# `AshBackendDrivenOptimisticUi.Change`
[🔗](https://github.com/Undeadlol1/ash_backend_driven_optimistic_ui/blob/v0.1.0/lib/ash_backend_driven_optimistic_ui/change.ex#L1)

Lets a change module describe its effect for the client, once, next to its
code. The client cannot run Elixir, so without this every `optimistic` block
of every action that uses the change repeats the effect as a `set`.

    defmodule MyApp.Changes.Bump do
      use Ash.Resource.Change
      use AshBackendDrivenOptimisticUi.Change

      @impl Ash.Resource.Change
      def change(changeset, opts, _context),
        do: Ash.Changeset.atomic_update(changeset, :count, expr(count + ^opts[:by]))

      @impl AshBackendDrivenOptimisticUi.Change
      def optimistic_sets(opts), do: [count: expr(count + ^opts[:by])]
    end

`optimistic_sets/1` gets the opts the action passes
(`change {MyApp.Changes.Bump, by: 2}`) and returns attributes with an
expression or a literal. Each one becomes a derived `set` in every optimistic
action that uses the change. A change guarded by `where` is skipped, and a
`set` written in the rule for the same attribute wins. The values follow the
rules of any `set`: no relationship, no actor, only functions the client can
run. `optimistic` in tests fails when the description and the change disagree.

# `optimistic_sets`

```elixir
@callback optimistic_sets(opts :: keyword()) :: keyword()
```

---

*Consult [api-reference.md](api-reference.md) for complete listing*
