Add user_gesture bit back to Mojo PermissionService methods.

Add user_gesture bit back to Mojo PermissionService methods.

We want to track metrics of which permission requests are made with
gestures and their acceptance rates. This change starts putting the
plumbing to do so back in place.

This change is essentially a revert of
https://crrev.com/27b97a28935fa4cb2bbb9e087933629ba49d899b

BUG=614599
Committed: https://crrev.com/891d092c74b16480b11b79723d6b2b98bc748e64
Cr-Commit-Position: refs/heads/master@{#399722}

[benwells, 2016-05-26 04:27:48]

benwells@chromium.org changed reviewers:
+ mlamouri@chromium.org

[benwells, 2016-05-26 04:27:49]

Mounir, as this is kind of reverting your change, wanna do a first pass?

[benwells, 2016-06-02 06:58:05]

On 2016/05/26 04:27:49, benwells wrote:
> Mounir, as this is kind of reverting your change, wanna do a first pass?

friendly ping ...

Mounir, I don't need a full review, just want to know you're ok with the general
idea of this before sending to another reviewer.

[mlamouri (slow - plz ping), 2016-06-02 16:12:09]

Sorry for the delay...

It seems unfortunate to add all this plumbing just to check the acceptance rate
based on the user gesture bit. Could this be done from the renderer to avoid
carrying over the info by any chance?

[benwells, 2016-06-03 04:06:54]

On 2016/06/02 16:12:09, Mounir Lamouri (slow) wrote:
> Sorry for the delay...
> 
> It seems unfortunate to add all this plumbing just to check the acceptance
rate
> based on the user gesture bit. Could this be done from the renderer to avoid
> carrying over the info by any chance?

We want to get data on permission acceptance rates of gestured permissions
versus non-gestured permissions. I don't think we can do that from the blink
layer, as we need to know exactly when the prompts are shown, vs. when a content
setting enables it (for example).

Once we have data we can make a decision on whether to limit permission prompts,
or permission use, to gestures. Depending on that decision we may be able to rip
this all out again.

[benwells, 2016-06-09 00:21:11]

On 2016/06/03 04:06:54, benwells wrote:
> On 2016/06/02 16:12:09, Mounir Lamouri (slow) wrote:
> > Sorry for the delay...
> > 
> > It seems unfortunate to add all this plumbing just to check the acceptance
> rate
> > based on the user gesture bit. Could this be done from the renderer to avoid
> > carrying over the info by any chance?
> 
> We want to get data on permission acceptance rates of gestured permissions
> versus non-gestured permissions. I don't think we can do that from the blink
> layer, as we need to know exactly when the prompts are shown, vs. when a
content
> setting enables it (for example).
> 
> Once we have data we can make a decision on whether to limit permission
prompts,
> or permission use, to gestures. Depending on that decision we may be able to
rip
> this all out again.

friendly ping - not sure if silence means you disagree, agree, are too busy, or
something else... :)

[mlamouri (slow - plz ping), 2016-06-10 11:00:11]

Don't assign to malice what could be caused by incompetence :) I'm just very
busy with other things and I have been avoiding non-{urgent,essential,trivial}
reviews.

I don't think it will be useful to do these checks and it's a lot of code
clutter but I'm happy to be proven wrong so I will happily lgtm this CL :)

[benwells, 2016-06-11 09:43:53]

On 2016/06/10 11:00:11, Mounir Lamouri (slow) wrote:
> Don't assign to malice what could be caused by incompetence :) I'm just very
> busy with other things and I have been avoiding non-{urgent,essential,trivial}
> reviews.
> 
> I don't think it will be useful to do these checks and it's a lot of code
> clutter but I'm happy to be proven wrong so I will happily lgtm this CL :)

Great, that's what this is about - getting data to hopefully prove the gesture
gating idea right or wrong :)

Thanks for getting back to me.

[benwells, 2016-06-11 09:45:24]

benwells@chromium.org changed reviewers:
+ dcheng@chromium.org, jochen@chromium.org

[benwells, 2016-06-11 09:45:24]

+jochen for full review
+dcheng for mojom

[jochen (gone - plz use gerrit), 2016-06-13 15:21:44]

lgtm

[dcheng, 2016-06-13 17:32:32]

dcheng@chromium.org changed reviewers:
+ kenrb@chromium.org

[dcheng, 2016-06-13 17:32:33]

+kenrb to review, since he was also looking at some of the user gesture stuff in
the context of OOPI.

[kenrb, 2016-06-13 20:26:52]

ipc lgtm

[benwells, 2016-06-14 15:06:25]

The CQ bit was checked by benwells@chromium.org

[commit-bot: I haz the power, 2016-06-14 15:06:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2009773002/1

[commit-bot: I haz the power, 2016-06-14 16:26:27]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-06-14 16:26:29]

CQ bit was unchecked

[commit-bot: I haz the power, 2016-06-14 16:29:45]

Description was changed from

==========
Add user_gesture bit back to Mojo PermissionService methods.

We want to track metrics of which permission requests are made with
gestures and their acceptance rates. This change starts putting the
plumbing to do so back in place.

This change is essentially a revert of
https://crrev.com/27b97a28935fa4cb2bbb9e087933629ba49d899b

BUG=614599
==========

to

==========
Add user_gesture bit back to Mojo PermissionService methods.

We want to track metrics of which permission requests are made with
gestures and their acceptance rates. This change starts putting the
plumbing to do so back in place.

This change is essentially a revert of
https://crrev.com/27b97a28935fa4cb2bbb9e087933629ba49d899b

BUG=614599

Committed: https://crrev.com/891d092c74b16480b11b79723d6b2b98bc748e64
Cr-Commit-Position: refs/heads/master@{#399722}
==========

[commit-bot: I haz the power, 2016-06-14 16:29:47]

Patchset 1 (id:??) landed as
https://crrev.com/891d092c74b16480b11b79723d6b2b98bc748e64
Cr-Commit-Position: refs/heads/master@{#399722}