Update mojo sdk to rev 8d45c89c30b230843c5bd6dd0693a555750946c0

Update mojo sdk to rev 8d45c89c30b230843c5bd6dd0693a555750946c0

Adapt to de-clienting of Surface.

Have nacl_listener pass a SimplePlatformSupport instance to mojo::embedder::Init
to satisfy the DCHECK introduced in
https://codereview.chromium.org/898623002/diff/40001/mojo/edk/embedder/embedder.cc.

[blundell, 2015-02-05 16:33:35]

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

[blundell, 2015-02-05 16:33:37]

Hi Trung,

Not sure if you (or someone else) was looking to do a roll. This one's almost
ready; it looks like there's just one more problem having to do with changes to
CreateChannel
(http://build.chromium.org/p/tryserver.chromium.win/builders/win8_chromium_rel...).
If anyone was looking to do a roll today, feel free to take over this CL.
Otherwise, I'll finish it up tomorrow.

[blundell, 2015-02-05 16:35:14]

rockot: fyi.

[Ken Rockot(use gerrit already), 2015-02-05 16:38:13]

On 2015/02/05 16:35:14, blundell wrote:
> rockot: fyi.

Ah thanks for starting this. I was planning to do a roll today. I'll look at
CreateChannel stuff.

[viettrungluu, 2015-02-05 18:33:05]

On 2015/02/05 16:38:13, Ken Rockot wrote:
> On 2015/02/05 16:35:14, blundell wrote:
> > rockot: fyi.
> 
> Ah thanks for starting this. I was planning to do a roll today. I'll look at
> CreateChannel stuff.

Let me know if you figure anything out (unfortunately, the stack trace isn't
very enlightening, and I currently don't have a working Windows machine).

Note that CreateChannel() has since changed some more.

[Ken Rockot(use gerrit already), 2015-02-06 00:46:19]

On 2015/02/05 18:33:05, viettrungluu wrote:
> On 2015/02/05 16:38:13, Ken Rockot wrote:
> > On 2015/02/05 16:35:14, blundell wrote:
> > > rockot: fyi.
> > 
> > Ah thanks for starting this. I was planning to do a roll today. I'll look at
> > CreateChannel stuff.
> 
> Let me know if you figure anything out (unfortunately, the stack trace isn't
> very enlightening, and I currently don't have a working Windows machine).
> 
> Note that CreateChannel() has since changed some more.

I just got a working Windows environment set up again, so hopefully I'll have
more information soon.

[Ken Rockot(use gerrit already), 2015-02-06 04:17:59]

On 2015/02/06 00:46:19, Ken Rockot wrote:
> On 2015/02/05 18:33:05, viettrungluu wrote:
> > On 2015/02/05 16:38:13, Ken Rockot wrote:
> > > On 2015/02/05 16:35:14, blundell wrote:
> > > > rockot: fyi.
> > > 
> > > Ah thanks for starting this. I was planning to do a roll today. I'll look
at
> > > CreateChannel stuff.
> > 
> > Let me know if you figure anything out (unfortunately, the stack trace isn't
> > very enlightening, and I currently don't have a working Windows machine).
> > 
> > Note that CreateChannel() has since changed some more.
> 
> I just got a working Windows environment set up again, so hopefully I'll have
> more information soon.

Well, nevermind. I need to set up a win8 machine now. Calling it a day. :|

[blundell, 2015-02-06 13:37:07]

On 2015/02/06 04:17:59, Ken Rockot wrote:
> On 2015/02/06 00:46:19, Ken Rockot wrote:
> > On 2015/02/05 18:33:05, viettrungluu wrote:
> > > On 2015/02/05 16:38:13, Ken Rockot wrote:
> > > > On 2015/02/05 16:35:14, blundell wrote:
> > > > > rockot: fyi.
> > > > 
> > > > Ah thanks for starting this. I was planning to do a roll today. I'll
look
> at
> > > > CreateChannel stuff.
> > > 
> > > Let me know if you figure anything out (unfortunately, the stack trace
isn't
> > > very enlightening, and I currently don't have a working Windows machine).
> > > 
> > > Note that CreateChannel() has since changed some more.
> > 
> > I just got a working Windows environment set up again, so hopefully I'll
have
> > more information soon.
> 
> Well, nevermind. I need to set up a win8 machine now. Calling it a day. :|

Hi,

The culprit is https://codereview.chromium.org/899993002/: I reverted that CL in
PS4 and all the trybots are green. Given the backtrace and the delta in that CL,
my guess is that the call to MakeChannelId() at line 122 of
https://codereview.chromium.org/899993002/diff/1/mojo/edk/embedder/embedder.cc
is causing the CHECK at line 43 to fire. I have no real explanation for why that
would happen though, so maybe I'm wrong. I don't have a Windows machine to do
any local testing on.

[viettrungluu, 2015-02-06 20:58:57]

On 2015/02/06 13:37:07, blundell wrote:
> On 2015/02/06 04:17:59, Ken Rockot wrote:
> > On 2015/02/06 00:46:19, Ken Rockot wrote:
> > > On 2015/02/05 18:33:05, viettrungluu wrote:
> > > > On 2015/02/05 16:38:13, Ken Rockot wrote:
> > > > > On 2015/02/05 16:35:14, blundell wrote:
> > > > > > rockot: fyi.
> > > > > 
> > > > > Ah thanks for starting this. I was planning to do a roll today. I'll
> look
> > at
> > > > > CreateChannel stuff.
> > > > 
> > > > Let me know if you figure anything out (unfortunately, the stack trace
> isn't
> > > > very enlightening, and I currently don't have a working Windows
machine).
> > > > 
> > > > Note that CreateChannel() has since changed some more.
> > > 
> > > I just got a working Windows environment set up again, so hopefully I'll
> have
> > > more information soon.
> > 
> > Well, nevermind. I need to set up a win8 machine now. Calling it a day. :|
> 
> Hi,
> 
> The culprit is https://codereview.chromium.org/899993002/: I reverted that CL
in
> PS4 and all the trybots are green. Given the backtrace and the delta in that
CL,
> my guess is that the call to MakeChannelId() at line 122 of
> https://codereview.chromium.org/899993002/diff/1/mojo/edk/embedder/embedder.cc
> is causing the CHECK at line 43 to fire. I have no real explanation for why
that
> would happen though, so maybe I'm wrong. I don't have a Windows machine to do
> any local testing on.

msw@ helped me out, and I have a fix pending. (The problem is between the
channel_info.release() and channel_info->channel_id, with the order of
evaluation undefined.) Oops!

You can either try rolling to ToT (once I've landed the fix) or backporting the
fix.... (Or maybe I'll take over your change and do so myself.)

[viettrungluu, 2015-02-06 21:02:37]

On 2015/02/06 20:58:57, viettrungluu wrote:
> On 2015/02/06 13:37:07, blundell wrote:
> > On 2015/02/06 04:17:59, Ken Rockot wrote:
> > > On 2015/02/06 00:46:19, Ken Rockot wrote:
> > > > On 2015/02/05 18:33:05, viettrungluu wrote:
> > > > > On 2015/02/05 16:38:13, Ken Rockot wrote:
> > > > > > On 2015/02/05 16:35:14, blundell wrote:
> > > > > > > rockot: fyi.
> > > > > > 
> > > > > > Ah thanks for starting this. I was planning to do a roll today. I'll
> > look
> > > at
> > > > > > CreateChannel stuff.
> > > > > 
> > > > > Let me know if you figure anything out (unfortunately, the stack trace
> > isn't
> > > > > very enlightening, and I currently don't have a working Windows
> machine).
> > > > > 
> > > > > Note that CreateChannel() has since changed some more.
> > > > 
> > > > I just got a working Windows environment set up again, so hopefully I'll
> > have
> > > > more information soon.
> > > 
> > > Well, nevermind. I need to set up a win8 machine now. Calling it a day. :|
> > 
> > Hi,
> > 
> > The culprit is https://codereview.chromium.org/899993002/: I reverted that
CL
> in
> > PS4 and all the trybots are green. Given the backtrace and the delta in that
> CL,
> > my guess is that the call to MakeChannelId() at line 122 of
> >
https://codereview.chromium.org/899993002/diff/1/mojo/edk/embedder/embedder.cc
> > is causing the CHECK at line 43 to fire. I have no real explanation for why
> that
> > would happen though, so maybe I'm wrong. I don't have a Windows machine to
do
> > any local testing on.
> 
> msw@ helped me out, and I have a fix pending. (The problem is between the
> channel_info.release() and channel_info->channel_id, with the order of
> evaluation undefined.) Oops!
> 
> You can either try rolling to ToT (once I've landed the fix) or backporting
the
> fix.... (Or maybe I'll take over your change and do so myself.)

https://codereview.chromium.org/898423002

[viettrungluu, 2015-02-06 21:15:57]

On 2015/02/06 21:02:37, viettrungluu wrote:
> On 2015/02/06 20:58:57, viettrungluu wrote:
> > On 2015/02/06 13:37:07, blundell wrote:
> > > On 2015/02/06 04:17:59, Ken Rockot wrote:
> > > > On 2015/02/06 00:46:19, Ken Rockot wrote:
> > > > > On 2015/02/05 18:33:05, viettrungluu wrote:
> > > > > > On 2015/02/05 16:38:13, Ken Rockot wrote:
> > > > > > > On 2015/02/05 16:35:14, blundell wrote:
> > > > > > > > rockot: fyi.
> > > > > > > 
> > > > > > > Ah thanks for starting this. I was planning to do a roll today.
I'll
> > > look
> > > > at
> > > > > > > CreateChannel stuff.
> > > > > > 
> > > > > > Let me know if you figure anything out (unfortunately, the stack
trace
> > > isn't
> > > > > > very enlightening, and I currently don't have a working Windows
> > machine).
> > > > > > 
> > > > > > Note that CreateChannel() has since changed some more.
> > > > > 
> > > > > I just got a working Windows environment set up again, so hopefully
I'll
> > > have
> > > > > more information soon.
> > > > 
> > > > Well, nevermind. I need to set up a win8 machine now. Calling it a day.
:|
> > > 
> > > Hi,
> > > 
> > > The culprit is https://codereview.chromium.org/899993002/: I reverted that
> CL
> > in
> > > PS4 and all the trybots are green. Given the backtrace and the delta in
that
> > CL,
> > > my guess is that the call to MakeChannelId() at line 122 of
> > >
> https://codereview.chromium.org/899993002/diff/1/mojo/edk/embedder/embedder.cc
> > > is causing the CHECK at line 43 to fire. I have no real explanation for
why
> > that
> > > would happen though, so maybe I'm wrong. I don't have a Windows machine to
> do
> > > any local testing on.
> > 
> > msw@ helped me out, and I have a fix pending. (The problem is between the
> > channel_info.release() and channel_info->channel_id, with the order of
> > evaluation undefined.) Oops!
> > 
> > You can either try rolling to ToT (once I've landed the fix) or backporting
> the
> > fix.... (Or maybe I'll take over your change and do so myself.)
> 
> https://codereview.chromium.org/898423002

OK, taking this over. https://codereview.chromium.org/904103003 since apparently
I can't upload changes to someone else's CL (fair enough, I guess).

[blundell, 2015-02-09 09:35:12]

Thanks for finishing this off, Trung. Closing this out.