Open MFlisar opened 3 years ago
Definitely! I already planned to integrate JetPack DataStore, however, I wanted to wait while it's still in beta. Already having an abstraction for storage providers makes a lot of sense though, so I appreciate the work you've done and would be happy if you opened a PR. I've looked at the code already and it looks mostly good, I'll probably still have some nitpicks to be adressed before merging :smile: Let's discuss those in the PR though.
Done, rest can be discussed in the PR
I prefer to keep this open until the PR is merged ^^
Any plans on implementing this soon? Or one working on my PR (together if you want)?
Definitely still planned, but I'm absolutely packed with work (bachelor thesis) right now. Your PR looks good, but there are some things I'd like to change about it. I'll revisit it not later than May, after I finished my bachelor thesis.
Ok, no problem. Good luck with your exams.
Just as a heads-up: I'll try to get this in soon, but probably not in the next version. Want to get some fixes out first.
No need to hurry if this is just "for me".
I've written my own library in the meantime because I wanted quite some more changes (a core that works with async operation via kotlin flows based on suspend functions, preference declarations via delegates, abstract storage with a data store based default implementation, the ability to observe change events of the storage, abstract dependencies between preferences, ...) - my screen module does even use a few small parts of your library as I've written into my credits region, so thanks for this here as well.
I've made a fork here https://github.com/MFlisar/ModernAndroidPreferences and implemented an arbitrary storage layer here
What I did:
Storage
SharedPreferencesStorage
which works just as your current solutionSharedPreferences
inside your library with theStorage
instance (builder, dsl, ...)Why?
In my case I need multi process support and therefore I have written a room based preference library and I want to plug in this library into yours. Additionally, preferences are not the most modern solution for preferences, datastore is the new way - with my solution you can easily provide a storage implementation that uses the
DataStore
instead of theSharedPreferences
in the future and still let your users choose. And of course, any user can replace the storage implementation with their ownQuestion
Do you think such a change (it's small and does not add any complexity to your library) does make sense and should be merged into your library? I did not adjust any tests yet but if you like the idea, I'll make those adjustments and create a pull request.