Call CDMWrapper from PpapiDecryptor.

Call CDMWrapper from PpapiDecryptor.

This CL is for testing purpose only. No need to review.

This CL is based on http://codereview.chromium.org/10545036/#ps78002 (Patch Set 25).

Now that PpapiDecryptor::GenerateKeyRequest will call into the CDMWrapper and KeyMessage is called from the plugin into the PluginInstance. With this, we don't need to hack from WebMediaPlayerImpl anymore.

To run the test:
src/out/Debug/chrome --enable-video-track --enable-media-source --enable-encrypted-media --no-sandbox --register-pepper-plugins="src/out/Debug/lib/libppapi_example_content_decryptor.so;application/x-ppapi-content-encryption" --vmodule=pipeline*=1,proxy*=3,aes*=3,ppapi_de*=3

BUG=none
TEST=none

[xhwang, 2012-08-01 21:35:23]

On 2012/07/31 23:41:57, xhwang wrote:

+dmichael as this CL can help understand what we want to achieve with the
CDMWrapper.

This patch is based on Tom's patch: http://codereview.chromium.org/10545036/

FYI/test only. Comments welcome, but no need to review yet.

Tom: 

I remember you asked about where does the key events go (e.g. KeyMessage). In
Patch Set 2, you can see how it's forwarded back to the DecryptorClient. Also, a
callback is passed in with Decrypt(). When DeliverBlock() is called, the
callback is fired with the decrypted data.

This patch is still pretty hacky, but should be able to let you see the ultimate
data/events flow.

[xhwang, 2012-08-07 00:53:48]

http://codereview.chromium.org/10836038/diff/5001/ppapi/api/dev/pp_decryption...
File ppapi/api/dev/pp_decryption_buffer_dev.idl (right):

http://codereview.chromium.org/10836038/diff/5001/ppapi/api/dev/pp_decryption...
ppapi/api/dev/pp_decryption_buffer_dev.idl:19: struct PP_DecryptionBuffer_Dev {
dmichael@: this is what I talked about offline. It works, but not really sure if
this is the best approach. PTAL and let me know your opinion.

[dmichael (off chromium), 2012-08-09 16:24:52]

It looks like you're still debugging, and this is all mixed up with tomfinegan's
CL, so I'm guessing you aren't looking for an in-depth review yet?

Anyway, here are a handful of comments.

http://codereview.chromium.org/10836038/diff/5001/ppapi/api/dev/pp_decryption...
File ppapi/api/dev/pp_decryption_buffer_dev.idl (right):

http://codereview.chromium.org/10836038/diff/5001/ppapi/api/dev/pp_decryption...
ppapi/api/dev/pp_decryption_buffer_dev.idl:16: * The data structure for
decryption buffer.
This doesn't tell me any more than the class name does. Could you elaborate on
what the decryption buffer is? Same for the SubsampleEntry. Referencing what
this maps to in the spec might be a good way to go.

http://codereview.chromium.org/10836038/diff/5001/ppapi/api/dev/pp_decryption...
ppapi/api/dev/pp_decryption_buffer_dev.idl:24: PP_Resource data;
I think everything else is okay, but I'd rather you separate out the data. We
don't really have a good way to embed a reference-counted type in a struct.

http://codereview.chromium.org/10836038/diff/5001/webkit/media/crypto/ppapi_d...
File webkit/media/crypto/ppapi_decryptor.cc (right):

http://codereview.chromium.org/10836038/diff/5001/webkit/media/crypto/ppapi_d...
webkit/media/crypto/ppapi_decryptor.cc:28: ->message_loop_proxy()) {
nit: indentation is off.

Would it fit if you break after the "(" and do a 4-space indent?

[xhwang, 2012-08-21 21:01:35]

This CL was experimental and now is obsolete. It will be replaced by a series of
updated CLs.