Add support for the 'webview_android' policy configuration

Add support for the 'webview_android' policy configuration
    
Allows specifying which policies are supported on Webview.
Supported policies will display the webview restriction
name for policy 'Foo' as 'com.android.browser:Foo'

BUG=536733
r201

[dgn, 2015-09-29 10:42:24]

dgn@chromium.org changed reviewers:
+ bartfab@chromium.org

[Bernhard Bauer, 2015-09-29 12:27:41]

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

[Bernhard Bauer, 2015-09-29 12:27:41]

lgtm

https://codereview.chromium.org/1372953002/diff/20001/grit/format/policy_temp...
File grit/format/policy_templates/writer_configuration.py (right):

https://codereview.chromium.org/1372953002/diff/20001/grit/format/policy_temp...
grit/format/policy_templates/writer_configuration.py:45: 'webview_name': 'System
WebView Google',
The official name (according to
https://play.google.com/store/apps/details?id=com.google.android.webview) is
"Android System WebView".

[dgn, 2015-09-29 15:52:59]

Thanks!

https://codereview.chromium.org/1372953002/diff/20001/grit/format/policy_temp...
File grit/format/policy_templates/writer_configuration.py (right):

https://codereview.chromium.org/1372953002/diff/20001/grit/format/policy_temp...
grit/format/policy_templates/writer_configuration.py:45: 'webview_name': 'System
WebView Google',
On 2015/09/29 12:27:41, Bernhard Bauer wrote:
> The official name (according to
> https://play.google.com/store/apps/details?id=com.google.android.webview) is
> "Android System WebView".

Done.

[dgn, 2015-10-27 21:06:50]

dgn@chromium.org changed reviewers:
+ thakis@chromium.org

[dgn, 2015-10-27 21:06:51]

PTAL

[Bernhard Bauer, 2015-10-28 09:52:32]

lgtm

[Nico, 2015-11-02 16:41:10]

code lgtm except a few nits

I don't really understand what the bug is about though. It sounds like we
explicitly want to have different policy configurations for Chromium and WebView
– I would've thought we would want to keep them identical so that WebView
requires less maintenance. If there's some good reason why we wouldn't want it
and it doesn't take too long to explain, I'd find that interesting :-)

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
File grit/format/policy_templates/writers/template_writer.py (left):

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:102: is_supported =
lambda x: platform in x['platforms']
This returns a bool…

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
File grit/format/policy_templates/writers/template_writer.py (right):

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:104: if not platform in
supported_on['platforms']:
nit: `platform not in` instead of `not platform in`

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:106: if product and not
product in supported_on['product']:
same nit

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:108: return supported_on
…but this returns an object. Since it's used with filter() which talks about
boolean values, I think returning False and True instead of None and
supported_on is easier to understand.

Is this function body equivalent to

  return platform in so[] and (!product or product in so[])

? (Technically, that also needs an `and bool(supported_on)`, but you already use
`in supported_on`, so if supported_on was None or empty that'd already fail)

[dgn, 2015-11-02 17:55:33]

Description was changed from

==========
Add support for the Android WebView policy configuration

Webview has the platform 'android' and the product 'webview',
its description specifies that the name is the same as the one
for android but prefixed with $android_webview_restriction_prefix

BUG=536733
==========

to

==========
Add support for the 'webview_android' policy configuration
    
Allows specifying which policies are supported on Webview.
Supported policies will display the webview restriction
name for policy 'Foo' as 'com.android.browser:Foo'

BUG=536733
==========

[dgn, 2015-11-02 18:19:50]

> I don't really understand what the bug is about though. It sounds like
> we explicitly want to have different policy configurations for Chromium 
> and WebView – I would've thought we would want to keep them identical so
> that WebView requires less maintenance. If there's some good reason why
> we wouldn't want it and it doesn't take too long to explain, I'd find
> that interesting :-)

When we added policy support to WebView, we prefixed the app restiction names
to reduce the chances of conflicting with restrictions declared by the
embedding app. Because of that, the policy "Foo" has the app restriction name
"Foo" in Chrome and "com.android.browser:Foo" in Webview.
The CL is to display an extra line that says from which version of webview the
policy is supported and show the full restriction name that has to be declared:
https://screenshot.googleplex.com/rUTBrLEcZ3d.png

https://codereview.chromium.org/1368373002/ introduces the visible changes
to the policy list, enabled by the current CL

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
File grit/format/policy_templates/writers/template_writer.py (right):

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:104: if not platform in
supported_on['platforms']:
On 2015/11/02 16:41:10, Nico wrote:
> nit: `platform not in` instead of `not platform in`

Acknowledged.

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:106: if product and not
product in supported_on['product']:
On 2015/11/02 16:41:10, Nico wrote:
> same nit

Acknowledged.

https://codereview.chromium.org/1372953002/diff/60001/grit/format/policy_temp...
grit/format/policy_templates/writers/template_writer.py:108: return supported_on
On 2015/11/02 16:41:10, Nico wrote:
> …but this returns an object. Since it's used with filter() which talks about
> boolean values, I think returning False and True instead of None and
> supported_on is easier to understand.
> 
> Is this function body equivalent to
> 
>   return platform in so[] and (!product or product in so[])
> 
> ? (Technically, that also needs an `and bool(supported_on)`, but you already
use
> `in supported_on`, so if supported_on was None or empty that'd already fail)

Done.

[Nico, 2015-11-03 19:20:57]

landed in grit r201

[dgn, 2015-11-03 21:24:32]

On 2015/11/03 19:20:57, Nico wrote:
> landed in grit r201

Thanks a lot!

[dgn, 2015-11-03 21:24:53]

On 2015/11/03 21:24:32, dgn wrote:
> On 2015/11/03 19:20:57, Nico wrote:
> > landed in grit r201
> 
> Thanks a lot!

Do I abandon the CL now?

[Nico, 2015-11-03 21:25:51]

Description was changed from

==========
Add support for the 'webview_android' policy configuration
    
Allows specifying which policies are supported on Webview.
Supported policies will display the webview restriction
name for policy 'Foo' as 'com.android.browser:Foo'

BUG=536733
==========

to

==========
Add support for the 'webview_android' policy configuration
    
Allows specifying which policies are supported on Webview.
Supported policies will display the webview restriction
name for policy 'Foo' as 'com.android.browser:Foo'

BUG=536733
r201
==========

[Nico, 2015-11-03 21:26:11]

Probably enough to close it ("edit issue" -> check "closed"). I just did that.

[dgn, 2015-11-03 21:26:45]

On 2015/11/03 21:26:11, Nico wrote:
> Probably enough to close it ("edit issue" -> check "closed"). I just did that.

Thanks