Implement receiving transferred ports from JS

This CL addresses a long standing TODO to implement transferring ports from JS to Java.

Note that this is the last piece of MessageChannel puzzle. With this, we can consider basic implementation done. There is a missing feature to implement posting messages to frames given with their names, but this has other dependency.

We can now start further testing, optimization and refactorings.

BUG=393291
Committed: https://crrev.com/758671bdf97f247de84ce6c814e852b02bdf8366
Cr-Commit-Position: refs/heads/master@{#320198}

[sgurun-gerrit only, 2015-03-10 21:43:14]

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

[sgurun-gerrit only, 2015-03-10 21:43:31]

On 2015/03/10 21:43:14, sgurun wrote:
> mailto:sgurun@chromium.org changed reviewers:
> + mailto:hush@chromium.org

ptal

[sgurun-gerrit only, 2015-03-10 21:44:28]

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

[sgurun-gerrit only, 2015-03-10 21:47:11]

On 2015/03/10 21:43:31, sgurun wrote:
> On 2015/03/10 21:43:14, sgurun wrote:
> > mailto:sgurun@chromium.org changed reviewers:
> > + mailto:hush@chromium.org
> 
> ptal

nasko: please review content. This is the last piece of the message channel
puzzle. When a message port is transferred from JS to Java, its delegate needs
to be updated. Unfortunately, this requires exposing one other api of
messageportservice to public.

hush: review everything else

[hush (inactive), 2015-03-10 22:31:36]

https://codereview.chromium.org/995663002/diff/20001/android_webview/java/src...
File
android_webview/java/src/org/chromium/android_webview/AwMessagePortService.java
(right):

https://codereview.chromium.org/995663002/diff/20001/android_webview/java/src...
android_webview/java/src/org/chromium/android_webview/AwMessagePortService.java:148:
for (int i = 0; i < ports.length; i++) {
Do you need to null check ports here?

https://codereview.chromium.org/995663002/diff/20001/android_webview/java/src...
android_webview/java/src/org/chromium/android_webview/AwMessagePortService.java:149:
if (messagePorts == null) {
Move it out of the loop so that you can avoid checking it for each iteration?

https://codereview.chromium.org/995663002/diff/20001/android_webview/javatest...
File
android_webview/javatests/src/org/chromium/android_webview/test/PostMessageTest.java
(right):

https://codereview.chromium.org/995663002/diff/20001/android_webview/javatest...
android_webview/javatests/src/org/chromium/android_webview/test/PostMessageTest.java:755:
+ "            myPort.postMessage('from window', [mc.port2]);"
The other scenario (that does not work) is JS trying to postMessage like this:
e.source.postMessage("from window", [mc.port2]);
This would be posting to AwContents directly.
Will this fail silently or crash?

https://codereview.chromium.org/995663002/diff/20001/android_webview/native/a...
File android_webview/native/aw_message_port_service_impl.cc (right):

https://codereview.chromium.org/995663002/diff/20001/android_webview/native/a...
android_webview/native/aw_message_port_service_impl.cc:100: for (const auto&
iter : sent_message_port_ids) {
do you really need to use auto? It is just int?

[sgurun-gerrit only, 2015-03-10 23:20:13]

https://codereview.chromium.org/995663002/diff/20001/android_webview/java/src...
File
android_webview/java/src/org/chromium/android_webview/AwMessagePortService.java
(right):

https://codereview.chromium.org/995663002/diff/20001/android_webview/java/src...
android_webview/java/src/org/chromium/android_webview/AwMessagePortService.java:148:
for (int i = 0; i < ports.length; i++) {
On 2015/03/10 22:31:36, hush wrote:
> Do you need to null check ports here?

it is guaranteed not to be null at this point (could be empty).

https://codereview.chromium.org/995663002/diff/20001/android_webview/java/src...
android_webview/java/src/org/chromium/android_webview/AwMessagePortService.java:149:
if (messagePorts == null) {
On 2015/03/10 22:31:36, hush wrote:
> Move it out of the loop so that you can avoid checking it for each iteration?

this is to prevent creating a new if ports array is empty.

https://codereview.chromium.org/995663002/diff/20001/android_webview/javatest...
File
android_webview/javatests/src/org/chromium/android_webview/test/PostMessageTest.java
(right):

https://codereview.chromium.org/995663002/diff/20001/android_webview/javatest...
android_webview/javatests/src/org/chromium/android_webview/test/PostMessageTest.java:755:
+ "            myPort.postMessage('from window', [mc.port2]);"
On 2015/03/10 22:31:36, hush wrote:
> The other scenario (that does not work) is JS trying to postMessage like this:
> e.source.postMessage("from window", [mc.port2]);
> This would be posting to AwContents directly.
> Will this fail silently or crash?
let's write a test in a separate CL.

https://codereview.chromium.org/995663002/diff/20001/android_webview/native/a...
File android_webview/native/aw_message_port_service_impl.cc (right):

https://codereview.chromium.org/995663002/diff/20001/android_webview/native/a...
android_webview/native/aw_message_port_service_impl.cc:100: for (const auto&
iter : sent_message_port_ids) {
On 2015/03/10 22:31:36, hush wrote:
> do you really need to use auto? It is just int?

either works.

[hush (inactive), 2015-03-10 23:32:17]

okay. 
a_w/ lgtm

[nasko, 2015-03-11 23:08:42]

LGTM with a nit.

https://codereview.chromium.org/995663002/diff/40001/content/public/browser/m...
File content/public/browser/message_port_provider.h (right):

https://codereview.chromium.org/995663002/diff/40001/content/public/browser/m...
content/public/browser/message_port_provider.h:70: // routing numbers equal to
the message port numbers.
nit: Let's be consistent and put the requirement of this being called on the IO
thread as part of the comment.

[sgurun-gerrit only, 2015-03-11 23:29:00]

https://codereview.chromium.org/995663002/diff/40001/content/public/browser/m...
File content/public/browser/message_port_provider.h (right):

https://codereview.chromium.org/995663002/diff/40001/content/public/browser/m...
content/public/browser/message_port_provider.h:70: // routing numbers equal to
the message port numbers.
On 2015/03/11 23:08:42, nasko wrote:
> nit: Let's be consistent and put the requirement of this being called on the
IO
> thread as part of the comment.

Done.

[sgurun-gerrit only, 2015-03-11 23:29:09]

The CQ bit was checked by sgurun@chromium.org

[sgurun-gerrit only, 2015-03-11 23:29:09]

The patchset sent to the CQ was uploaded after l-g-t-m from hush@chromium.org,
nasko@chromium.org
Link to the patchset: https://codereview.chromium.org/995663002/#ps60001 (title:
"address nit from nasko")

[commit-bot: I haz the power, 2015-03-11 23:30:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/995663002/60001

[commit-bot: I haz the power, 2015-03-12 01:11:54]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-03-12 01:12:30]

Patchset 4 (id:??) landed as
https://crrev.com/758671bdf97f247de84ce6c814e852b02bdf8366
Cr-Commit-Position: refs/heads/master@{#320198}