Add helper class for tests using browser process hit testing

Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Committed: https://crrev.com/e975655cd86a5d47b0cfdb5c82039bcb5a561665
Cr-Commit-Position: refs/heads/master@{#378044}

[kenrb, 2016-02-26 17:59:37]

Description was changed from

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
==========

to

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[kenrb, 2016-02-26 18:06:10]

Description was changed from

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[kenrb, 2016-02-26 18:06:11]

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

[nasko, 2016-02-26 18:33:51]

LGTM with some nits.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
File content/browser/site_per_process_browsertest.cc (right):

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:177: class
SurfaceHitTestReadyNotifier {
nit: Shouldn't this be in content_browser_test_utils_internal.*? It might be
useful beyond this file as we write more tests. Alternatively it can be moved
later if needed, though it will lower discoverability being here.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:199:
base::TimeDelta::FromMilliseconds(10));
TestTimeouts::tiny_timeout()?

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:207: bool containsSurfaceId() {
ContainsSurfaceId

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:218: cc::SurfaceId
getRootSurfaceId() {
GetRootSurfaceId

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:238: base::Closure quit_;
Unused?

[kenrb, 2016-02-26 21:33:19]

nasko: Would you mind taking another look? I had to make some significant
changes for this to work on Mac.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
File content/browser/site_per_process_browsertest.cc (right):

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:177: class
SurfaceHitTestReadyNotifier {
On 2016/02/26 18:33:51, nasko wrote:
> nit: Shouldn't this be in content_browser_test_utils_internal.*? It might be
> useful beyond this file as we write more tests. Alternatively it can be moved
> later if needed, though it will lower discoverability being here.

Done.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:199:
base::TimeDelta::FromMilliseconds(10));
On 2016/02/26 18:33:51, nasko wrote:
> TestTimeouts::tiny_timeout()?

Done.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:207: bool containsSurfaceId() {
On 2016/02/26 18:33:51, nasko wrote:
> ContainsSurfaceId

Done.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:218: cc::SurfaceId
getRootSurfaceId() {
On 2016/02/26 18:33:51, nasko wrote:
> GetRootSurfaceId

This method is removed as part of the refactor.

https://codereview.chromium.org/1734403003/diff/1/content/browser/site_per_pr...
content/browser/site_per_process_browsertest.cc:238: base::Closure quit_;
On 2016/02/26 18:33:51, nasko wrote:
> Unused?

Done.

[nasko, 2016-02-26 22:11:28]

Still LGTM

[kenrb, 2016-02-26 22:40:32]

The CQ bit was checked by kenrb@chromium.org

[commit-bot: I haz the power, 2016-02-26 22:41:46]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1734403003/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1734403003/60001

[commit-bot: I haz the power, 2016-02-27 00:53:57]

Description was changed from

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

[commit-bot: I haz the power, 2016-02-27 00:53:58]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-02-27 00:56:36]

Description was changed from

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation
==========

to

==========
Add helper class for tests using browser process hit testing

Tests that rely on browser process hit testing are challenging to write
because hit testing will not work until a compositor frame has been
received from an out-of-process iframe's renderer, and a subsequent
compositor frame is received from its parent renderer that contains
the SurfaceId used by the OOPIF.

In trying to write a test that hit tests through nested OOPIFs, trying
to reason it out just based on compositor frame swap events becomes
the kind of concurrency problem that you studied in school and thought,
"Thankfully there are primitives to handle this sort of thing so I
won't have to remember any of this after the exam."

This CL adds a helper class that handles the problem by reaching into
the compositor's Surface code and signalling when a desired Surface
is available for hit testing.

BUG=589572
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Committed: https://crrev.com/e975655cd86a5d47b0cfdb5c82039bcb5a561665
Cr-Commit-Position: refs/heads/master@{#378044}
==========

[commit-bot: I haz the power, 2016-02-27 00:56:37]

Patchset 4 (id:??) landed as
https://crrev.com/e975655cd86a5d47b0cfdb5c82039bcb5a561665
Cr-Commit-Position: refs/heads/master@{#378044}