Drop Stalwart 0.15 support

ihasmail spoke to two generations of Stalwart that are less alike than
their version numbers suggest: 0.16 replaced the REST management API with
JMAP registry objects, changed the shape of FileNode, split its rights
up, and moved configuration into the store. Carrying both meant 34 branch
points across nine files, a 92-line compatibility shim whose only job was
telling them apart, a parallel REST implementation of every credential
operation, and a mock that had to model both.

The branches were not the real cost. The cost was that a wrong answer
about which generation had answered always had somewhere to fall back to,
so it failed quietly rather than loudly: one capability looked for in the
wrong place downgraded every real 0.16 server onto the 0.15 path, which
posted the current password to an endpoint 0.16 had removed, reported the
wrong generation on About, and ran Files on the older code. It reached
production and was recorded as verified when it was not. The mock
mirrored the same wrong placement, which is why the tests agreed.

Removed: the filenode compatibility shim, the dual "registry" | "legacy"
backend in account.ts, the pre-0.16 generation in AccountInfo and
everything that read it, the mock's LEGACY mode and dev:mock:legacy, and
the three test files that existed only to pin 0.15 behaviour.

Sign-in now refuses an older server by name, once, rather than letting
Files, the account locale and credentials each fail in their own way with
nothing connecting them. It says the credentials were fine -- someone
hitting this has typed a correct password, and telling them otherwise
sends them round in circles -- and names the tag to build from. Four
tests cover it, including that no session cookie is minted and that bad
credentials on such a server are still a plain 401.

Two fallbacks went that were not strictly about 0.15, and both for the
same reason the removal is happening. Files no longer answers a refused
filter or sort by fetching every node in the account, which would hide a
real fault behind a performance cliff nobody would notice. And the app
folder lookups now filter on parentId/isTopLevel alone and match names
client-side, since `name` is not a filter Stalwart is known to implement
and one it does not know fails the whole query rather than being ignored.

The last release that runs on 0.15 is tagged stalwart-0.15-support.

Verified against the mock end to end: sign-in, the Files tree on the 0.16
path with the app folder hidden, and self-service credentials over the
registry. 226 web + 75 server tests pass; typecheck and build clean.
This commit is contained in:
2026-08-26 09:51:03 -07:00
parent f41e3d2631
commit 94bf42cfda
20 changed files with 315 additions and 1036 deletions
+7 -59
View File
@@ -1,6 +1,6 @@
import { create } from "zustand";
import { CAP, JmapMethodError, client, setErrorMessage } from "@/jmap/client";
import { directoryCreate, fileCreate, fileNodeProps, normalizeFileNodes, queryOmitsDirectories } from "@/lib/filenode";
import { CAP, client, setErrorMessage } from "@/jmap/client";
import { directoryCreate, fileCreate, fileNodeProps } from "@/lib/filenode";
import { isAppFolder } from "@/lib/appFolder";
import type { FileNode, GetResponse, Id, QueryResponse, SetResponse } from "@/jmap/types";
import { useSession } from "./session";
@@ -26,10 +26,6 @@ interface FilesState {
}
/** Whether the server supports parentId/isTopLevel query filters (detected at runtime). */
let filtersSupported = true;
const byName = (a: FileNode, b: FileNode) => (a.nodeType === b.nodeType ? a.name.localeCompare(b.name, undefined, { numeric: true, sensitivity: "base" }) : a.nodeType === "directory" ? -1 : 1);
/**
* Drop the client's own `ihasmail` folder, and everything inside it, from a
@@ -56,42 +52,6 @@ export function withoutAppFolder(nodes: FileNode[]): FileNode[] {
return nodes.filter((n) => !hidden.has(n.id));
}
/** Fetch all nodes (paged, no filter) and rebuild the full children map. */
async function loadAllNodes(accountId: Id, set: (fn: (s: FilesState) => Partial<FilesState>) => void): Promise<void> {
const all: FileNode[] = [];
if (queryOmitsDirectories()) {
// Query would hand back files only, so every folder — including one just
// created — would be missing with nothing to say why. Ask get for the lot.
const res = await client.call<GetResponse<FileNode>>("FileNode/get", { accountId, ids: null, properties: fileNodeProps() });
all.push(...normalizeFileNodes(res.list));
} else {
let position = 0;
for (let guard = 0; guard < 100; guard++) {
const res = await client.chain([
["FileNode/query", { accountId, position, limit: 500, calculateTotal: true }, "q"],
["FileNode/get", { accountId, "#ids": { resultOf: "q", name: "FileNode/query", path: "/ids" }, properties: fileNodeProps() }, "g"],
]);
const q = res.get("q")?.[0] as unknown as QueryResponse;
const g = res.get("g")?.[0] as unknown as GetResponse<FileNode>;
all.push(...normalizeFileNodes(g.list));
position += q.ids.length;
if (!q.ids.length || (q.total != null && position >= q.total)) break;
}
}
// After the whole collection, not per page: the folder and its contents can
// land in different pages, and a half-filtered pass would spill the rest.
const visible = withoutAppFolder(all);
const nodes: Record<Id, FileNode> = {};
const children: Record<string, Id[]> = { root: [] };
for (const n of visible) nodes[n.id] = n;
for (const n of visible.sort(byName)) {
const key = n.parentId && nodes[n.parentId] ? n.parentId : "root";
(children[key] ??= []).push(n.id);
}
for (const n of visible) children[n.id] ??= [];
set(() => ({ nodes, children, loading: false, error: null }));
}
export const useFiles = create<FilesState>((set, get) => ({
accountId: null,
available: false,
@@ -113,10 +73,6 @@ export const useFiles = create<FilesState>((set, get) => ({
if (!accountId) return;
set({ loading: true });
try {
if (!filtersSupported || queryOmitsDirectories()) {
await loadAllNodes(accountId, set);
return;
}
const filter = parentId ? { parentId } : { isTopLevel: true };
const res = await client.chain([
["FileNode/query", { accountId, filter, sort: [{ property: "nodeType", isAscending: false }, { property: "name", isAscending: true }], limit: 1000 }, "q"],
@@ -124,7 +80,7 @@ export const useFiles = create<FilesState>((set, get) => ({
]);
const q = res.get("q")?.[0] as unknown as QueryResponse;
const g = res.get("g")?.[0] as unknown as GetResponse<FileNode>;
const listed = withoutAppFolder(normalizeFileNodes(g.list));
const listed = withoutAppFolder(g.list);
const keep = new Set(listed.map((n) => n.id));
set((s) => {
const nodes = { ...s.nodes };
@@ -132,18 +88,10 @@ export const useFiles = create<FilesState>((set, get) => ({
return { nodes, children: { ...s.children, [parentId ?? "root"]: q.ids.filter((id) => keep.has(id)) }, loading: false, error: null };
});
} catch (err) {
// Older Stalwart releases don't support parentId / isTopLevel filters: fall back to
// fetching every node and building the tree client-side.
if (err instanceof JmapMethodError && (err.type === "unsupportedFilter" || err.type === "unsupportedSort")) {
filtersSupported = false;
try {
await loadAllNodes(accountId, set);
return;
} catch (err2) {
set({ loading: false, error: (err2 as Error).message });
return;
}
}
// There used to be a fallback here that abandoned filters and fetched
// every node in the account, because 0.15 refused parentId/isTopLevel.
// 0.16 supports them, and quietly loading the whole tree instead would
// hide a real fault behind a performance cliff nobody would notice.
set({ loading: false, error: (err as Error).message });
}
},