Added webdriver support for App v2.

Added webdriver support for App v2.

This change allows testing using Chrome webdriver on App v2 windows.  It exposes (via window_handles) and allows switching control to App v2 windows for testing.

BUG=6801177
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=151328

[Danh Nguyen, 2012-08-07 16:50:19]

Hi Ken,

As we talked last week, this CL adds supports for testing Apps v2, such as
Pulsar app, using Chrome webdriver in google3.  I hope you could give me some
time reviewing this change.

James,

When I got Ken's approval on webdriver related changes, could you take a quick
look and grant your OWNERS' approval?

Thanks,
Danh

[James Hawkins, 2012-08-07 16:57:07]

Yes, just ping me when it's ready.

[kkania, 2012-08-07 19:57:44]

can you add a test in chrome/test/webdriver/test/chromedriver_tests.py in the
Extension test class?

Also, make sure you run chrome/test/webdriver/test/run_chromedriver_tests.py
before you submit. It doesn't run on the try/commit bots

http://codereview.chromium.org/10854026/diff/1/chrome/test/automation/automat...
File chrome/test/automation/automation_json_requests.h (right):

http://codereview.chromium.org/10854026/diff/1/chrome/test/automation/automat...
chrome/test/automation/automation_json_requests.h:94: // Returns whether this ID
refers to an App v2 view.
This class should eventually be removed and replaced with just an AutomationID.
Instead of adding this method, just get the ID and check its type at the caller
site.

[Danh Nguyen, 2012-08-07 20:33:09]

Thanks Ken for your quick response.  I'm working on it.

Thanks,
Danh


On Tue, Aug 7, 2012 at 12:57 PM, <kkania@chromium.org> wrote:

> can you add a test in chrome/test/webdriver/test/**chromedriver_tests.py
> in the
> Extension test class?
>
> Also, make sure you run chrome/test/webdriver/test/**
> run_chromedriver_tests.py
> before you submit. It doesn't run on the try/commit bots
>
>
> http://codereview.chromium.**org/10854026/diff/1/chrome/**
>
test/automation/automation_**json_requests.h<http://codereview.chromium.org/10854026/diff/1/chrome/test/automation/automation_json_requests.h>
> File chrome/test/automation/**automation_json_requests.h (right):
>
> http://codereview.chromium.**org/10854026/diff/1/chrome/**
>
test/automation/automation_**json_requests.h#newcode94<http://codereview.chromium.org/10854026/diff/1/chrome/test/automation/automation_json_requests.h#newcode94>
> chrome/test/automation/**automation_json_requests.h:94: // Returns whether
>
> this ID refers to an App v2 view.
> This class should eventually be removed and replaced with just an
> AutomationID. Instead of adding this method, just get the ID and check
> its type at the caller site.
>
>
http://codereview.chromium.**org/10854026/<http://codereview.chromium.org/108...
>

[Danh Nguyen, 2012-08-09 18:21:47]

Hi Ken,

I've added a test for packaged app in webdriver/test.  Also removed the changes
to automation_json_requests.h/cc as you suggested.

Could you take a look again?

Thanks a lot for your help.

Danh

http://codereview.chromium.org/10854026/diff/1/chrome/test/automation/automat...
File chrome/test/automation/automation_json_requests.h (right):

http://codereview.chromium.org/10854026/diff/1/chrome/test/automation/automat...
chrome/test/automation/automation_json_requests.h:94: // Returns whether this ID
refers to an App v2 view.
On 2012/08/07 19:57:44, kkania wrote:
> This class should eventually be removed and replaced with just an
AutomationID.
> Instead of adding this method, just get the ID and check its type at the
caller
> site. 

Done.

[kkania, 2012-08-09 22:04:49]

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver.py (right):

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver.py:25: EXTENSION_PACKAGED_APP = 6
is this the same thing as an app shell? Is there a reason for the two different
names?  Is this the same thing as a packaged app?  A packaged app doesn't
necessarily have to have a app shell window, does it?

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver_tests.py (right):

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver_tests.py:1147: app =
driver.find_element_by_xpath("//div[@title='Packaged app extension']")
is this finding the app icon on the ntp page? explicitly navigate to the
chrome://newtab after installing the extension for 2 reasons:
1) we might change the default url from ntp in the future
2) I'm not sure the app update in the ntp is synchronous with extension
installation...

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/packaged_app_extension/manifest.json (right):

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
chrome/test/webdriver/test/packaged_app_extension/manifest.json:6: "background":
{
just to make sure I understand, unlike extensions, app background pages aren't
started until the user clicks the app icon?

[Danh Nguyen, 2012-08-10 16:35:09]

Thanks Ken for reviewing and your comments.  I hope I've addressed them all. 
Could you take a look at the CL again?

Thanks,
Danh

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver.py (right):

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver.py:25: EXTENSION_PACKAGED_APP = 6
On 2012/08/09 22:04:49, kkania wrote:
> is this the same thing as an app shell? Is there a reason for the two
different
> names?  Is this the same thing as a packaged app?  A packaged app doesn't
> necessarily have to have a app shell window, does it?

No it doesn't.  Thanks Ken for pointing this out.  Changed all instances of
'packaged_app' to 'app_shell'.

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver_tests.py (right):

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver_tests.py:1147: app =
driver.find_element_by_xpath("//div[@title='Packaged app extension']")
On 2012/08/09 22:04:49, kkania wrote:
> is this finding the app icon on the ntp page? explicitly navigate to the
> chrome://newtab after installing the extension for 2 reasons:
> 1) we might change the default url from ntp in the future
> 2) I'm not sure the app update in the ntp is synchronous with extension
> installation...

Yes, explicit navigation to the NTP page was added.  Thanks Ken.  I overlooked
that.

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/packaged_app_extension/manifest.json (right):

http://codereview.chromium.org/10854026/diff/7001/chrome/test/webdriver/test/...
chrome/test/webdriver/test/packaged_app_extension/manifest.json:6: "background":
{
On 2012/08/09 22:04:49, kkania wrote:
> just to make sure I understand, unlike extensions, app background pages aren't
> started until the user clicks the app icon?

True, they're not loaded until the user launches the app.  Chrome then informs
the add to open its window.

[kkania, 2012-08-10 16:58:15]

lgtm

[James Hawkins, 2012-08-11 16:33:17]

LGTM with semi-question nit.

http://codereview.chromium.org/10854026/diff/3002/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver.py (right):

http://codereview.chromium.org/10854026/diff/3002/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver.py:25: APP_SHELL = 6
6? What happened to 5?

[Danh Nguyen, 2012-08-13 15:28:42]

Thanks James for reviewing the changes and for your approval.

Danh

http://codereview.chromium.org/10854026/diff/3002/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver.py (right):

http://codereview.chromium.org/10854026/diff/3002/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver.py:25: APP_SHELL = 6
On 2012/08/11 16:33:17, James Hawkins wrote:
> 6? What happened to 5?

The values here mirror the enum AutomationId::Type in
chrome/common/automation_id.h, in which 1 is kTypeTab, 4 is
kTypeExtensionInfoBar etc.  The value 5 is kTypeExtension, which doesn't have a
test in this file, so it's not defined here.

[James Hawkins, 2012-08-13 15:29:31]

OK, SLGTM, but please add comment where that value would be explaining this.

[Danh Nguyen, 2012-08-13 17:05:28]

Just added the comment.  Thanks, James.

Danh

http://codereview.chromium.org/10854026/diff/3002/chrome/test/webdriver/test/...
File chrome/test/webdriver/test/chromedriver.py (right):

http://codereview.chromium.org/10854026/diff/3002/chrome/test/webdriver/test/...
chrome/test/webdriver/test/chromedriver.py:25: APP_SHELL = 6
On 2012/08/13 15:28:42, Danh Nguyen wrote:
> On 2012/08/11 16:33:17, James Hawkins wrote:
> > 6? What happened to 5?
> 
> The values here mirror the enum AutomationId::Type in
> chrome/common/automation_id.h, in which 1 is kTypeTab, 4 is
> kTypeExtensionInfoBar etc.  The value 5 is kTypeExtension, which doesn't have
a
> test in this file, so it's not defined here.

Comments added.

[commit-bot: I haz the power, 2012-08-13 17:05:50]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/danhn@chromium.org/10854026/2004

[commit-bot: I haz the power, 2012-08-13 19:53:11]

Change committed as 151328