Make playback fail if redirected to a different origin.

Make playback fail if redirected to a different origin.

Fixing spec compliance.
http://dev.w3.org/html5/spec/Overview.html#concept-media-load-resource

BUG=72578
TEST=BufferedResourceLoaderTest.HasSingleOrigin(), SimpleDataSourceTest.HasSingleOrigin()

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=76064

[acolwell GONE FROM CHROMIUM, 2011-02-24 00:04:56]

Fixing spec compliance for redirects to other origins.

http://dev.w3.org/html5/spec/Overview.html#concept-media-load-resource says to
"fetch the current media resource, from the media element's Document's origin,
with the force same-origin flag set"

http://dev.w3.org/html5/spec/Overview.html#fetch says "If the force same-origin
flag is set and the URL of the target of the redirect does not have the same
origin as the URL for which the fetch algorithm was invoked
Abort these steps and return failure from this algorithm, as if the remote host
could not be contacted."

This changes make bring the code into compliance.

[scherkus (not reviewing), 2011-02-25 01:40:54]

LGTM w/ spelling nit

few sanity check questions:
  - Can we have a layout test test for this case?  Seems beneficial
  - Having duplicated security-ish code make me sad but I have no
off-the-top-of-my-head ideas on how to de-duplicate (might require large
refactoring or something)

http://codereview.chromium.org/6580014/diff/1/webkit/glue/media/simple_data_s...
File webkit/glue/media/simple_data_source_unittest.cc (right):

http://codereview.chromium.org/6580014/diff/1/webkit/glue/media/simple_data_s...
webkit/glue/media/simple_data_source_unittest.cc:262: // Test redirct to the
same domain and then to a different domain.
redirct -> redirect

[acolwell GONE FROM CHROMIUM, 2011-02-25 17:42:42]

On 2011/02/25 01:40:54, scherkus wrote:
> LGTM w/ spelling nit
> 
> few sanity check questions:
>   - Can we have a layout test test for this case?  Seems beneficial
We'd have to be able to set up a web server that redirects to a different
domain. Is that possible in the current layout test code?

>   - Having duplicated security-ish code make me sad but I have no
> off-the-top-of-my-head ideas on how to de-duplicate (might require large
> refactoring or something)
I agree. I don't really like the duplication either, but since we have to pretty
separate implementations here I don't really feel like we have a choice without
some major rework. It doesn't seem worth it at this point.

[acolwell GONE FROM CHROMIUM, 2011-02-25 17:45:02]

Fixed nit

http://codereview.chromium.org/6580014/diff/1/webkit/glue/media/simple_data_s...
File webkit/glue/media/simple_data_source_unittest.cc (right):

http://codereview.chromium.org/6580014/diff/1/webkit/glue/media/simple_data_s...
webkit/glue/media/simple_data_source_unittest.cc:262: // Test redirct to the
same domain and then to a different domain.
On 2011/02/25 01:40:54, scherkus wrote:
> redirct -> redirect

Done.

[scherkus (not reviewing), 2011-02-25 18:57:57]

On 2011/02/25 17:42:42, acolwell wrote:
> On 2011/02/25 01:40:54, scherkus wrote:
> > LGTM w/ spelling nit
> > 
> > few sanity check questions:
> >   - Can we have a layout test test for this case?  Seems beneficial
> We'd have to be able to set up a web server that redirects to a different
> domain. Is that possible in the current layout test code?

Take a look at the xmlhttprequest tests, i.e,.
http://trac.webkit.org/browser/trunk/LayoutTests/http/tests/xmlhttprequest/ac...

What they do is redirect from localhost to 127.0.0.1 heh