search

5HT / fs

📁 FS: Windows, Linux, Mac Driver
https://fs.n2o.dev
Other
234 stars 69 forks source link

Log error, but don't crash when backend is not working #69

Closed dominicletz closed 2 years ago

dominicletz commented 2 years ago

I've got an application that is deployed on many different machines and architectures. I would like to ensure that when the file monitoring tool is not working correctly (e.g. it crashes) then that is not bringing down the whole app. This change just tries to restart the file monitoring backend with an exponential back-off in those cases.

But I'm not sure if this is breaking other use cases? Also, I was thinking to add a fallback polling fs backend in this case (e.g. after 5 attempts to restart the backend), but that seems like it should require configuration?

Anyway, let me know if you have any thoughts on this.

dominicletz commented 2 years ago

My local version is too old, give me a bit here to rebase and resolve the conflict...

5HT commented 2 years ago

It's fine. I've done 👍

5HT commented 2 years ago

Just verify manually, if master is fine, then close this PR.

dominicletz commented 2 years ago

Looks good!

5HT commented 2 years ago

Just published 8.6.1: https://hex.pm/packages/fs/8.6.1