Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Web image serves index.html without Cache-Control and answers missing assets with index.html, so the app loads blank after an upgrade

Open Beginner friendly
#1,771 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
2/5
Estimated time
1-3 hours
Newbie friendliness
79/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
docker, typescript
Domain
devops, frontend

Research direction

Start with apps/web/Caddyfile and the reproduction steps in the issue; they identify the cache-header and missing-asset behaviors to verify. Build and run the web image with docker compose up -d web caddy, then use the stated curl -I checks for / and a nonexistent /assets/ path. Done when the app entry point is revalidated and missing assets return 404; verify hashed assets retain long-lived immutable caching if implementing the suggested configuration.

Written by the indexing model from the issue text.

Description

Area: Infrastructure Bug Difficulty: Low Priority: Medium

After an upgrade, the web app can load a blank page for returning users. The web image's Caddy serves index.html with Last-Modified and ETag but no Cache-Control. Browsers therefore cache it heuristically: RFC 9111 §4.2.2 allows a response to be treated as fresh for a fraction of its age, typically 10% of the time since Last-Modified. index.html is rewritten by import-meta-env when the container starts, so a web container that has been running for 30 days hands out an index.html that browsers may reuse for about 3 days without asking the server. Once a new image is deployed, a user who opens the app from a bookmark gets the old cached index.html. Its <script type="module" src="/assets/index-<old hash>.js"> no longer exists in the new image. try_files {path} /index.html then answers that request with 200 text/html, the browser refuses to run HTML as a module, and the page stays blank until the heuristic expires or the user force-reloads. The same fallback breaks any tab left open across an upgrade. Route chunks are code-split (autoCodeSplitting: true), so the next navigation imports an old chunk and gets HTML back instead of a 404. #936 tracks the same caching problem for the playground image. This issue is the clinician-facing web image.

Where

apps/web/Caddyfile:1-6:

:80 {
	root * /srv
	encode gzip zstd
	try_files {path} /index.html
	file_server
}

apps/web/Dockerfile:30 rewrites index.html at every container start:

CMD [ "sh", "-c", "import-meta-env -x .env.public -p index.html && caddy run --config /etc/caddy/Caddyfile --adapter caddyfile" ]

Reproduce

  1. Run caddy:2.7-alpine with apps/web/Caddyfile over a /srv containing index.html and assets/index-NEW.js.
  2. Run curl -I / and curl -I /assets/index-OLD.js.

Actual (verified locally): / returns 200 with Last-Modified and ETag and no Cache-Control. /assets/index-OLD.js, which does not exist, returns 200 Content-Type: text/html with the body of index.html.
Expected: index.html is sent with Cache-Control: no-cache, so every navigation revalidates it, and a missing /assets/* file returns 404.

Tests

No unit or Playwright test runs the production Caddy image. The e2e suite drives the Vite dev server. Verify the fix with the curl checks above against the built image, using docker compose up -d web caddy and curl -I http://localhost:$APP_PORT/ and /assets/does-not-exist.js.

Suggested fix

In apps/web/Caddyfile, give /assets/* its own handle with file_server and no try_files fallback, and set header Cache-Control "public, max-age=31536000, immutable" there, since those files are content-hashed. In the SPA fallback handle, set header Cache-Control "no-cache" so index.html is always revalidated. With the 404 in place, apps/web can also listen for Vite's vite:preloadError and reload, which covers tabs left open across an upgrade.

Dominant language
TypeScript
Stars
119
Forks
19
Avg merge
1d 2h
Merged PRs (30d)
56

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from DouglasNeuroInformatics/OpenDataCapture

All issues in DouglasNeuroInformatics/OpenDataCapture

Similar issues

More TypeScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.