Audit and improve service worker loading strategy

Right now, our service worker configuration is great for its original intent, which is to make Popcode available offline if a network connection is unavailable. The network-first strategy also has the advantage of making it reasonably simple to ensure that the most recent version of Popcode is always going to load.

However, if I’m not mistaken, there is some overhead to loading assets from a service worker, so I want to make sure we’re also using them in the most performant possible way. In particular, there is absolutely no reason to go network-first for versioned assets that I can think of.

Overall though this isn’t as much of a concrete proposal as a call for more investigation, into what options the offline-plugin gives us and how we can get the best performance from service workers. A couple of specific things to keep an eye on:

• I’m not really clear on how service workers interact with normal browser caching. We currently have something like a 2 hour expiry on assets in the browser. Obviously this means we’re expiring versioned assets too soon. But a good strategy at the service worker level might make this less of a concern
• Anything we do should make the development experience no worse. Ideally we would also not need to have substantially different service worker settings in development, although this may be unrealistic

Comienza revisando la configuración existente del service worker y las opciones de offline-plugin, y luego compáralas con las directrices de carga de alto rendimiento del service worker enlazadas en el issue. Investiga la caché de assets versionados, la expiración de dos horas del navegador y el comportamiento en desarrollo; se considera terminado cuando se selecciona y aplica una estrategia que preserve el acceso offline y las actualizaciones fiables, al tiempo que mejora el rendimiento de carga de los assets.