## applyCanon · function

Temporarily revert all outstanding predictions, optionally run the provided function
(which will generally make authoritative changes to the data based on a server response),
and then attempt to reapply the predictions on top of the new canonical state, dropping
any predictions that can no longer be applied cleanly (the data has been modified) or
that were specified in `dropPredictions`.

All of this is done such that redraws are only triggered if the overall effect is an
actual change to an `Observable`.

**Signature:** `(canonFunc?: () => void, dropPredictions?: Patch[]) => void`

**Parameters:**

- `canonFunc?: () => void` - The function to run without any predictions applied. This will typically
make authoritative changes to the data, based on a server response.
- `dropPredictions: Array<Patch>` (optional) - An optional list of predictions (as returned by `applyPrediction`)
to undo. Typically, when a server response for a certain request is being handled,
you'd want to drop the prediction that was done for that request.