Metro has a multi-layered cache: you can set up multiple caches to be used by Metro instead of one. This has several advantages, on this page we will explain how the caches work.
Caches give big performance benefits, they can increase the speed of a bundler with more than tenfold. However, many systems use a non-persistent cache. With Metro we have a more sophisticated way of caching with a layer system. For example, we can store the cache on a server. Because of this all bundlers connected to the same server can use the shared cache. As a result the initial build time for CI servers and local development become significantly lower.
We want to store caches in multiple places as to always have a cache to fallback to. That's why there is a multi-layered cache system.
Cache Fetching & Saving
There is an ordering mechanism to determine which cache to use. For retrieving a cache we go through the caches from top to bottom until we find a result, for saving a cache we do the same until we find a store that has the cache.
Let's say you have two cache stores: one on a server and one on your local file system. You would specify that in this way:
Metro will first look into the
FileStore when we retrieve a cache. If it can't find the cache there it will check
NetworkStore, and so on. Finally if there's no cache there it will generate a new cache itself. As soon as the cache has been generated, Metro will go again from top to bottom to store the cache in all stores. This also happens if a cache is found. For example, if Metro finds a cache in the
NetworkStore it will store it in
FileStore as well.