[DO NOT MERGE]: Rehydrate time trial results.

This commit fails tests because it intentionally uses
`remotedev-serialize` without our escaping wrapper, to collect
performance data about the wrapper.

See the parent commit for info on this experiment. Here's the data:

On 9e78659 (current upstream/master; this branch is rebased on that
commit):

850
844
856
843
843
841
840
839
850
844

On the parent commit, `remotedev-serialize` WITH our escaping
wrapper:

1038
965
951
960
960
965
966
962
962
953

At this commit, `remotedev-serialize` WITHOUT our escaping wrapper:

913
902
883
888
874
971
892
896
878
890

Any increase in rehydrate time from the status quo to our use of
`remotedev-serialize` is expected and hard to avoid.

I would guess that we're seeing a statistically significant increase
in rehydrate time associated with our escaping wrapper, as opposed
to just using `remotedev-serialize` without the wrapper. If the
delay is long enough, maybe we'd like to look at a more efficient
implementation of the wrapper.

src/boot/store.js

@@ -2,6 +2,7 @@
 import { applyMiddleware, compose, createStore } from 'redux';
 import type { Store } from 'redux';
 import Immutable from 'immutable';
+import * as Serialize from 'remotedev-serialize';
 import { persistStore, autoRehydrate } from '../third/redux-persist';
 import type { Config } from '../third/redux-persist';
 
@@ -184,12 +185,7 @@ const customReviver = (key, value, defaultReviver) => {
   return defaultReviver(key, value);
 };
 
-const { stringify, parse } = SerializeEscaped.immutable(
-  Immutable,
-  null,
-  customReplacer,
-  customReviver,
-);
+const { stringify, parse } = Serialize.immutable(Immutable, null, customReplacer, customReviver);
 
 const reduxPersistConfig: Config = {
   whitelist: [...storeKeys, ...cacheKeys],