On 1 Nov 2011, at 10:21, Jean-Noël Colin wrote:

> Returning cookies from remote side is one thing, another thing is to send 
> available cookies to that remote side as well. 

Given the origin issue (see other email) I'm not sure either is a good idea :(

> 
> About inserting websites into widgets, it also sounds to me like a bad idea. 
> The way I see it is that the widget would provide the UI and possibly some 
> logic, while the remote side would act as a backend. Do you share this view? 

I think thats the basic model most people think of.

> 
> Is there anything in the widgets specs that forbids inserting websites into 
> widgets (basically having a widget with just an iframe with src parameter 
> pointing to a remote site)

No, not really. However a widget isn't expected to navigate, so if you pull in 
a site that contains links its unlikely to work, especially on mobile widget 
runtimes. I think for Wookie the problem is mainly if you try to proxify it 
first, as it will then have the wrong origin. But just pulling in a page as an 
iframe should be OK, and can be a reasonable quick hack to get a widget created 
by wrapping an existing "widget-like" web application.

> 
> Cheers
> 
> --
> Jean-Noël 
> 
> 
> On 29 Oct 2011, at 11:12, Scott Wilson <[email protected]> wrote:
> 
>> 
>> On 29 Oct 2011, at 10:10, Ross Gardler wrote:
>> 
>>> Whilst I agree about proxy not being used to insert websites into widgets,
>>> it is sometimes necessary to send and receive cookies, e.g. basic
>>> authentication. For me this is the important pay of this query.
>> 
>> +1 on returning complete headers including cookies
>> 
>>> 
>>> Sent from my mobile device, please forgive errors and brevity.
>>> On Oct 29, 2011 9:38 AM, "Scott Wilson" <[email protected]>
>>> wrote:
>> 

Reply via email to