Properly queue messages sent to message ports that are transferred to a service worker.

Properly queue messages sent to message ports that are transferred to a service worker.

There can be some delay between the message port being transferred to the
browser process and the renderer for the service worker being available, so
this change makes sure that during this period messages sent to these message
ports are queued, rather than get lost/cause assertions to fail.

Additionally, ports that are send to these ports while they are queued will be similarly delayed, and thus also need to be put in this special state.

Finally when for some reason launching the service worker fails all these ports (that are not associated with any renderer) need to be cleaned up, which might have to happen asynchronously, since the source renderer process might not have finished sending the message queue for this port yet.

Some layout tests for this are in https://codereview.chromium.org/729923004/

BUG=432678
Committed: https://crrev.com/ddb153288a99ad6be14b88244bfa070e612580f9
Cr-Commit-Position: refs/heads/master@{#305120}

[Marijn Kruisselbrink, 2014-11-18 23:38:25]

Patchset #2 (id:20001) has been deleted

[Marijn Kruisselbrink, 2014-11-19 00:46:30]

Patchset #3 (id:60001) has been deleted

[Marijn Kruisselbrink, 2014-11-19 01:03:08]

mek@chromium.org changed reviewers:
+ jam@chromium.org, michaeln@chromium.org

[Marijn Kruisselbrink, 2014-11-19 01:03:09]

jam: I don't know who the best reviewer/owner is for these MessagePortService
changes, but you seemed to have reviewed some of that code in the past. Feel
free to redirect to somebody else if you know of a better reviewer.

michaeln: If there is a better person from the service worker side also feel
free to redirect.

This really could do with more tests than the somewhat fragile layout tests in
https://codereview.chromium.org/729923004/.
The layout tests in that CL try to cover most cases, but since a lot of this is
dependent on timing/race conditions there is no guarantee that any of the
problematic cases is actually tested.

[Marijn Kruisselbrink, 2014-11-19 19:04:40]

mek@chromium.org changed reviewers:
+ falken@chromium.org

[Marijn Kruisselbrink, 2014-11-19 19:04:41]

+falken, since apparently michaeln is on vacation

[falken, 2014-11-20 03:42:43]

lgtm for service worker

https://codereview.chromium.org/737833002/diff/80001/content/browser/message_...
File content/browser/message_port_service.h (right):

https://codereview.chromium.org/737833002/diff/80001/content/browser/message_...
content/browser/message_port_service.h:50: // code doing isn't able to
immediately update the message port with a new
"code doing that"?

[Marijn Kruisselbrink, 2014-11-20 22:14:44]

jam: can you please take a look, or redirect to somebody that can?

https://codereview.chromium.org/737833002/diff/80001/content/browser/message_...
File content/browser/message_port_service.h (right):

https://codereview.chromium.org/737833002/diff/80001/content/browser/message_...
content/browser/message_port_service.h:50: // code doing isn't able to
immediately update the message port with a new
On 2014/11/20 03:42:43, falken wrote:
> "code doing that"?

Done.

[jam, 2014-11-20 22:36:48]

On 2014/11/20 22:14:44, Marijn Kruisselbrink wrote:
> jam: can you please take a look, or redirect to somebody that can?
lgtm

https://codereview.chromium.org/737833002/diff/120001/content/browser/message...
File content/browser/message_port_service.cc (right):

https://codereview.chromium.org/737833002/diff/120001/content/browser/message...
content/browser/message_port_service.cc:286: {
nit: wrong line

[Marijn Kruisselbrink, 2014-11-20 23:13:31]

Patchset #6 (id:140001) has been deleted

[Marijn Kruisselbrink, 2014-11-20 23:15:27]

mek@chromium.org changed reviewers:
+ tsepez@chromium.org

[Marijn Kruisselbrink, 2014-11-20 23:15:27]

+tsepez for RPC OWNERS

https://codereview.chromium.org/737833002/diff/120001/content/browser/message...
File content/browser/message_port_service.cc (right):

https://codereview.chromium.org/737833002/diff/120001/content/browser/message...
content/browser/message_port_service.cc:286: {
On 2014/11/20 22:36:47, jam wrote:
> nit: wrong line

Done.

[Tom Sepez, 2014-11-20 23:23:48]

Messages LGTM.

[Marijn Kruisselbrink, 2014-11-20 23:28:27]

The CQ bit was checked by mek@chromium.org

[commit-bot: I haz the power, 2014-11-20 23:29:21]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/737833002/160001

[commit-bot: I haz the power, 2014-11-21 00:31:29]

Committed patchset #6 (id:160001)

[commit-bot: I haz the power, 2014-11-21 00:32:15]

Patchset 6 (id:??) landed as
https://crrev.com/ddb153288a99ad6be14b88244bfa070e612580f9
Cr-Commit-Position: refs/heads/master@{#305120}