Serverless App demo.

[ghost]

Is it possible to get a SvelteKit (or NextJS, Nuxt ) PostgreSQL / Supabase demo?

[ghost]

Also a GraphQL example would be cool.

[aboodman]

GraphQL and Replicache don't really go together. The Replicache sync protocol takes the spot that GraphQL usually does.

[ghost]

Interesting. For example, I used FaunaDB for an application and I interact with it using GraphQL. Does this mean that I cannot use Replicache for offline real time syncing?

[aboodman]

You would need to put an application server in front of fauna (perhaps a lambda would do) to implement the Replicache push and pull endpoints. See: https://doc.replicache.dev/how-it-works#the-big-picture. This is fairly easy to do. Additionally for each Replicache mutator (https://doc.replicache.dev/how-it-works#push) you need to implement the corresponding mutation code for Fauna. Most apps don't have a very large number of mutators so this isn't a big deal.

[aboodman]

Most people find Replicache easier to use than GraphQL 😂 so they don't really mourn the loss. Once you have sync setup you can kind of forget the server exists much of the time and program entirely on the client.

[ghost]

You would need to put an application server in front of fauna (perhaps a lambda would do) to implement the Replicache push and pull endpoints. See: https://doc.replicache.dev/how-it-works#the-big-picture. This is fairly easy to do. Additionally for each Replicache mutator (https://doc.replicache.dev/how-it-works#push) you need to implement the corresponding mutation code for Fauna. Most apps don't have a very large number of mutators so this isn't a big deal.

Okay but as far as I understood these docs, it would be possible in my situation. Is this correct?

I could implement pull and push endpoints and then just run graphql queries instead of sql on the server side.

[aboodman]

Yep, you could do that.