To facilitate apatrick's work on the 3D renderer for Pepper,...

To facilitate apatrick's work on the 3D renderer for Pepper,
I am splitting out the pepper-specific portions of npapi.h.
This is because his code currently relies on a version
of npapi.h different from third_party\npapi\bindings.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=30671

[sehr (please use chromium), 2009-10-30 23:22:46]

Please have a look.

[apatrick, 2009-10-30 23:34:28]

I'm not sure which style guide we should use in pepper.h. Probably the same as
npapi.h since this stuff will end up in there eventually so I'm not worried
about the lint errors for placement of opening {.

http://codereview.chromium.org/343069/diff/4008/4011
File webkit/glue/pepper/pepper.h (right):

http://codereview.chromium.org/343069/diff/4008/4011#newcode17
Line 17: #define NPNVPepperExtensions static_cast<NPNVariable>(4000)
I think it should be one of these:

// C++
const NPNVariable NPNVPepperExtensions =
    static_cast<NPNVariable>(4000);

// C compatible
#define NPNVPepperExtensions ((NPNVariable)(4000))

http://codereview.chromium.org/343069/diff/4008/4011#newcode148
Line 148: NPFlushRenderContextCallbackPtr callback,
80 chars

[sehr (please use chromium), 2009-10-30 23:41:34]

Thanks for the review, Al.

http://codereview.chromium.org/343069/diff/4008/4011
File webkit/glue/pepper/pepper.h (right):

http://codereview.chromium.org/343069/diff/4008/4011#newcode17
Line 17: #define NPNVPepperExtensions static_cast<NPNVariable>(4000)
On 2009/10/30 23:34:28, apatrick wrote:
> I think it should be one of these:
> 
> // C++
> const NPNVariable NPNVPepperExtensions =
>     static_cast<NPNVariable>(4000);
> 
> // C compatible
> #define NPNVPepperExtensions ((NPNVariable)(4000))

You're right.  NPAPI is a C API, so it should be the C cast.

http://codereview.chromium.org/343069/diff/4008/4011#newcode148
Line 148: NPFlushRenderContextCallbackPtr callback,
On 2009/10/30 23:34:28, apatrick wrote:
> 80 chars

This code is intended to use the NPAPI conventions, which don't care about line
length and like alignment, the best I can tell.

[apatrick, 2009-10-30 23:44:23]

LGTM with the C style cast change.

[jam, 2009-10-31 00:49:42]

lgtm

http://codereview.chromium.org/343069/diff/7001/8003
File webkit/glue/pepper/pepper.h (right):

http://codereview.chromium.org/343069/diff/7001/8003#newcode1
Line 1: // Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
nit: should be 2009

[darin (slow to review), 2009-10-31 22:34:05]

What was the problem exactly?  What was conflicting?

It was nice when npapi.h held the pepper revisions behind an #ifdef since it
makes it possible to see the API grow in-place.  What are the conditions for
putting it back into npapi.h?

[sehr (please use chromium), 2009-10-31 22:40:35]

Darin,

I totally agree with your position, and only did this to get Al unstuck on
the 3D api.  The issue was that the GPU plugin and O3D use a copy of npapi
other than Chrome's.  Fwiw, I'm currently trying to get NaCl to use the
Chrome version as well, rather than the copy we have.

David

On Sat, Oct 31, 2009 at 3:34 PM, <darin@chromium.org> wrote:

> What was the problem exactly?  What was conflicting?
>
> It was nice when npapi.h held the pepper revisions behind an #ifdef since
> it
> makes it possible to see the API grow in-place.  What are the conditions
> for
> putting it back into npapi.h?
>
>
> http://codereview.chromium.org/343069
>

[darin (slow to review), 2009-10-31 22:47:18]

I see.
-Darin

On Sat, Oct 31, 2009 at 3:40 PM, David Sehr <sehr@google.com> wrote:

> Darin,
>
> I totally agree with your position, and only did this to get Al unstuck on
> the 3D api.  The issue was that the GPU plugin and O3D use a copy of npapi
> other than Chrome's.  Fwiw, I'm currently trying to get NaCl to use the
> Chrome version as well, rather than the copy we have.
>
> David
>
>
> On Sat, Oct 31, 2009 at 3:34 PM, <darin@chromium.org> wrote:
>
>> What was the problem exactly?  What was conflicting?
>>
>> It was nice when npapi.h held the pepper revisions behind an #ifdef since
>> it
>> makes it possible to see the API grow in-place.  What are the conditions
>> for
>> putting it back into npapi.h?
>>
>>
>> http://codereview.chromium.org/343069
>>
>
>