This was originally raised in whatwg/fetch#527
Do we need to treat requests for dedicated workers as non-subresource requests?
In the current HTML/fetch spec, dedicated workers are treated as non-subresources. This means they can have their own service worker. However, in the current Chromium implementation, dedicated workers don't have their own service worker and instead they are controlled by document's service worker because of an implementation reason. AFAICS, Gecko does the same behavior. If dedicated workers are treated as subresources, we could easily resolve this non-conformance.
(I'm a bit concerned about nested workers (#411). With this change, how do we determine a controller for a dedicated worker created from a service worker?)
CC: @annevk @jakearchibald @mikewest @mattto
This was originally raised in whatwg/fetch#527
Do we need to treat requests for dedicated workers as non-subresource requests?
In the current HTML/fetch spec, dedicated workers are treated as non-subresources. This means they can have their own service worker. However, in the current Chromium implementation, dedicated workers don't have their own service worker and instead they are controlled by document's service worker because of an implementation reason. AFAICS, Gecko does the same behavior. If dedicated workers are treated as subresources, we could easily resolve this non-conformance.
(I'm a bit concerned about nested workers (#411). With this change, how do we determine a controller for a dedicated worker created from a service worker?)
CC: @annevk @jakearchibald @mikewest @mattto