What about immutable data structures ? about react-cursor HOT 5 CLOSED

We played around with that. We feel that mori has a significant cost - it adds a cross cutting concern of "Which type is the array that I'm looking at?", and the benefit does not outweigh that cost. React.addons.update is plenty fast for the type of application most people are building; the people who actually need full mori performance are the type of team that is already enjoying ClojureScript in the first place.

from react-cursor.

Not sure I understand the cross-cutting concerns problem.

And what about the Facebook lib ? My understanding is that it's meant to play well with React.

Thanks for the answer by the way.

from react-cursor.

It means at any given time, our objects can either be an immutable array or a JS array, and we don't have a type system to help us know the difference. Try it out in a nontrivial js app and you will see what I mean.

from react-cursor.

Oh, I get it. Basically you're saying it makes mori usable only where you have total control of your data. Or you pollute your code with lot of tests/conversion.

Strange, I just looked again at the mori docs and it says most function works on JS types as well.

from react-cursor.

This is a problem that's been bugging me recently, so I'm taking a stab at a library that will allow you to use arbitrary data structures as cursors: https://github.com/j201/react-precursor. In addition to mori, I'm interested in using data structures from Facebook's immutable library.

Anyway, I'm not sure it'll be possible to apply some of the aggressive optimizations that are in react-cursor, but it should provide a basic usable implementation of mori cursors.

from react-cursor.