Cleanup InterfaceRegistry + adds Serialize()

Cleans up InterfaceRegistry, making name_to_binder_ be a persistent source of truth of the total set of interfaces that could be bound, with binding controlled /only/ at GetInterface() time by exposed_interfaces_, populated at Bind().

This seems less brittle.

Also adds a Serialize() function which serializes the contents of the local & remote specs, as well as the name_to_binder_ map.

Committed: https://crrev.com/cfe99fb237d62af9535ee41c4469062299523fc9
Cr-Commit-Position: refs/heads/master@{#428649}

[Ben Goodger (Google), 2016-10-27 06:57:03]

ben@chromium.org changed reviewers:
+ rockot@chromium.org

[Ben Goodger (Google), 2016-10-27 06:57:04]

Here's one for discussion. Now that the client library does 100% of capability
enforcement (the service manager is now only responsible for being the
authoritative source of static capability data), it seems that InterfaceRegistry
is a candidate for the place to do dynamic capability grants.

Downside of doing it in the client: While this might work for content's use case
today (the code doing the granting lives in-process), in the future "granting" a
capability might involve granting to N InterfaceRegistries in different
services. This doesn't solve that. Having something central (ServiceManager?)
might allow for that. However it's not clear to me that the ServiceManager is in
a place to know who to tell about capability changes since after it brokers a
connection between A&B it no longer tracks that relationship...

[Ben Goodger (Google), 2016-10-28 20:27:27]

The CQ bit was checked by ben@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-28 20:27:49]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Ben Goodger (Google), 2016-10-28 20:32:32]

Description was changed from

==========
Add methods to grant/revoke a capability via a InterfaceRegistry dynamically.

Also adds the method HasCapability() to IR, allowing a service to inspect the
capabilities of the bound remote.
==========

to

==========
Cleans up InterfaceRegistry, making name_to_binder_ be a persistent source of
truth of the total set of interfaces that could be bound, with binding
controlled /only/ at GetInterface() time by exposed_interfaces_, populated at
Bind().

This seems less brittle.

Also adds a Serialize() function which serializes the contents of the local &
remote specs, as well as the name_to_binder_ map.
==========

[Ben Goodger (Google), 2016-10-28 20:37:29]

The CQ bit was checked by ben@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-28 20:38:03]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Ben Goodger (Google), 2016-10-28 20:52:23]

OK I removed the capability granting/revoking and pivoted this to a cleanup
change. Now ready for review.

[Ben Goodger (Google), 2016-10-28 20:53:31]

The CQ bit was checked by ben@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-28 20:53:55]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-28 22:00:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-10-28 22:00:07]

Dry run: This issue passed the CQ dry run.

[Ken Rockot(use gerrit already), 2016-10-28 23:52:03]

lgtm

[Ben Goodger (Google), 2016-10-31 02:30:27]

The CQ bit was checked by ben@chromium.org

[commit-bot: I haz the power, 2016-10-31 02:30:36]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-31 03:29:27]

Description was changed from

==========
Cleans up InterfaceRegistry, making name_to_binder_ be a persistent source of
truth of the total set of interfaces that could be bound, with binding
controlled /only/ at GetInterface() time by exposed_interfaces_, populated at
Bind().

This seems less brittle.

Also adds a Serialize() function which serializes the contents of the local &
remote specs, as well as the name_to_binder_ map.
==========

to

==========
Cleans up InterfaceRegistry, making name_to_binder_ be a persistent source of
truth of the total set of interfaces that could be bound, with binding
controlled /only/ at GetInterface() time by exposed_interfaces_, populated at
Bind().

This seems less brittle.

Also adds a Serialize() function which serializes the contents of the local &
remote specs, as well as the name_to_binder_ map.
==========

[commit-bot: I haz the power, 2016-10-31 03:29:28]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-10-31 03:31:17]

Description was changed from

==========
Cleans up InterfaceRegistry, making name_to_binder_ be a persistent source of
truth of the total set of interfaces that could be bound, with binding
controlled /only/ at GetInterface() time by exposed_interfaces_, populated at
Bind().

This seems less brittle.

Also adds a Serialize() function which serializes the contents of the local &
remote specs, as well as the name_to_binder_ map.
==========

to

==========
Cleans up InterfaceRegistry, making name_to_binder_ be a persistent source of
truth of the total set of interfaces that could be bound, with binding
controlled /only/ at GetInterface() time by exposed_interfaces_, populated at
Bind().

This seems less brittle.

Also adds a Serialize() function which serializes the contents of the local &
remote specs, as well as the name_to_binder_ map.

Committed: https://crrev.com/cfe99fb237d62af9535ee41c4469062299523fc9
Cr-Commit-Position: refs/heads/master@{#428649}
==========

[commit-bot: I haz the power, 2016-10-31 03:31:18]

Patchset 6 (id:??) landed as
https://crrev.com/cfe99fb237d62af9535ee41c4469062299523fc9
Cr-Commit-Position: refs/heads/master@{#428649}