fix: persistent random Snowflake worker ID + clean reaction API

Two fixes addressing architectural review feedback:

1. Snowflake ID collisions: Replace process.pid-based worker ID with a
   cryptographically random value (0-1023) generated once at first boot
   and persisted to instance_settings.worker_id. Eliminates deterministic
   ID collisions between Docker instances that all run as PID 1.

2. Reaction API leak: Revert addReaction/removeReaction signatures to
   (messageId, emoji) — the store now resolves the channel internally by
   scanning its message cache, keeping routing logic out of the UI layer.
This commit is contained in:
Jannis Braun
2026-03-03 00:03:29 +01:00
parent 5194dbef25
commit 72e07c1cc1
6 changed files with 76 additions and 10 deletions
+21 -1
View File
@@ -10,15 +10,35 @@
* - ~139 years of IDs from epoch
* - 1024 workers
* - 4096 IDs per millisecond per worker
*
* IMPORTANT: The worker ID MUST be unique per instance to prevent ID
* collisions in a federation setup. It is generated randomly at first boot
* and persisted to the database. Call setWorkerId() before generating any IDs.
*/
const EPOCH = 1704067200000n; // Jan 1, 2024 00:00:00 UTC
const WORKER_ID = BigInt(process.pid % 1024);
let WORKER_ID: bigint | null = null;
let sequence = 0n;
let lastTimestamp = -1n;
/**
* Set the worker ID for this instance. Must be called once during server
* startup, after the database is initialized, before any IDs are generated.
* The value is persisted in instance_settings.worker_id.
*/
export function setWorkerId(id: number): void {
if (id < 0 || id > 1023) {
throw new Error(`Worker ID must be 0-1023, got ${id}`);
}
WORKER_ID = BigInt(id);
}
export function generateSnowflake(): string {
if (WORKER_ID === null) {
throw new Error('Snowflake worker ID not initialized — call setWorkerId() during startup');
}
let timestamp = BigInt(Date.now());
if (timestamp === lastTimestamp) {