Convert ppapi/shared to use TrackedCallback.

Convert ppapi/shared to use TrackedCallback.

Convert all classes in ppapi/shared_impl to use the new shared TrackedCallback
class for the PP_CompletionCallbacks.

This required a few changes to some shared objects to make them resources.

BUG=
TEST=

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=116467

[viettrungluu, 2012-01-03 22:43:46]

As noted below, "Post...AndClear()" is wrong; things should remain "in progress"
until the callbacks have actually been called. (This is why TrackedCallback has
the "completed()" method.)

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_file_io_s...
File ppapi/shared_impl/ppb_file_io_shared.cc (right):

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_file_io_s...
ppapi/shared_impl/ppb_file_io_shared.cc:60: for (size_t i = 0; i <
callbacks_.size(); i++)
I suppose you could also use PostAbortForResource()? (Slightly slower, but
slightly cleaner?)

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_video_dec...
File ppapi/shared_impl/ppb_video_decoder_shared.cc (right):

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_video_dec...
ppapi/shared_impl/ppb_video_decoder_shared.cc:61: if (reset_callback_.get())
&& !reset_callback_->completed() ?

(Though I suppose you always do ClearAndRun().)

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_video_dec...
File ppapi/shared_impl/ppb_video_decoder_shared.h (right):

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_video_dec...
ppapi/shared_impl/ppb_video_decoder_shared.h:32:
PPB_VideoDecoder_Shared(PP_Instance instance);
explicit

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_video_dec...
ppapi/shared_impl/ppb_video_decoder_shared.h:33: PPB_VideoDecoder_Shared(const
HostResource& host_resource);
"

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/private/tcp_s...
File ppapi/shared_impl/private/tcp_socket_private_impl.cc (right):

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/private/tcp_s...
ppapi/shared_impl/private/tcp_socket_private_impl.cc:173:
PostAbortAndClearIfNecessary(&connect_callback_);
This actually violates the standard semantics that things are "in progress"
until the callbacks are actually run (possibly abortively).

Doing the right thing would mean changing all the |foo_callback_.get()| checks
to |foo_callback_.get() && !foo_callback_->completed()|, and *not* clearing when
posting (in particular, just "PostAbortIfNecessary()"). (Synchronous clearing
and running is fine.)

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/private/udp_s...
File ppapi/shared_impl/private/udp_socket_private_impl.cc (right):

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/private/udp_s...
ppapi/shared_impl/private/udp_socket_private_impl.cc:120:
PostAbortAndClearIfNecessary(&bind_callback_);
Ditto what I said for TCP....

[brettw, 2012-01-04 00:27:28]

New snap up.

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_file_io_s...
File ppapi/shared_impl/ppb_file_io_shared.cc (right):

http://codereview.chromium.org/9053003/diff/1/ppapi/shared_impl/ppb_file_io_s...
ppapi/shared_impl/ppb_file_io_shared.cc:60: for (size_t i = 0; i <
callbacks_.size(); i++)
Actually I can just delete this code, it's already done by the resource tracker.

[viettrungluu, 2012-01-04 01:15:13]

LGTM w/comments (that you can act on or ignore, as you like).

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/ppb_graphi...
File ppapi/shared_impl/ppb_graphics_3d_shared.cc (right):

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/ppb_graphi...
ppapi/shared_impl/ppb_graphics_3d_shared.cc:98: return !!swap_callback_.get();
TrackedCallback::IsPending()?

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/ppb_graphi...
ppapi/shared_impl/ppb_graphics_3d_shared.cc:98: return !!swap_callback_.get();
"

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/private/tc...
File ppapi/shared_impl/private/tcp_socket_private_impl.cc (right):

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/private/tc...
ppapi/shared_impl/private/tcp_socket_private_impl.cc:273:
scoped_refptr<TrackedCallback>* callback) {
Hmmm, this could now just take a scoped_refptr -- no need for a pointer to a
scoped_refptr.... <shrug>

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/private/ud...
File ppapi/shared_impl/private/udp_socket_private_impl.cc (right):

http://codereview.chromium.org/9053003/diff/6001/ppapi/shared_impl/private/ud...
ppapi/shared_impl/private/udp_socket_private_impl.cc:185: void
UDPSocketPrivateImpl::PostAbortIfNecessary(
"