Switch IPC::ChannelProxy to use MessageLoopProxy instead of MessageLoop. This allows us to remov...

Switch IPC::ChannelProxy to use MessageLoopProxy instead of MessageLoop.  This allows us to remove usage of the IOThread object, and generally makes IPC::ChannelProxy more robust for future uses.
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=83741

[sanjeevr, 2011-05-02 17:27:37]

A couple of comments. Also, in general, there are many places where the type has
been changed to base::MessageLoopPoxy but the variable name is still foo_loop.
This is not a big deal, just wanted to point out that it creates an
inconsistency with the naming pattern we have used elsewhere. That said, given
the side of the code change, I would be fine with leaving stuff as is.

http://codereview.chromium.org/6901146/diff/7062/ppapi/proxy/ppb_flash_file_p...
File ppapi/proxy/ppb_flash_file_proxy.cc (right):

http://codereview.chromium.org/6901146/diff/7062/ppapi/proxy/ppb_flash_file_p...
ppapi/proxy/ppb_flash_file_proxy.cc:199:
DCHECK(base::MessageLoopProxy::CreateForCurrentThread() == main_thread_);
This should be main_thread_->BelongsToCurrentThread()

http://codereview.chromium.org/6901146/diff/7062/ppapi/proxy/ppb_flash_file_p...
ppapi/proxy/ppb_flash_file_proxy.cc:277: if
(base::MessageLoopProxy::CreateForCurrentThread() == main_thread_) {
Same here (and everywhere else where we replace the "MessageLoop::current() ==
foo" check)

[jam, 2011-05-02 17:43:47]

http://codereview.chromium.org/6901146/diff/7062/ppapi/proxy/ppb_flash_file_p...
File ppapi/proxy/ppb_flash_file_proxy.cc (right):

http://codereview.chromium.org/6901146/diff/7062/ppapi/proxy/ppb_flash_file_p...
ppapi/proxy/ppb_flash_file_proxy.cc:199:
DCHECK(base::MessageLoopProxy::CreateForCurrentThread() == main_thread_);
On 2011/05/02 17:27:37, sanjeevr wrote:
> This should be main_thread_->BelongsToCurrentThread()

Done.

http://codereview.chromium.org/6901146/diff/7062/ppapi/proxy/ppb_flash_file_p...
ppapi/proxy/ppb_flash_file_proxy.cc:277: if
(base::MessageLoopProxy::CreateForCurrentThread() == main_thread_) {
On 2011/05/02 17:27:37, sanjeevr wrote:
> Same here (and everywhere else where we replace the "MessageLoop::current() ==
> foo" check)

Done (didn't see any other uses).

[jam, 2011-05-02 17:44:33]

On 2011/05/02 17:27:37, sanjeevr wrote:
> A couple of comments. Also, in general, there are many places where the type
has
> been changed to base::MessageLoopPoxy but the variable name is still foo_loop.
> This is not a big deal, just wanted to point out that it creates an
> inconsistency with the naming pattern we have used elsewhere. That said, given
> the side of the code change, I would be fine with leaving stuff as is.

oh, I wasn't aware that there was an existing pattern for this.  in places where
i have proxy code (non-ML), I've tended to drop the proxy from the member
variable name

[sanjeevr, 2011-05-02 17:49:07]

LGTM