[slightlyoff/ServiceWorker] Handling race conditions - API for accessing pending requests? (#959)

There's a couple of race conditions arising in the SW I'm working on relating to prefetch/preload:

1. Once a user has navigated to a page I send off a request to prefetch the next page, this comes back with a load of link headers for css & js resources, which I also prefetch. There is a chance that the user will navigate to this next page before all those requests have responded. How can I force the new request (and all its related preload requests for the links) to use the requests that are already in flight?
2. I have a css file containing my critical path css. On the first page load this is inlined, but a link=preload header also tells the browser to fetch it as a standalone file so that when the sw kicks in it can safely request the page without inlined css. But in the SW I don't know if the browser supports preload so, to be on the safe side, I may want to manually request all the files in the link headers. For browsers that _do_ support preload this will mean unnecessary doubling of requests. 

It'd be useful to have a pending requests API, with similar request matching rules to the cache API. Could it even be built into the Cache API? e.g.

```js
cache.open('mycache')
  .then(() => {
    cache.set(request, fetch(request).then(res => {
        // optionally either throw to reject, or return res to put in cache
    })
  }
```
If the promise rejects, or resolves with anything other than a response then the response would not be put in the cache. `cache.get(request)` would resolve/reject when the promise resolves/rejects

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/slightlyoff/ServiceWorker/issues/959