feat: DEP #32 "Remember me" for Password-based Connectors

[daemonfire300]

Overview

kind/enhancement

Build concept/solution for #32

What this PR does / why we need it

For details refer to docs/enhancements/remember-me-2025-10-19-#32.md

Closes #32

Special notes for your reviewer

[daemonfire300]

@nabokihms @sagikazarmark I think you have to add a label for me because I cannot do that. Would be appreciated.

[daemonfire300]

TODO: Remove this file, this was just my scratch pad to gather my thoughts.

[nabokihms]

I looked through this PR briefly. As was mentioned, this feature requires the sessions concept and the way of storing them. There are other possible features that can benefit from it, like SSO or 2FA, but moving towards a simple in-memory store will make this thing unextendible in the future.Let's discuss the session implementation first and then move forward with the remember me feature.

[daemonfire300]

I looked through this PR briefly. As was mentioned, this feature requires the sessions concept and the way of storing them. There are other possible features that can benefit from it, like SSO or 2FA, but moving towards a simple in-memory store will make this thing unextendible in the future.Let's discuss the session implementation first and then move forward with the remember me feature.

In general I agree with your notion, so it's a yes but I would have the following suggestion(s):

1. Intentionally did it in memory only to avoid complexity, because this did not cost that much and does not introduce such a big maintenance burden. So I would be in favor to have a two phase solution, if possible, where we have a in memory remember version out first (soon-ish) and have 80% of the benefit of 20% of the work and then use that cushion to work on a generic/extended session concept.
2. Would you like to have an initial Teamspeak/Discord/Signal/Matrix/Google Meet call between us or more people to get on common ground and then start writing proposals?