[ServiceWorker] Sends the blob uuid of the request body to the ServiceWorker.

Sends the blob uuid of the request body to the ServiceWorker.

This changes the followings:
 - ResourceDispatcherHostImpl passes the reference of ResourceRequestBody to ServiceWorkerControlleeRequestHandler via ServiceWorkerProviderHost.
 - ServiceWorkerControlleeRequestHandler keeps the reference of ResourceRequestBody and pass it to ServiceWorkerURLRequestJob when MaybeCreateJob() is called.
 - ServiceWorkerURLRequestJob keeps the reference of ResourceRequestBody and creates BlobDataHandle and ServiceWorkerFetchRequest from URLRequest and ResourceRequestBody when StartRequest() is called.
 - ServiceWorkerURLRequestJob passes the ServiceWorkerFetchRequest to ServiceWorkerFetchDispatcher in StartRequest() and ServiceWorkerVersion will send the blob uuid and the size of the blob using IPC to the renederer process.
 - BlobDataHandle is kept during the lifetime of ServiceWorkerURLRequestJob.

 - In the ServiceWorker side, WebServiceWorkerRequest::setBlob() will be called with the blob uuid and the size of the blob.
   WebServiceWorkerRequest::setBlob() is introduced in https://codereview.chromium.org/483603003/


This cl depends on https://codereview.chromium.org/483603003/ and https://codereview.chromium.org/483613004/ .

BUG=402387
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=290795

[horo, 2014-08-18 08:25:31]

falken@
Could you please review this?
Thank you.

[falken, 2014-08-19 03:37:39]

I don't see additional tests. I assume this will be tested Blink-side when
hooked up?

https://codereview.chromium.org/475333002/diff/210001/content/common/service_...
File content/common/service_worker/service_worker_types.h (right):

https://codereview.chromium.org/475333002/diff/210001/content/common/service_...
content/common/service_worker/service_worker_types.h:46: // we'll probably want
to have response struct/class too.
While you're here can you delete this outdated TODO.

[horo, 2014-08-19 03:50:33]

Yes fetch-event.html in https://codereview.chromium.org/475533005/ will test
this codes.

https://codereview.chromium.org/475333002/diff/210001/content/common/service_...
File content/common/service_worker/service_worker_types.h (right):

https://codereview.chromium.org/475333002/diff/210001/content/common/service_...
content/common/service_worker/service_worker_types.h:46: // we'll probably want
to have response struct/class too.
On 2014/08/19 03:37:39, falken wrote:
> While you're here can you delete this outdated TODO.

Done.

[falken, 2014-08-19 04:07:09]

Ah, I hadn't gotten to that review yet :) lgtm

[horo, 2014-08-19 04:29:21]

nasko@
Could you please review content/common/service_worker/service_worker_messages.h?

[horo, 2014-08-19 04:30:40]

mmenke@
Could you please review content/browser/loader/resource_dispatcher_host_impl.cc?

[mmenke, 2014-08-19 14:06:13]

On 2014/08/19 04:30:40, horo wrote:
> mmenke@
> Could you please review
content/browser/loader/resource_dispatcher_host_impl.cc?

content/browser/loader/resource_dispatcher_host_impl.cc LGTM (Note that I'm not
familiar with blob handling, and did not review any other code).

[nasko, 2014-08-19 16:58:48]

IPC LGTM

[michaeln1, 2014-08-19 22:03:26]

https://codereview.chromium.org/475333002/diff/250001/content/browser/service...
File content/browser/service_worker/service_worker_controllee_request_handler.cc
(right):

https://codereview.chromium.org/475333002/diff/250001/content/browser/service...
content/browser/service_worker/service_worker_controllee_request_handler.cc:33:
body_(body),
It's pretty hidden that the body has a collection of
webkit_blob::BlobDataHandles attached to it using the userdata mechanism. Those
handles ensure the blobs don't get freed up for as long as the
ResourceRequestBody instance is around. ResolveBlobReference in the
UploadDataStreamBuilder sets that up.

It might be worth a comment. Also might be worth hoisting the SetUserData calls
up from the UploadDataStreamBuilder to somewhere closer to IPC message
deserialization (maybe RDH::BeginRequest) since there are two consumers
depending on that behavior now.

[horo, 2014-08-20 02:28:02]

https://codereview.chromium.org/475333002/diff/250001/content/browser/service...
File content/browser/service_worker/service_worker_controllee_request_handler.cc
(right):

https://codereview.chromium.org/475333002/diff/250001/content/browser/service...
content/browser/service_worker/service_worker_controllee_request_handler.cc:33:
body_(body),
On 2014/08/19 22:03:25, michaeln1 wrote:
> It's pretty hidden that the body has a collection of
> webkit_blob::BlobDataHandles attached to it using the userdata mechanism.
Those
> handles ensure the blobs don't get freed up for as long as the
> ResourceRequestBody instance is around. ResolveBlobReference in the
> UploadDataStreamBuilder sets that up.
> 
> It might be worth a comment. Also might be worth hoisting the SetUserData
calls
> up from the UploadDataStreamBuilder to somewhere closer to IPC message
> deserialization (maybe RDH::BeginRequest) since there are two consumers
> depending on that behavior now.

Added comments in service_worker_url_request_job.h.
I will create another cl to move SetUserData calls.

[horo, 2014-08-20 04:39:10]

The CQ bit was checked by horo@chromium.org

[commit-bot: I haz the power, 2014-08-20 04:40:19]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/horo@chromium.org/475333002/270001

[horo, 2014-08-20 05:13:39]

The CQ bit was unchecked by horo@chromium.org

[horo, 2014-08-20 07:59:30]

The CQ bit was checked by horo@chromium.org

[commit-bot: I haz the power, 2014-08-20 08:00:33]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/horo@chromium.org/475333002/290001

[horo, 2014-08-20 08:09:58]

The CQ bit was unchecked by horo@chromium.org

[horo, 2014-08-20 08:19:25]

The CQ bit was checked by horo@chromium.org

[commit-bot: I haz the power, 2014-08-20 08:20:50]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/horo@chromium.org/475333002/290001

[commit-bot: I haz the power, 2014-08-20 09:27:47]

Committed patchset #4 (290001) as 290795