- From: Anne van Kesteren <annevk@annevk.nl>
- Date: Thu, 6 Aug 2015 11:20:51 +0200
- To: Brian Smith <brian@briansmith.org>
- Cc: Mike West <mkwst@google.com>, Brad Hill <hillbrad@gmail.com>, Wendy Seltzer <wseltzer@w3.org>, Dan Veditz <dveditz@mozilla.com>, "public-webappsec@w3.org" <public-webappsec@w3.org>
On Wed, Aug 5, 2015 at 6:54 PM, Brian Smith <brian@briansmith.org> wrote: > On Wed, Aug 5, 2015 at 3:48 AM, Mike West <mkwst@google.com> wrote: >> Last I heard, Anne was going to decide whether we would end up defining >> the "Is X a passthrough request?" property in MIX or Fetch. I don't have a >> strong opinion either way. > > IMO, it is better to do in Fetch, because then (a) more specifications can > reference it, (b) people working on changing Fetch and Fetch-based things > may more easily notice that it is something that needs to be considered. and > (c) it is easy to tweak the Fetch spec to improve the definition later, if > necessary, than it is to improve MIX. (c) seems like a bug. (b) rings true. As for (a), nobody could think of any other specifications. The other problem I have is how I would go about defining this. It's quite easy to construct a new request from an existing one. I guess for the purposes of this that would not necessarily make it passthrough. It seems in order to define "passthrough" service workers would have to set a bit on the request they associate with the fetch event. And then when you invoke fetch() or new Request() and pass in such a request and don't modify anything else we'd keep the bit set? (Though we would modify what CSP policy applies and such as part of the normal request normalization procedures.) -- https://annevankesteren.nl/
Received on Thursday, 6 August 2015 09:21:16 UTC