Pass the render process id to the FrameNavigationState.

Pass the render process id to the FrameNavigationState.

It's used in addition to frame ids to identify frames correctly.

Also expose the processId via webNavigation.getFrame and getAllFrames

BUG=136090
TEST=ExtensionApiTest.WebNavigationGetFrame

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=148074

[jochen (gone - plz use gerrit), 2012-07-23 19:45:46]

plz review

this is the second last step. In the last step, I'll change how old frames are
discarded during a navigation. Currently FrameNavigationState::TrackFrame
discards old states if a main frame navigates.

Instead, old frames should be discarded when a main frame is committed.

Also, when old frames are discarded, we should send onErrorOccurred events for
ongoing navigations that we'll now ignore

[Matt Perry, 2012-07-23 21:28:10]

lgtm

https://chromiumcodereview.appspot.com/10805066/diff/1/chrome/browser/extensi...
File chrome/browser/extensions/api/web_navigation/frame_navigation_state.h
(right):

https://chromiumcodereview.appspot.com/10805066/diff/1/chrome/browser/extensi...
chrome/browser/extensions/api/web_navigation/frame_navigation_state.h:24:
typedef std::pair<int64, int> FrameID;
How much more work would it be to define your own struct with named variables?
"first" and "second" could easily get confusing as to which is which.

https://chromiumcodereview.appspot.com/10805066/diff/1/chrome/browser/extensi...
File chrome/browser/extensions/api/web_navigation/web_navigation_api.h (right):

https://chromiumcodereview.appspot.com/10805066/diff/1/chrome/browser/extensi...
chrome/browser/extensions/api/web_navigation/web_navigation_api.h:53: int64
frame_no,
optional nit: abbreviate as num instead of no

[jochen (gone - plz use gerrit), 2012-07-23 22:04:05]

http://codereview.chromium.org/10805066/diff/1/chrome/browser/extensions/api/...
File chrome/browser/extensions/api/web_navigation/frame_navigation_state.h
(right):

http://codereview.chromium.org/10805066/diff/1/chrome/browser/extensions/api/...
chrome/browser/extensions/api/web_navigation/frame_navigation_state.h:24:
typedef std::pair<int64, int> FrameID;
On 2012/07/23 21:28:10, Matt Perry wrote:
> How much more work would it be to define your own struct with named variables?
> "first" and "second" could easily get confusing as to which is which.

not too much. I'd have to specialize the std::less template, so I can insert it
into maps and sets. wdyt?

[Matt Perry, 2012-07-23 22:31:29]

http://codereview.chromium.org/10805066/diff/1/chrome/browser/extensions/api/...
File chrome/browser/extensions/api/web_navigation/frame_navigation_state.h
(right):

http://codereview.chromium.org/10805066/diff/1/chrome/browser/extensions/api/...
chrome/browser/extensions/api/web_navigation/frame_navigation_state.h:24:
typedef std::pair<int64, int> FrameID;
On 2012/07/23 22:04:05, jochen wrote:
> On 2012/07/23 21:28:10, Matt Perry wrote:
> > How much more work would it be to define your own struct with named
variables?
> > "first" and "second" could easily get confusing as to which is which.
> 
> not too much. I'd have to specialize the std::less template, so I can insert
it
> into maps and sets. wdyt?

Yeah I think that's worth it. Or maybe just define operator<