Choose the Stalwart by the domain somebody signs in with

One ihasmail in front of several Stalwarts, from #238. STALWART_URL stays
required and stays the default, so an installation that sets nothing behaves
exactly as it always has -- the mapping only adds domains that go elsewhere.

An unlisted domain goes to the default. So does a bare username, which
Stalwart accepts and which has no domain to map at all.

A listed domain never falls back. If its server is unreachable that sign-in
fails rather than retrying against the default, because falling back would
authenticate somebody against a server their domain was deliberately routed
away from -- and if the same account name existed there, they would land in
another tenant's mailbox. The fallback is a decision about unmapped domains,
taken before any network call, not a recovery path.

Smaller than it sounds because only four places read config.stalwartUrl, all
in upstream.ts. The upstream session now records which server issued it, since
the relative URLs inside it only mean anything against that server, and every
route already holding a session gets the right upstream without a second
lookup. The client is untouched: it talks to one proxy and never learns there
is more than one server behind it, which is exactly why this is small and
several-servers-at-once is not.

The upstream is derived from the username rather than stored on the session,
so a mapping change takes effect on restart instead of being frozen into
sessions that outlive it.

Validated at boot the way the settings policy is: malformed JSON, a duplicate
domain once normalised, a missing file or a value that is not an http(s) URL
all stop the server. Domains are lower-cased and stripped of a trailing dot,
because that is how one arrives off a username and comparing them any other
way means a mapping that silently never matches. The servers themselves are
not contacted -- a mapping is a routing table, not a health check, and one
customer's outage must not stop ihasmail starting for the other four.

Eight tests on the routing, two on the shipped example, and the four refusals
checked by hand against a real config load.

server/src/account.ts

@@ -65,7 +65,7 @@ function accountId(ctx: Ctx): string {
 type Invocation = [string, Record<string, unknown>, string];
 
 async function jmap(ctx: Ctx, methodCalls: Invocation[]): Promise<{ methodResponses?: [string, unknown, string][] }> {
-  const res = await fetch(absoluteUpstream(ctx.session.apiUrl), {
+  const res = await fetch(absoluteUpstream(ctx.session.apiUrl, ctx.session.baseUrl), {
     method: "POST",
     headers: { authorization: ctx.authorization, "content-type": "application/json", accept: "application/json" },
     body: JSON.stringify({ using: [JMAP_CORE, STALWART_CAP], methodCalls }),

server/src/app.ts

@@ -16,6 +16,7 @@ import {
   forgetUpstreamSession,
   getAccountInfo,
   getUpstreamSession,
+  upstreamFor,
   localizeSession,
 } from "./upstream.js";
 import {
@@ -237,7 +238,7 @@ export function createApp(basePath = config.basePath): Hono<Env> {
     const effectivePassword = totp ? `${password}$${totp}` : password;
     const authorization = `Basic ${Buffer.from(`${username}:${effectivePassword}`, "utf8").toString("base64")}`;
     try {
-      const upstream = await fetchUpstreamSession(authorization);
+      const upstream = await fetchUpstreamSession(authorization, upstreamFor(username));
       // ihasmail requires Stalwart 0.16 or newer. Refuse here, once and
       // clearly, rather than signing someone in and letting Files, the account
       // locale and self-service credentials each fail in their own way with
@@ -311,7 +312,7 @@ export function createApp(basePath = config.basePath): Hono<Env> {
   api.get("/auth/session", requireSession, async (c) => {
     const session = c.get("session");
     try {
-      const upstream = await getUpstreamSession(session.id, session.authorization, c.req.query("refresh") === "1");
+      const upstream = await getUpstreamSession(session.id, session.authorization, upstreamFor(session.username), c.req.query("refresh") === "1");
       const info = await getAccountInfo(session.id, session.authorization, upstream);
       return c.json(localizeSession(upstream, sessionExtras(session, info)));
     } catch (err) {
@@ -528,8 +529,8 @@ export function createApp(basePath = config.basePath): Hono<Env> {
       return c.json({ error: "unsupported_media_type" }, 415);
     }
     try {
-      const upstream = await getUpstreamSession(session.id, session.authorization);
-      const res = await fetch(absoluteUpstream(upstream.apiUrl), {
+      const upstream = await getUpstreamSession(session.id, session.authorization, upstreamFor(session.username));
+      const res = await fetch(absoluteUpstream(upstream.apiUrl, upstream.baseUrl), {
         method: "POST",
         headers: {
           authorization: session.authorization,
@@ -562,8 +563,8 @@ export function createApp(basePath = config.basePath): Hono<Env> {
     // suggestion; count the bytes as they go past.
     const body = c.req.raw.body ? c.req.raw.body.pipeThrough(byteCap(config.maxUploadBytes)) : null;
     try {
-      const upstream = await getUpstreamSession(session.id, session.authorization);
-      const url = absoluteUpstream(expandTemplate(upstream.uploadUrl, { accountId }));
+      const upstream = await getUpstreamSession(session.id, session.authorization, upstreamFor(session.username));
+      const url = absoluteUpstream(expandTemplate(upstream.uploadUrl, { accountId }), upstream.baseUrl);
       const res = await fetch(url, {
         method: "POST",
         headers: {
@@ -588,8 +589,8 @@ export function createApp(basePath = config.basePath): Hono<Env> {
     const accept = c.req.query("accept") ?? "application/octet-stream";
     const inline = c.req.query("inline") === "1";
     try {
-      const upstream = await getUpstreamSession(session.id, session.authorization);
-      const url = absoluteUpstream(expandTemplate(upstream.downloadUrl, { accountId, blobId, name, type: accept }));
+      const upstream = await getUpstreamSession(session.id, session.authorization, upstreamFor(session.username));
+      const url = absoluteUpstream(expandTemplate(upstream.downloadUrl, { accountId, blobId, name, type: accept }), upstream.baseUrl);
       const res = await fetch(url, {
         // Ask for the bytes as they are. undici would otherwise negotiate gzip
         // on our behalf and hand back a decompressed body whose content-length
@@ -639,8 +640,8 @@ export function createApp(basePath = config.basePath): Hono<Env> {
     const closeafter = c.req.query("closeafter") ?? "no";
     const ping = c.req.query("ping") ?? "30";
     try {
-      const upstream = await getUpstreamSession(session.id, session.authorization);
-      const url = absoluteUpstream(expandTemplate(upstream.eventSourceUrl, { types, closeafter, ping }));
+      const upstream = await getUpstreamSession(session.id, session.authorization, upstreamFor(session.username));
+      const url = absoluteUpstream(expandTemplate(upstream.eventSourceUrl, { types, closeafter, ping }), upstream.baseUrl);
       const controller = new AbortController();
       c.req.raw.signal.addEventListener("abort", () => controller.abort());
       const res = await fetch(url, {

server/src/config.ts

@@ -187,6 +187,59 @@ function readSettingsPolicy(): { defaults: Record<string, unknown>; enforced: Re
   };
 }
 
+/**
+ * Which Stalwart a domain signs in to.
+ *
+ * `STALWART_URL` stays required and stays the default; this only adds domains
+ * that go somewhere else (#238). An installation that sets nothing behaves
+ * exactly as it always has.
+ *
+ * Read once at boot and never written, so it mounts read-only and costs
+ * nothing in immutability -- the same shape as the settings policy.
+ *
+ * Servers are deliberately **not** probed here. A mapping is a routing table,
+ * not a health check, and refusing to boot because one of five customers is
+ * having an outage would take the other four down with it. What happens when
+ * one is unreachable is a sign-in question, answered in #239.
+ */
+function readStalwartServers(): Record<string, string> {
+  const file = process.env.STALWART_SERVERS_FILE;
+  if (!file) return {};
+  if (!existsSync(file)) throw new Error(`STALWART_SERVERS_FILE does not exist: ${file}`);
+
+  let raw: unknown;
+  try {
+    raw = JSON.parse(readFileSync(file, "utf8"));
+  } catch (err) {
+    throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): ${(err as Error).message}`);
+  }
+  if (!raw || typeof raw !== "object" || Array.isArray(raw)) {
+    throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): expected an object of domain to URL`);
+  }
+
+  const out: Record<string, string> = {};
+  for (const [rawDomain, rawUrl] of Object.entries(raw as Record<string, unknown>)) {
+    /* Lower-cased and stripped of the root dot, because that is how a domain
+       taken off a username will arrive and comparing them any other way means
+       a mapping that silently never matches. */
+    const domain = rawDomain.trim().toLowerCase().replace(/\.$/, "");
+    if (!domain) throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): a domain key is empty`);
+    if (domain in out) throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): "${domain}" appears twice once normalised`);
+    if (typeof rawUrl !== "string") throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): "${domain}" is not a URL`);
+    let parsed: URL;
+    try {
+      parsed = new URL(rawUrl);
+    } catch {
+      throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): "${domain}" is not an absolute URL`);
+    }
+    if (parsed.protocol !== "http:" && parsed.protocol !== "https:") {
+      throw new Error(`Invalid STALWART_SERVERS_FILE (${file}): "${domain}" must be http or https`);
+    }
+    out[domain] = rawUrl.replace(/\/+$/, "");
+  }
+  return out;
+}
+
 export const config = {
   isProd,
   appName: env("APP_NAME", "ihasmail"),
@@ -222,6 +275,7 @@ export const config = {
    */
   basePath: normalizeBasePath(process.env.BASE_PATH),
   stalwartUrl,
+  stalwartServers: readStalwartServers(),
   appSecret,
   trustProxy: bool("TRUST_PROXY", true),
   /**

server/src/domainrouting.test.ts

@@ -0,0 +1,67 @@
+import { test } from "node:test";
+import assert from "node:assert/strict";
+
+process.env.STALWART_URL = "https://default.example";
+
+const { upstreamFor } = await import("./upstream.js");
+const { config } = await import("./config.js");
+
+/**
+ * Which Stalwart a username goes to (#238).
+ *
+ * `STALWART_URL` is required and is the default. The mapping only adds domains
+ * that go elsewhere, so an installation with no mapping behaves exactly as it
+ * always has -- which is what these first cases pin.
+ */
+
+test("with no mapping at all, everything goes to the default", () => {
+  assert.deepEqual(config.stalwartServers, {});
+  assert.equal(upstreamFor("[email protected]"), "https://default.example");
+  assert.equal(upstreamFor("[email protected]"), "https://default.example");
+});
+
+test("a bare username has no domain to map, so it goes to the default", () => {
+  // Stalwart accepts a login with no domain at all.
+  assert.equal(upstreamFor("demo"), "https://default.example");
+  assert.equal(upstreamFor(""), "https://default.example");
+});
+
+test("a mapped domain goes to its own server", () => {
+  config.stalwartServers["mapped.test"] = "https://mail.mapped.test";
+  try {
+    assert.equal(upstreamFor("[email protected]"), "https://mail.mapped.test");
+  } finally {
+    delete config.stalwartServers["mapped.test"];
+  }
+});
+
+test("an unmapped domain still goes to the default while others are mapped", () => {
+  config.stalwartServers["mapped.test"] = "https://mail.mapped.test";
+  try {
+    assert.equal(upstreamFor("[email protected]"), "https://default.example");
+  } finally {
+    delete config.stalwartServers["mapped.test"];
+  }
+});
+
+test("the domain is matched however it was typed", () => {
+  // Keys are normalised on load; the username has to be normalised the same
+  // way or a mapping silently never matches.
+  config.stalwartServers["mapped.test"] = "https://mail.mapped.test";
+  try {
+    assert.equal(upstreamFor("[email protected]"), "https://mail.mapped.test");
+    assert.equal(upstreamFor("[email protected]."), "https://mail.mapped.test", "root dot");
+    assert.equal(upstreamFor("someone@ mapped.test "), "https://mail.mapped.test", "stray spaces");
+  } finally {
+    delete config.stalwartServers["mapped.test"];
+  }
+});
+
+test("an address with an @ in the local part maps on the last one", () => {
+  config.stalwartServers["mapped.test"] = "https://mail.mapped.test";
+  try {
+    assert.equal(upstreamFor('"odd@name"@mapped.test'), "https://mail.mapped.test");
+  } finally {
+    delete config.stalwartServers["mapped.test"];
+  }
+});

server/src/settingspolicy.test.ts

@@ -53,3 +53,30 @@ test("the example's commentary cannot be mistaken for a section", () => {
     assert.ok(real.has(key) || key.startsWith("_"), `unexpected top-level key ${key}`);
   }
 });
+
+/**
+ * The shipped server-mapping example, checked the same way and for the same
+ * reason: an example that no longer loads is worse than no example, because
+ * the first experience of the feature is a server that refuses to start.
+ */
+const SERVERS = fileURLToPath(new URL("../../stalwart-servers.example.json", import.meta.url));
+
+test("the example server mapping is valid JSON", () => {
+  assert.doesNotThrow(() => JSON.parse(readFileSync(SERVERS, "utf8")));
+});
+
+test("every entry in the example mapping is a domain and an http(s) URL", () => {
+  const m = JSON.parse(readFileSync(SERVERS, "utf8")) as Record<string, unknown>;
+  const seen = new Set<string>();
+  for (const [key, value] of Object.entries(m)) {
+    if (key.startsWith("_")) continue;
+    const domain = key.trim().toLowerCase().replace(/\.$/, "");
+    assert.ok(domain, "a domain key is empty");
+    assert.ok(!seen.has(domain), `${domain} appears twice once normalised`);
+    seen.add(domain);
+    assert.equal(typeof value, "string", `${domain} is not a string`);
+    const url = new URL(value as string);
+    assert.ok(url.protocol === "http:" || url.protocol === "https:", `${domain} must be http or https`);
+  }
+  assert.ok(seen.size > 0, "the example should show at least one mapping");
+});

server/src/upstream.ts

@@ -10,6 +10,15 @@ export interface UpstreamSession {
   uploadUrl: string;
   eventSourceUrl: string;
   state: string;
+  /**
+   * Which Stalwart this document came from.
+   *
+   * Recorded rather than looked up again, because the relative URLs inside it
+   * -- apiUrl, uploadUrl and the rest -- only mean anything against the server
+   * that issued them. Anything holding a session already knows where to send
+   * the next request. Not part of the JMAP session resource; ours.
+   */
+  baseUrl: string;
 }
 
 export class UpstreamError extends Error {
@@ -24,16 +33,37 @@ export class UpstreamError extends Error {
 const sessionCache = new Map<string, { session: UpstreamSession; fetchedAt: number }>();
 const SESSION_CACHE_MS = 5 * 60_000;
 
-export function wellKnownUrl(): string {
-  return `${config.stalwartUrl}/.well-known/jmap`;
+/**
+ * The Stalwart a username belongs to.
+ *
+ * `STALWART_URL` is the default and is always the answer for a domain nobody
+ * mapped -- and for a bare username, which Stalwart accepts and which has no
+ * domain to map (#238).
+ *
+ * A *mapped* domain never falls back. If its server is unreachable that
+ * sign-in fails, because falling back would authenticate somebody against a
+ * server their domain was deliberately routed away from -- and if the same
+ * account name exists there, they would land in another tenant's mailbox. The
+ * fallback is a decision about unmapped domains, taken before any network
+ * call, not a recovery path.
+ */
+export function upstreamFor(username: string): string {
+  const at = username.lastIndexOf("@");
+  if (at < 0) return config.stalwartUrl;
+  const domain = username.slice(at + 1).trim().toLowerCase().replace(/\.$/, "");
+  return config.stalwartServers[domain] ?? config.stalwartUrl;
+}
+
+export function wellKnownUrl(base: string = config.stalwartUrl): string {
+  return `${base}/.well-known/jmap`;
 }
 
 /**
  * Fetch the JMAP session resource from Stalwart using the given Authorization
  * header. Throws UpstreamError(401) on bad credentials.
  */
-export async function fetchUpstreamSession(authorization: string): Promise<UpstreamSession> {
-  const res = await fetch(wellKnownUrl(), {
+export async function fetchUpstreamSession(authorization: string, base: string = config.stalwartUrl): Promise<UpstreamSession> {
+  const res = await fetch(wellKnownUrl(base), {
     headers: { authorization, accept: "application/json" },
     redirect: "follow",
     signal: AbortSignal.timeout(config.upstreamTimeout),
@@ -46,13 +76,13 @@ export async function fetchUpstreamSession(authorization: string): Promise<Upstr
   }
   const session = (await res.json()) as UpstreamSession;
   if (!session.apiUrl) throw new UpstreamError("Upstream returned an invalid JMAP session", 502);
-  return session;
+  return { ...session, baseUrl: base };
 }
 
-export async function getUpstreamSession(sessionId: string, authorization: string, force = false) {
+export async function getUpstreamSession(sessionId: string, authorization: string, base: string = config.stalwartUrl, force = false) {
   const cached = sessionCache.get(sessionId);
   if (!force && cached && Date.now() - cached.fetchedAt < SESSION_CACHE_MS) return cached.session;
-  const session = await fetchUpstreamSession(authorization);
+  const session = await fetchUpstreamSession(authorization, base);
   sessionCache.set(sessionId, { session, fetchedAt: Date.now() });
   return session;
 }
@@ -210,9 +240,9 @@ function localeOf(call: [string, Record<string, unknown>, string] | undefined):
  * Which edition the server is running. Stalwart deliberately does not publish
  * its version number to clients, but 0.16 does report its edition here.
  */
-async function fetchEdition(authorization: string): Promise<string | null> {
+async function fetchEdition(authorization: string, base: string): Promise<string | null> {
   try {
-    const res = await fetch(`${config.stalwartUrl}/api/account`, {
+    const res = await fetch(`${base}/api/account`, {
       headers: { authorization, accept: "application/json" },
       signal: AbortSignal.timeout(config.upstreamTimeout),
     });
@@ -230,7 +260,7 @@ export async function getAccountInfo(sessionId: string, authorization: string, s
   let info = EMPTY_INFO;
   try {
     info = await fetchAccountInfo(authorization, session);
-    info = { ...info, edition: await fetchEdition(authorization) };
+    info = { ...info, edition: await fetchEdition(authorization, session.baseUrl) };
   } catch {
     /* all of this is a nicety - never fail the session over it */
   }
@@ -258,9 +288,9 @@ export function localizeSession(s: UpstreamSession, extras: Record<string, unkno
 }
 
 /** Resolve a possibly-relative upstream URL template against STALWART_URL. */
-export function absoluteUpstream(url: string): string {
+export function absoluteUpstream(url: string, base: string = config.stalwartUrl): string {
   try {
-    return new URL(url, config.stalwartUrl).toString();
+    return new URL(url, base).toString();
   } catch {
     return url;
   }