Skip to content
Sign up

flagon-io/g1t

public
0
251 linesCodeBlameRaw
1import { createRequestHandler } from "react-router";
2
3import { identityClient, isNamespaceShaped } from "@g1t/contracts";
4
5import { addressesFor } from "../app/lib/addresses";
6import { visitorAsksFirst } from "../app/lib/analytics-consent";
7import { hardenRegistryHeaders } from "../app/lib/content-safety";
8import { withSiteHeaders } from "../app/lib/page-headers";
9import { finishResponse, withRequestPerf } from "../app/lib/perf.server";
10import { READ_DEADLINE_MS, withDeadline } from "../app/lib/deadline";
11import { gitLimited, pageLimited, usercontentLimited } from "../app/lib/front-door-limits";
12import { goImport } from "../app/lib/go-get";
13import { repositoryOfPage, stillPublic } from "../app/lib/public-cache";
14import { registryWorkspace, servicePath } from "../app/lib/registry-paths";
15import { usercontentPath } from "../app/lib/usercontent";
16import { TOKEN_CHALLENGE } from "../app/lib/website-token";
17import { serveUsercontent } from "./usercontent";
18
19const loadBuild = () => import("virtual:react-router/server-build");
20
21/**
22 * Given a function for the build, React Router derives everything from it
23 * again on every request: each route wrapped for the timings in
24 * entry.server.tsx, then the route table flattened and ranked, about a
25 * millisecond and a half of CPU a page. The build never changes while an
26 * isolate lives, so outside development it is loaded once and the handler
27 * made once. Development keeps the function, so a changed file is picked up.
28 * Only the made handler is kept, never a promise of one, so no request
29 * waits on another's (a build that fails to load is loaded again next time).
30 */
31let handler: ReturnType<typeof createRequestHandler> | undefined;
32async function requestHandler(request: Request): Promise<Response> {
33 if (import.meta.env.DEV) return createRequestHandler(loadBuild, import.meta.env.MODE)(request);
34 handler ??= createRequestHandler(await loadBuild(), import.meta.env.MODE);
35 return handler(request);
36}
37
38const DOCS = "https://docs.g1t.sh";
39
40/** Where the documentation pages that used to live under /docs are now. */
41const MOVED_DOCS: Record<string, string> = {
42 "/docs": "/quickstart/",
43 "/docs/concepts": "/concepts/overview/",
44 "/docs/authentication": "/guides/authentication/",
45 "/docs/git": "/guides/git/",
46 "/docs/g1t-agents": "/guides/working-with-g1t/",
47 "/docs/agents": "/guides/bring-your-own-agent/",
48 "/docs/api": "/reference/api/",
49 "/docs/api/reference": "/reference/api/",
50};
51
52export default {
53 async fetch(request, env, ctx) {
54 // Repository files and avatars, on their own origin
55 // (g1tusercontent.com): answered before anything of the site's runs,
56 // so nothing there reads or sets the session cookie.
57 const usercontent = usercontentPath(new URL(request.url), addressesFor(env).usercontent);
58 if (usercontent !== null) {
59 return (await usercontentLimited(env, request, usercontent)) ?? serveUsercontent(env, ctx, request, usercontent);
60 }
61 // Every answer: no sniffing, a referrer of the origin alone, and no
62 // framing of pages (app/lib/page-headers.ts).
63 return withSiteHeaders(await site(request, env, ctx));
64 },
65} satisfies ExportedHandler<Env>;
66
67async function site(request: Request, env: Env, ctx: ExecutionContext): Promise<Response> {
68 const { pathname } = new URL(request.url);
69 // Git over HTTPS shares this hostname but belongs to the repos service.
70 // Its answer goes back to the git client as it is: a repository under a
71 // renamed workspace's old name answers with a 301, which git follows and
72 // must see, so the redirect is never followed here.
73 // The container registry (`docker login g1t.sh`) and the npm registry
74 // (`g1t.sh/-/npm/`) are the packages
75 // service's, handed over the same way.
76 // `go get g1t.sh/<workspace>/<repo>`: where its code is, from the
77 // address alone, so it costs nothing and caches.
78 const go = request.method === "GET" ? goImport(new URL(request.url)) : null;
79 if (go) {
80 return new Response(go, {
81 headers: { "content-type": "text/html; charset=utf-8", "cache-control": "public, max-age=3600" },
82 });
83 }
84 const service = servicePath(pathname);
85 if (service === "git") {
86 // Per address without credentials, per credential with them
87 // (app/lib/front-door-limits.ts): anonymous clones are not free to
88 // the repository's owner.
89 return (await gitLimited(env, request)) ?? proxyGit(env, request);
90 }
91 if (service === "packages") {
92 return proxyPackages(env, request);
93 }
94 // The documentation is its own site.
95 if (pathname === "/docs" || pathname.startsWith("/docs/")) {
96 const page = pathname.endsWith("/") ? pathname.slice(0, -1) : pathname;
97 const target = MOVED_DOCS[page] ?? "/";
98 return Response.redirect(DOCS + target, 301);
99 }
100 // Pages and data requests, limited per address signed out and per
101 // session signed in (app/lib/front-door-limits.ts).
102 const limited = await pageLimited(env, request, pathname);
103 if (limited) return limited;
104 // Every page and data request says where its time went (Server-Timing)
105 // and keeps the reader's D1 bookmarks (app/lib/perf.server.ts).
106 const render = () => withRequestPerf(request, async () => finishResponse(request, await requestHandler(request)));
107 if (anonymousPage(request, pathname)) return servePublic(env, request, ctx, render);
108 const answer = await render();
109 // A token the site refused (app/lib/website-token.ts) is told so as the
110 // API tells it: React Router drops the headers of what middleware throws.
111 if (answer.status === 401 && request.headers.has("authorization")) answer.headers.set("www-authenticate", TOKEN_CHALLENGE);
112 return answer;
113}
114
115/**
116 * Public pages as someone signed out sees them: the same for every such
117 * visitor, so kept in this data centre's cache. Reserved first segments
118 * (settings, sign-in, invitations and the like) and workspace pages (`-`)
119 * are never kept. A repository's kept page is served only while repos
120 * says the repository is still public: one made private or deleted is
121 * never served from any data centre's copy.
122 */
123const PUBLIC_TOP = /^\/(?:|_root\.data|pricing|explore|security|support|policies(?:\/[a-z-]+)?)(?:\.data)?$/;
124const PUBLIC_PROJECT =
125 /^\/(?!(?:settings|u|auth|oauth|integrations|new|invite|workspaces|device|verify|confirm-email|login|register|logout|forgot|reset|search|status|avatars|docs)\/)[^/]+\/(?!-\/|-$)[^/]+(?:\/(?:code|commits|issues|pulls|pull\/\d+|issues\/\d+|commit\/[0-9a-f]+|tree\/.+|blob\/.+))?(?:\.data)?$/;
126/** Fresh for this long; then served once more while a new copy is made. */
127const PUBLIC_FRESH_SECONDS = 30;
128const PUBLIC_STALE_SECONDS = 300;
129/** The longest a cache lookup holds up a public page. */
130const CACHE_DEADLINE_MS = 2_000;
131
132function anonymousPage(request: Request, pathname: string): boolean {
133 if (request.method !== "GET") return false;
134 if (/(?:^|;\s*)g1t_session=/.test(request.headers.get("cookie") ?? "")) return false;
135 // A token signs the request in (app/lib/website-token.ts): never kept, never served a kept page.
136 if (request.headers.has("authorization")) return false;
137 return PUBLIC_TOP.test(pathname) || PUBLIC_PROJECT.test(pathname);
138}
139
140async function servePublic(env: Env, request: Request, ctx: ExecutionContext, render: () => Promise<Response>): Promise<Response> {
141 const cache = (caches as unknown as { default: Cache }).default;
142 // Visitors asked about analytics first get a page that says so, kept apart.
143 const keyUrl = new URL(request.url);
144 if (visitorAsksFirst(request)) keyUrl.searchParams.set("_g1t_consent", "1");
145 const key = new Request(keyUrl, { method: "GET" });
146 const repository = PUBLIC_PROJECT.test(new URL(request.url).pathname) ? repositoryOfPage(new URL(request.url).pathname) : null;
147 // Asked alongside the cache, so a hit waits for one indexed read at most.
148 // A cache that does not answer is a miss: the page is drawn afresh.
149 const [cached, visible] = await Promise.all([
150 withDeadline("cache", CACHE_DEADLINE_MS, () => cache.match(key)).catch(() => undefined),
151 repository ? isStillPublic(env, repository) : Promise.resolve(true),
152 ]);
153 if (cached && !visible) {
154 ctx.waitUntil(cache.delete(key).then(() => undefined, () => undefined));
155 return render();
156 }
157 const keptAt = Number(cached?.headers.get("x-g1t-kept-at") ?? 0);
158 const age = Math.round((Date.now() - keptAt) / 1000);
159 const refresh = async () => {
160 const fresh = await render();
161 // Only a plain answer for everyone: nothing that sets a cookie or says
162 // it is private.
163 const cacheable =
164 (fresh.status === 200 || fresh.status === 404) &&
165 !fresh.headers.has("set-cookie") &&
166 !/private|no-store/.test(fresh.headers.get("cache-control") ?? "");
167 if (cacheable) {
168 const copy = new Response(fresh.clone().body, fresh);
169 copy.headers.set("x-g1t-kept-at", String(Date.now()));
170 copy.headers.set("cache-control", `public, max-age=${PUBLIC_STALE_SECONDS}`);
171 ctx.waitUntil(cache.put(key, copy));
172 }
173 return fresh;
174 };
175 if (cached && keptAt > 0 && age < PUBLIC_STALE_SECONDS) {
176 if (age >= PUBLIC_FRESH_SECONDS) ctx.waitUntil(refresh().then(() => undefined, () => undefined));
177 const answer = new Response(cached.body, cached);
178 answer.headers.delete("x-g1t-kept-at");
179 answer.headers.delete("cache-control");
180 answer.headers.set("server-timing", `cache;desc="hit, ${age}s old"`);
181 return answer;
182 }
183 return refresh();
184}
185
186/** Whether the repository is there and public; anything else, including no answer, is no. */
187async function isStillPublic(env: Env, repository: string): Promise<boolean> {
188 try {
189 return await withDeadline("repos visibility", READ_DEADLINE_MS, async (signal) => {
190 const answer = await env.REPOS.fetch("https://service/rpc/visibility", {
191 method: "POST",
192 headers: { "content-type": "application/json" },
193 body: JSON.stringify({ paths: [repository] }),
194 signal,
195 });
196 return answer.ok && stillPublic(await answer.json(), repository);
197 });
198 } catch {
199 return false;
200 }
201}
202
203/**
204 * A git request, answered by the repos service. Its `Server-Timing` header
205 * gains `repos`: how long the answer took to start from here, so the time
206 * between this Worker and the repos service shows beside the steps the
207 * repos service reports.
208 */
209async function proxyGit(env: Env, request: Request): Promise<Response> {
210 const started = Date.now();
211 const answer = await env.REPOS.fetch(new Request(request, { redirect: "manual" }));
212 const response = new Response(answer.body, answer);
213 response.headers.append("server-timing", `repos;dur=${Date.now() - started}`);
214 return response;
215}
216
217/**
218 * A registry request, answered by the packages service as it is: its
219 * redirects (a large blob sent to storage) go back to the client, which
220 * follows them itself. One that found nothing under a workspace's old name
221 * or an alias staff set (`g1t` for `flagon-io`) is sent to the same path
222 * under the workspace's name: only the not-found answer pays for the lookup.
223 * Every answer runs nothing in a browser (app/lib/content-safety.ts): what
224 * a registry serves is its publisher's, on this origin.
225 */
226async function proxyPackages(env: Env, request: Request): Promise<Response> {
227 const started = Date.now();
228 const answer = await env.PACKAGES.fetch(new Request(request, { redirect: "manual" }));
229 const moved = answer.status === 404 ? await registryMoved(env, request) : null;
230 const response = moved ?? new Response(answer.body, answer);
231 hardenRegistryHeaders(response.headers);
232 response.headers.append("server-timing", `packages;dur=${Date.now() - started}`);
233 return response;
234}
235
236/** Where a registry request under an alias or old name goes now, or null. */
237async function registryMoved(env: Env, request: Request): Promise<Response | null> {
238 const url = new URL(request.url);
239 const named = registryWorkspace(url.pathname);
240 if (!named || !isNamespaceShaped(named.slug)) return null;
241 let current: string | null = null;
242 try {
243 current = await identityClient(env.IDENTITY).resolveSlug(named.slug);
244 } catch {
245 return null;
246 }
247 if (!current || current === named.slug) return null;
248 const get = request.method === "GET" || request.method === "HEAD";
249 // 308 keeps a publish a PUT, for the clients that follow it.
250 return new Response(null, { status: get ? 301 : 308, headers: { location: named.under(current) + url.search } });
251}