Scroll position is part of structs.Page, which means that it isn't remembered if pages aren't cached. But this is bad UX when navigating back and forth on long un-cached pages. An example of this is Astrobotany, which uses client certs (and is therefore not cached according to current logic), and has a long page that lists all the plants.

Possible solutions

• Enable caching for client cert pages, and find some other solution to the problem of caching dynamic pages
• Store scroll data elsewhere, like in a new array of structs for history ✔️
• ...