Update mojo sdk to rev 3d23dae011859a2aae49f1d1adde705c8e85d819

Update mojo sdk to rev 3d23dae011859a2aae49f1d1adde705c8e85d819

Highlights:
 - mojo::ChannelInit has been replaced with
   content::ChannelInit.
 - ScopedIPCSupport has been added so Mojo consumers can
   ensure the EDK is initialized.
 - single process mode now uses some evil tricks to get
   child threads to create their mojo client channels on
   the browser I/O thread.
 - several Android bits adapted to new interfaces
 - a number of tests have been adapted to work properly
   in spite of unconventional process arrangements

BUG=None

Committed: https://crrev.com/cbca72f3c9e31309a28b98624e1e0147c5dca7a1
Cr-Commit-Position: refs/heads/master@{#318883}

[Ken Rockot(use gerrit already), 2015-02-24 23:24:22]

Patchset #6 (id:100001) has been deleted

[Ken Rockot(use gerrit already), 2015-02-24 23:24:41]

Patchset #6 (id:100001) has been deleted

[Ken Rockot(use gerrit already), 2015-02-28 00:57:16]

rockot@chromium.org changed reviewers:
+ jamesr@chromium.org, morrita@chromium.org

[Ken Rockot(use gerrit already), 2015-02-28 00:57:19]

Alright I think this is ready to go.

James, I had to hack out a VS warning in data_pipe.h. I can upstream this later
if it lgty but maybe there's a better option?

Morrita, please review changes to IPC code. Due to some EDK changes a lot of
tests were broken because it's no longer possible simultaneously host multiple
channels on different threads in the same process. Also added explicit channel
closing to IPCChannelMojoTest clients to ensure they clean up on the right task
runner.

[jamesr, 2015-02-28 01:22:54]

jamesr@chromium.org changed reviewers:
+ viettrungluu@chromium.org

[jamesr, 2015-02-28 01:22:56]

Trung or Morrita would probably know better about that alignment thing.

[viettrungluu, 2015-02-28 02:23:20]

https://codereview.chromium.org/954643002/diff/300001/third_party/mojo/src/mo...
File third_party/mojo/src/mojo/edk/system/data_pipe.h (right):

https://codereview.chromium.org/954643002/diff/300001/third_party/mojo/src/mo...
third_party/mojo/src/mojo/edk/system/data_pipe.h:32: #pragma warning(disable :
4324)
I believe you only need this for one or two declarations.

You should probably use MSVC_SUPPRESS_WARNING() from base/compiler_specific.h.

I'm also not sure why you're applying this to this file. It's not clear to me
what needs it, or would trigger that warning. (On ToT Mojo, I'd expect it for a
couple for a struct or two in data_pipe_impl.h, but they don't exist in this
revision AFAICT.)

[viettrungluu, 2015-02-28 02:26:05]

On 2015/02/28 02:23:20, viettrungluu wrote:
>
https://codereview.chromium.org/954643002/diff/300001/third_party/mojo/src/mo...
> File third_party/mojo/src/mojo/edk/system/data_pipe.h (right):
> 
>
https://codereview.chromium.org/954643002/diff/300001/third_party/mojo/src/mo...
> third_party/mojo/src/mojo/edk/system/data_pipe.h:32: #pragma warning(disable :
> 4324)
> I believe you only need this for one or two declarations.
> 
> You should probably use MSVC_SUPPRESS_WARNING() from base/compiler_specific.h.
> 
> I'm also not sure why you're applying this to this file. It's not clear to me
> what needs it, or would trigger that warning. (On ToT Mojo, I'd expect it for
a
> couple for a struct or two in data_pipe_impl.h, but they don't exist in this
> revision AFAICT.)

Now looking at it more carefully, it makes more sense (an aligned member was
added to DataPipe). Using MSVS_SUPPRESS_WARNING() is probably still the right
thing to do.

[qsr, 2015-03-02 09:09:48]

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

[qsr, 2015-03-02 09:09:52]

https://codereview.chromium.org/954643002/diff/340001/mojo/android/system/src...
File mojo/android/system/src/org/chromium/mojo/system/impl/CoreImpl.java
(right):

https://codereview.chromium.org/954643002/diff/340001/mojo/android/system/src...
mojo/android/system/src/org/chromium/mojo/system/impl/CoreImpl.java:213: public
RunLoop createDefaultRunLoop() {
I a missing something. Why don't you get the CoreImpl (and also other files)
implementation from the mojo repo instead of forking those files?

[Ken Rockot(use gerrit already), 2015-03-02 17:34:30]

On 2015/02/28 02:23:20, viettrungluu wrote:
> 
> Now looking at it more carefully, it makes more sense (an aligned member was
> added to DataPipe). Using MSVS_SUPPRESS_WARNING() is probably still the right
> thing to do.

Done. I added an MSVC_SUPPRESS_WARNING to validate_options_ only.

On a related note, I also had to apply NON_EXPORTED_BASE to ConnectionManager
base declarations for MasterConnectionManager and SlaveConnectionManager,
assuming you don't want to export ConnectionManager.

On 2015/03/02 09:09:52, qsr wrote:
>
https://codereview.chromium.org/954643002/diff/340001/mojo/android/system/src...
> File mojo/android/system/src/org/chromium/mojo/system/impl/CoreImpl.java
> (right):
> 
>
https://codereview.chromium.org/954643002/diff/340001/mojo/android/system/src...
> mojo/android/system/src/org/chromium/mojo/system/impl/CoreImpl.java:213:
public
> RunLoop createDefaultRunLoop() {
> I a missing something. Why don't you get the CoreImpl (and also other files)
> implementation from the mojo repo instead of forking those files?

I know of no good reason not to do this, it just hasn't been done yet. The
mojo->chrome roll is minimal. We should discuss this further before the next
roll.

[viettrungluu, 2015-03-02 17:47:58]

On 2015/03/02 17:34:30, Ken Rockot wrote:
> On 2015/02/28 02:23:20, viettrungluu wrote:
> > 
> > Now looking at it more carefully, it makes more sense (an aligned member was
> > added to DataPipe). Using MSVS_SUPPRESS_WARNING() is probably still the
right
> > thing to do.
> 
> Done. I added an MSVC_SUPPRESS_WARNING to validate_options_ only.

Thanks.

> 
> On a related note, I also had to apply NON_EXPORTED_BASE to ConnectionManager
> base declarations for MasterConnectionManager and SlaveConnectionManager,
> assuming you don't want to export ConnectionManager.

Hmmm, I don't really want to export very much (I don't really want to export
{Master,Slave}ConnectionManager either, except that I think I need to for tests
to build in component builds), but it might be simpler/more consistent if we
exported ConnectionManager too.

> 
> On 2015/03/02 09:09:52, qsr wrote:
> >
>
https://codereview.chromium.org/954643002/diff/340001/mojo/android/system/src...
> > File mojo/android/system/src/org/chromium/mojo/system/impl/CoreImpl.java
> > (right):
> > 
> >
>
https://codereview.chromium.org/954643002/diff/340001/mojo/android/system/src...
> > mojo/android/system/src/org/chromium/mojo/system/impl/CoreImpl.java:213:
> public
> > RunLoop createDefaultRunLoop() {
> > I a missing something. Why don't you get the CoreImpl (and also other files)
> > implementation from the mojo repo instead of forking those files?
> 
> I know of no good reason not to do this, it just hasn't been done yet. The
> mojo->chrome roll is minimal. We should discuss this further before the next
> roll.

[Ken Rockot(use gerrit already), 2015-03-02 17:53:35]

On 2015/03/02 17:47:58, viettrungluu wrote:
> Hmmm, I don't really want to export very much (I don't really want to export
> {Master,Slave}ConnectionManager either, except that I think I need to for
tests
> to build in component builds), but it might be simpler/more consistent if we
> exported ConnectionManager too.
> 

OK, SGTM and done.

[Ken Rockot(use gerrit already), 2015-03-02 20:18:23]

Patchset #18 (id:360001) has been deleted

[Ken Rockot(use gerrit already), 2015-03-02 20:34:17]

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

[Ken Rockot(use gerrit already), 2015-03-02 20:36:24]

also adding davidben@ for changes to crypto gyp(i); nacl win64 builds now need
crypto::RandomBytes.

[Hajime Morrita, 2015-03-02 23:19:52]

lgtm for ipc/mojo.
It'd be nice to have some explanation why we need ScopedIPCSupport.
At the first glance, it looks a mysterious entity.

[Hajime Morrita, 2015-03-02 23:27:26]

(still lgtm)

https://codereview.chromium.org/954643002/diff/440001/content/browser/mojo/mo...
File content/browser/mojo/mojo_application_host.cc (right):

https://codereview.chromium.org/954643002/diff/440001/content/browser/mojo/mo...
content/browser/mojo/mojo_application_host.cc:66:
scoped_refptr<base::TaskRunner> io_task_runner;
We could possibly pass the task runner to the ctor of this class.
We can make that change later though. Let's land this first.

[viettrungluu, 2015-03-02 23:27:41]

On 2015/03/02 17:53:35, Ken Rockot wrote:
> On 2015/03/02 17:47:58, viettrungluu wrote:
> > Hmmm, I don't really want to export very much (I don't really want to export
> > {Master,Slave}ConnectionManager either, except that I think I need to for
> tests
> > to build in component builds), but it might be simpler/more consistent if we
> > exported ConnectionManager too.
> > 
> 
> OK, SGTM and done.

Thanks. LGTM for the changes to the edk.

Could you please send me an upstream change?

[rvargas (doing something else), 2015-03-03 01:47:25]

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

[rvargas (doing something else), 2015-03-03 01:47:30]

crypto lgtm

[Ken Rockot(use gerrit already), 2015-03-03 02:37:29]

rockot@chromium.org changed reviewers:
+ jam@chromium.org
- davidben@chromium.org

[Ken Rockot(use gerrit already), 2015-03-03 02:37:32]

John, would you mind taking a look at content? Thanks!

[jam, 2015-03-03 04:32:22]

lgtm

https://codereview.chromium.org/954643002/diff/460001/content/browser/rendere...
File content/browser/renderer_host/render_process_host_impl.cc (right):

https://codereview.chromium.org/954643002/diff/460001/content/browser/rendere...
content/browser/renderer_host/render_process_host_impl.cc:675: if
(base::CommandLine::ForCurrentProcess()->HasSwitch(
use run_renderer_in_process()

https://codereview.chromium.org/954643002/diff/460001/content/common/mojo/cha...
File content/common/mojo/channel_init.h (right):

https://codereview.chromium.org/954643002/diff/460001/content/common/mojo/cha...
content/common/mojo/channel_init.h:25: class CONTENT_EXPORT ChannelInit {
why export?

[Ken Rockot(use gerrit already), 2015-03-03 06:13:31]

https://codereview.chromium.org/954643002/diff/460001/content/browser/rendere...
File content/browser/renderer_host/render_process_host_impl.cc (right):

https://codereview.chromium.org/954643002/diff/460001/content/browser/rendere...
content/browser/renderer_host/render_process_host_impl.cc:675: if
(base::CommandLine::ForCurrentProcess()->HasSwitch(
On 2015/03/03 04:32:22, jam wrote:
> use run_renderer_in_process()

Done.

https://codereview.chromium.org/954643002/diff/460001/content/common/mojo/cha...
File content/common/mojo/channel_init.h (right):

https://codereview.chromium.org/954643002/diff/460001/content/common/mojo/cha...
content/common/mojo/channel_init.h:25: class CONTENT_EXPORT ChannelInit {
On 2015/03/03 04:32:22, jam wrote:
> why export?

needed to link browser tests, unfortunately

[Ken Rockot(use gerrit already), 2015-03-03 07:01:38]

The CQ bit was checked by rockot@chromium.org

[Ken Rockot(use gerrit already), 2015-03-03 07:01:40]

The patchset sent to the CQ was uploaded after l-g-t-m from
morrita@chromium.org, viettrungluu@chromium.org, jam@chromium.org,
rvargas@chromium.org
Link to the patchset: https://codereview.chromium.org/954643002/#ps480001
(title: "use run_renderer_in_process()")

[commit-bot: I haz the power, 2015-03-03 07:04:21]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/954643002/480001

[commit-bot: I haz the power, 2015-03-03 14:36:09]

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

[commit-bot: I haz the power, 2015-03-03 14:36:12]

Try jobs failed on following builders:
  android_rel_tests_recipe on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/android_rel_tes...)

[tonyg, 2015-03-03 15:48:53]

The CQ bit was checked by tonyg@chromium.org

[commit-bot: I haz the power, 2015-03-03 15:49:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/954643002/480001

[commit-bot: I haz the power, 2015-03-03 16:31:25]

Committed patchset #23 (id:480001)

[commit-bot: I haz the power, 2015-03-03 16:31:59]

Patchset 23 (id:??) landed as
https://crrev.com/cbca72f3c9e31309a28b98624e1e0147c5dca7a1
Cr-Commit-Position: refs/heads/master@{#318883}