Tests of ctrl-click can't postMessage to opener - using custom text instead.

Tests of ctrl-click can't postMessage to opener - using custom text instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
communicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386
Review-Url: https://codereview.chromium.org/2689483007
Cr-Commit-Position: refs/heads/master@{#463400}
Committed: https://chromium.googlesource.com/chromium/src/+/69e056006f20f2973eafb2e95b7ea4395f3065db

[Łukasz Anforowicz, 2017-02-10 17:57:44]

Description was changed from

==========
Ctrl-Click cannot postMessage to opener - using setCustomTextOutput instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
comminicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386
==========

to

==========
Tests of ctrl-click can't postMessage to opener - using custom text instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
communicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386
==========

[Łukasz Anforowicz, 2017-02-10 17:57:55]

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

[commit-bot: I haz the power, 2017-02-10 17:58:35]

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

[Łukasz Anforowicz, 2017-02-10 18:01:58]

Description was changed from

==========
Tests of ctrl-click can't postMessage to opener - using custom text instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
communicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386
==========

to

==========
Tests of ctrl-click can't postMessage to opener - using custom text instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
communicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386
==========

[commit-bot: I haz the power, 2017-02-10 19:22:00]

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

[commit-bot: I haz the power, 2017-02-10 19:22:01]

Dry run: This issue passed the CQ dry run.

[Łukasz Anforowicz, 2017-04-10 18:05:26]

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

[commit-bot: I haz the power, 2017-04-10 18:06:12]

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

[Łukasz Anforowicz, 2017-04-10 18:10:06]

lukasza@chromium.org changed reviewers:
+ alexmos@chromium.org

[Łukasz Anforowicz, 2017-04-10 18:10:06]

alexmos@, could you PTAL?

1. The two modified tests simulate clicking on an anchor/link while holding down
a shift key.
2. Currently this puts the newly opened window in the same process as the old
frame and allows postMessage back to the opener.
3. #2 will change after fixing https://crbug.com/658386 in
https://codereview.chromium.org/2686943002 and
https://codereview.chromium.org/2680353005.

This CL makes sure the tests still work after #3.


To make sure that custom text dump changes propagate correctly, I had to fix up
the logic in blink_test_runner.cc (this logic was put there to avoid sending
updates that just come from initialization of blink test runner state).

[commit-bot: I haz the power, 2017-04-10 20:09:08]

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

[commit-bot: I haz the power, 2017-04-10 20:09:09]

Dry run: This issue passed the CQ dry run.

[alexmos, 2017-04-10 21:18:08]

Thanks, LGTM.  I agree that propagating LayoutTestRuntimeFlags changes from
secondary test windows is reasonable/desirable.

[Łukasz Anforowicz, 2017-04-10 21:20:16]

The CQ bit was checked by lukasza@chromium.org

[commit-bot: I haz the power, 2017-04-10 21:20:54]

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, 2017-04-10 21:30:03]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1491859216462610,
"parent_rev": "1d0058ca68075ed731ef2d3c150b9c3aba1534f6", "commit_rev":
"69e056006f20f2973eafb2e95b7ea4395f3065db"}

[commit-bot: I haz the power, 2017-04-10 21:30:59]

Description was changed from

==========
Tests of ctrl-click can't postMessage to opener - using custom text instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
communicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386
==========

to

==========
Tests of ctrl-click can't postMessage to opener - using custom text instead.

This CL fixes/refactors two layout tests that used to depend on ability
to find the opener via window.open, even in case of ctrl-click or
shift-click.  This dependency is incorrect and will be broken once
https://crbug.com/658386 is fixed.

The tests had to be refactored so that they don't depend on ability to
communicate back with the main test window.  After this CL the tests
report their results via testRunner.setCustomTextOutput (rather than
sending the results back to the main window using window.postMessage).

In the 2 tests here, the custom text dump is being set from a secondary
test window - the CL also needs to relax the condition that limited
propagation of LayoutTestRuntimeFlags to the main test window.  The
intent of the original condition should have covered both secondary
and main windows (i.e. the intent was to just check if the renderer
has received a LayoutTestRuntimeFlags).  It seems desirable overall
to also propagate LayoutTestRuntimeFlags changes originating from
secondary test windows.

BUG=658386

Review-Url: https://codereview.chromium.org/2689483007
Cr-Commit-Position: refs/heads/master@{#463400}
Committed:
https://chromium.googlesource.com/chromium/src/+/69e056006f20f2973eafb2e95b7e...
==========

[commit-bot: I haz the power, 2017-04-10 21:31:04]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/chromium/src/+/69e056006f20f2973eafb2e95b7e...