Add the array output C API.

Add the array output C API.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=125674

[dmichael (off chromium), 2012-03-07 03:56:41]

So was the consensus from the meeting that (even though we don't love the C
interface), nobody can think of anything better? :-)

Should we consider putting this in Dev for a bit until we've tried it with some
real interfaces? Or are you wanting to push it in for Gamepad?

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.idl
File ppapi/api/pp_array_output.idl (right):

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.id...
ppapi/api/pp_array_output.idl:34: * @param element_size The size of each
element.
" in bytes"

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.id...
ppapi/api/pp_array_output.idl:46: * browser can not allocate memory on behalf of
the plugin.
The last sentence could be confusing, because "so" could be read to mean "so
that" instead of "and therefore". Maybe something like:
"The browser can not allocate memory on behalf of the plugin because the plugin
and browser may use different allocators."

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.id...
ppapi/api/pp_array_output.idl:54: * <code>
Apparently, we're using @code now for blocks of code like this (but <code> for
in-line).  Somehow the doxygen formatting is different.

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.id...
ppapi/api/pp_array_output.idl:67: *     output->data = NULL;
I know size == 0 is a weird case, but shouldn't we still set
output->element_count?

[brettw, 2012-03-07 05:56:20]

Maybe we won't do gamepad, although that's what I was thinking.

Plus, when I land the C++ stuff it can't realistically be "dev" because it will
need to be in the completion callback header.

And for this type of thing, it's only as dev (or not) as the interfaces we use
it on. If we don't use it on any non-dev interfaces, we can always change it
without breaking compatibility.

New snap up.

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.idl
File ppapi/api/pp_array_output.idl (right):

http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.id...
ppapi/api/pp_array_output.idl:67: *     output->data = NULL;
On 2012/03/07 03:56:41, dmichael wrote:
> I know size == 0 is a weird case, but shouldn't we still set
> output->element_count?

It's set above.

[dmichael (off chromium), 2012-03-07 16:45:36]

On 2012/03/07 05:56:20, brettw wrote:
> Maybe we won't do gamepad, although that's what I was thinking.
> 
> Plus, when I land the C++ stuff it can't realistically be "dev" because it
will
> need to be in the completion callback header.
> 
> And for this type of thing, it's only as dev (or not) as the interfaces we use
> it on. If we don't use it on any non-dev interfaces, we can always change it
> without breaking compatibility.
I know, but in general, I'd rather not have some cruft in the public interfaces
that we're not using. It's just more noise for developers who are trying to
learn the API.

> 
> New snap up.
> 
> http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.idl
> File ppapi/api/pp_array_output.idl (right):
> 
>
http://codereview.chromium.org/9592033/diff/3001/ppapi/api/pp_array_output.id...
> ppapi/api/pp_array_output.idl:67: *     output->data = NULL;
> On 2012/03/07 03:56:41, dmichael wrote:
> > I know size == 0 is a weird case, but shouldn't we still set
> > output->element_count?
> 
> It's set above.
Sorry I overlooked that.

It's "Mostly harmless" and will likely solve problems for us in the near future,
so lgtm

[commit-bot: I haz the power, 2012-03-08 18:42:01]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/brettw@chromium.org/9592033/4

[commit-bot: I haz the power, 2012-03-08 20:43:19]

Change committed as 125674