Is it okay to use WatermelonDB with MobX? about watermelondb

[radex]

Sure, I think using MobX as a local state management system + WatermelonDB as the database should work just fine.

[sbtly]

Interesting!
But wouldn't it cause related component to be updated twice every time some value changes?

for example,

// mobX part
@observable myColor

// watermelon part
...color.value = myColor...

user changes @observable myColor ->
mobX will update related component ->
mobX's changed observable value will update watermelon's color.value ->
watermelon will update related component again

This is what I think might happen when using MobX and WatermelonDB together.
Is this wrong or is this perfectly okay?

[radex]

I don't get it. Why do you want to use both MobX and Watermelon to manage the same data? I get using MobX as local-state-only library, but why go through MobX to modify persistent data?

[sbtly]

oh, now I see.
This is my first time trying backend myself.
So I got all mixed up about how to deal with data.

I'll try using MobX + WatermelonDB then!
Thanks for the quick reply :)

[awgv]

This thread is the first search result, so I’ll post it here. Unless I’m missing something—withObservables() can’t be combined with observer() HOC—when you have multiple HOCs, you’re supposed to apply observer() first, but observer() memoizes components, and you should memoize after withObservables(). If you apply observer() last, an error is thrown; if you apply observer() first, everything is now memoized, and changes to WatermelonDB models don’t rerender components. As far as I can tell, there’s no way to work around this—at least not in a way that’s mentally tolerable.

I don’t know how any of this work, but perhaps, it’ll become possible when and if WatermelonDB switches to a hook. I’d also like someone to confirm this, and until someone does, it would be best to take this message with a grain of salt and set up a basic WatermelonDB+MobX project to see the issues for yourself.