Transfer pending async script loader to new document.

Transfer pending async script loader to new document.

If a script element is moved to a new document and it has a pending
async script being loaded in the old document, transfer the pending
script load to the new document.

R=japhet@chromium.org
BUG=407824
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181081

[sof, 2014-08-28 08:50:00]

sigbjornf@opera.com changed reviewers:
+ haraken@chromium.org, morrita@chromium.org, tkent@chromium.org

[sof, 2014-08-28 08:50:00]

Please take a look.

The test is essentially morrita-san's from http://crbug.com/361011

[sof, 2014-08-28 11:10:46]

sigbjornf@opera.com changed reviewers:
+ abarth@chromium.org

[sof, 2014-08-28 11:10:46]

+abarth (accidentally left out initially.)

[haraken, 2014-08-28 11:52:46]

On 2014/08/28 11:10:46, sof wrote:
> +abarth (accidentally left out initially.)

Yeah, abarth would be the best reviewer.

Can you explain why it's guaranteed to be safe to execute the async script in a
new document? I can't think up of any particular exploits but I'm not sure of
security implications of this change.

[sof, 2014-08-28 12:18:48]

On 2014/08/28 11:52:46, haraken wrote:
> On 2014/08/28 11:10:46, sof wrote:
> > +abarth (accidentally left out initially.)
> 
> Yeah, abarth would be the best reviewer.
> 
> Can you explain why it's guaranteed to be safe to execute the async script in
a
> new document? I can't think up of any particular exploits but I'm not sure of
> security implications of this change.

[sof, 2014-08-28 12:24:00]

On 2014/08/28 12:18:48, sof wrote:
> On 2014/08/28 11:52:46, haraken wrote:
> > On 2014/08/28 11:10:46, sof wrote:
> > > +abarth (accidentally left out initially.)
> > 
> > Yeah, abarth would be the best reviewer.
> > 
> > Can you explain why it's guaranteed to be safe to execute the async script
in
> a
> > new document? I can't think up of any particular exploits but I'm not sure
of
> > security implications of this change.

I can explain my assumption: I'm assuming that the security check on
contentDocument access (in the testcase's case) will prevent transplanting
documents that shouldn't be allowed.

[sof, 2014-08-28 19:46:03]

On 2014/08/28 11:52:46, haraken wrote:
> On 2014/08/28 11:10:46, sof wrote:
> > +abarth (accidentally left out initially.)
> 
> Yeah, abarth would be the best reviewer.

To avoid overloading abarth too much - japhet, could you take a look, please?

[Nate Chapin, 2014-08-28 21:39:44]

japhet@chromium.org changed reviewers:
+ japhet@chromium.org

[Nate Chapin, 2014-08-28 21:39:44]

LGTM

Should there be a test variant that's expected to fail the transfer to the new
document, too?

[sof, 2014-08-29 08:52:46]

On 2014/08/28 21:39:44, Nate Chapin wrote:
> LGTM
> 
> Should there be a test variant that's expected to fail the transfer to the new
> document, too?

Thanks for the review. Good idea to have coverage for the load error code path
also; added such a variant.

[sof, 2014-08-29 08:53:21]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2014-08-29 08:53:51]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sigbjornf@opera.com/496443008/80001

[commit-bot: I haz the power, 2014-08-29 09:28:44]

Committed patchset #5 (id:80001) as 181081