@astryxdesign/cli 0.1.2 → 0.1.3

package/src/api/blog.mjs

@@ -0,0 +1,192 @@
+// Copyright (c) Meta Platforms, Inc. and affiliates.
+
+/**
+ * @file Programmatic API for the blog command.
+ *
+ * The blog is read the same way any feed reader reads it: over the published
+ * RSS feed. Listing parses the feed; reading a post fetches the plaintext
+ * (.txt) alternate the feed advertises for each item. Nothing here touches the
+ * blog's source files — the CLI is just a consumer of the public feed, so the
+ * blog's structure can change freely without touching the CLI.
+ */
+
+import {AstryxError} from './error.mjs';
+import {ERROR_CODES} from '../lib/error-codes.mjs';
+import {SITE_URL, SITE_ORIGIN} from '../lib/site.mjs';
+
+/** Abort a feed/post fetch that hangs, and cap how much we'll read. */
+const FETCH_TIMEOUT_MS = 15000;
+const MAX_BYTES = 5 * 1024 * 1024; // 5 MB — a blog feed is never larger.
+
+const FEED_URL = new URL('/rss.xml', SITE_URL).toString();
+
+async function fetchText(url) {
+  const controller = new AbortController();
+  const timer = setTimeout(() => controller.abort(), FETCH_TIMEOUT_MS);
+  let res;
+  try {
+    res = await fetch(url, {
+      signal: controller.signal,
+      redirect: 'follow',
+    });
+  } catch (e) {
+    clearTimeout(timer);
+    const reason = e.name === 'AbortError' ? 'timed out' : e.message;
+    throw new AstryxError(
+      `Could not reach ${url}: ${reason}`,
+      [],
+      ERROR_CODES.ERR_FETCH_FAILED,
+    );
+  }
+  clearTimeout(timer);
+  if (!res.ok) {
+    throw new AstryxError(
+      `Request to ${url} failed with ${res.status}`,
+      [],
+      ERROR_CODES.ERR_FETCH_FAILED,
+    );
+  }
+  const body = await res.text();
+  if (body.length > MAX_BYTES) {
+    throw new AstryxError(
+      `Response from ${url} exceeded ${MAX_BYTES} bytes`,
+      [],
+      ERROR_CODES.ERR_FETCH_FAILED,
+    );
+  }
+  return body;
+}
+
+/**
+ * Defense in depth: a post's plaintext URL comes from feed content. Require it
+ * to live on the canonical origin so even a tampered feed can't redirect the
+ * CLI to another host.
+ */
+function assertCanonicalOrigin(target) {
+  let targetOrigin;
+  try {
+    targetOrigin = new URL(target).origin;
+  } catch {
+    throw new AstryxError(
+      `Post has an invalid plaintext URL: "${target}"`,
+      [],
+      ERROR_CODES.ERR_FETCH_FAILED,
+    );
+  }
+  if (targetOrigin !== SITE_ORIGIN) {
+    throw new AstryxError(
+      `Refusing to fetch post text from a non-canonical origin (${targetOrigin})`,
+      [],
+      ERROR_CODES.ERR_FETCH_FAILED,
+    );
+  }
+}
+
+function unescapeXml(value) {
+  return value
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&quot;/g, '"')
+    .replace(/&apos;/g, "'")
+    .replace(/&amp;/g, '&');
+}
+
+function tag(item, name) {
+  const m = item.match(new RegExp(`<${name}[^>]*>([\\s\\S]*?)</${name}>`));
+  return m ? unescapeXml(m[1].trim()) : '';
+}
+
+function tagAll(item, name) {
+  const out = [];
+  const re = new RegExp(`<${name}[^>]*>([\\s\\S]*?)</${name}>`, 'g');
+  let m;
+  while ((m = re.exec(item))) out.push(unescapeXml(m[1].trim()));
+  return out;
+}
+
+/** Extract the plaintext alternate href from an <item>. */
+function textHref(item) {
+  // Match the atom:link alternate regardless of attribute order/quoting; then
+  // confirm it's the text/plain alternate before trusting the href.
+  const re = /<atom:link\b[^>]*?\/?>/g;
+  let m;
+  while ((m = re.exec(item))) {
+    const el = m[0];
+    if (/rel\s*=\s*["']alternate["']/.test(el) &&
+        /type\s*=\s*["']text\/plain["']/.test(el)) {
+      const href = el.match(/href\s*=\s*["']([^"']+)["']/);
+      if (href) return unescapeXml(href[1]);
+    }
+  }
+  return null;
+}
+
+/** Derive a slug from a post link (last path segment). */
+function slugFromLink(link) {
+  try {
+    const path = new URL(link).pathname.replace(/\/$/, '');
+    return path.slice(path.lastIndexOf('/') + 1);
+  } catch {
+    return link;
+  }
+}
+
+function parseFeed(xml) {
+  const items = [];
+  const re = /<item>([\s\S]*?)<\/item>/g;
+  let m;
+  while ((m = re.exec(xml))) {
+    const raw = m[1];
+    const link = tag(raw, 'link');
+    items.push({
+      slug: slugFromLink(link),
+      title: tag(raw, 'title'),
+      description: tag(raw, 'description'),
+      date: tag(raw, 'pubDate'),
+      type: tag(raw, 'category'),
+      authors: tagAll(raw, 'author'),
+      link,
+      textUrl: textHref(raw),
+    });
+  }
+  return items;
+}
+
+/**
+ * List posts (from the feed), or read one post (via its .txt alternate).
+ * Both envelopes carry `feedUrl` so a caller can hit the RSS feed directly.
+ * The feed is always the canonical site — there is no user-supplied URL.
+ *
+ * @param {string} [slug]
+ * @returns {Promise<{type: string, data: unknown}>}
+ */
+export async function blog(slug) {
+  const xml = await fetchText(FEED_URL);
+  const posts = parseFeed(xml);
+
+  if (!slug) {
+    return {type: 'blog.list', data: {feedUrl: FEED_URL, posts}};
+  }
+
+  const normalized = slug.toLowerCase();
+  const post = posts.find(p => p.slug.toLowerCase() === normalized);
+  if (!post) {
+    throw new AstryxError(
+      `No blog post with slug "${slug}"`,
+      posts.map(p => ({name: p.slug, reason: 'available post'})),
+      ERROR_CODES.ERR_UNKNOWN_POST,
+    );
+  }
+
+  if (!post.textUrl) {
+    throw new AstryxError(
+      `Post "${slug}" has no plaintext alternate in the feed`,
+      [],
+      ERROR_CODES.ERR_FETCH_FAILED,
+    );
+  }
+
+  assertCanonicalOrigin(post.textUrl);
+  const text = await fetchText(post.textUrl);
+  return {type: 'blog.detail', data: {...post, feedUrl: FEED_URL, text}};
+}

package/src/api/blog.test.mjs

@@ -0,0 +1,144 @@
+// Copyright (c) Meta Platforms, Inc. and affiliates.
+
+/**
+ * @file Tests for the blog API — reads the blog over the canonical RSS feed
+ * and fetches each post's plaintext (.txt) alternate. `fetch` is stubbed so
+ * the tests never hit the network. The feed origin is fixed (not
+ * user-configurable), so the tests assert the canonical URLs directly.
+ */
+
+import {describe, it, expect, beforeEach, afterEach, vi} from 'vitest';
+import {blog} from './blog.mjs';
+import {AstryxError} from './error.mjs';
+import {SITE_URL} from '../lib/site.mjs';
+
+const FEED = `<?xml version="1.0" encoding="UTF-8"?>
+<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
+  <channel>
+    <title>Astryx Blog</title>
+    <link>${SITE_URL}/blog</link>
+    <item>
+      <title>How Astryx works</title>
+      <link>${SITE_URL}/blog/how-astryx-works</link>
+      <guid isPermaLink="true">${SITE_URL}/blog/how-astryx-works</guid>
+      <description>Under the hood &amp; more</description>
+      <category>engineering</category>
+      <author>cvkxx</author>
+      <author>cixzhang</author>
+      <pubDate>Mon, 29 Jun 2026 00:00:00 GMT</pubDate>
+      <atom:link rel="alternate" type="text/plain" href="${SITE_URL}/blog/how-astryx-works.txt" />
+    </item>
+    <item>
+      <title>Introducing Astryx</title>
+      <link>${SITE_URL}/blog/introducing-astryx</link>
+      <guid isPermaLink="true">${SITE_URL}/blog/introducing-astryx</guid>
+      <description>The launch</description>
+      <category>update</category>
+      <author>cvkxx</author>
+      <pubDate>Thu, 18 Jun 2026 00:00:00 GMT</pubDate>
+      <atom:link rel="alternate" type="text/plain" href="${SITE_URL}/blog/introducing-astryx.txt" />
+    </item>
+  </channel>
+</rss>`;
+
+const POST_TEXT = '# How Astryx works\n\nThe body of the post.';
+const FEED_URL = `${SITE_URL}/rss.xml`;
+const TXT_URL = `${SITE_URL}/blog/how-astryx-works.txt`;
+
+/** Build a fetch stub from a url→{status,body} map. */
+function stubFetch(routes) {
+  return vi.fn(async (url, opts) => {
+    const u = String(url);
+    const r = routes[u];
+    if (!r) return {ok: false, status: 404, text: async () => 'not found'};
+    if (r.throw) throw r.throw;
+    return {ok: r.status < 400, status: r.status, text: async () => r.body};
+  });
+}
+
+beforeEach(() => {
+  vi.stubGlobal('fetch', stubFetch({
+    [FEED_URL]: {status: 200, body: FEED},
+    [TXT_URL]: {status: 200, body: POST_TEXT},
+    [`${SITE_URL}/blog/introducing-astryx.txt`]: {status: 200, body: 'launch body'},
+  }));
+});
+
+afterEach(() => {
+  vi.unstubAllGlobals();
+});
+
+describe('blog API', () => {
+  it('lists posts parsed from the canonical feed', async () => {
+    const res = await blog();
+    expect(res.type).toBe('blog.list');
+    expect(res.data.feedUrl).toBe(FEED_URL);
+    expect(res.data.posts.map(p => p.slug)).toEqual([
+      'how-astryx-works',
+      'introducing-astryx',
+    ]);
+    expect(res.data.posts[0].authors).toEqual(['cvkxx', 'cixzhang']);
+    expect(res.data.posts[0].textUrl).toBe(TXT_URL);
+    // Entity decoding works (&amp; -> &).
+    expect(res.data.posts[0].description).toBe('Under the hood & more');
+    // The list never fetches post bodies.
+    expect(res.data.posts[0].text).toBeUndefined();
+  });
+
+  it('reads a post via its plaintext alternate', async () => {
+    const res = await blog('how-astryx-works');
+    expect(res.type).toBe('blog.detail');
+    expect(res.data.text).toBe(POST_TEXT);
+    expect(res.data.feedUrl).toBe(FEED_URL);
+    expect(fetch).toHaveBeenCalledWith(TXT_URL, expect.any(Object));
+  });
+
+  it('is case-insensitive on the slug', async () => {
+    const res = await blog('How-Astryx-Works');
+    expect(res.data.slug).toBe('how-astryx-works');
+  });
+
+  it('throws ERR_UNKNOWN_POST with suggestions for a bad slug', async () => {
+    await expect(blog('does-not-exist')).rejects.toMatchObject({
+      code: 'ERR_UNKNOWN_POST',
+    });
+    try {
+      await blog('does-not-exist');
+    } catch (e) {
+      expect(e).toBeInstanceOf(AstryxError);
+      expect(Array.isArray(e.suggestions)).toBe(true);
+      expect(e.suggestions.length).toBeGreaterThan(0);
+    }
+  });
+
+  it('throws ERR_FETCH_FAILED when the feed request fails', async () => {
+    vi.stubGlobal('fetch', stubFetch({
+      [FEED_URL]: {status: 500, body: 'boom'},
+    }));
+    await expect(blog()).rejects.toMatchObject({code: 'ERR_FETCH_FAILED'});
+  });
+
+  it('refuses a post whose plaintext URL is on a different origin (SSRF guard)', async () => {
+    const evilFeed = FEED.replace(
+      `${SITE_URL}/blog/how-astryx-works.txt`,
+      'http://169.254.169.254/latest/meta-data',
+    );
+    vi.stubGlobal('fetch', stubFetch({
+      [FEED_URL]: {status: 200, body: evilFeed},
+      'http://169.254.169.254/latest/meta-data': {status: 200, body: 'SECRETS'},
+    }));
+    await expect(blog('how-astryx-works')).rejects.toMatchObject({
+      code: 'ERR_FETCH_FAILED',
+    });
+    // The internal host must never be fetched.
+    expect(fetch).not.toHaveBeenCalledWith(
+      'http://169.254.169.254/latest/meta-data',
+      expect.anything(),
+    );
+  });
+
+  it('always reads from the canonical feed URL', async () => {
+    await blog();
+    expect(fetch).toHaveBeenCalledWith(FEED_URL, expect.any(Object));
+  });
+});