SteveLillis
commented
2 years ago @nboughton Ok, I've given this some thought and I think that the most sensible approach for me to enable users to persist their information whilst retaining the support for offline:
- The same sort of flow for persisting data as the google approach (a lazy sync)
- Instead of using Google Drive, using Azure Functions over Cosmos DB (1million azure function executions are free per month, 1million ops at 100GB container for Cosmos ~$8 per month)
- Use Auth0 for sign-in (its free up to 7k active users then ~$23 per month)
So all in all it'd be about $30 per month to host. How does that sound to you? Is it outside of what you were thinking for Stargazer? If so, let me know and I'll give some more thought to how I might make Google work for this. Or perhaps find another free alternative.
Either way, I think it would be good to merge this removal of the half-implemented google sync, so that users can have access to all of the improvements that have otherwise been added.
Edit Another idea I had that would do away with race conditions (or at least, put the emphasis on the user), is if I simplify the Google Drive approach so that instead of trying to keep devices in sync, we simply let the user manually upload/download campaigns themselves (i.e. duplicate the export functionality but in a "export to Google Drive" sense)
nboughton
@nboughton The Google drive based sync approach feels too brittle to me due to a number of difficult-to-circumvent race conditions, and I didn't like that it was blocking the release of all the recent quality of life improvements.
This PR removes the Google functionality so that the app can be released when appropriate, while I rethink the approach to perhaps use a Dexie syncable or similar.
Dexie Cloud looks promising but isn't publicly available yet.
Otherwise I may just cave and host a small server somewhere, it would certainly make things simpler.