Expose privacy-relevant preferences via Chrome's extension API.

Expose privacy-relevant preferences in Chrome's extension APIs.

Adds a `privacy` permission to hold toggles for features that a) send information to external services (google or otherwise), b) send information to first-party websites, and c) effect Chrome's handling of the network in general.

(Metrics are still a TODO).

* privacy
    * network
    * networkPredictionEnabled (ChromeSetting)
  * services
    * alternateErrorPagesEnabled (ChromeSetting)
    * autofillEnabled (ChromeSetting)
    * instantEnabled (ChromeSetting)
    * safeBrowsingEnabled (ChromeSetting)
    * searchSuggestEnabled (ChromeSetting)
    * translateEnabled (ChromeSetting)
  * websites
    * blockThirdPartyCookies (ChromeSetting)
    * hyperlinkAuditingEnabled (ChromeSetting)
    * referrersEnabled (ChromeSetting)

BUG=88030, 55059, 54853
TEST=browser_tests

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=98396

[Mike West, 2011-07-01 13:39:15]

bauerb: I won't commit this until we've worked out what to do about the UMA bits
and pieces, and clarified the question of access to these permissions in
general.  I'd appreciate you reviewing it in the meantime anyway, but no rush.

Dominic, Jochen, Markus, Pam: FYI. If you see any privacy-related toggles that
I've missed, please let me know. :)

[Mike West, 2011-07-26 15:56:39]

This is a second strawman I put together in the hopes of kickstarting the
conversation around how we'd like to present these preferences to users. I've
implemented the suggestions I made to the chrome-extensions-team@ list last
week, which are reflected in the new description.

This isn't ready to commit yet:

* There's not agreement on the API, which is the biggest blocker. :)
* Two `contentSettings` tests are failing, and I'll need to move them to the new
`permission` test set once we agree how it should look.
* The permission strings are things I pulled out of thin air: we should discuss
them.
* UMA isn't implemented yet, as it's a bit more work.

Still, it's a good place to start.

@bauerb: do you see anything in the implementation that's worth fixing now?
@everyone else: WDYT?

[Bernhard Bauer, 2011-07-27 11:41:33]

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6353:
"thirdPartyCookiesAllowed": {
Could we keep thirdPartyCookiesAllowed in contentSettings? It has a much closer
interaction with those than with network settings in general.

http://codereview.chromium.org/7298005/diff/6001/chrome/test/data/extensions/...
File chrome/test/data/extensions/api_test/content_settings/standard/test.html
(right):

http://codereview.chromium.org/7298005/diff/6001/chrome/test/data/extensions/...
chrome/test/data/extensions/api_test/content_settings/standard/test.html:66:
cs.global[pref].get({}, expectFalse(message));
These prefs are not in global anymore.

[Mike West, 2011-07-27 11:48:01]

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6353:
"thirdPartyCookiesAllowed": {
On 2011/07/27 11:41:34, Bernhard Bauer wrote:
> Could we keep thirdPartyCookiesAllowed in contentSettings? It has a much
closer
> interaction with those than with network settings in general.

Let's talk this afternoon. :)

http://codereview.chromium.org/7298005/diff/6001/chrome/test/data/extensions/...
File chrome/test/data/extensions/api_test/content_settings/standard/test.html
(right):

http://codereview.chromium.org/7298005/diff/6001/chrome/test/data/extensions/...
chrome/test/data/extensions/api_test/content_settings/standard/test.html:66:
cs.global[pref].get({}, expectFalse(message));
On 2011/07/27 11:41:34, Bernhard Bauer wrote:
> These prefs are not in global anymore.

Lots of ContentSettings tests are broken in this CL at the moment. I'm waiting
until we actually have agreement on the namespaces before I fix all of them,
just so I don't end up throwing things away, and then bringing them back if we
change our minds. :)

[battre, 2011-07-27 12:24:20]

http://codereview.chromium.org/7298005/diff/6001/chrome/browser/extensions/ex...
File chrome/browser/extensions/extension_preference_apitest.cc (right):

http://codereview.chromium.org/7298005/diff/6001/chrome/browser/extensions/ex...
chrome/browser/extensions/extension_preference_apitest.cc:28:
pref_service->SetBoolean(prefs::kSearchSuggestEnabled, false);
these defaults not match the extensio_api.json

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6316: "description": "This
preference's value is a boolean, defaulting to <code>true</code>."
What is an alternate error page? Will there always be one? (-> api stability)

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6321: "description": "This
preference's value is a boolean, defaulting to <code>true</code>."
I'd find it great, if each element was shortly described.

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6331: "description": "This
preference's value is a boolean, defaulting to <code>true</code>."
defaults to true?

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6346: "description": "This
preference's value is a boolean, defaulting to <code>true</code>."
the setting enabled that translate is *offered*, it does not enable it

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6356: "description": "Whether
third party cookies should be allowed. The value of this preference is of type
boolean, and the default value is true."
what does "allowed" mean? Setting? Reading?

[Mike West, 2011-07-27 18:04:14]

Added descriptions, moved tests around so that they match the current structure.
Anticipate moving them again as the namespace discussion continues. :)

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6321: "description": "This
preference's value is a boolean, defaulting to <code>true</code>."
On 2011/07/27 12:24:20, battre wrote:
> I'd find it great, if each element was shortly described.

I've just copy/pasted all of these. I'm waiting until we have agreement on how
things look before fleshing it out.

That said, regardless of the namespace, these need descriptions... so. I'll do
something about it.

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6346: "description": "This
preference's value is a boolean, defaulting to <code>true</code>."
On 2011/07/27 12:24:20, battre wrote:
> the setting enabled that translate is *offered*, it does not enable it

Done.

http://codereview.chromium.org/7298005/diff/6001/chrome/common/extensions/api...
chrome/common/extensions/api/extension_api.json:6356: "description": "Whether
third party cookies should be allowed. The value of this preference is of type
boolean, and the default value is true."
On 2011/07/27 12:24:20, battre wrote:
> what does "allowed" mean? Setting? Reading?

Done.

[Mike West, 2011-08-23 12:39:06]

Hi Erik!

I've taken your last suggestion (`chrome.privacy.*`) and implemented it here for
discussion.

Bernhard and Dominic have already been through most of the implementation bits,
so I'd appreciate you taking a look at the extension bits (or delegating to
someone on the extensions team).

Thanks!

-Mike

[Mihai Parparita -not on Chrome, 2011-08-23 22:24:55]

Ben is going be looking at this from the extensions side.

[Ben Olmstead, 2011-08-24 19:51:46]

+kathyw for the documentation changes

http://codereview.chromium.org/7298005/diff/13001/chrome/app/generated_resour...
File chrome/app/generated_resources.grd (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/app/generated_resour...
chrome/app/generated_resources.grd:3768: TODO: Settings for features that rely
on network services that send information to Google and your default search
provider.
This shouldn't go in yet.

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7750: "description": "Settings
that effect Chrome's handling of network connections in general.",
effect -> affect

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7804: "description": "Settings
that effect what information Chrome sends when requesting websites.",
effect -> affect

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ex...
File chrome/common/extensions/extension_permission_set.h (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ex...
chrome/common/extensions/extension_permission_set.h:49: kPrivacy,
The changes to extension_permission_set.{cc,h} should not go in until the
privacy API is actually moving out of experimental.  Otherwise, an extension
with "privacy" in its permission set in the manifest will show the "TODO:"
placeholder in its permission dialog.

[kathyw, 2011-08-24 20:10:18]

LGTM

It'd be nice to see this staged somewhere, but the doc comments seem fine to me.
(Just one small thing I noticed, besides the affect/effect thing Ben pointed
out.)

-k-

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7814: "description": "If
enabled, Chrome sends auditing pings when requested by a website (<code>&lt;a
ping&gt;</code>). The value of this preference is of type boolean, and the
default value is true."
true -> <code>true</code>

[Mike West, 2011-08-25 08:18:25]

Thanks for the feedback, I've dropped the permission.

Docs somewhat staged at
http://www.corp.google.com/~mkwst/docs/experimental.privacy.html (no style, but
I can't figure out how to easily do that without copying half the world to my
home directory... Kathy? I assume there's an easy way that I've missed? :) )

http://codereview.chromium.org/7298005/diff/13001/chrome/app/generated_resour...
File chrome/app/generated_resources.grd (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/app/generated_resour...
chrome/app/generated_resources.grd:3768: TODO: Settings for features that rely
on network services that send information to Google and your default search
provider.
On 2011/08/24 19:51:51, Ben Olmstead wrote:
> This shouldn't go in yet.

Done.

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7750: "description": "Settings
that effect Chrome's handling of network connections in general.",
On 2011/08/24 19:51:51, Ben Olmstead wrote:
> effect -> affect

Gah. English.

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7804: "description": "Settings
that effect what information Chrome sends when requesting websites.",
On 2011/08/24 19:51:51, Ben Olmstead wrote:
> effect -> affect

Done.

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7814: "description": "If
enabled, Chrome sends auditing pings when requested by a website (<code>&lt;a
ping&gt;</code>). The value of this preference is of type boolean, and the
default value is true."
On 2011/08/24 20:10:18, kathyw wrote:
> true -> <code>true</code>

Done.

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ex...
File chrome/common/extensions/extension_permission_set.h (right):

http://codereview.chromium.org/7298005/diff/13001/chrome/common/extensions/ex...
chrome/common/extensions/extension_permission_set.h:49: kPrivacy,
On 2011/08/24 19:51:51, Ben Olmstead wrote:
> The changes to extension_permission_set.{cc,h} should not go in until the
> privacy API is actually moving out of experimental.  Otherwise, an extension
> with "privacy" in its permission set in the manifest will show the "TODO:"
> placeholder in its permission dialog.

Done.

[kathyw, 2011-08-25 16:14:47]

> Docs somewhat staged at
> http://www.corp.google.com/~mkwst/docs/experimental.privacy.html (no style,
> but
> I can't figure out how to easily do that without copying half the world to
> my
> home directory... Kathy? I assume there's an easy way that I've missed? :) )

Here's my brain-dead script for copying my files for staging
(substitute the right directory path wherever you see '...').

#! /bin/sh

cd /Chromium/chromium-trunk/src/chrome/common/extensions/docs

echo 'Copying from /Chromium/chromium-trunk/src/chrome/common/extensions/docs'
echo 'to http://www/.../Chrome/trunk-docs/'
echo

## Keep a copy
rm -rf .../Chrome/trunk-docs.OLD
mv .../Chrome/trunk-docs .../Chrome/trunk-docs.OLD
mkdir .../Chrome/trunk-docs

## Copy the files over
cp *html .../Chrome/trunk-docs
cp -r images/ .../Chrome/trunk-docs/
cp -r css/ .../Chrome/trunk-docs/
cp -r js/ .../Chrome/trunk-docs

## And make them readable
chmod -R 755 .../Chrome/trunk-docs/

-k-

[kathyw, 2011-08-25 16:19:03]

LGTM

(though I noticed one more unformatted "true")

http://codereview.chromium.org/7298005/diff/21001/chrome/common/extensions/ap...
File chrome/common/extensions/api/extension_api.json (right):

http://codereview.chromium.org/7298005/diff/21001/chrome/common/extensions/ap...
chrome/common/extensions/api/extension_api.json:7824: "description": "If
disabled, Chrome blocks third-party sites from setting cookies. The value of
this preference is of type boolean, and the default value is true."
true -> <code>true</code>

[Ben Olmstead, 2011-08-25 20:37:57]

LGTM

[commit-bot: I haz the power, 2011-08-25 21:57:10]

Can't process patch for file
chrome/test/data/extensions/api_test/preference/session_only_incognito/test.html.
A +

[Mike West, 2011-08-25 22:34:36]

On 2011/08/25 21:57:10, I haz the power (commit-bot) wrote:
> Can't process patch for file
>
chrome/test/data/extensions/api_test/preference/session_only_incognito/test.html.
> A +

Filed http://crbug.com/94330 for the commit bot weirdness; I'll try to land this
manually in a moment.