Interactive static assets stop loading after the browser stops the idle service worker, which keeps them only in memory
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 76/100
- Issue type
- Bug
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- typescript
- Domain
- backend
Research direction
Start with packages/runtime-internal/src/interactive/worker.js and bootstrap.js to trace how static assets are registered and fetched; then read the named worker test and its Cache Storage test globals. Implement persistent asset storage across worker restarts, and verify the unit test plus the CDP end-to-end test in testing/src/specs/instrument-completion.spec.ts pass, including after a stopped worker is restarted.
Written by the indexing model from the issue text.
Description
An interactive instrument's static assets stop loading once the browser stops its idle service worker. worker.js keeps the assets only in a module-level Map. Browsers terminate a service worker that has handled no event for about 30 seconds (Chrome and Firefox both use a 30 s idle timeout), even while a page it controls is open, and the next fetch starts a fresh copy of the script with an empty map. bootstrap.js posts the assets only once, before render. So any asset first requested after a pause of more than 30 s with no network activity misses the map and goes to the network, where it does not exist. Typical cases are a stimulus image shown after an instruction screen, or a sound played on a later trial. The participant sees a broken image or hears nothing, partway through a task whose data is then recorded as if the stimulus had been shown.
Where
packages/runtime-internal/src/interactive/worker.js:5-6 and 53-77:
/** @type {Map<string, string>} */
const staticAssets = new Map();
// ...
addEventListener('message', (event) => {
// ...
Object.entries(data.staticAssets).forEach(([key, value]) => {
staticAssets.set(key, value);
});
event.ports[0]?.postMessage({ type: 'STATIC_ASSETS_READY' });
});
worker.addEventListener('fetch', (_event) => {
// ...
if (staticAssets.has(pathname)) {
packages/runtime-internal/src/interactive/bootstrap.js:49-74 sends STATIC_ASSETS once, at startup, and never again.
Reproduce
- Use the playground example "Interactive Instrument with Static Assets", changed so the
<img src="smiley.png">is inserted on a button click instead of being in the initial HTML. - Open it in
apps/webin Chrome. - Wait more than 30 s without interacting, or click "stop" for the worker under DevTools > Application > Service workers, which does the same as an idle termination. Then click the button.
Actual: the restarted worker has an empty staticAssets map, /runtime/v1/@opendatacapture/runtime-internal/interactive/smiley.png goes to the server, and the image is broken.
Expected: the asset is served however long the task has been running.
Tests
packages/runtime-internal/src/interactive/__tests__/worker.test.ts: it('should still serve registered assets after the worker script is re-evaluated, so an idle termination does not lose them'). Load the assets, call vi.resetModules() and re-import worker.js to simulate a restart, then fetch. This needs a stubbed caches (Cache Storage) in the test globals.
e2e, in testing/src/specs/instrument-completion.spec.ts: open an interactive instrument with static assets, stop its service worker through CDP (ServiceWorker.stopWorker), then trigger an asset request and assert it loads.
Suggested fix
Persist the assets where a restarted worker can read them. On STATIC_ASSETS, write each one as a Response into a named Cache Storage cache (caches.open(...)), and acknowledge only after the writes resolve. In fetch, respond from caches.match(pathname) and fall back to the network on a miss. Clear the cache whenever a new STATIC_ASSETS message arrives, so one instrument's assets are never served to the next.
- Dominant language
- TypeScript
- Stars
- 119
- Forks
- 19
- Avg merge
- 1d 2h
- Merged PRs (30d)
- 56
Getting set up
- Ships a Dockerfile or Docker Compose file
- No pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from DouglasNeuroInformatics/OpenDataCapture
-
Area: Playground Bug Difficulty: Low Good First Issue Priority: Low
Difficulty 2/5 1-3 hours Newbie friendliness 85/100
DouglasNeuroInformatics/OpenDataCapture#1805 ·
Maintainers usually reply within 1 day
-
Area: Instruments Bug Difficulty: Low Priority: Low
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
DouglasNeuroInformatics/OpenDataCapture#1801 ·
Maintainers usually reply within 1 day
-
Area: Instruments Bug Difficulty: Low Good First Issue Priority: Low
Difficulty 2/5 1-3 hours Newbie friendliness 85/100
DouglasNeuroInformatics/OpenDataCapture#1800 ·
Maintainers usually reply within 1 day
-
Area: Instruments Bug Difficulty: Low Good First Issue Priority: Low
Difficulty 2/5 1-3 hours Newbie friendliness 84/100
DouglasNeuroInformatics/OpenDataCapture#1799 ·
Maintainers usually reply within 1 day
-
Area: Instruments Bug Difficulty: Low Performance Priority: Medium
Difficulty 2/5 1-3 hours Newbie friendliness 83/100
DouglasNeuroInformatics/OpenDataCapture#1795 ·
Maintainers usually reply within 1 day
All issues in DouglasNeuroInformatics/OpenDataCapture
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
better-auth/better-auth#11723 · 1 reaction ·
Maintainers usually reply within 1 day
-
area: docs difficulty: easy documentation good first issue
Difficulty 2/5 1-3 hours Newbie friendliness 82/100
Maintainers usually reply within 1 day
-
Worker-Erstellung bricht langsame Git-Worktree-Checkouts nach 20 Sekunden abPossibly taken @KajoVanBuyten claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 69/100
AgentSystemLabs/agent-office#367 ·
Maintainers usually reply within 1 day
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
pierrecomputer/pierre#1187 ·
Maintainers usually reply within 1 day
-
area: frontend bug
Difficulty 1/5 1-3 hours Newbie friendliness 85/100
snapotter-hq/SnapOtter#2425 ·
Maintainers usually reply within 1 day