Fix PPB_VideoDecoder_Impl::NotifyEndOfBitstreamBuffer to use correct ID.

Fix PPB_VideoDecoder_Impl::NotifyEndOfBitstreamBuffer to use correct ID.

Enhanced gles2.cc sample plugin to have multipe decodes outstanding at a time,
and assert that we get back from the decode API exactly the bitstream buffers we
sent to it.

This CL is relative to http://codereview.chromium.org/7200033/ which must land first.

BUG=86235
TEST=gles2 sample plugin completes correctly even with concurrent Decode()s.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=89779

[Ami GONE FROM CHROMIUM, 2011-06-20 20:27:28]

brettw: OWNERS for webkit/plugins/ppapi please
scherkus: enjoy!
vjain: please verify that this works for you in lieu of your fix.

cfreeman/vhiremath: FYI.

[brettw, 2011-06-20 21:51:06]

ppapi LGTM

http://codereview.chromium.org/7204038/diff/1/webkit/plugins/ppapi/ppb_video_...
File webkit/plugins/ppapi/ppb_video_decoder_impl.h (right):

http://codereview.chromium.org/7204038/diff/1/webkit/plugins/ppapi/ppb_video_...
webkit/plugins/ppapi/ppb_video_decoder_impl.h:78: typedef std::map<int32,
PP_CompletionCallback> CallbackById;
Can you clarify what the "id" is that this is mapping from and what these
callbacks are for?

[Ami GONE FROM CHROMIUM, 2011-06-20 22:13:02]

Thanks for the fast review!

http://codereview.chromium.org/7204038/diff/1/webkit/plugins/ppapi/ppb_video_...
File webkit/plugins/ppapi/ppb_video_decoder_impl.h (right):

http://codereview.chromium.org/7204038/diff/1/webkit/plugins/ppapi/ppb_video_...
webkit/plugins/ppapi/ppb_video_decoder_impl.h:78: typedef std::map<int32,
PP_CompletionCallback> CallbackById;
On 2011/06/20 21:51:06, brettw wrote:
> Can you clarify what the "id" is that this is mapping from and what these
> callbacks are for?

Done.

[scherkus (not reviewing), 2011-06-21 01:22:27]

LGTM

I think I was overly cautious about the assert-in-release-mode so feel free to
ignore.

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.cc
File ppapi/examples/gles2/gles2.cc (right):

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.c...
ppapi/examples/gles2/gles2.cc:28: // include-guards), make sure this is the last
file #include'd in this file.
:(

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.c...
ppapi/examples/gles2/gles2.cc:201: assert(!bitstream_ids_at_decoder_.size());
empty?

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.c...
ppapi/examples/gles2/gles2.cc:253:
assert(bitstream_buffers_by_id_.insert(std::make_pair(id, buffer)).second);
I'm guessing assert doesn't get optimized out in release modes if you're doing
this, right?

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.c...
ppapi/examples/gles2/gles2.cc:258:
assert(bitstream_ids_at_decoder_.insert(id).second);
ditto here

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.c...
ppapi/examples/gles2/gles2.cc:273:
assert(buffers_by_id_.insert(std::make_pair(id, buffer)).second);
etc

[Ami GONE FROM CHROMIUM, 2011-06-21 01:35:22]

Right, assert() is being kept even in NDEBUG builds.

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.cc
File ppapi/examples/gles2/gles2.cc (right):

http://codereview.chromium.org/7204038/diff/2003/ppapi/examples/gles2/gles2.c...
ppapi/examples/gles2/gles2.cc:201: assert(!bitstream_ids_at_decoder_.size());
On 2011/06/21 01:22:27, scherkus wrote:
> empty?

Done.