Several changes to reduce flakiness in the wrangler remote-binding and dev e2e tests:

Consistent waitFor/waitForLong helpers

• Extract shared waitFor() and waitForLong() helpers that wrap vi.waitFor() with tuned defaults
  • waitFor(): 100ms interval, 5s timeout — for polling synchronous state (e.g. console output)
  • waitForLong(): 500ms interval, 10s timeout — for polling HTTP endpoints
• Add ESLint rule to enforce using these helpers instead of bare vi.waitFor() in e2e tests
• Migrate all e2e tests to use the shared helpers

Retry transient API failures in remote preview

• Wrap createPreviewSession() and createWorkerPreview() in retryOnAPIFailure so transient 5xx errors are retried automatically (up to 3 attempts with linear backoff)
• Add an optional abortSignal parameter to retryOnAPIFailure so backoff delays can be cancelled immediately when a new bundle arrives

Increase e2e timeouts for remote preview and add per-request API timeout

• startWorker: use waitForLong (10s) instead of waitFor (5s) for remote reload polling, matching the convention for HTTP endpoint polling
• start-worker-remote-bindings: increase beforeAll deploy timeout from 35s to 60s for slow Windows CI
• dev.test Workers+Assets: increase waitForReady/waitForReload from 15s to 30s since remote mode involves session creation + asset upload + bundle upload to edge-preview
• create-worker-preview: add 30s per-request timeout via AbortSignal.any so a hung Cloudflare API response doesn't block the reload indefinitely
• dev-remote-bindings multi-worker test: increase waitForReady from 15s to 30s since two serialised remote proxy sessions must complete before "Ready on" appears
• dev-remote-bindings error-log tests: use waitForLong instead of waitFor since error messages depend on Cloudflare API validation round-trips

Use shared preserve-e2e-* workers for remote-binding tests

• Add ensureWorkerDeployed() helper to WranglerE2ETestHelper that checks whether a worker is already live (by fetching its workers.dev URL) and only deploys if it returns 404
• Migrate dev-remote-bindings.test.ts, start-worker-remote-bindings.test.ts, and remote-bindings-api.test.ts to use fixed preserve-e2e-wrangler-remote-worker / preserve-e2e-wrangler-remote-worker-alt names instead of deploying fresh workers with random names on every run
• These workers persist across test runs and are excluded from the periodic e2e cleanup job by their preserve-e2e- prefix

Fix EADDRINUSE port races (TOCTOU elimination)

• stop() now awaits process exit — previously only waited for the kill signal to be sent, not for the process to actually terminate. Ports held by dying processes could cause EADDRINUSE in the next test.
• Remove get-port TOCTOU pattern from e2e tests — spawnLocalWorker and dev-registry pages dev tests now rely on --port 0 (OS-assigned at bind time) instead of get-port (check-then-use race). Extended getWranglerCommand to also auto-add --port 0 --inspector-port 0 for wrangler pages dev commands.
• Fix getPort() TOCTOU in production code — start-remote-proxy-session.ts used get-port to allocate a port for the nested remote proxy DevEnv. Replaced with port: 0 so the OS assigns the port atomically at bind time, eliminating EADDRINUSE during wrangler dev with remote bindings.

Make workers.dev domain configurable for e2e tests

• Add E2E_ACCOUNT_WORKERS_DEV_DOMAIN env var (defaults to devprod-testing7928.workers.dev) so anyone can run the e2e suite against their own Cloudflare account
• Replace all hardcoded devprod-testing7928.workers.dev references across wrangler, vite-plugin-cloudflare, create-cloudflare, and get-platform-proxy-remote-bindings e2e tests
• Add the env var to turbo.json so Turbo invalidates cache when it changes
• Document the env var in packages/wrangler/e2e/README.md with instructions for running against a personal account

Follow-up items

• pages-dev.test.ts still uses module-level getPort() shared across sequential tests — lower risk (single allocation at module load), but could be migrated to --port 0 in a follow-up
• startWorker.test.ts keeps getPort() for its port-switch test — the TOCTOU window is small (in-process programmatic API) and the test genuinely needs distinct known ports to verify port changes

• Tests
  • Tests included/updated
  • Automated tests not possible - manual testing has been completed as follows:
  • Additional testing not necessary because:
• Public documentation
  • Cloudflare docs PR(s):
  • Documentation not necessary because: test infrastructure and internal retry logic only

A picture of a cute animal (not mandatory, but encouraged)