[chromium-dev] Re: Request for comments: Feed preview work

2009-05-07 Thread Aaron Boodman
On Thu, May 7, 2009 at 12:03 AM, Adam Barth wrote: > On Wed, May 6, 2009 at 11:55 PM, Aaron Boodman wrote: >> What about a setup where the content rendered in the tab area is >> running on chrome://, but contains a frame that hosts the actual feed >> running on http://foo.com? > > Sure, we could

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-07 Thread Darin Fisher
I see. That probably simplifies things. Hmm... -Darin On Wed, May 6, 2009 at 10:51 PM, Adam Barth wrote: > I don't think we want these feed previews to run with foo.com's > authority. I'd rather they ran with no one's authority. > > Adam > > > On Wed, May 6, 2009 at 8:42 PM, Darin Fisher wr

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-07 Thread Darin Fisher
chrome:// pages cannot load HTTP-based sub-resources. We don't want to taint the processes that render Chrome UI. -Darin On Wed, May 6, 2009 at 11:55 PM, Aaron Boodman wrote: > Let's say the feed is http://foo.com/feed.xml > > What about a setup where the content rendered in the tab area is >

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-07 Thread Itai
Minor comment, but I assume you will be triggering the feed-preview when people click on a link to the feed as well. Many sites have an RSS link to the feed's XML file because it is previewable in many browsers without autodiscovery. Currently, Chrome shows unformated XML. Doing CTRL-U after shows

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-07 Thread Adam Barth
On Wed, May 6, 2009 at 11:55 PM, Aaron Boodman wrote: > What about a setup where the content rendered in the tab area is > running on chrome://, but contains a frame that hosts the actual feed > running on http://foo.com? Sure, we could do that. Or even better is if the outer page is a chrome-e

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Aaron Boodman
Let's say the feed is http://foo.com/feed.xml What about a setup where the content rendered in the tab area is running on chrome://, but contains a frame that hosts the actual feed running on http://foo.com? The subscribe UI runs on the other page, so it is the only thing that needs elevated priv

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Ben Goodger (Google)
Regardless of whose authority they run at, it is somewhat desirable to have the feed URL display in the address bar, since that's the content that's being loaded. I would like to keep the flow in page as much as possible. We should be able to come up with some solution here that doesn't involve e

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Adam Barth
On Wed, May 6, 2009 at 8:45 PM, Mike Beltzner wrote: > FWIW, Firefox has had several security issues crop up with the mixed-content > feed preview implementation. Placing privileged controls so close to web > content should be avoided, IMO, if you want to keep this from being a > problem for Chro

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Adam Barth
I don't think we want these feed previews to run with foo.com's authority. I'd rather they ran with no one's authority. Adam On Wed, May 6, 2009 at 8:42 PM, Darin Fisher wrote: > WebKit does not support nested schemes.  It would fail in so many places to > recognize that the authority of such

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Mike Beltzner
On 6-May-09, at 9:56 PM, Adam Barth wrote: > From a security point of view, we'd ideally like to render feeds with > JavaScript and plug-ins disabled, as well as in a noAccess > SecurityOrigin. This is easier if the feed preview lives in its own > scheme. I'm happy to help out with the security

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Darin Fisher
WebKit does not support nested schemes. It would fail in so many places to recognize that the authority of such an URL is actually foo.com. (However, we could perhaps support this as we do view-source, where WebKit never actually sees the view-source URL.) -Darin On Wed, May 6, 2009 at 6:56 PM,

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Adam Barth
I think Darin had some strong opinions about whether we should do nested schemes like feed-view:http://foo.com/bar. >From a security point of view, we'd ideally like to render feeds with JavaScript and plug-ins disabled, as well as in a noAccess SecurityOrigin. This is easier if the feed preview

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Ben Goodger (Google)
On Wed, May 6, 2009 at 6:13 PM, Evan Martin wrote: > - Some existing practice on the web is to use > "feed://hostname/etc.xml", which drops the protocol (and should be > interpreted as HTTP).  Ideally you should redirect these into > view-feed:http://hostname/etc.xml so our view-feed works with h

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Aaron Boodman
Why bother with view-feed://? Why not just have the feed be styled more nicely, similar to the way that XML is styled more nicely by default in most browsers? - a On Wed, May 6, 2009 at 6:13 PM, Evan Martin wrote: > > On Wed, May 6, 2009 at 5:36 PM, Finnur Thorarinsson > wrote: >> I have alre

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Evan Martin
On Wed, May 6, 2009 at 5:36 PM, Finnur Thorarinsson wrote: > I have already added an API for PageActions and have a working RSS > PageAction extension, which does feed auto-detection on the page. Now it is > time to look into Feed previews. > I have spoken briefly to AdamB and EvanM about feed pr

[chromium-dev] Re: Request for comments: Feed preview work

2009-05-06 Thread Aaron Boodman
The way that makes most sense to me to implement this in the extensions system is: a) In C++ use content sniffing to make sure that the content type is always correct for feeds b) Add a feature to content scripts in extensions, so that they can match by content type c) Have a content script that