[Bug] Cannot run alive-server using podman

[edumserrano]

Trying to run alive-server from a container using podman on Windows throws an error.

Steps to reproduce

1. Install podman on Windows. For instance: winget install --id RedHat.Podman.
2. Run a container with podman in interactive mode with a shell: podman run --entrypoint /bin/sh -it node:22.13.1.
3. At this point you should be in a shell in the container. Run alive-server: mkdir /test && cd /test && npx --yes alive-server .
4. You get an error.

Expected behavior

After step 3 what you should see is

Serving "/test" at http://127.0.0.1:8080
Ready for changes

Actual behavior

This is the error you get:

Serving "/test" at http://127.0.0.1:8080
node:events:502
      throw er; // Unhandled 'error' event
      ^

Error: spawn /mnt/c/Windows/System32/WindowsPowerShell/v1.0/powershell.exe ENOENT
    at ChildProcess._handle.onexit (node:internal/child_process:285:19)
    at onErrorNT (node:internal/child_process:483:16)
    at process.processTicksAndRejections (node:internal/process/task_queues:90:21)
Emitted 'error' event on ChildProcess instance at:
    at ChildProcess._handle.onexit (node:internal/child_process:291:12)
    at onErrorNT (node:internal/child_process:483:16)
    at process.processTicksAndRejections (node:internal/process/task_queues:90:21) {
  errno: -2,
  code: 'ENOENT',
  syscall: 'spawn /mnt/c/Windows/System32/WindowsPowerShell/v1.0/powershell.exe',
  path: '/mnt/c/Windows/System32/WindowsPowerShell/v1.0/powershell.exe',
  spawnargs: [
    '-NoProfile',
    '-NonInteractive',
    '–ExecutionPolicy',
    'Bypass',
    '-EncodedCommand',
    'UwB0AGEAcgB0ACAAIgBoAHQAdABwADoALwAvADEAMgA3AC4AMAAuADAALgAxADoAOAAwADgAMAAiAA=='
  ]
}

Node.js v22.13.1

Additional context

This only happens with podman, I've tried both rancher and docker desktop and those work fine.

I believe this is the same problem as I reported in microsoft/playwright#34737. In that issue I described some research I did a suggestion that the problem might be related with the is-wsl dependency of the open package, which alive-server also uses.

As with the reported issue, alive-server is also using version 8 of the open package where the latest version available is 10. In the Playwright issue above it seems that indeed updating the package solved the problem.

[edumserrano]

I've also found that besides the workaround I mentioned in microsoft/playwright#34737, which also works for this bug, if I run alive-server with the --no-browser option than I don't get the error.

This is expected because with --no-browser the open npm package is not invoked. Using --no-browser is likely enough for this scenario and should be used when running in a container because I guess you never want the browser to open in that case, you want to access the alive-server from a browser in the host instead.

However, I still think there would be value in fixing this and updating the open package if possible.

---

EDIT: actually the --no-browser doesn't always help with this problem. It seems that there's a difference between:

alive-server --port=41300 --open=test-reporters/jest/html-report --no-browser

and

alive-server --port=41300 --no-browser --open=test-reporters/jest/html-report

If I use --no-browser before --open then the alive-server doesn't try to open the browser, but if --no-browser comes after --open then it still tries to open the browser and I get the error.

The reason this is a problem for me is that I have an npm script like:

"my-open": "alive-server --port=41300 --open=test-reporters/jest/html-report"

Which I use when NOT running in a container, and then when running in the container I do:

npm run my-open -- --no-browser

Which gets me the reported error because alive-server still tries to open the page.