This file contains information about the laboratory itself and design choices inside.
The laboratory uses a simple routing system with no outside dependencies. The simple router is stateless. Components do not need to know about the router since everything is handled.
When the redux state changes, we can then serialize it and save it in the url. This is handled by a middleware called routerMiddleware.
This is reactive and won't change the redux store.
- An action happened (LOAD_STATE actions are ignored).
The a component uses the routerListener to provide access to dispatching actions. It will perform a diff to see if an the store needs to be updated. If so, it will dispatch an event.
If save data is not specified, the redux store should not be cleared.
- Load the state from the query string and navigate
- Only navigate. Don't load state
- Page initially loads: Result 1
- Hash change with query present: Result 1
- Hash change with empty query: Result 2
User navigation should always occur through hash links. This is so that users can open links in new tabs as well as right clicking to copy the link.
Although these events are user initiated, these are considered "inbound" changes since it is changing the hash which is outside of redux.
The only code creating actions for routing.js should be the routerListener.
Within the laboratory, tools are interconnected and will want to link to each other. For example, in Transaction Builder, it is convenient to have a button to sign the built transaction.
The linkBuilder interface provides simple functions that build the desired url using the storeSerializer. Components that want to link to others don't need knowledge of how the storeSerializer works. They only need to use the simple functions that the linkBuilder provides.