Expose requireAsync method and make it more robust

Add mojoPrivate.requireAsync method to chrome IDL.
Add unit tests for requireAsync.

R=mfoltz@chromium.org,sammc@chromium.org
CC=wez@chromium.org,haibinlu@chromium.org
BUG=

Committed: https://crrev.com/151e70d6fbf2c524959bb225d86a557204571105
Cr-Commit-Position: refs/heads/master@{#313403}

[Kevin M, 2015-01-21 20:51:28]

kmarshall@chromium.org changed reviewers:
+ mfoltz@chromium.org, sammc@chromium.org

[Sam McNally, 2015-01-22 07:42:11]

https://codereview.chromium.org/862793003/diff/20001/extensions/renderer/modu...
File extensions/renderer/module_system.cc (right):

https://codereview.chromium.org/862793003/diff/20001/extensions/renderer/modu...
extensions/renderer/module_system.cc:551: if ((source.IsEmpty() ||
source->IsUndefined()) &&
I'm not a fan of this. It introduces ordering dependencies:
define('foo', ...);
requireAsync('foo');
will succeed, but
requireAsync('foo');
define('foo', ...);
will fail.

[Kevin M, 2015-01-22 18:23:15]

The motivation behind this CL was to allow extensions to do feature testing and
error handling in the event that the Mojo service dependencies aren't met.

I understand that isn't a problem when all the code that uses Mojo is contained
within the same browser build, but as extensions can deployed and updated
independently, versioning issues might arise.

What would you think about a "mojoPrivate.isDefined()" function that's
restricted to the Mojo services registered in service_registration_manager?

[Kevin M, 2015-01-26 22:02:25]

OK, I read through the Mojo versioning proposal and think that it might be a
better solution to the problem I'm framing. I removed the module missing error
handling from requireAsync. Please take a look. Thanks!

[Sam McNally, 2015-01-27 04:06:29]

LGTM with a couple test nits.

https://codereview.chromium.org/862793003/diff/60001/extensions/test/data/moj...
File extensions/test/data/mojo_private_unittest.js (right):

https://codereview.chromium.org/862793003/diff/60001/extensions/test/data/moj...
extensions/test/data/mojo_private_unittest.js:47:
mojoPrivate.requireAsync('dependency').then(
Does this need to be inside a define()ed module?

https://codereview.chromium.org/862793003/diff/60001/extensions/test/data/moj...
extensions/test/data/mojo_private_unittest.js:48: test.succeed,
test.callbackPass()

[Kevin M, 2015-01-27 18:07:57]

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

[Kevin M, 2015-01-27 18:10:09]

+rockot for:

  extensions/common/api/mojo_private.idl
  extensions/renderer/api/mojo_private/mojo_private_unittest.cc
  extensions/renderer/resources/mojo_private_custom_bindings.js


Thanks!

https://codereview.chromium.org/862793003/diff/60001/extensions/test/data/moj...
File extensions/test/data/mojo_private_unittest.js (right):

https://codereview.chromium.org/862793003/diff/60001/extensions/test/data/moj...
extensions/test/data/mojo_private_unittest.js:47:
mojoPrivate.requireAsync('dependency').then(
On 2015/01/27 04:06:28, Sam McNally wrote:
> Does this need to be inside a define()ed module?

Done.

https://codereview.chromium.org/862793003/diff/60001/extensions/test/data/moj...
extensions/test/data/mojo_private_unittest.js:48: test.succeed,
On 2015/01/27 04:06:28, Sam McNally wrote:
> test.callbackPass()

Done.

[mark a. foltz, 2015-01-27 21:58:54]

Superficial review, but lgtm

https://codereview.chromium.org/862793003/diff/80001/extensions/test/data/moj...
File extensions/test/data/mojo_private_unittest.js (right):

https://codereview.chromium.org/862793003/diff/80001/extensions/test/data/moj...
extensions/test/data/mojo_private_unittest.js:36: console.log('requireAsync');
Is this necessary?

[Ken Rockot(use gerrit already), 2015-01-27 22:03:41]

nothing to add to previous comments, lgtm

[Kevin M, 2015-01-27 22:41:43]

The CQ bit was checked by kmarshall@chromium.org

[commit-bot: I haz the power, 2015-01-27 22:43:40]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/862793003/100001

[commit-bot: I haz the power, 2015-01-27 23:42:38]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2015-01-27 23:43:48]

Patchset 6 (id:??) landed as
https://crrev.com/151e70d6fbf2c524959bb225d86a557204571105
Cr-Commit-Position: refs/heads/master@{#313403}