Introduce the PushSubscription.curve25519dh attribute (3/3)

Introduce the PushSubscription.curve25519dh attribute (3/3)

This CL introduces the actual curve25519dh attribute (guarded behind the
PushMessagingData runtime flag), exposing the ECDH encryption key
associated with the subscription. This will still be an empty buffer
for non-layout test usages.

This CL is part of a three-sided patch:
  [1] https://codereview.chromium.org/1222903007
  [2] https://codereview.chromium.org/1210283005
  [3] This CL.

BUG=486040
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=198949

[Peter Beverloo, 2015-07-13 13:57:25]

peter@chromium.org changed reviewers:
+ mvanouwerkerk@chromium.org

[Peter Beverloo, 2015-07-13 13:57:26]

+Michaël

[johnme, 2015-07-14 16:49:50]

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

[johnme, 2015-07-14 16:49:50]

https://codereview.chromium.org/1237673002/diff/1/LayoutTests/http/tests/push...
File LayoutTests/http/tests/push_messaging/subscribe-encryption-public-key.html
(right):

https://codereview.chromium.org/1237673002/diff/1/LayoutTests/http/tests/push...
LayoutTests/http/tests/push_messaging/subscribe-encryption-public-key.html:46:
assert_equals(pushSubscription.curve25519dh.byteLength,
Seems you might as well add whichever of assert_equals(encryptionKey,
pushSubscription.curve25519dh) or assert_not_equals(encryptionKey,
pushSubscription.curve25519dh) is supposed to be true.

https://codereview.chromium.org/1237673002/diff/1/LayoutTests/http/tests/push...
LayoutTests/http/tests/push_messaging/subscribe-encryption-public-key.html:49:
return service_worker_unregister_and_done(test, workerScope);
Please also test that after unsubscribing and re-subscribing the curve25519dh is
different.

https://codereview.chromium.org/1237673002/diff/1/public/platform/modules/pus...
File public/platform/modules/push_messaging/WebPushSubscription.h (right):

https://codereview.chromium.org/1237673002/diff/1/public/platform/modules/pus...
public/platform/modules/push_messaging/WebPushSubscription.h:15: // The
|endpoint| must be unique for each subscription.
Nit: "The |endpoint| and |curve25519dh| must both be unique for each
subscription."

[Peter Beverloo, 2015-07-14 17:06:06]

Thanks!

https://codereview.chromium.org/1237673002/diff/1/LayoutTests/http/tests/push...
File LayoutTests/http/tests/push_messaging/subscribe-encryption-public-key.html
(right):

https://codereview.chromium.org/1237673002/diff/1/LayoutTests/http/tests/push...
LayoutTests/http/tests/push_messaging/subscribe-encryption-public-key.html:46:
assert_equals(pushSubscription.curve25519dh.byteLength,
On 2015/07/14 16:49:50, johnme wrote:
> Seems you might as well add whichever of assert_equals(encryptionKey,
> pushSubscription.curve25519dh) or assert_not_equals(encryptionKey,
> pushSubscription.curve25519dh) is supposed to be true.

The contents of the ArrayBuffer will be identical, but they will be represented
by different objects. I've serialized the keys to base64 so that we can do a
string comparison between subscribe() and getSubscription().

https://codereview.chromium.org/1237673002/diff/1/LayoutTests/http/tests/push...
LayoutTests/http/tests/push_messaging/subscribe-encryption-public-key.html:49:
return service_worker_unregister_and_done(test, workerScope);
On 2015/07/14 16:49:50, johnme wrote:
> Please also test that after unsubscribing and re-subscribing the curve25519dh
is
> different.

This is the value returned by the LayoutTestPushMessagingService. The goal of
this test is to cover the plumbing, not generation of the keys (which is a
concern shared between //chrome and //components/gcm_driver). Additional tests
there will cover your concern :).

https://codereview.chromium.org/1237673002/diff/1/public/platform/modules/pus...
File public/platform/modules/push_messaging/WebPushSubscription.h (right):

https://codereview.chromium.org/1237673002/diff/1/public/platform/modules/pus...
public/platform/modules/push_messaging/WebPushSubscription.h:15: // The
|endpoint| must be unique for each subscription.
On 2015/07/14 16:49:50, johnme wrote:
> Nit: "The |endpoint| and |curve25519dh| must both be unique for each
> subscription."

Done.

[johnme, 2015-07-14 17:45:14]

lgtm

[Peter Beverloo, 2015-07-15 09:58:35]

The CQ bit was checked by peter@chromium.org

[Peter Beverloo, 2015-07-15 09:58:36]

The patchset sent to the CQ was uploaded after l-g-t-m from johnme@chromium.org
Link to the patchset: https://codereview.chromium.org/1237673002/#ps40001
(title: " ")

[commit-bot: I haz the power, 2015-07-15 09:58:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1237673002/40001

[commit-bot: I haz the power, 2015-07-15 11:11:07]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=198949