search

ArctosDB / arctos

Arctos is a museum collections management system
https://arctos.database.museum
60 stars 13 forks source link

Feature Request - merge cf_user_data #4117

Closed dustymc closed 2 years ago

dustymc commented 2 years ago

Is your feature request related to a problem? Please describe.

Arbitrary split is more difficult than it needs to be

Describe what you're trying to accomplish

Merge cf_user_data and cf_users

Describe the solution you'd like

do it

Describe alternatives you've considered

don't do it

Additional context

Need to define the "profile" that's required to download - suggest email address is a suitable replacement.

dustymc commented 2 years ago

Upping priority, do as part of https://github.com/ArctosDB/internal/issues/184

dustymc commented 2 years ago

Going active - I think we have a good solution in /internal/, this could quickly become problematic esp. with off-site work potentially exposing more data, and I don't want to release anything while SPNHC is meeting which should provide a sufficient timeline.

dustymc commented 2 years ago

This is now running at test, and could use testing - core back-end functionality has changed significantly.

You should NOT need to change your password, it will be magically updated when you first log in. All of my accounts worked, but I sort of expect this to break something - just logging in to https://arctos-test.tacc.utexas.edu would be a useful test.

I can't find anywhere I messed up the encryption change, but it was a few thousand updates and I'd almost be surprised if I didn't miss something somewhere - click that link I haven't!

Merging the user tables might have broken some customization or setting somewhere.

Part of the technology behind the new encryption is that it is purposefully computationally expensive. (A malicious actor - even a State - is unlikely to have the resources to brute-force anything.) I've tuned that such that logging in should take about a half-second (on production, slightly slower on test) which I think should be mostly unnoticeable but please let me know if you get radically different results.