conflict.

* refactor(console): model group revisions

Carry server revisions explicitly so console writes can reject
changes based on an outdated group snapshot.

* fix(console): send loaded route revisions

Send the revision actually loaded with the route list and adopt the
server response after a save. Reload on conflict instead of silently
overwriting another editor.

* fix(console): guard fragment saves by revision

Fragments share a group version with other configuration writes.
Submit the loaded revision and reload on conflict to preserve
concurrent edits.

* fix(console): save groups with guarded writes

Whole-list saves can overwrite unrelated groups. Use per-group
revision-aware writes, retain rename results and confirm deletion
on the exact group, routes and version shown to the user.

* fix(console): reject edits of stale routes

A conflict reload can leave an older route editor open.
Compare its original value before saving so it cannot overwrite
the newly loaded route.

* fix(console): show overview loading failures

Failed overview requests must not look like an empty dashboard or
a successful refresh. Show the reason while retaining prior data.

* fix(console): explain rejected install bindings

Explain why an installation binding was refused instead of silently
returning to the console with no visible outcome.

* fix(console): allow selecting a group named all

The all sentinel collides with a valid group id.
Use a value that cannot be a group id so that group stays selectable.

* fix(console): discard stale log detail responses

An older detail request can finish after a newer selection or close.
Discard it so the dialog never shows the previously clicked record.

* fix(console): retain legacy group owners

Groups predating member roles store owners in adminIds.
Load those as owners so adding a member never drops all old owners.

* fix(console): lock member edits during saves

Edits made during a save are overwritten by its result.
Disable member controls and guard handlers until the write settles.

* fix(console): show invite loading failures

Failed invite loads must not look like no pending invites.
Show the error and discard responses from a previously shown group.

* fix(console): retain legacy discord log targets

A stored log target without platform means Discord.
Keep that meaning in the editor instead of clearing the target.

* fix(console): retain undisplayed forge sources

An editor that only displays some providers must preserve the others.
Keep stored custom sources when saving unrelated group settings.

* fix(console): lock the group editor while saving

Prevent edits while a group save is in flight and make super-only
installation settings readonly instead of silently ignoring edits.

* fix(console): lock route edits and show failures

Lock the route editor while saving and show fragment load failures
with a retry action rather than presenting a false empty list.

* fix(config): expose the legal contact override

Declare the public runtime key so the documented environment
override reaches legal pages in deployed Nuxt builds.

* build(deploy): migrate d1 before deployment

Apply pending D1 migrations before deploying code that requires
them. Stop deployment when migration fails instead of running
against an incomplete schema.

* docs(worker): align guides with runtime behavior

Synchronize both language guides, API references and examples with
the implemented authorization, storage and delivery behavior.
Document migration requirements and accepted operational limits.

* fix(storage): repair orphaned KV routes

Repair orphaned KV routes before serving KV-only configuration.

* fix(security): fully escape forge HTML

---------

Co-authored-by: RhenCloud <i@rhen.cloud>
2026-10-24 16:19:06 +08:00

522 lines
20 KiB
TypeScript

import type { Env, Group } from "../types";
import {
loadGroups,
updateGroup,
memberMatches,
memberRole,
memberRoleStatement,
normalizeGroupMembers,
} from "./groups";
import {
acceptedKey,
claimOnce,
confirmClaim,
idempotencyStore,
recordStatement,
} from "../lib/idempotency";
import { canUseD1 } from "../storage/d1";
import { ConfigConflictError } from "../storage/config-store";
import { log } from "../lib/log";
export interface Invite {
groupId: string;
/** Invited users can never be owners; only admins or viewers. */
role: "admin" | "viewer";
expiresAt: number;
createdBy: string;
note?: string;
/**
* The invited group's immutable `instance` id. A group later re-created
* under the same id has another instance, so the invite never admits anyone
* to it. Absent on invites created before instances existed.
*/
groupInstance?: string;
/**
* Who issued it: a super admin, or a member who must still be an owner of
* the group when the invite is used — an owner removed meanwhile cannot
* have it let anyone in. Absent on older invites.
*/
issuer?: { super: true } | { userId: string; login: string };
}
/** The invite belongs to an earlier group that had the same id. */
class StaleInviteError extends Error {}
/**
* The login is still on a member entry bound to another account (a member
* who renamed since). Adding a second entry with that login would be dropped
* as a duplicate — a "successful" join without access — so the owner must
* clear the stale entry first.
*/
class MemberConflictError extends Error {}
function revokedKey(token: string): string {
return `invite-revoked:${token}`;
}
/** How long an invite stays claimed while one acceptance is being processed. */
const INVITE_CLAIM_SECONDS = 120;
/** Invites are valid for seven days; their KV keys live exactly that long. */
export const INVITE_LIFETIME_MS = 7 * 86400_000;
const INDEX_TTL = INVITE_LIFETIME_MS / 1000;
/** KV TTL (seconds) for an invite expiring at `expiresAt` (KV minimum: 60). */
function ttlFor(expiresAt: number): number {
return Math.max(60, Math.ceil((expiresAt - Date.now()) / 1000));
}
function generateInviteToken(): string {
const bytes = new Uint8Array(32);
crypto.getRandomValues(bytes);
return Array.from(bytes)
.map((b) => b.toString(16).padStart(2, "0"))
.join("");
}
function inviteKey(token: string): string {
return `invite:${token}`;
}
/**
* Per-group index of invite tokens. Listing pending invites reads this key
* instead of `kv.list({ prefix })`, which is eventually consistent and can
* lag behind a fresh write by minutes — the index keeps listing reliable.
*/
function indexKey(groupId: string): string {
return `invite:group:${groupId}`;
}
/**
* Tokens of a group's index. A failed read propagates: taken for an empty
* index, it would be written back without the existing tokens, leaving those
* invites valid but unlisted (and so impossible to revoke from the console).
*/
async function readIndex(kv: KVNamespace, groupId: string): Promise<string[]> {
const raw = await kv.get<string[]>(indexKey(groupId), "json");
return Array.isArray(raw) ? raw : [];
}
async function writeIndex(kv: KVNamespace, groupId: string, tokens: string[]): Promise<void> {
await kv.put(indexKey(groupId), JSON.stringify(tokens), { expirationTtl: INDEX_TTL });
}
/**
* Per-invite membership key, written on its own (no read-modify-write): two
* invites created at once can both lose their index entry to the other's
* index write, but never their marker. Listing reads both — the index for
* fresh writes, the markers (an eventually consistent `list`) for durability.
*/
function markerPrefix(groupId: string): string {
return `invite:by-group:${groupId}:`;
}
function listMarkers(kv: KVNamespace, groupId: string): Promise<string[]> {
return listPrefix(kv, markerPrefix(groupId));
}
/**
* Per-invite key under the issuing group's instance: it never moves, so the
* group finds its invites even when moving them to a new id after a rename
* failed halfway.
*/
function instancePrefix(instance: string): string {
return `invite:by-instance:${instance}:`;
}
async function listPrefix(kv: KVNamespace, prefix: string): Promise<string[]> {
const tokens: string[] = [];
let cursor: string | undefined;
do {
const page = await kv.list({ prefix, cursor });
for (const k of page.keys) {
if (k.name.startsWith(prefix)) tokens.push(k.name.slice(prefix.length));
}
cursor = page.list_complete ? undefined : (page as { cursor?: string }).cursor;
} while (cursor);
return tokens;
}
/** Every token listed for a group: by its index, its id markers or its instance markers. */
async function groupTokens(kv: KVNamespace, groupId: string, instance?: string): Promise<string[]> {
const [indexed, marked, owned] = await Promise.all([
readIndex(kv, groupId),
listMarkers(kv, groupId),
instance ? listPrefix(kv, instancePrefix(instance)) : Promise.resolve([]),
]);
return [...new Set([...indexed, ...marked, ...owned])];
}
/**
* Write the index without the given tokens — only when it lists any of them:
* KV writes are the scarcest quota (Workers Free), and an unchanged index
* needs no rewrite.
*/
async function dropFromIndex(kv: KVNamespace, groupId: string, gone: string[]): Promise<void> {
const current = await readIndex(kv, groupId);
const kept = current.filter((t) => !gone.includes(t));
if (kept.length !== current.length) await writeIndex(kv, groupId, kept);
}
/** Remove every listing link of an invite that is gone (index, id and instance markers). */
async function removeFromIndex(kv: KVNamespace, invite: Invite, token: string): Promise<void> {
await kv.delete(`${markerPrefix(invite.groupId)}${token}`);
if (invite.groupInstance) await kv.delete(`${instancePrefix(invite.groupInstance)}${token}`);
await dropFromIndex(kv, invite.groupId, [token]);
}
export async function createInvite(kv: KVNamespace, invite: Invite): Promise<string> {
const token = generateInviteToken();
const ttl = { expirationTtl: ttlFor(invite.expiresAt) };
await kv.put(inviteKey(token), JSON.stringify(invite), ttl);
await kv.put(`${markerPrefix(invite.groupId)}${token}`, "1", ttl);
if (invite.groupInstance) {
await kv.put(`${instancePrefix(invite.groupInstance)}${token}`, "1", ttl);
}
const tokens = await readIndex(kv, invite.groupId);
tokens.push(token);
await writeIndex(kv, invite.groupId, tokens);
return token;
}
/** The invite, or null when it does not exist (any more). Read errors propagate. */
export async function getInvite(kv: KVNamespace, token: string): Promise<Invite | null> {
const raw = await kv.get<Invite>(inviteKey(token), "json");
if (!raw) return null;
if (Date.now() > raw.expiresAt) {
await kv.delete(inviteKey(token));
await removeFromIndex(kv, raw, token);
return null;
}
return raw;
}
export async function consumeInvite(kv: KVNamespace, token: string): Promise<void> {
const raw = await kv.get<Invite>(inviteKey(token), "json");
await kv.delete(inviteKey(token));
if (raw) await removeFromIndex(kv, raw, token);
}
/**
* Whether an invite was issued by this very group, not by an earlier group
* that had the same id (its index may outlive a failed cleanup). An invite from
* before instances belongs to the group that had its id back then (`legacyId`).
*/
function issuedBy(
invite: Invite,
group: Pick<Group, "instance" | "legacyId"> | undefined,
): boolean {
if (!group?.instance) return true;
if (invite.groupInstance) return invite.groupInstance === group.instance;
return group.legacyId === invite.groupId;
}
/**
* All pending (unexpired) invites of a group, via the per-group index. With
* the group given, invites of an earlier group under the same id are left
* out (and removed): their links must not reach the new group's owners.
*/
export async function listInvites(
kv: KVNamespace,
groupId: string,
group?: Pick<Group, "instance" | "legacyId">,
): Promise<Array<Invite & { token: string }>> {
// Read errors propagate (the console shows a failure, not "no invites"),
// and only tokens whose invite is known to be gone are pruned.
// A pre-instance group may still have invites filed under the id it had
// back then (a rename whose invite move failed).
const formerId = group?.legacyId && group.legacyId !== groupId ? group.legacyId : undefined;
const tokens = [
...new Set([
...(await groupTokens(kv, groupId, group?.instance)),
...(formerId ? await groupTokens(kv, formerId) : []),
]),
];
const out: Array<Invite & { token: string }> = [];
const stale: string[] = [];
for (const token of tokens) {
const invite = await getInvite(kv, token);
// Also the group's own invites still filed under a former id (a rename
// whose invite move failed): they are usable, so they must be listed.
const ours =
invite != null &&
issuedBy(invite, group) &&
(invite.groupId === groupId ||
(group?.instance != null && invite.groupInstance === group.instance) ||
(formerId != null && !invite.groupInstance && invite.groupId === formerId));
if (invite && ours) {
out.push({ ...invite, token });
} else {
// Not this group's (gone, or issued by a former holder of the id — which
// may live on under another id): only the listing link to this id goes;
// the invite itself is not ours to revoke and expires on its own.
await kv.delete(`${markerPrefix(groupId)}${token}`);
// A gone invite's marker under this group's instance goes too, or every
// later listing would find (and clean up after) it again until it expires.
if (!invite && group?.instance) {
await kv.delete(`${instancePrefix(group.instance)}${token}`);
}
stale.push(token);
}
}
// Re-read: an invite created meanwhile must stay listed.
if (stale.length > 0) await dropFromIndex(kv, groupId, stale);
return out;
}
/**
* Revoke an invite. With `db`, the revocation is first recorded in the
* idempotency store (for the invite's lifetime): an acceptance already past
* its KV read checks it inside its member write, so it cannot still join.
*/
export async function revokeInvite(
kv: KVNamespace,
token: string,
db?: D1Database,
guards?: (db: D1Database) => D1PreparedStatement[],
): Promise<void> {
const raw = await kv.get<Invite>(inviteKey(token), "json");
if (db) {
const ttl = ttlFor(raw?.expiresAt ?? Date.now() + INVITE_LIFETIME_MS);
if (guards && canUseD1(db)) {
// The revocation record commits only while the guards hold (e.g. the
// revoking owner still is one): a request authorized before a role
// change can no longer act after it.
await db.batch([...guards(db), recordStatement(db, revokedKey(token), ttl)]);
} else {
await idempotencyStore(db, kv).claim(revokedKey(token), ttl);
}
}
await kv.delete(inviteKey(token));
if (raw) await removeFromIndex(kv, raw, token);
}
/**
* Re-point every pending invite of a group to its new id (group rename).
* Best-effort: a failure leaves the old invites in place (they will be
* rejected as group-missing after the rename).
*/
export async function migrateInvites(
kv: KVNamespace,
from: string,
to: string,
group?: Pick<Group, "instance" | "legacyId">,
): Promise<void> {
try {
const tokens = await groupTokens(kv, from);
if (tokens.length === 0) {
await kv.delete(indexKey(from));
return;
}
const moved: string[] = [];
for (const token of tokens) {
const invite = await kv.get<Invite>(inviteKey(token), "json");
// Only the renamed group's own invites move (a former group's leftovers
// stay behind); they get its instance, so they keep working under the
// new id even when they date from before instances.
if (invite && invite.groupId === from && issuedBy(invite, group)) {
const groupInstance = invite.groupInstance ?? group?.instance;
const ttl = { expirationTtl: ttlFor(invite.expiresAt) };
await kv.put(
inviteKey(token),
JSON.stringify({ ...invite, groupId: to, groupInstance }),
ttl,
);
await kv.put(`${markerPrefix(to)}${token}`, "1", ttl);
if (groupInstance) await kv.put(`${instancePrefix(groupInstance)}${token}`, "1", ttl);
await kv.delete(`${markerPrefix(from)}${token}`);
moved.push(token);
}
}
const kept = (await readIndex(kv, to)).filter((t) => !moved.includes(t));
await kv.put(indexKey(to), JSON.stringify([...kept, ...moved]), {
expirationTtl: INDEX_TTL,
});
await kv.delete(indexKey(from));
} catch (err) {
log.warn({ err, from, to }, "Failed to migrate invites on group rename");
}
}
/** Revoke every pending invite of a deleted group. */
export async function deleteGroupInvites(
kv: KVNamespace,
groupId: string,
group?: Pick<Group, "instance" | "legacyId">,
): Promise<void> {
const instance = group?.instance;
try {
for (const token of await groupTokens(kv, groupId, instance)) {
// Only the deleted group's own invites are revoked: one listed under
// its id may have been issued by a group that lives on under another.
const invite = await kv.get<Invite>(inviteKey(token), "json");
if (
invite &&
issuedBy(invite, group) &&
(invite.groupId === groupId || (instance != null && invite.groupInstance === instance))
) {
await kv.delete(inviteKey(token));
}
await kv.delete(`${markerPrefix(groupId)}${token}`);
if (instance) await kv.delete(`${instancePrefix(instance)}${token}`);
}
await kv.delete(indexKey(groupId));
} catch (err) {
log.warn({ err, groupId }, "Failed to delete invites of a removed group");
}
}
/**
* Adds the accepting user to the invited group (or upgrades their role when
* they are already a viewer) and consumes the invite. The invited role is
* never an owner — ownership stays with the inviter's discretion.
*
* The invite is claimed atomically before the membership changes, so two
* people accepting the same link at once cannot both join; the member write is
* a compare-and-swap on the group, so it never reverts a concurrent edit.
*/
export async function acceptInvite(
env: Pick<Env, "KV" | "DB">,
token: string,
userId: string,
login: string,
/** The login identity checks may use ("" when it is not trusted as this user's). */
matchLogin = login,
): Promise<
| { ok: true; groupId: string; groupInstance?: string; role: "admin" | "viewer" }
| { ok: false; reason: "invalid" | "group-missing" | "conflict" }
> {
const kv = env.KV;
const invite = await getInvite(kv, token);
if (!invite) return { ok: false, reason: "invalid" };
const claims = idempotencyStore(env.DB, kv);
const claimKey = `invite-used:${token}`;
// A short claim while joining, confirmed for the invite's lifetime once
// used: an attempt that fails and cannot give the claim back blocks the
// invite for a couple of minutes, not until it expires.
if ((await claimOnce(claims, claimKey, INVITE_CLAIM_SECONDS)) !== "claimed") {
return { ok: false, reason: "invalid" };
}
if (await claims.has(revokedKey(token))) {
await claims.release(claimKey).catch(() => undefined);
return { ok: false, reason: "invalid" };
}
let joined = false;
let joinedId: string | undefined;
let joinedInstance: string | undefined;
/** The member who issued the invite is no longer an owner (it is void then). */
const issuerGone = async (): Promise<boolean> => {
const by = invite.issuer && "userId" in invite.issuer ? invite.issuer : null;
if (!by) return false;
const group = (await loadGroups(kv, { fresh: true })).find((g) =>
invite.groupInstance ? g.instance === invite.groupInstance : g.legacyId === invite.groupId,
);
return !group || memberRole(group, by.userId, by.login) !== "owner";
};
try {
// The invited group is the instance the invite was issued for, under
// whatever id it has now. An invite from before instances only applies to
// a group that existed back then, never to one created since.
const groups = await loadGroups(kv, { fresh: true });
const target = invite.groupInstance
? groups.find((g) => g.instance === invite.groupInstance)
: groups.find((g) => g.legacyId === invite.groupId);
// Its group instance is gone for good: the invite can never be used.
if (!target && invite.groupInstance) throw new StaleInviteError();
if (!target) return { ok: false, reason: "group-missing" };
// The invite is confirmed as used in the same transaction that adds the
// member — only when that write went through (`changes()`: the member
// update right before; a lost compare-and-swap changes nothing) and the
// member is there after it: no crash between the two can leave a member
// added and the invite still open, nor a retried write consume it unused.
const confirmWithJoin = (db: D1Database): D1PreparedStatement[] => [
db
.prepare(
`INSERT INTO dedup_keys (key, claimed_at, expires_at)
SELECT ?, ?, ? WHERE changes() > 0 AND EXISTS (
SELECT 1 FROM (SELECT id, CASE WHEN json_valid(data) THEN data ELSE '{}' END AS data FROM d1_groups) g, json_each(g.data, '$.members') m
WHERE g.id = ? AND json_extract(m.value, '$.userId') = ?)
ON CONFLICT(key) DO UPDATE SET expires_at = excluded.expires_at`,
)
.bind(acceptedKey(claimKey), Date.now(), invite.expiresAt, target.id, userId),
];
// And it only commits while the invite is unused and not revoked: an
// earlier acceptance whose answer was lost (its claim given back, its use
// committed), or a revocation that landed after the KV read, aborts it.
const issuer = invite.issuer && "userId" in invite.issuer ? invite.issuer : null;
const unusedGuard = (db: D1Database): D1PreparedStatement[] => [
db
.prepare(
"SELECT CASE WHEN EXISTS (SELECT 1 FROM dedup_keys WHERE key IN (?, ?) AND expires_at > ?) THEN json('invite used') ELSE 1 END AS ok",
)
.bind(acceptedKey(claimKey), revokedKey(token), Date.now()),
...(issuer
? [memberRoleStatement(db, target.id, issuer.userId, issuer.login, ["owner"])]
: []),
];
const updated = await updateGroup(
kv,
target.id,
(group) => {
const sameInstance = invite.groupInstance
? group.instance === invite.groupInstance
: group.legacyId === invite.groupId;
if (!sameInstance) throw new StaleInviteError();
const members = normalizeGroupMembers(group);
const idx = members.findIndex((m) => memberMatches(m, userId, matchLogin));
if (idx >= 0) {
if (invite.role === "admin" && members[idx]!.role === "viewer") {
members[idx]!.role = "admin";
}
members[idx]!.userId ??= userId;
} else {
if (members.some((m) => m.login.toLowerCase() === login.toLowerCase())) {
throw new MemberConflictError();
}
// Bound to the account that accepted, not to its (renamable) login.
members.push({ login, role: invite.role, userId });
}
return {
...group,
members,
adminIds: members.filter((m) => m.role === "owner").map((m) => m.login),
};
},
{ guards: unusedGuard, after: confirmWithJoin },
);
joined = updated != null;
joinedId = updated?.id;
joinedInstance = updated?.instance;
} catch (err) {
// The invite stays usable once the owner resolved the conflict.
if (err instanceof MemberConflictError) return { ok: false, reason: "conflict" };
// Used meanwhile (by an acceptance whose answer was lost).
if (
err instanceof ConfigConflictError &&
((await claims.has(acceptedKey(claimKey))) ||
(await claims.has(revokedKey(token))) ||
(await issuerGone()))
) {
return { ok: false, reason: "invalid" };
}
if (!(err instanceof StaleInviteError)) throw err;
await consumeInvite(kv, token);
return { ok: false, reason: "invalid" };
} finally {
if (!joined) await claims.release(claimKey).catch(() => undefined);
}
if (!joined) return { ok: false, reason: "group-missing" };
await confirmClaim(claims, claimKey, ttlFor(invite.expiresAt)).catch((err: unknown) =>
log.warn({ err }, "Failed to confirm an invite claim"),
);
await consumeInvite(kv, token);
return {
ok: true,
groupId: joinedId ?? invite.groupId,
groupInstance: joinedInstance,
role: invite.role,
};
}