Open fowl2 opened 2 years ago
Oh and in the interests of reducing header bloat, it also might make sense for these requests to stop implicitly adding the various Accept
headers? That would be a behaviour change though.
They can terminate the fetch from the processResponse callback already. I'm not sure it's worth automating that through an additional request boolean.
<a ping>
(Hyperlink auditing),navigator.sendBeacon()
, and thekeepalive
flag (once the parent document has been unloaded) have a concept of terminating the request once the response body starts to be received as these are "fire and forget" requests and there is nowhere for a response to go.eg. in Hyperlink auditing:
Would it make sense to add a flag and make this explicit? I'm not proposing this be exposed to web content because it doesn't seem that useful, but I guess it might be?
related: https://github.com/whatwg/fetch/issues/679
Aside: I guess the wait is for the start of the body rather than the status to allow for any cookies, etc. to be set by the response?