Hacker News new | past | comments | ask | show | jobs | submit login

> I just fundamentally don't see how any update manager (be it chocolatey, nuget, windows, or linux etc) can efficiently update a client without knowing what's already on the client.

Pretty simple. Instead of sending information about what's on the machine to the update server, the client would download the list of available updates from the server, use that list to determine what updates it needs, and request those specific updates from the server.

That's e.g. how apt - the Debian family package manager - works. You `apt update` to update the local cache that lists available packages and their versions; then, you `apt upgrade` to download and install any or all packages that need updating.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: