I have been doing both, with pros and cons.
All in one
pros:
- all in one, easier deployment
- no need for cors
cons:
- bind to backend, what if You want to reuse?
Separate
pros:
- easier to switch and reuse
- –no-html --no-webpack
cons:
- not really a cons, but maybe You need to configure cors.
- need to add phoenix.js as dependency
- need two servers, more complex to deploy
I prefer to separate, but use only one server… and build into priv/static of the Phoenix server.
BTW You can do server side react rendering.
No, You can also use websocket.
1 Like
