Move NewDownload out of AwContentsIoThreadClientImpl to AwContentsClientBridge

Move NewDownload out of AwContentsIoThreadClientImpl

The IoThreadClient has a map that's keyed by RenderFrame IDs. That doesn't work with plznavigate where navigation requests only have a frame tree node id. This requires another map.

Ideally it would be good to not have to use that map and remove IoThreadClient completely. However, that is not possible. But most of the code in IoThreadClient already posts directly to UI thread and not need to be plumbed from there.

This is first of the patches to move code that does not necessarily has to use IoThreadClientImpl out of it.

BUG=645983
Committed: https://crrev.com/d465957ff8f649c59fe174fcabc6320c233b22b8
Cr-Commit-Position: refs/heads/master@{#426521}

[sgurun-gerrit only, 2016-10-19 20:51:38]

Description was changed from

==========
NOT EVEN TESTED Move NewDownload out of AwContentsIoThreadClientImpl

BUG=
==========

to

==========
Move NewDownload out of AwContentsIoThreadClientImpl

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983
==========

[sgurun-gerrit only, 2016-10-19 20:57:05]

sgurun@chromium.org changed reviewers:
+ boliu@chromium.org

[sgurun-gerrit only, 2016-10-19 20:57:06]

PTAL,

[boliu, 2016-10-19 20:59:38]

better CL description describing what's actually in the CL please

and motivation too, if it's not already in the bug

[sgurun-gerrit only, 2016-10-19 21:15:32]

Description was changed from

==========
Move NewDownload out of AwContentsIoThreadClientImpl

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983
==========

to

==========
Move NewDownload out of AwContentsIoThreadClientImpl

The IoThreadClient has a map that's keyed by RenderFrame IDs. That doesn't work
with plznavigate where navigation requests only have a frame tree node id. This
requires another map. 

Ideally it would be good to not have to use that map and remove IoThreadClient
completely. However, that is not possible. But most of the code in
IoThreadClient already posts directly to UI thread and not need to be plumbed
from there.

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983
==========

[sgurun-gerrit only, 2016-10-19 21:15:51]

On 2016/10/19 20:59:38, boliu wrote:
> better CL description describing what's actually in the CL please
> 
> and motivation too, if it's not already in the bug

updated.

[boliu, 2016-10-20 16:23:40]

lgtm

say something like moving onDownloadStart to AwContentsClientBridge in the
description

[sgurun-gerrit only, 2016-10-20 16:47:09]

On 2016/10/20 16:23:40, boliu wrote:
> lgtm
> 
> say something like moving onDownloadStart to AwContentsClientBridge in the
> description

done

[sgurun-gerrit only, 2016-10-20 16:51:13]

The CQ bit was checked by sgurun@chromium.org

[sgurun-gerrit only, 2016-10-20 16:51:13]

The patchset sent to the CQ was uploaded after l-g-t-m from boliu@chromium.org
Link to the patchset: https://codereview.chromium.org/2425423004/#ps40001
(title: "changed "even if" to "unless"")

[commit-bot: I haz the power, 2016-10-20 16:51:36]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-10-20 17:33:53]

Description was changed from

==========
Move NewDownload out of AwContentsIoThreadClientImpl

The IoThreadClient has a map that's keyed by RenderFrame IDs. That doesn't work
with plznavigate where navigation requests only have a frame tree node id. This
requires another map. 

Ideally it would be good to not have to use that map and remove IoThreadClient
completely. However, that is not possible. But most of the code in
IoThreadClient already posts directly to UI thread and not need to be plumbed
from there.

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983
==========

to

==========
Move NewDownload out of AwContentsIoThreadClientImpl

The IoThreadClient has a map that's keyed by RenderFrame IDs. That doesn't work
with plznavigate where navigation requests only have a frame tree node id. This
requires another map. 

Ideally it would be good to not have to use that map and remove IoThreadClient
completely. However, that is not possible. But most of the code in
IoThreadClient already posts directly to UI thread and not need to be plumbed
from there.

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983
==========

[commit-bot: I haz the power, 2016-10-20 17:33:54]

Committed patchset #3 (id:40001)

[jam, 2016-10-20 18:15:21]

jam@chromium.org changed reviewers:
+ jam@chromium.org

[jam, 2016-10-20 18:15:22]

thanks for the fix, one comment to make it work with PlzNavigate

https://codereview.chromium.org/2425423004/diff/40001/android_webview/browser...
File
android_webview/browser/renderer_host/aw_resource_dispatcher_host_delegate.cc
(right):

https://codereview.chromium.org/2425423004/diff/40001/android_webview/browser...
android_webview/browser/renderer_host/aw_resource_dispatcher_host_delegate.cc:323:
request_info->GetRenderFrameID(), url, user_agent,
To make this code work with PlzNavigate, use GetWebContentsGetterForRequest()
instead of GetChildID and GetRenderFrameID, then use
AwContentsClientBridgeBase::FromWebContents on the UI thread from the returned
WebContents. (GetChildId/GetRenderFrameID will return -1 in the PlzNavigate
case).

[sgurun-gerrit only, 2016-10-20 18:33:26]

https://codereview.chromium.org/2425423004/diff/40001/android_webview/browser...
File
android_webview/browser/renderer_host/aw_resource_dispatcher_host_delegate.cc
(right):

https://codereview.chromium.org/2425423004/diff/40001/android_webview/browser...
android_webview/browser/renderer_host/aw_resource_dispatcher_host_delegate.cc:323:
request_info->GetRenderFrameID(), url, user_agent,
On 2016/10/20 18:15:22, jam wrote:
> To make this code work with PlzNavigate, use GetWebContentsGetterForRequest()
> instead of GetChildID and GetRenderFrameID, then use
> AwContentsClientBridgeBase::FromWebContents on the UI thread from the returned
> WebContents. (GetChildId/GetRenderFrameID will return -1 in the PlzNavigate
> case).

will do, thanks!

[commit-bot: I haz the power, 2016-10-21 13:19:50]

Description was changed from

==========
Move NewDownload out of AwContentsIoThreadClientImpl

The IoThreadClient has a map that's keyed by RenderFrame IDs. That doesn't work
with plznavigate where navigation requests only have a frame tree node id. This
requires another map. 

Ideally it would be good to not have to use that map and remove IoThreadClient
completely. However, that is not possible. But most of the code in
IoThreadClient already posts directly to UI thread and not need to be plumbed
from there.

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983
==========

to

==========
Move NewDownload out of AwContentsIoThreadClientImpl

The IoThreadClient has a map that's keyed by RenderFrame IDs. That doesn't work
with plznavigate where navigation requests only have a frame tree node id. This
requires another map.

Ideally it would be good to not have to use that map and remove IoThreadClient
completely. However, that is not possible. But most of the code in
IoThreadClient already posts directly to UI thread and not need to be plumbed
from there.

This is first of the patches to move code that does not necessarily has to use
IoThreadClientImpl out of it.

BUG=645983

Committed: https://crrev.com/d465957ff8f649c59fe174fcabc6320c233b22b8
Cr-Commit-Position: refs/heads/master@{#426521}
==========

[commit-bot: I haz the power, 2016-10-21 13:19:51]

Patchset 3 (id:??) landed as
https://crrev.com/d465957ff8f649c59fe174fcabc6320c233b22b8
Cr-Commit-Position: refs/heads/master@{#426521}