background-fetch
background-fetch copied to clipboard
API proposal for background downloading/uploading
The spec defines outlines the matching algorithm here: https://w3c.github.io/ServiceWorker/#request-matches-cached-item A Background Fetch should have the same set of restrictions (duplicate URLs allowed only with an appropriate `Vary` header).
We got here from w3ctag/design-reviews#279, looking at this in a breakout within the TAG meeting. One thing I noticed is that this seems, on some platforms, to add the ability...
1. I guess this name is only for local use as it's oddly generic given the niche semantics? 2. I'm a little concerned that this doesn't match implementations when multiple...
At some point @domenic asserted that all new fetches would use CORS. Is that no longer Google's policy?
Should `HEAD` be specialcased as well? Typically that's what Fetch does for these two methods. Perhaps we should even introduce a joint name for them to encourage this.
"Background fetch response" refers directly to an IDL enum from prose, expecting the reader to expand it. I'm not aware of any precedent for that and interweaving IDL with prose...
Since this is done in parallel, it seems two threads could read settledFetches at a similar time, then increment it locally, and write out the result, leading to it only...
This looks bad.
For Shaka Player, the service worker will need metadata from the page in order to process the responses from the background fetch. As BG fetch is currently spec'd/implemented, it looks...