Initial support for web-extent background pages

Initial support for web-extent background pages.

This patch adds a new RVH container: BackgroundContents. The idea is that apps can open a live web-page as a "background" page using window.open('<url>', '<name>', 'background');

If 'background' is specified and the opener is within the app's extent, a BackgroundContents will be used. Otherwise, the 'background' feature is ignored and it is treated as a regular popup call.

Note that as of this patch the following are explicitly not-yet addressed:

1) Session storage for BackgroundContents
2) SSL (or other failures) requiring UI
3) Javascript messages (alert, etc...)
4) Session restore

TEST=All tests should pass
BUG=41275
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=46544

[Aaron Boodman, 2010-04-29 16:57:52]

Did not review BackgroundContents yet, but I think the GetAs* methods should be
removed.

http://codereview.chromium.org/1734014/diff/2001/3005
File chrome/browser/renderer_host/render_view_host_delegate.h (right):

http://codereview.chromium.org/1734014/diff/2001/3005#newcode475
chrome/browser/renderer_host/render_view_host_delegate.h:475: virtual
ExtensionHost* GetAsExtensionHost();
I really don't like these GetAs* methods. Also, RenderViewHost is not supposed
to know about browser/... I thought, so this only works because the forward
declaration is the only reference. Why is it needed?

http://codereview.chromium.org/1734014/diff/2001/3008
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/2001/3008#newcode99
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:99:
new_tab_contents->render_view_host()->Init();
Ok, I see now. In this case, it seems like you're just using the GetAs() methods
to get access to RPH and RVH->Init(). It seems like these should both be
accessible to this code without casting. RVH seems like it should be accessible
to something called RVHDelegateHelper. And I bet RPH is accessile through RVH.

http://codereview.chromium.org/1734014/diff/2001/3008#newcode101
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:101:
new_rvh_delegate->GetAsBackgroundContents()->render_view_host()->Init();
This is the reason I don't like these GetAs* methods. When the next person comes
along and adds an implementation of RVHD, they won't know that they need to
update this code and we will crash.

http://codereview.chromium.org/1734014/diff/2001/3008#newcode118
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:118: if
(contents->GetAsBackgroundContents()) {
It seems like in this case, it should just be the RVHD that handles this. They
are going to have different implemetnations.

[rafaelw, 2010-04-30 20:44:01]

Ok, I've addressed your concerns (thanks to Matt for helping me brainstorm a
good approach to this).

The appropriate behavior is no relegated to being implemented in the RVHD
implementations and no changes to the RVHD interface are required.

Note that RVHD->GetAsTabContents() is still present and used. However, I think
this reasonable, since there are now two types of views that opened with
window.open: those with a visible view and those without.

Have another look.

http://codereview.chromium.org/1734014/diff/2001/3005
File chrome/browser/renderer_host/render_view_host_delegate.h (right):

http://codereview.chromium.org/1734014/diff/2001/3005#newcode475
chrome/browser/renderer_host/render_view_host_delegate.h:475: virtual
ExtensionHost* GetAsExtensionHost();
Removed.

On 2010/04/29 16:57:52, Aaron Boodman wrote:
> I really don't like these GetAs* methods. Also, RenderViewHost is not supposed
> to know about browser/... I thought, so this only works because the forward
> declaration is the only reference. Why is it needed?

http://codereview.chromium.org/1734014/diff/2001/3008
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/2001/3008#newcode99
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:99:
new_tab_contents->render_view_host()->Init();
Excellent point. RVHDelegateHelper now keeps track of its created RVHs (rather
than delegates).

On 2010/04/29 16:57:52, Aaron Boodman wrote:
> Ok, I see now. In this case, it seems like you're just using the GetAs()
methods
> to get access to RPH and RVH->Init(). It seems like these should both be
> accessible to this code without casting. RVH seems like it should be
accessible
> to something called RVHDelegateHelper. And I bet RPH is accessile through RVH.

http://codereview.chromium.org/1734014/diff/2001/3008#newcode101
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:101:
new_rvh_delegate->GetAsBackgroundContents()->render_view_host()->Init();
On 2010/04/29 16:57:52, Aaron Boodman wrote:
> This is the reason I don't like these GetAs* methods. When the next person
comes
> along and adds an implementation of RVHD, they won't know that they need to
> update this code and we will crash.

Done.

http://codereview.chromium.org/1734014/diff/2001/3008#newcode118
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:118: if
(contents->GetAsBackgroundContents()) {
On 2010/04/29 16:57:52, Aaron Boodman wrote:
> It seems like in this case, it should just be the RVHD that handles this. They
> are going to have different implemetnations.

Done.

[Aaron Boodman, 2010-05-01 01:05:00]

Yes, this seems better now. Will look at BackgroundContents later.

http://codereview.chromium.org/1734014/diff/10001/11002
File chrome/browser/notifications/balloon_host.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11002#newcode94
chrome/browser/notifications/balloon_host.cc:94:
DOMUIFactory::GetDOMUIType(balloon_->notification().content_url()), NULL,
why doesn't this pass |this| here, like ExtensionHost does?

http://codereview.chromium.org/1734014/diff/10001/11005
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11005#newcode29
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:29: GURL
opening_url,
Nit: opener_url would be a better name for this param since "opener" is a
well-known term.

http://codereview.chromium.org/1734014/diff/10001/11005#newcode37
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:37: return
(extensions_service->GetExtensionByURL(opening_url) ||
You should also check that the source renderer is hosting the matching app.

Also, do you want to implement the "background" manifest permission in this CL
or a subsequent one. That check would go here too.

http://codereview.chromium.org/1734014/diff/10001/11006
File chrome/browser/tab_contents/render_view_host_delegate_helper.h (right):

http://codereview.chromium.org/1734014/diff/10001/11006#newcode48
chrome/browser/tab_contents/render_view_host_delegate_helper.h:48: virtual
RenderWidgetHostView* GetCreatedWidget(int route_id);
Nit: add a blank line between each of these declarations. It would also be nice
to add comments to the other methods as long as you're here.

[Aaron Boodman, 2010-05-01 01:35:15]

On Fri, Apr 30, 2010 at 6:05 PM,  <[email protected]> wrote:
> Yes, this seems better now. Will look at BackgroundContents later.

BTW, to finish the thought, the reason why GetAsTabContents() here
seems OK to me is that the callsites actually want a TabContents and
can't do anything with other RVHD's. It's the logic where we go: are
you a tab contents? oh, well then you must be a x... that seems
dangerous to me.

- a

[Aaron Boodman, 2010-05-01 01:51:10]

http://codereview.chromium.org/1734014/diff/10001/11003
File chrome/browser/tab_contents/background_contents.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11003#newcode53
chrome/browser/tab_contents/background_contents.cc:53: bool app_closing_non_mac
= *Details<bool>(details).ptr();
The non_mac naming here seem superfluous -- we know it is non-mac by the
#ifdefs.

http://codereview.chromium.org/1734014/diff/10001/11003#newcode60
chrome/browser/tab_contents/background_contents.cc:60: // Don't follow the
behavior of having the last browser window closed
I don't understand this comment.

http://codereview.chromium.org/1734014/diff/10001/11003#newcode88
chrome/browser/tab_contents/background_contents.cc:88: //
RenderView::decidePolicyForNaviation. If BackgroundCotents become available
sp: BackgroundCotents

http://codereview.chromium.org/1734014/diff/10001/11003#newcode191
chrome/browser/tab_contents/background_contents.cc:191: // Windows not attached
to any browser window have an id of -1.
Says who? Link to reference? Or is there a constant that can be used instead?

http://codereview.chromium.org/1734014/diff/10001/11004
File chrome/browser/tab_contents/background_contents.h (right):

http://codereview.chromium.org/1734014/diff/10001/11004#newcode22
chrome/browser/tab_contents/background_contents.h:22: // Need description
Need description :)

http://codereview.chromium.org/1734014/diff/10001/11004#newcode41
chrome/browser/tab_contents/background_contents.h:41: 
kill this blank line?

http://codereview.chromium.org/1734014/diff/10001/11004#newcode101
chrome/browser/tab_contents/background_contents.h:101: RenderViewHost*
render_view_host() { return render_view_host_; }
Is this a part of JavaScriptMessageBoxClient? It's hard to tell. If not, maybe
add a "public" section.

http://codereview.chromium.org/1734014/diff/10001/11004#newcode104
chrome/browser/tab_contents/background_contents.h:104: 
No blank line here.

[rafaelw, 2010-05-03 23:36:02]

http://codereview.chromium.org/1734014/diff/10001/11003
File chrome/browser/tab_contents/background_contents.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11003#newcode53
chrome/browser/tab_contents/background_contents.cc:53: bool app_closing_non_mac
= *Details<bool>(details).ptr();
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> The non_mac naming here seem superfluous -- we know it is non-mac by the
> #ifdefs.

Done.

http://codereview.chromium.org/1734014/diff/10001/11003#newcode60
chrome/browser/tab_contents/background_contents.cc:60: // Don't follow the
behavior of having the last browser window closed
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> I don't understand this comment.

Agree. Removed.

http://codereview.chromium.org/1734014/diff/10001/11003#newcode88
chrome/browser/tab_contents/background_contents.cc:88: //
RenderView::decidePolicyForNaviation. If BackgroundCotents become available
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> sp: BackgroundCotents

Done.

http://codereview.chromium.org/1734014/diff/10001/11003#newcode191
chrome/browser/tab_contents/background_contents.cc:191: // Windows not attached
to any browser window have an id of -1.
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> Says who? Link to reference? Or is there a constant that can be used instead?

Done.

http://codereview.chromium.org/1734014/diff/10001/11004
File chrome/browser/tab_contents/background_contents.h (right):

http://codereview.chromium.org/1734014/diff/10001/11004#newcode22
chrome/browser/tab_contents/background_contents.h:22: // Need description
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> Need description :)

Woops. Done.

http://codereview.chromium.org/1734014/diff/10001/11004#newcode41
chrome/browser/tab_contents/background_contents.h:41: 
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> kill this blank line?

Done.

http://codereview.chromium.org/1734014/diff/10001/11004#newcode101
chrome/browser/tab_contents/background_contents.h:101: RenderViewHost*
render_view_host() { return render_view_host_; }
Nope, tehe RenderViewHostDelegateViewHelper needs access. I've added a comment
and moved this to the top.

On 2010/05/01 01:51:10, Aaron Boodman wrote:
> Is this a part of JavaScriptMessageBoxClient? It's hard to tell. If not, maybe
> add a "public" section.

http://codereview.chromium.org/1734014/diff/10001/11004#newcode104
chrome/browser/tab_contents/background_contents.h:104: 
On 2010/05/01 01:51:10, Aaron Boodman wrote:
> No blank line here.

Done.

[Aaron Boodman, 2010-05-03 23:43:34]

lgtm

http://codereview.chromium.org/1734014/diff/18001/19003
File chrome/browser/extensions/extension_tabs_module_constants.h (right):

http://codereview.chromium.org/1734014/diff/18001/19003#newcode72
chrome/browser/extensions/extension_tabs_module_constants.h:72: 
nop

[rafaelw, 2010-05-03 23:49:47]

http://codereview.chromium.org/1734014/diff/18001/19003
File chrome/browser/extensions/extension_tabs_module_constants.h (right):

http://codereview.chromium.org/1734014/diff/18001/19003#newcode72
chrome/browser/extensions/extension_tabs_module_constants.h:72: 
On 2010/05/03 23:43:35, Aaron Boodman wrote:
> nop

Thanks. removed

[rafaelw, 2010-05-04 00:57:13]

almost forgot about these earlier comments. sorry.

http://codereview.chromium.org/1734014/diff/10001/11002
File chrome/browser/notifications/balloon_host.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11002#newcode94
chrome/browser/notifications/balloon_host.cc:94:
DOMUIFactory::GetDOMUIType(balloon_->notification().content_url()), NULL,
On 2010/05/01 01:05:01, Aaron Boodman wrote:
> why doesn't this pass |this| here, like ExtensionHost does?

Done. Thanks for the catch.

http://codereview.chromium.org/1734014/diff/10001/11005
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11005#newcode29
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:29: GURL
opening_url,
On 2010/05/01 01:05:01, Aaron Boodman wrote:
> Nit: opener_url would be a better name for this param since "opener" is a
> well-known term.

Done.

http://codereview.chromium.org/1734014/diff/10001/11005#newcode37
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:37: return
(extensions_service->GetExtensionByURL(opening_url) ||
On 2010/05/01 01:05:01, Aaron Boodman wrote:
> You should also check that the source renderer is hosting the matching app.

Done.

> 
> Also, do you want to implement the "background" manifest permission in this CL
> or a subsequent one. That check would go here too.

I'll do that in a follow-on CL.

http://codereview.chromium.org/1734014/diff/10001/11006
File chrome/browser/tab_contents/render_view_host_delegate_helper.h (right):

http://codereview.chromium.org/1734014/diff/10001/11006#newcode48
chrome/browser/tab_contents/render_view_host_delegate_helper.h:48: virtual
RenderWidgetHostView* GetCreatedWidget(int route_id);
On 2010/05/01 01:05:01, Aaron Boodman wrote:
> Nit: add a blank line between each of these declarations. It would also be
nice
> to add comments to the other methods as long as you're here.

Done.

[rafaelw, 2010-05-04 01:35:48]

http://codereview.chromium.org/1734014/diff/10001/11005
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/10001/11005#newcode37
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:37: return
(extensions_service->GetExtensionByURL(opening_url) ||

> Also, do you want to implement the "background" manifest permission in this CL
> or a subsequent one. That check would go here too.

No. wait. The other thing: I'll do it in this CL. (Done).

[Aaron Boodman, 2010-05-04 02:29:31]

Still LGTM

[darin (slow to review), 2010-05-04 23:43:05]

http://codereview.chromium.org/1734014/diff/35001/36006
File chrome/browser/tab_contents/background_contents.cc (right):

http://codereview.chromium.org/1734014/diff/35001/36006#newcode34
chrome/browser/tab_contents/background_contents.cc:34:
render_view_host_->PopupNotificationVisibilityChanged(false);
nit: it seems like this method is poorly named if it requires a comment to say
what it really means to call this method.  Maybe RenderViewHost should have a
method named AllowScriptToClose() that internally does the same thing as
PopupNotificationVisibilityChanged(false)?

http://codereview.chromium.org/1734014/diff/35001/36006#newcode36
chrome/browser/tab_contents/background_contents.cc:36: #if defined(OS_WIN) ||
defined(OS_LINUX)
it seems a shame that we have different notifications for this :-/  that's not
something you should fix in this CL of course.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode82
chrome/browser/tab_contents/background_contents.cc:82: switch
(params.transition) {
transition type is a bit-field.  maybe you should just test for the
{AUTO,MANUAL}_SUBFRAME bits directly?

http://codereview.chromium.org/1734014/diff/35001/36006#newcode88
chrome/browser/tab_contents/background_contents.cc:88: // Note: because
BackgroundContents are only available to extension apps,
"extension apps"?  is that how we are referring to chrome apps?

http://codereview.chromium.org/1734014/diff/35001/36006#newcode93
chrome/browser/tab_contents/background_contents.cc:93: // origin. Note: if the
first navigation is a to a URL outside the app's
nit: "is a to a" -> "is to a"

http://codereview.chromium.org/1734014/diff/35001/36006#newcode94
chrome/browser/tab_contents/background_contents.cc:94: // extent a background
page will be opened but will remain at about://blank.
nit: about://blank -> about:blank

http://codereview.chromium.org/1734014/diff/35001/36006#newcode110
chrome/browser/tab_contents/background_contents.cc:110: // TODO(rafaelw):
Implement, The JavaScriptModalDialog needs to learn about
what TabContents will we parent the alert over?

http://codereview.chromium.org/1734014/diff/35001/36006#newcode148
chrome/browser/tab_contents/background_contents.cc:148: true);  // is_dom_ui
why set is_dom_ui to true?

http://codereview.chromium.org/1734014/diff/35001/36008
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/35001/36008#newcode25
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:25: 
nit: no need for extra new line

[Aaron Boodman, 2010-05-05 00:14:04]

http://codereview.chromium.org/1734014/diff/35001/36006
File chrome/browser/tab_contents/background_contents.cc (right):

http://codereview.chromium.org/1734014/diff/35001/36006#newcode88
chrome/browser/tab_contents/background_contents.cc:88: // Note: because
BackgroundContents are only available to extension apps,
On 2010/05/04 23:43:05, darin wrote:
> "extension apps"?  is that how we are referring to chrome apps?

Yes, to distinguish from the old style application shortcuts, which are referred
to sometimes as "apps" in the codebase.

[rafaelw, 2010-05-05 18:05:30]

http://codereview.chromium.org/1734014/diff/35001/36006
File chrome/browser/tab_contents/background_contents.cc (right):

http://codereview.chromium.org/1734014/diff/35001/36006#newcode34
chrome/browser/tab_contents/background_contents.cc:34:
render_view_host_->PopupNotificationVisibilityChanged(false);
On 2010/05/04 23:43:05, darin wrote:
> nit: it seems like this method is poorly named if it requires a comment to say
> what it really means to call this method.  Maybe RenderViewHost should have a
> method named AllowScriptToClose() that internally does the same thing as
> PopupNotificationVisibilityChanged(false)?

Done.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode36
chrome/browser/tab_contents/background_contents.cc:36: #if defined(OS_WIN) ||
defined(OS_LINUX)
Agree.

On 2010/05/04 23:43:05, darin wrote:
> it seems a shame that we have different notifications for this :-/  that's not
> something you should fix in this CL of course.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode82
chrome/browser/tab_contents/background_contents.cc:82: switch
(params.transition) {
On 2010/05/04 23:43:05, darin wrote:
> transition type is a bit-field.  maybe you should just test for the
> {AUTO,MANUAL}_SUBFRAME bits directly?

Done.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode88
chrome/browser/tab_contents/background_contents.cc:88: // Note: because
BackgroundContents are only available to extension apps,

On 2010/05/05 00:14:05, Aaron Boodman wrote:
> On 2010/05/04 23:43:05, darin wrote:
> > "extension apps"?  is that how we are referring to chrome apps?
> 
> Yes, to distinguish from the old style application shortcuts, which are
referred
> to sometimes as "apps" in the codebase.

That's why I've been using term. I'm happy to use whatever term is best.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode93
chrome/browser/tab_contents/background_contents.cc:93: // origin. Note: if the
first navigation is a to a URL outside the app's
On 2010/05/04 23:43:05, darin wrote:
> nit: "is a to a" -> "is to a"

Done.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode94
chrome/browser/tab_contents/background_contents.cc:94: // extent a background
page will be opened but will remain at about://blank.
On 2010/05/04 23:43:05, darin wrote:
> nit: about://blank -> about:blank

Done.

http://codereview.chromium.org/1734014/diff/35001/36006#newcode110
chrome/browser/tab_contents/background_contents.cc:110: // TODO(rafaelw):
Implement, The JavaScriptModalDialog needs to learn about
BrowserList::LastActiveWithProfile? To be honest, I'm not sure yet.

On 2010/05/04 23:43:05, darin wrote:
> what TabContents will we parent the alert over?

http://codereview.chromium.org/1734014/diff/35001/36006#newcode148
chrome/browser/tab_contents/background_contents.cc:148: true);  // is_dom_ui
is_dom_ui forces javascript and image loading on. I was thinking that the
background page is pretty meaningless without javascript, but now that you point
it out, (a) no matter what, it does seem like a bad idea to pass is_dom_ui ==
true in case anyone changes what that means in the future, and (2) now that i
think about it more, javascript had to be running on the same domain in order to
get a background page, so this probably doesn't really make any different. And
further more, if the user wanted to do this to himself, it's probably not worth
preventing it.

changed to false. also, in extension_host.cc, i've changed is_dom_ui to false
and set load_images_automatically and javascript_enabled to be true outside of
this call.

On 2010/05/04 23:43:05, darin wrote:
> why set is_dom_ui to true?

http://codereview.chromium.org/1734014/diff/35001/36008
File chrome/browser/tab_contents/render_view_host_delegate_helper.cc (right):

http://codereview.chromium.org/1734014/diff/35001/36008#newcode25
chrome/browser/tab_contents/render_view_host_delegate_helper.cc:25: 
On 2010/05/04 23:43:05, darin wrote:
> nit: no need for extra new line

Done.

[darin (slow to review), 2010-05-05 18:19:27]

LGTM