@ai-ecoverse/slicc-shared-web 1.7.1 → 1.8.0

package/README.md

@@ -66,10 +66,11 @@ test('boots', async (t) => {
 | `stallAfter` | How long one evaluation may run before the harness probes the test's targets, in ms | `30000` |
 
 `chrome.page(t)` opens a tab and returns the page API: `goto(path)`, `reload()`, `evaluate(fn, ...args)`, `until(fn, ...args)`, `press(key, ...modifiers)` (a key name like `Enter` or `F5`, or a single character typed with its US-layout key code), `type(text)`, `insert(text)`, `enter()`, `init(fn)` (runs in every new document), `expose(name, handler)` (a binding the page calls with a JSON string), `screenshot(file)`, `tab()` (another tab in the same context), `close()`, `errors`, `responses` and `dir` (the test's artifact directory). `goto` and `reload` wait until the new document is complete and no longer `about:blank`, and retry the navigation once if that never happens. `until` polls until the function returns a truthy value (not only strict `true`) and resolves with that value; the timeout error names the last result. `within(ms, fn, ...args)` is `until` with its own deadline, for a stage that should take a known time. Both stop at the deadline even if one evaluation never answers. Every CDP call has a deadline: `connect(url).send(method, params, sessionId, ms)` rejects with `<method>: no answer in <ms> ms` (60 s unless given), so a stuck page fails its test instead of hanging the run. When one evaluation runs past `stallAfter`, the harness probes once, without interrupting the wait:
+- it adds a `pending` line to `console.log` for every navigation and request of the page that has no answer yet, with its age. Chrome holds CDP messages to a page while a navigation is waiting for its response, so an evaluation that hangs while the page is idle usually means one of these;
 - it saves `stall-<n>.png` (a screenshot of the page, if the renderer still answers) in the test's artifact directory;
 - it saves a CPU profile sample of every page and worker in the test, including workers started by other workers, as `stall-<n>-<target>-<session>.cpuprofile`;
 - it adds a line per target to `console.log`, saying whether that target is running or paused;
-- on the first stall, it records a 10 s browser trace as `stall-1-trace.json` (open it in DevTools' Performance panel or Perfetto). The browser collects it with V8 CPU samples per thread, so it shows what a page's main thread is doing even when that thread no longer answers CDP. It also logs each process's CPU time over the window;
+- on the first stall, it records a 10 s browser trace as `stall-1-trace.json` (open it in DevTools' Performance panel or Perfetto). The browser collects it with V8 CPU samples per thread, task, IPC, Blink and loading events, so it shows what a page's main thread is doing even when that thread no longer answers CDP. It logs each process's CPU time over the window, and for each renderer its main thread's task count, the call chain of its longest or still open task (function, file, line and URL of a request), and where its CPU samples fall. The targets are sampled after the trace starts, so their new profiles carry their call frames into it. A call that started before the trace shows only as a node id, since its frames were named before; a thread that is blocked keeps `Tracing.start` waiting, which the harness logs and traces through;
 - it adds a line for every debugger pause still held in the browser, in any target, with its id, URL, start time, reason and top call frames.
 
 A pause in one target blocks every page in the same renderer process, so those pause lines name the culprit even when it isn't the page under test. The harness resumes its own coverage pauses (`beforeunload`, worker exits) within 10 s, even when collecting coverage fails. Downloads of intercepted URLs time out after 60 s and are tried up to three times; a `4xx` is not retried. `chrome.requests` lists the paths the server saw in this test, `chrome.overrides` is a `Map` of paths to bodies served instead of files, and `chrome.cdn` takes `status` (answer every intercepted request with it), `corrupt` (flip the last byte) and lists the intercepted `requests`. All of them reset with each test. `serve(options)` starts the same server on its own, for an `npm start`.

package/harness/chrome.mjs

@@ -140,9 +140,9 @@ export async function launch({
       const tab = async () => {
         const opened = page(cdp, await record.open(browserContextId), server, timeout, {
           stallAfter,
-          stall: async (tag, png) => {
+          stall: async (tag, png, pending) => {
             if (png) await writeFile(new URL(`${tag}.png`, dir), png).catch(ignore);
-            await bounded(record.stall(dir, tag), tag, 45000);
+            await bounded(record.stall(dir, tag, pending), tag, 45000);
           },
         });
         pages.push(opened);

package/harness/flight.mjs

@@ -0,0 +1,34 @@
+const done = new Set(['Network.loadingFinished', 'Network.loadingFailed']);
+const landed = new Set([
+  'Page.frameNavigated',
+  'Page.frameStoppedLoading',
+  'Page.navigatedWithinDocument',
+]);
+
+export function flights(now = Date.now) {
+  const requests = new Map();
+  const navigations = new Map();
+
+  function watch({ method, params }) {
+    if (method === 'Network.requestWillBeSent') {
+      const { request, type = 'Other', requestId } = params;
+      requests.set(requestId, { what: `${request.method} ${request.url} (${type})`, since: now() });
+    }
+    if (done.has(method)) requests.delete(params.requestId);
+    if (method === 'Page.frameStartedNavigating') {
+      const what = `navigation to ${params.url} (${params.navigationType})`;
+      navigations.set(params.frameId, { what, since: now() });
+    }
+    if (landed.has(method)) navigations.delete(params.frame?.id ?? params.frameId);
+  }
+
+  function pending() {
+    const at = now();
+    return [...navigations.values(), ...requests.values()]
+      .sort((a, b) => a.since - b.since)
+      .slice(0, 20)
+      .map(({ what, since }) => `pending ${what} for ${((at - since) / 1000).toFixed(1)}s`);
+  }
+
+  return { watch, pending };
+}

package/harness/page.mjs

@@ -1,4 +1,5 @@
 import { writeFile } from 'node:fs/promises';
+import { flights } from './flight.mjs';
 import { sleep } from './util.mjs';
 
 const masks = { alt: 1, ctrl: 2, meta: 4, shift: 8 };
@@ -99,8 +100,11 @@ export function page(
   const errors = [];
   const responses = [];
   const bindings = new Map();
-  const dispose = cdp.on(({ method, params, sessionId: from }) => {
+  const flying = flights();
+  const dispose = cdp.on((message) => {
+    const { method, params, sessionId: from } = message;
     if (from !== sessionId) return;
+    flying.watch(message);
     if (method === 'Runtime.exceptionThrown') {
       const { exception, text } = params.exceptionDetails;
       errors.push(exception?.description ?? text);
@@ -112,8 +116,9 @@ export function page(
   async function capture() {
     stalls += 1;
     const tag = `stall-${stalls}`;
+    const pending = flying.pending();
     const shot = await send('Page.captureScreenshot', { format: 'png' }, 10000).catch(() => null);
-    await stall(tag, shot?.data ? Buffer.from(shot.data, 'base64') : null);
+    await stall(tag, shot?.data ? Buffer.from(shot.data, 'base64') : null, pending);
   }
 
   function probe() {

package/harness/recorder.mjs

@@ -2,6 +2,7 @@ import { mkdir, readFile, writeFile } from 'node:fs/promises';
 import { basename, relative } from 'node:path';
 import { cwd } from 'node:process';
 import { pathToFileURL } from 'node:url';
+import { trace } from './trace.mjs';
 import { bounded, ignore, raw, samples, sleep, TIMED_OUT } from './util.mjs';
 
 const profiled = new Set(['page', 'worker', 'service_worker', 'shared_worker']);
@@ -104,56 +105,6 @@ function pauseOf(target = {}, params) {
   };
 }
 
-const TRACED = [
-  'toplevel',
-  'devtools.timeline',
-  'disabled-by-default-devtools.timeline',
-  'v8.execute',
-  'disabled-by-default-v8.cpu_profiler',
-];
-
-async function cpuTimes(cdp) {
-  const { processInfo = [] } = await cdp
-    .send('SystemInfo.getProcessInfo', {}, undefined, 5000)
-    .catch(() => ({}));
-  return new Map(processInfo.map((info) => [`${info.type} ${info.id}`, info.cpuTime]));
-}
-
-export async function trace(cdp, dir, tag, ms = 10000) {
-  const events = [];
-  const done = Promise.withResolvers();
-  const off = cdp.on(({ method, params }) => {
-    if (method === 'Tracing.dataCollected') events.push(...params.value);
-    if (method === 'Tracing.tracingComplete') done.resolve();
-  });
-  try {
-    const before = await cpuTimes(cdp);
-    await cdp.send(
-      'Tracing.start',
-      { traceConfig: { includedCategories: TRACED }, transferMode: 'ReportEvents' },
-      undefined,
-      10000
-    );
-    await sleep(ms);
-    await cdp.send('Tracing.end', {}, undefined, 10000);
-    await bounded(done.promise, 'trace', 20000);
-    const after = await cpuTimes(cdp);
-    await writeFile(new URL(`${tag}-trace.json`, dir), JSON.stringify({ traceEvents: events }));
-    const busy = [...after]
-      .map(([name, time]) => [name, time - (before.get(name) ?? 0)])
-      .filter(([, used]) => used > 0.05)
-      .map(([name, used]) => `${name} ${used.toFixed(1)}s`);
-    return [
-      `trace ${tag}-trace.json (${events.length} events)`,
-      `cpu over ${ms / 1000}s: ${busy.join(', ') || 'none'}`,
-    ];
-  } catch (error) {
-    return [`trace failed: ${error.message}`];
-  } finally {
-    off();
-  }
-}
-
 function claim(run) {
   if (!run || run.traced) return false;
   run.traced = true;
@@ -272,16 +223,14 @@ export function recorder(cdp, server, { coverage = ['/'], exits = new Map() } =
     }
   }
 
-  async function stall(dir, tag) {
+  async function stall(dir, tag, pending = []) {
     const first = claim(run);
     const current = [...sessions].filter(
       ([sessionId]) => contextOf(sessions, parents, sessionId) === run?.context
     );
-    const [lines, traced] = await Promise.all([
-      Promise.all(current.map((entry) => sample(cdp, pauses, entry, dir, tag))),
-      first ? trace(cdp, dir, tag) : [],
-    ]);
-    const all = [...lines, ...traced, ...held(pauses)];
+    const sampled = () => Promise.all(current.map((entry) => sample(cdp, pauses, entry, dir, tag)));
+    const [lines, traced] = first ? await trace(cdp, dir, tag, sampled) : [await sampled(), []];
+    const all = [...pending, ...lines, ...traced, ...held(pauses)];
     for (const line of all) run?.console.push(`${tag}: ${line}`);
     return all;
   }

package/harness/trace.mjs

@@ -0,0 +1,192 @@
+import { writeFile } from 'node:fs/promises';
+import { bounded, sleep } from './util.mjs';
+
+export const TRACED = [
+  'toplevel',
+  'ipc',
+  'disabled-by-default-ipc.flow',
+  'mojom',
+  'blink',
+  'loading',
+  'navigation',
+  'devtools.timeline',
+  'disabled-by-default-devtools.timeline',
+  'v8.execute',
+  'disabled-by-default-v8.cpu_profiler',
+];
+
+function where({ functionName, url, lineNumber }) {
+  const path = URL.canParse(url) ? new URL(url).pathname : url;
+  const name = functionName || '(anonymous)';
+  return path ? `${name} ${path}:${lineNumber + 1}` : name;
+}
+
+function threads(events) {
+  const names = new Map();
+  const processes = new Map();
+  for (const { name, pid, tid, args } of events) {
+    if (name === 'thread_name') names.set(`${pid}:${tid}`, args.name);
+    if (name === 'process_name') processes.set(pid, args.name);
+  }
+  return { names, processes };
+}
+
+function detail(event) {
+  const data = event.args?.data ?? event.args?.beginData ?? {};
+  const tag = event.args?.chrome_mojo_event_info?.mojo_interface_tag;
+  if (typeof data.functionName === 'string') {
+    return where({ ...data, lineNumber: (data.lineNumber ?? 1) - 1 });
+  }
+  return data.url ?? event.args?.url ?? tag ?? '';
+}
+
+function labels(slices) {
+  const named = slices.map((event) => `${event.name} ${detail(event)}`.trim());
+  return named
+    .filter((name, i) => name !== named[i - 1])
+    .slice(0, 10)
+    .join(' > ');
+}
+
+function inside(slices, task) {
+  const end = task.ts + task.dur;
+  return slices
+    .filter(
+      ({ ph, ts, dur }) => ph === 'X' && ts >= task.ts && ts + dur <= end && dur >= task.dur / 2
+    )
+    .sort((a, b) => a.ts - b.ts || b.dur - a.dur);
+}
+
+function unmatched(slices) {
+  const stacks = new Map();
+  for (const event of [...slices].sort((a, b) => a.ts - b.ts)) {
+    const stack = stacks.get(event.tid) ?? [];
+    if (event.ph === 'B') stack.push(event);
+    if (event.ph === 'E') stack.pop();
+    stacks.set(event.tid, stack);
+  }
+  return [...stacks.values()].flat();
+}
+
+function main(pid, slices, last) {
+  const tasks = slices.filter(({ name, ph }) => name === 'RunTask' && ph === 'X');
+  const longest = tasks.reduce((most, task) => (task.dur > (most?.dur ?? 0) ? task : most), null);
+  const open = unmatched(slices).filter(({ ts }) => ts < last - 50000);
+  const head = `renderer ${pid} main thread: ${tasks.length} tasks`;
+  if (open.length) return `${head}, still in ${labels(open)}`;
+  if (!longest) return head;
+  const ms = (longest.dur / 1000).toFixed(1);
+  return longest.dur < 50000
+    ? `${head}, longest ${ms} ms`
+    : `${head}, longest ${ms} ms in ${labels(inside(slices, longest))}`;
+}
+
+function mains(events, { names, processes }) {
+  const found = new Map();
+  const last = events.reduce((most, { ts = 0 }) => Math.max(most, ts), 0);
+  for (const event of events) {
+    const ours = names.get(`${event.pid}:${event.tid}`) === 'CrRendererMain';
+    if (!ours || processes.get(event.pid) !== 'Renderer' || event.ph === 'M') continue;
+    found.set(event.pid, [...(found.get(event.pid) ?? []), event]);
+  }
+  return [...found].map(([pid, slices]) => main(pid, slices, last));
+}
+
+function stack(nodes, id) {
+  const frames = [];
+  for (let node = nodes.get(id); node && frames.length < 3; node = nodes.get(node.parent)) {
+    frames.push(where(node.callFrame));
+  }
+  return frames.join(' < ');
+}
+
+function profiles(events) {
+  const streams = new Map();
+  for (const event of events) {
+    if (event.name !== 'ProfileChunk') continue;
+    const key = `${event.pid} ${event.id}`;
+    const stream = streams.get(key) ?? { pid: event.pid, nodes: new Map(), counts: new Map() };
+    const { nodes = [], samples = [] } = event.args.data.cpuProfile ?? {};
+    for (const node of nodes) stream.nodes.set(node.id, node);
+    for (const id of samples) stream.counts.set(id, (stream.counts.get(id) ?? 0) + 1);
+    streams.set(key, stream);
+  }
+  return [...streams.values()].filter(({ counts }) => counts.size > 0).map(busiest);
+}
+
+function busiest({ pid, nodes, counts }) {
+  const total = [...counts.values()].reduce((sum, n) => sum + n, 0);
+  const ranked = [...counts].sort((a, b) => b[1] - a[1]);
+  const share = (n) => `${Math.round((100 * n) / total)}%`;
+  const [id, most] = ranked[0];
+  const named = nodes.has(id) ? stack(nodes, id) : `node ${id}, named before the trace began`;
+  const line = `samples in ${pid}: ${share(most)} of ${total} samples in ${named}`;
+  const script = ranked.find(
+    ([n]) => nodes.has(n) && !nodes.get(n).callFrame.functionName.startsWith('(')
+  );
+  if (!script || script[0] === id || script[1] * 100 < total) return line;
+  return `${line}, ${share(script[1])} in ${stack(nodes, script[0])}`;
+}
+
+export function summary(events) {
+  const known = threads(events);
+  return [...mains(events, known), ...profiles(events)];
+}
+
+async function cpuTimes(cdp) {
+  const { processInfo = [] } = await cdp
+    .send('SystemInfo.getProcessInfo', {}, undefined, 5000)
+    .catch(() => ({}));
+  return new Map(processInfo.map((info) => [`${info.type} ${info.id}`, info.cpuTime]));
+}
+
+async function begin(cdp) {
+  const traceConfig = { includedCategories: TRACED };
+  try {
+    await cdp.send(
+      'Tracing.start',
+      { traceConfig, transferMode: 'ReportEvents' },
+      undefined,
+      10000
+    );
+    return [];
+  } catch (error) {
+    if (!/no answer/.test(error.message)) throw error;
+    return [`${error.message}, a renderer may be blocked; tracing anyway`];
+  }
+}
+
+export async function trace(cdp, dir, tag, during = () => [], ms = 10000) {
+  const events = [];
+  let ran = null;
+  const done = Promise.withResolvers();
+  const off = cdp.on(({ method, params }) => {
+    if (method === 'Tracing.dataCollected') events.push(...params.value);
+    if (method === 'Tracing.tracingComplete') done.resolve();
+  });
+  try {
+    const before = await cpuTimes(cdp);
+    const late = await begin(cdp);
+    ran = during();
+    await sleep(ms);
+    await cdp.send('Tracing.end', {}, undefined, 10000);
+    await bounded(done.promise, 'trace', 20000);
+    const after = await cpuTimes(cdp);
+    await writeFile(new URL(`${tag}-trace.json`, dir), JSON.stringify({ traceEvents: events }));
+    const busy = [...after]
+      .map(([name, time]) => [name, time - (before.get(name) ?? 0)])
+      .filter(([, used]) => used > 0.05)
+      .map(([name, used]) => `${name} ${used.toFixed(1)}s`);
+    const lines = [
+      ...late,
+      `trace ${tag}-trace.json (${events.length} events)`,
+      `cpu over ${ms / 1000}s: ${busy.join(', ') || 'none'}`,
+      ...summary(events),
+    ];
+    return [await ran, lines];
+  } catch (error) {
+    return [await (ran ?? during()), [`trace failed: ${error.message}`]];
+  } finally {
+    off();
+  }
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ai-ecoverse/slicc-shared-web",
-  "version": "1.7.1",
+  "version": "1.8.0",
   "description": "Shared Renovate, Biome, TypeScript, lefthook and GitHub Actions configuration, plus the CDP integration-test harness, for SLICC's web repos",
   "type": "module",
   "license": "Apache-2.0",