- From: Marcos Caceres <marcosscaceres@gmail.com>
- Date: Mon, 20 Jun 2011 12:13:41 +0200
- To: Robin Berjon <robin.berjon@gmail.com>
- Cc: timeless <timeless@gmail.com>, public-webapps <public-webapps@w3.org>, Wojciech Masłowski <wmaslowski@opera.com>, Charles McCathieNevile <chaals@opera.com>
On Mon, Jun 20, 2011 at 11:27 AM, Robin Berjon <robin.berjon@gmail.com> wrote: > Hey, never too late to jump in I guess! > > On May 12, 2011, at 14:54 , Marcos Caceres wrote: >> On Thu, May 12, 2011 at 2:43 PM, timeless <timeless@gmail.com> wrote: >>> I really don't think relaxing the syntax is the right path forward. >> >> I'm ok with leaving it as is... but I guess we will have to see what >> runtimes end up doing. Opera has already willfully violated the spec. > > There's a very good reason for not loosening the syntax, and for not wilfully violating the spec. The reason the spec is strict here is for forward compatibility. If someone says http://w3.org/foo and you ignore the /foo in v1, but then in v2 we wish to add path matching (as has already been suggesting), then if we go ahead all of a sudden that old content acquires new behaviour. > In case of <access *origin*> you are only really interested in the *origin*. If we need to start limiting to resources (paths), we can introduce something else. However, there are few use cases for that AFAIK. It never came up with Opera Extensions or with people who have implemented widgets. > Of course the better solution is probably just to kill WARP and rely solely on CORS, but that's another kettle of fish altogether. > That will already work if your widgets run on HTTP, though they will need to be running on a proper origin. -- Marcos Caceres http://datadriven.com.au
Received on Monday, 20 June 2011 10:15:05 UTC