Introduce the PushSubscription.curve25519dh attribute (1/3)

Introduce the PushSubscription.curve25519dh attribute (1/3)

This CL changes the WebPushSubscription object in the Blink API to accept
a new curve25519dh parameter in its constructor, and carry the member.

Mozilla's straw man has been used as a base for the behavior, with the
difference that we are experimenting with using Curve25519 rather than
P-256 as the ECDH curve. This is being discussed over at the IETF.

https://github.com/w3c/push-api/pull/130

The following Intent to Implement covers the newly exposed API:

https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/_LQX_cjAATA

This CL is part of a three-sided patch:
  [1] This CL.
  [2] https://codereview.chromium.org/1210283005
  [3] https://codereview.chromium.org/1237673002

BUG=486040
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=198784

[Peter Beverloo, 2015-07-08 20:19:37]

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

[Peter Beverloo, 2015-07-08 20:19:38]

+johnme

[johnme, 2015-07-09 10:48:13]

Code lgtm. Is there a spec for how the encryption keys will be exposed in the
Push API? Couldn't find any mention in the spec or GitHub issues.

[johnme, 2015-07-09 11:00:15]

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
File public/platform/modules/push_messaging/WebPushSubscription.h (right):

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
public/platform/modules/push_messaging/WebPushSubscription.h:29:
WebVector<unsigned char> curve25519dh;
Actually, this name is a little cryptic. Presumably it's the (secret) public
encryption key with which sent messages should be encrypted.

[Peter Beverloo, 2015-07-09 14:28:34]

This is based on Mozilla's straw man, with s/p-256/curve25519/g applied.

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
File public/platform/modules/push_messaging/WebPushSubscription.h (right):

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
public/platform/modules/push_messaging/WebPushSubscription.h:29:
WebVector<unsigned char> curve25519dh;
On 2015/07/09 11:00:15, johnme wrote:
> Actually, this name is a little cryptic. Presumably it's the (secret) public
> encryption key with which sent messages should be encrypted.

I clarified the CLs description.

[Peter Beverloo, 2015-07-09 19:17:45]

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

[Peter Beverloo, 2015-07-09 19:17:45]

+mkwst for WebPushSubscription.cpp removal.

[Mike West, 2015-07-10 08:06:55]

LGTM.

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
File public/platform/modules/push_messaging/WebPushSubscription.h (right):

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
public/platform/modules/push_messaging/WebPushSubscription.h:16:
WebPushSubscription(const WebURL& endpoint, const WebVector<unsigned char>&
curve25519dh)
Might be worth adding a comment that |endpoint| must contain a subscription id?

[Peter Beverloo, 2015-07-13 13:58:14]

Thanks!

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
File public/platform/modules/push_messaging/WebPushSubscription.h (right):

https://codereview.chromium.org/1222903007/diff/1/public/platform/modules/pus...
public/platform/modules/push_messaging/WebPushSubscription.h:16:
WebPushSubscription(const WebURL& endpoint, const WebVector<unsigned char>&
curve25519dh)
On 2015/07/10 08:06:55, Mike West wrote:
> Might be worth adding a comment that |endpoint| must contain a subscription
id?

I've mentioned that it must be unique for each subscription. (Theoretically a
push service could work on something other than a subscription id.)

[Peter Beverloo, 2015-07-13 13:58:19]

The CQ bit was checked by peter@chromium.org

[Peter Beverloo, 2015-07-13 13:58:19]

The patchset sent to the CQ was uploaded after l-g-t-m from johnme@chromium.org,
mkwst@chromium.org
Link to the patchset: https://codereview.chromium.org/1222903007/#ps20001
(title: " ")

[commit-bot: I haz the power, 2015-07-13 13:58:58]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1222903007/20001

[commit-bot: I haz the power, 2015-07-13 15:16:19]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=198784