Missing files from previous CL with fixes

Missing files from previous CL with fixes

File adds got missed when landing https://codereview.appspot.com/12699047/, readded here.

There was a crash when embedder's pointers for tracing weren't
set, that's addressed here.

The import/export is changed to not rely on __declspec(dllimport)
and __declspec(dllexport), the setup function is extern "C" and 
included in the .def file so that the main project can 
GetProcAddress and not have a hard dependency on libGLESv2.dll.

third_party/trace_event is the same as previous CL.

R=apatrick@chromium.google.com,shannonwoods@google.com

[shannonwoods1, 2013-08-13 14:50:40]

On 2013/08/12 21:43:48, scottmg wrote:

Fix looks good to me. I can land this for you.

[scottmg, 2013-08-13 16:55:21]

OK, updated to include files missed in last landing and to remove the dll
futzing in favour of a function that can be GetProcAddress'd in chrome for
https://codereview.chromium.org/22744005

[scottmg, 2013-08-13 19:14:00]

On 2013/08/13 16:55:21, scottmg wrote:
> OK, updated to include files missed in last landing and to remove the dll
> futzing in favour of a function that can be GetProcAddress'd in chrome for
> https://codereview.chromium.org/22744005

(It looks like there's some problems here still, so hold off on
reviewing/landing for now)

[shannonwoods1, 2013-08-13 20:54:49]

On 2013/08/13 19:14:00, scottmg wrote:
> On 2013/08/13 16:55:21, scottmg wrote:
> > OK, updated to include files missed in last landing and to remove the dll
> > futzing in favour of a function that can be GetProcAddress'd in chrome for
> > https://codereview.chromium.org/22744005
> 
> (It looks like there's some problems here still, so hold off on
> reviewing/landing for now)

Reverting prior patch in the meantime.

[scottmg, 2013-08-13 23:22:03]

On 2013/08/13 20:54:49, shannonwoods1 wrote:
> On 2013/08/13 19:14:00, scottmg wrote:
> > On 2013/08/13 16:55:21, scottmg wrote:
> > > OK, updated to include files missed in last landing and to remove the dll
> > > futzing in favour of a function that can be GetProcAddress'd in chrome for
> > > https://codereview.chromium.org/22744005
> > 
> > (It looks like there's some problems here still, so hold off on
> > reviewing/landing for now)
> 
> Reverting prior patch in the meantime.

I was getting a black window on startup which I thought was related to this
change, but I just bisected and looks like it's caused by 

[afcb765...]d:\src\cr2\src\third_party\angle_dx11>git bisect good
d206c85e27e10636060f09b7feaba04f6d17665d is the first bad commit
commit d206c85e27e10636060f09b7feaba04f6d17665d
Author: Nicolas Capens <nicolas@transgaming.com>
Date:   Mon Aug 12 11:21:19 2013 -0400

    Eliminate bitfield enum members.

    TRAC #23653
    Signed-off-by: Shannon Woods
    Signed-off-by: Geoff Lang
    Author: Nicolas Capens

:040000 040000 7872af5be1b986e2f68054c56ae1660a41072a50
b3fab28b941825cfeb8bfd6d64a56a6fc498f611 M      src

[shannonwoods1, 2013-08-13 23:53:08]

On 2013/08/13 23:22:03, scottmg wrote:
> On 2013/08/13 20:54:49, shannonwoods1 wrote:
> > On 2013/08/13 19:14:00, scottmg wrote:
> > > On 2013/08/13 16:55:21, scottmg wrote:
> > > > OK, updated to include files missed in last landing and to remove the
dll
> > > > futzing in favour of a function that can be GetProcAddress'd in chrome
for
> > > > https://codereview.chromium.org/22744005
> > > 
> > > (It looks like there's some problems here still, so hold off on
> > > reviewing/landing for now)
> > 
> > Reverting prior patch in the meantime.
> 
> I was getting a black window on startup which I thought was related to this
> change, but I just bisected and looks like it's caused by 
> 
> [afcb765...]d:\src\cr2\src\third_party\angle_dx11>git bisect good
> d206c85e27e10636060f09b7feaba04f6d17665d is the first bad commit
> commit d206c85e27e10636060f09b7feaba04f6d17665d
> Author: Nicolas Capens <mailto:nicolas@transgaming.com>
> Date:   Mon Aug 12 11:21:19 2013 -0400
> 
>     Eliminate bitfield enum members.
> 
>     TRAC #23653
>     Signed-off-by: Shannon Woods
>     Signed-off-by: Geoff Lang
>     Author: Nicolas Capens
> 
> :040000 040000 7872af5be1b986e2f68054c56ae1660a41072a50
> b3fab28b941825cfeb8bfd6d64a56a6fc498f611 M      src

Odd-- I've been working against master branch without seeing that behavior. Does
applying a revert of that patch alleviate the black window? Are you starting up
with any particular flags?

[scottmg, 2013-08-14 00:15:26]

On 2013/08/13 23:53:08, shannonwoods1 wrote:
> On 2013/08/13 23:22:03, scottmg wrote:
> > On 2013/08/13 20:54:49, shannonwoods1 wrote:
> > > On 2013/08/13 19:14:00, scottmg wrote:
> > > > On 2013/08/13 16:55:21, scottmg wrote:
> > > > > OK, updated to include files missed in last landing and to remove the
> dll
> > > > > futzing in favour of a function that can be GetProcAddress'd in chrome
> for
> > > > > https://codereview.chromium.org/22744005
> > > > 
> > > > (It looks like there's some problems here still, so hold off on
> > > > reviewing/landing for now)
> > > 
> > > Reverting prior patch in the meantime.
> > 
> > I was getting a black window on startup which I thought was related to this
> > change, but I just bisected and looks like it's caused by 
> > 
> > [afcb765...]d:\src\cr2\src\third_party\angle_dx11>git bisect good
> > d206c85e27e10636060f09b7feaba04f6d17665d is the first bad commit
> > commit d206c85e27e10636060f09b7feaba04f6d17665d
> > Author: Nicolas Capens <mailto:nicolas@transgaming.com>
> > Date:   Mon Aug 12 11:21:19 2013 -0400
> > 
> >     Eliminate bitfield enum members.
> > 
> >     TRAC #23653
> >     Signed-off-by: Shannon Woods
> >     Signed-off-by: Geoff Lang
> >     Author: Nicolas Capens
> > 
> > :040000 040000 7872af5be1b986e2f68054c56ae1660a41072a50
> > b3fab28b941825cfeb8bfd6d64a56a6fc498f611 M      src
> 
> Odd-- I've been working against master branch without seeing that behavior.
Does
> applying a revert of that patch alleviate the black window? Are you starting
up
> with any particular flags?

No, no flags, but I am running an aura build, so ANGLE is compositing everything
I guess (no software rendering). If I run with --disable-gpu it's fine.

Yes, I built at afcb7653356d22151ecfbc0790711e733b44c9ee and it was fine, and
rolled forward to d206c85e27e10636060f09b7feaba04f6d17665d and I get no
painting.

[Alok Priyadarshi, 2013-08-14 04:44:10]

d206c85e27e1 indeed broke chrome. None of the shaders compile. I have verified
that reverting that patch brings everything back to normal. I think we should
revert it for now.

[scottmg, 2013-08-14 04:53:49]

On 2013/08/14 04:44:10, Alok Priyadarshi wrote:
> d206c85e27e1 indeed broke chrome. None of the shaders compile. I have verified
> that reverting that patch brings everything back to normal. I think we should
> revert it for now.

Shannon was going to revert d206c85e and land an updated version of this change
in the morning.