Use process/control messages (instead of RVH messages) for layout tests.

Use process IPCs (instead of routing via RVH) to print layout tests output.

This CL changes ...HostMsg_PrintMessage IPC message so it is a "control"
(process-level) message, rather than a message routed to an individual
RenderViewHost.  This means that we no longer drop test output when
WebTestDelegate::PrintMessage (accidentally) happens to be called on a
WebTestRunner with a swapped-out RenderView - after this CL we will
consistently deliver each PrintMessage to the browser and include it in
the test output.  This is important, because it means that messages will
work the same, regardless of which WebTestDelegate they go are sent
through - and *this* is important to fix UaF caused by using an
incorrect WebTestDelegate from WebFrameTestClient
(https://crbug.com/606594).

Obviously, after this CL, more test output gets through to the browser.
This is why this CL has to update test expectations.

BUG=595895

[Łukasz Anforowicz, 2016-03-17 22:29:49]

Description was changed from

==========
Use process/control messages to avoid swapped-out exceptions for RVH messages.

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

to

==========
Use process/control messages to avoid swapped-out exceptions for RVH messages.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, we have multiple singletons:
LayoutTestRenderProcessObserver, BlinkTestRunner, WebTestRunner,
WebTestInterfaces.

BUG=595895
==========

[Łukasz Anforowicz, 2016-03-17 22:32:10]

Description was changed from

==========
Use process/control messages to avoid swapped-out exceptions for RVH messages.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, we have multiple singletons:
LayoutTestRenderProcessObserver, BlinkTestRunner, WebTestRunner,
WebTestInterfaces.

BUG=595895
==========

to

==========
Use process/control messages to avoid swapped-out exceptions for RVH messages.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, LayoutTestRenderProcessObserver is a singleton
used to find other process-wide things (WebTestRunner, WebTestInterfaces,
BlinkTestRunner aka WebTestDelegate).

BUG=595895
==========

[Łukasz Anforowicz, 2016-03-17 22:34:28]

Description was changed from

==========
Use process/control messages to avoid swapped-out exceptions for RVH messages.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, LayoutTestRenderProcessObserver is a singleton
used to find other process-wide things (WebTestRunner, WebTestInterfaces,
BlinkTestRunner aka WebTestDelegate).

BUG=595895
==========

to

==========
Use process/control messages (instead of RVH messages) for layout tests.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, LayoutTestRenderProcessObserver is a singleton
used to find other process-wide things (WebTestRunner, WebTestInterfaces,
BlinkTestRunner aka WebTestDelegate).

BUG=595895
==========

[Łukasz Anforowicz, 2016-03-17 22:34:40]

Description was changed from

==========
Use process/control messages (instead of RVH messages) for layout tests.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, LayoutTestRenderProcessObserver is a singleton
used to find other process-wide things (WebTestRunner, WebTestInterfaces,
BlinkTestRunner aka WebTestDelegate).

BUG=595895
==========

to

==========
Use process/control messages (instead of RVH messages) for layout tests.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, LayoutTestRenderProcessObserver is a singleton
used to find other process-wide things (WebTestRunner, WebTestInterfaces,
BlinkTestRunner aka WebTestDelegate).

BUG=595895
==========

[Łukasz Anforowicz, 2016-09-16 23:23:51]

The CQ bit was checked by lukasza@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-16 23:24:12]

Dry run: 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, 2016-09-16 23:29:37]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-16 23:29:38]

Dry run: Try jobs failed on following builders:
  android_arm64_dbg_recipe on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_arm6...)
  cast_shell_android on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/cast_shell_a...)
  ios-device on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-device/builds...)
  ios-simulator on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-simulator/bui...)

[Łukasz Anforowicz, 2016-09-16 23:42:12]

Description was changed from

==========
Use process/control messages (instead of RVH messages) for layout tests.

This CL changes ...HostMsg_PrintMessage and ...HostMsg_LayoutDumpFlagsChanged
IPC messages so they are "control" (process-level) messages, rather than
messages
routed to individual RVH.  In the short-term this helps avoid a swapped-out
exception in SwappedOutMessages::CanSendWhileSwappedOut which means that
there should be no content -> content/shell dependencies left (except for
browser tests).  This should also be okay for the long-term - removing RVH
messages is consistent with the long-term goal of making RVH non-routable.

Note that even before this change BlinkTestController and
BlinkTestRunner were effectively process-wide singletons.  For example
NotifyDoneForwarder is finding BlinkTestController through
BlinkTestController::Get() and there are 8 other calls to Get.
Similarily, on the renderer side, LayoutTestRenderProcessObserver is a singleton
used to find other process-wide things (WebTestRunner, WebTestInterfaces,
BlinkTestRunner aka WebTestDelegate).

BUG=595895
==========

to

==========
Use process IPCs (instead of routing via RVH) to print layout tests output.

This CL changes ...HostMsg_PrintMessage IPC message so it is a "control"
(process-level) message, rather than a message routed to an individual
RenderViewHost.  This means that we no longer drop test output when
WebTestDelegate::PrintMessage (accidentally) happens to be called on a
WebTestRunner with a swapped-out RenderView - after this CL we will
consistently deliver each PrintMessage to the browser and include it in
the test output.  This is important, because it means that messages will
work the same, regardless of which WebTestDelegate they go are sent
through - and *this* is important to fix UaF caused by using an
incorrect WebTestDelegate from WebFrameTestClient
(https://crbug.com/606594).

Obviously, after this CL, more test output gets through to the browser.
This is why this CL has to update test expectations.

BUG=595895
==========

[Łukasz Anforowicz, 2016-09-17 00:21:00]

The CQ bit was checked by lukasza@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-09-17 00:21:26]

Dry run: 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, 2016-09-17 02:15:13]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-17 02:15:15]

Dry run: Try jobs failed on following builders:
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)