Implement a runtime headless mode for Linux

Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc: https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys90/edit#

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/b354f8865623b8ca8da43f7ef37332bdb586dd82
Cr-Commit-Position: refs/heads/master@{#438239}

[Sami, 2016-10-28 12:30:38]

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

[commit-bot: I haz the power, 2016-10-28 12:30:49]

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, 2016-10-28 12:37:58]

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

[commit-bot: I haz the power, 2016-10-28 12:37:59]

Dry run: Try jobs failed on following builders:
  chromium_presubmit on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[Sami, 2016-10-28 16:47:30]

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

[commit-bot: I haz the power, 2016-10-28 16:47:55]

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

[Sami, 2016-10-28 16:58:42]

Description was changed from

==========
DO NOT LAND: Extremely sketchy --headless runtime switch implementation
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a --headless command line
switch) that makes it possible to use a regular Chrome binary as a headless.
When the binary is launched with this switch, the main entrypoint calls into the
Headless Shell entrypoint, effectively starting that shell instead of Chrome.

To make this possible we must remove the dependency from Headless to Ozone,
because Ozone is a build-time feature which is generally not enabled for regular
Chrome builds. In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input entrypoints to do
something appropriate in headless mode.

BUG=612904
==========

[Sami, 2016-10-28 16:59:25]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a --headless command line
switch) that makes it possible to use a regular Chrome binary as a headless.
When the binary is launched with this switch, the main entrypoint calls into the
Headless Shell entrypoint, effectively starting that shell instead of Chrome.

To make this possible we must remove the dependency from Headless to Ozone,
because Ozone is a build-time feature which is generally not enabled for regular
Chrome builds. In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input entrypoints to do
something appropriate in headless mode.

BUG=612904
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds. In practice this means
implementing a new headless-specific
WindowTreeHost and modifying various graphics and input entrypoints to do
something appropriate in headless mode.

BUG=612904
==========

[Sami, 2016-10-28 17:00:21]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds. In practice this means
implementing a new headless-specific
WindowTreeHost and modifying various graphics and input entrypoints to do
something appropriate in headless mode.

BUG=612904
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

BUG=612904
==========

[commit-bot: I haz the power, 2016-10-28 17:05:29]

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

[commit-bot: I haz the power, 2016-10-28 17:05:30]

Dry run: Try jobs failed on following builders:
  linux_chromium_chromeos_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Eric Seckler, 2016-11-28 15:51:48]

eseckler@chromium.org changed reviewers:
+ eseckler@chromium.org

[Eric Seckler, 2016-11-28 15:51:49]

https://codereview.chromium.org/1991953002/diff/120001/content/gpu/gpu_main.cc
File content/gpu/gpu_main.cc (right):

https://codereview.chromium.org/1991953002/diff/120001/content/gpu/gpu_main.c...
content/gpu/gpu_main.cc:204: base::MessageLoop
main_message_loop(base::MessageLoop::TYPE_DEFAULT);
think this won't compile, needs to do main_message_loop.reset() as well

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
File headless/lib/browser/headless_browser_impl.cc (right):

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
headless/lib/browser/headless_browser_impl.cc:126: new
HeadlessWindowTreeHost(gfx::Rect(options()->window_size)));
I suspect that this is the reason for the screenshots not coming back, but I
can't really tell what's missing in HeadlessWindowTreeHost either :(

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
File headless/lib/browser/headless_window_tree_host.cc (right):

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.cc:43: bounds_ = bounds;
You might want to call OnHostMoved() / OnHostResized() here, see

https://cs.chromium.org/chromium/src/ui/aura/window_tree_host_x11.cc?sq=packa...

[Sami, 2016-11-28 18:00:37]

https://codereview.chromium.org/1991953002/diff/120001/content/gpu/gpu_main.cc
File content/gpu/gpu_main.cc (right):

https://codereview.chromium.org/1991953002/diff/120001/content/gpu/gpu_main.c...
content/gpu/gpu_main.cc:204: base::MessageLoop
main_message_loop(base::MessageLoop::TYPE_DEFAULT);
On 2016/11/28 15:51:48, Eric Seckler wrote:
> think this won't compile, needs to do main_message_loop.reset() as well

Hmm yeah, I think I messed this up while rebasing. Thanks.

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
File headless/lib/browser/headless_browser_impl.cc (right):

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
headless/lib/browser/headless_browser_impl.cc:126: new
HeadlessWindowTreeHost(gfx::Rect(options()->window_size)));
On 2016/11/28 15:51:48, Eric Seckler wrote:
> I suspect that this is the reason for the screenshots not coming back, but I
> can't really tell what's missing in HeadlessWindowTreeHost either :(

No worries, turns out the problem was that we need to return false from
ShouldCreateGpuCompositorFrameSink().

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
File headless/lib/browser/headless_window_tree_host.cc (right):

https://codereview.chromium.org/1991953002/diff/120001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.cc:43: bounds_ = bounds;
On 2016/11/28 15:51:48, Eric Seckler wrote:
> You might want to call OnHostMoved() / OnHostResized() here, see
> 
>
https://cs.chromium.org/chromium/src/ui/aura/window_tree_host_x11.cc?sq=packa...

Thanks, done!

[Sami, 2016-11-28 18:49:29]

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

[commit-bot: I haz the power, 2016-11-28 18:49:57]

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, 2016-11-28 19:23:51]

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

[commit-bot: I haz the power, 2016-11-28 19:23:51]

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

[Sami, 2016-11-29 17:42:43]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

BUG=612904
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
==========

[Sami, 2016-11-29 17:44:35]

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

[commit-bot: I haz the power, 2016-11-29 17:44:59]

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, 2016-11-29 18:01:15]

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

[commit-bot: I haz the power, 2016-11-29 18:01:16]

Dry run: Try jobs failed on following builders:
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[Sami, 2016-11-29 18:18:16]

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

[commit-bot: I haz the power, 2016-11-29 18:18:34]

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, 2016-11-29 18:34:51]

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

[commit-bot: I haz the power, 2016-11-29 18:34:52]

Dry run: Try jobs failed on following builders:
  chromeos_daisy_chromium_compile_only_ng on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_daisy_...)

[Sami, 2016-11-30 17:12:46]

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

[commit-bot: I haz the power, 2016-11-30 17:13:08]

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, 2016-11-30 17:32:31]

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

[commit-bot: I haz the power, 2016-11-30 17:32:32]

Dry run: Try jobs failed on following builders:
  linux_chromium_chromeos_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Sami, 2016-11-30 19:04:27]

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

[commit-bot: I haz the power, 2016-11-30 19:04:57]

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, 2016-11-30 19:42:15]

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

[commit-bot: I haz the power, 2016-11-30 19:42:17]

Dry run: Try jobs failed on following builders:
  mac_chromium_compile_dbg_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_comp...)

[Sami, 2016-12-01 16:20:43]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[Sami, 2016-12-01 16:20:50]

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

[commit-bot: I haz the power, 2016-12-01 16:21:04]

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, 2016-12-01 16:48:15]

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

[commit-bot: I haz the power, 2016-12-01 16:48:16]

Dry run: Try jobs failed on following builders:
  mac_optional_gpu_tests_rel on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_optional_gpu_...)

[Sami, 2016-12-01 18:15:52]

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

[commit-bot: I haz the power, 2016-12-01 18:16:13]

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

[Sami, 2016-12-01 18:50:28]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[Sami, 2016-12-01 18:50:28]

skyostil@chromium.org changed reviewers:
+ sadrul@chromium.org

[Sami, 2016-12-01 18:50:57]

Okay, the bots seem happier now. Eric, wanna look at headless/ again? Sadrul,
could you do the first pass over ui/ please?

[commit-bot: I haz the power, 2016-12-01 19:24:53]

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

[commit-bot: I haz the power, 2016-12-01 19:24:55]

Dry run: This issue passed the CQ dry run.

[Eric Seckler, 2016-12-02 10:00:56]

On 2016/12/01 18:50:57, Sami wrote:
> Okay, the bots seem happier now. Eric, wanna look at headless/ again? Sadrul,
> could you do the first pass over ui/ please?

headless/ lgtm % nits :)

[Eric Seckler, 2016-12-02 10:01:15]

and here go the nits ..

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
File headless/lib/browser/headless_window_tree_host.h (right):

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.h:10: #include <memory>
nit: unnecessary?

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.h:13: #include
"ui/aura/aura_export.h"
nit: unnecessary?

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.h:16: #include
"ui/gfx/geometry/insets.h"
nit: unnecessary?

[Sami, 2016-12-02 11:30:09]

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

[Sami, 2016-12-02 11:30:24]

Thanks Eric, all nits addressed.

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
File headless/lib/browser/headless_window_tree_host.h (right):

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.h:10: #include <memory>
On 2016/12/02 10:01:14, Eric Seckler wrote:
> nit: unnecessary?

Done.

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.h:13: #include
"ui/aura/aura_export.h"
On 2016/12/02 10:01:14, Eric Seckler wrote:
> nit: unnecessary?

Done.

https://codereview.chromium.org/1991953002/diff/260001/headless/lib/browser/h...
headless/lib/browser/headless_window_tree_host.h:16: #include
"ui/gfx/geometry/insets.h"
On 2016/12/02 10:01:14, Eric Seckler wrote:
> nit: unnecessary?

Done.

[commit-bot: I haz the power, 2016-12-02 11:30:32]

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, 2016-12-02 12:27:09]

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

[commit-bot: I haz the power, 2016-12-02 12:27:10]

Dry run: This issue passed the CQ dry run.

[Sami, 2016-12-06 10:40:52]

skyostil@chromium.org changed reviewers:
+ sky@chromium.org
- sadrul@chromium.org

[Sami, 2016-12-06 10:41:44]

Ah, looks like Sadrul is out. Scott, would you like to look at the changes under
ui/?

[sky, 2016-12-06 14:27:32]

https://codereview.chromium.org/1991953002/diff/280001/ui/aura/env.cc
File ui/aura/env.cc (right):

https://codereview.chromium.org/1991953002/diff/280001/ui/aura/env.cc#newcode178
ui/aura/env.cc:178: if (RunningInHeadlessMode())
Would it make more sense for a headless PlatformEventsource?

[Sami, 2016-12-07 11:49:54]

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

[commit-bot: I haz the power, 2016-12-07 11:50:15]

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

[Sami, 2016-12-07 11:51:26]

https://codereview.chromium.org/1991953002/diff/280001/ui/aura/env.cc
File ui/aura/env.cc (right):

https://codereview.chromium.org/1991953002/diff/280001/ui/aura/env.cc#newcode178
ui/aura/env.cc:178: if (RunningInHeadlessMode())
On 2016/12/06 14:27:31, sky wrote:
> Would it make more sense for a headless PlatformEventsource?

Great idea! Done.

[commit-bot: I haz the power, 2016-12-07 13:07:45]

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

[commit-bot: I haz the power, 2016-12-07 13:07:46]

Dry run: This issue passed the CQ dry run.

[Sami, 2016-12-08 15:43:58]

sky@: Friendly ping.

[sky, 2016-12-08 16:16:36]

LGTM

https://codereview.chromium.org/1991953002/diff/300001/headless/lib/browser/h...
File headless/lib/browser/headless_platform_event_source.h (right):

https://codereview.chromium.org/1991953002/diff/300001/headless/lib/browser/h...
headless/lib/browser/headless_platform_event_source.h:17: };
private: DISALLOW...

https://codereview.chromium.org/1991953002/diff/300001/ui/aura/env.cc
File ui/aura/env.cc (right):

https://codereview.chromium.org/1991953002/diff/300001/ui/aura/env.cc#newcode21
ui/aura/env.cc:21: #include "ui/gfx/switches.h"
Is this include still needed?

[Sami, 2016-12-08 16:57:00]

skyostil@chromium.org changed reviewers:
+ jochen@chromium.org

[Sami, 2016-12-08 16:58:16]

skyostil@chromium.org changed reviewers:
+ zmo@chromium.org

[Sami, 2016-12-08 16:58:36]

Thanks! Jochen, could you check chrome/ and content/ please? Also +zmo@ for
gpu/.

https://codereview.chromium.org/1991953002/diff/300001/headless/lib/browser/h...
File headless/lib/browser/headless_platform_event_source.h (right):

https://codereview.chromium.org/1991953002/diff/300001/headless/lib/browser/h...
headless/lib/browser/headless_platform_event_source.h:17: };
On 2016/12/08 16:16:35, sky wrote:
> private: DISALLOW...

Done.

https://codereview.chromium.org/1991953002/diff/300001/ui/aura/env.cc
File ui/aura/env.cc (right):

https://codereview.chromium.org/1991953002/diff/300001/ui/aura/env.cc#newcode21
ui/aura/env.cc:21: #include "ui/gfx/switches.h"
On 2016/12/08 16:16:35, sky wrote:
> Is this include still needed?

Oops, thanks. Removed.

[Sami, 2016-12-08 16:58:51]

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

[commit-bot: I haz the power, 2016-12-08 16:59:13]

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, 2016-12-08 20:12:08]

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

[commit-bot: I haz the power, 2016-12-08 20:12:09]

Dry run: Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[jochen (gone - plz use gerrit), 2016-12-09 08:25:00]

https://codereview.chromium.org/1991953002/diff/320001/chrome/DEPS
File chrome/DEPS (right):

https://codereview.chromium.org/1991953002/diff/320001/chrome/DEPS#newcode37
chrome/DEPS:37: "+headless",
please move this to more specific files, so that e.g. chrome/browser can only
include headless/public/browser or whatever the right path is

https://codereview.chromium.org/1991953002/diff/320001/headless/BUILD.gn
File headless/BUILD.gn (right):

https://codereview.chromium.org/1991953002/diff/320001/headless/BUILD.gn#newc...
headless/BUILD.gn:242: "//content/public/renderer",
depending on both browser and renderer looks fishy, how does that work?

[Sami, 2016-12-09 14:05:19]

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

[commit-bot: I haz the power, 2016-12-09 14:05:37]

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

[Sami, 2016-12-09 14:06:46]

https://codereview.chromium.org/1991953002/diff/320001/chrome/DEPS
File chrome/DEPS (right):

https://codereview.chromium.org/1991953002/diff/320001/chrome/DEPS#newcode37
chrome/DEPS:37: "+headless",
On 2016/12/09 08:25:00, jochen wrote:
> please move this to more specific files, so that e.g. chrome/browser can only
> include headless/public/browser or whatever the right path is

Good point, done.

https://codereview.chromium.org/1991953002/diff/320001/headless/BUILD.gn
File headless/BUILD.gn (right):

https://codereview.chromium.org/1991953002/diff/320001/headless/BUILD.gn#newc...
headless/BUILD.gn:242: "//content/public/renderer",
On 2016/12/09 08:25:00, jochen wrote:
> depending on both browser and renderer looks fishy, how does that work?

Turns out these were some ancient things we forgot to clean up. The renderer dep
is now gone.

[jochen (gone - plz use gerrit), 2016-12-09 14:30:16]

lgtm with final comment addressed

https://codereview.chromium.org/1991953002/diff/340001/headless/DEPS
File headless/DEPS (right):

https://codereview.chromium.org/1991953002/diff/340001/headless/DEPS#newcode2
headless/DEPS:2: "+content/public",
could you restrict this to browser & common as well?

[Sami, 2016-12-09 14:36:50]

https://codereview.chromium.org/1991953002/diff/340001/headless/DEPS
File headless/DEPS (right):

https://codereview.chromium.org/1991953002/diff/340001/headless/DEPS#newcode2
headless/DEPS:2: "+content/public",
On 2016/12/09 14:30:16, jochen wrote:
> could you restrict this to browser & common as well?

Sure, done. (Needed app/ and test/ too.)

[Sami, 2016-12-12 11:06:25]

skyostil@chromium.org changed reviewers:
+ vmiura@chromium.org

[Sami, 2016-12-12 11:07:04]

Not sure if zmo@ is around...Victor, could you have a look at gpu/ please?

[Zhenyao Mo, 2016-12-12 18:21:43]

https://codereview.chromium.org/1991953002/diff/360001/gpu/ipc/service/gpu_in...
File gpu/ipc/service/gpu_init.cc (right):

https://codereview.chromium.org/1991953002/diff/360001/gpu/ipc/service/gpu_in...
gpu/ipc/service/gpu_init.cc:131: !command_line.HasSwitch(switches::kHeadless) &&
Can you explain why this is needed?

[Sami, 2016-12-12 20:04:13]

https://codereview.chromium.org/1991953002/diff/360001/gpu/ipc/service/gpu_in...
File gpu/ipc/service/gpu_init.cc (right):

https://codereview.chromium.org/1991953002/diff/360001/gpu/ipc/service/gpu_in...
gpu/ipc/service/gpu_init.cc:131: !command_line.HasSwitch(switches::kHeadless) &&
On 2016/12/12 18:21:43, Zhenyao Mo wrote:
> Can you explain why this is needed?

In Linux the watchdog thread has a dependency on the X11 display server, which
won't be there in headless mode. I decided the most straightforward way to deal
with that was to turn the watchdog off, since it doesn't really serve any
purpose in headless mode anyway (we're not currently trying to use any real GPU
there).

[Zhenyao Mo, 2016-12-12 21:30:34]

On 2016/12/12 20:04:13, Sami wrote:
>
https://codereview.chromium.org/1991953002/diff/360001/gpu/ipc/service/gpu_in...
> File gpu/ipc/service/gpu_init.cc (right):
> 
>
https://codereview.chromium.org/1991953002/diff/360001/gpu/ipc/service/gpu_in...
> gpu/ipc/service/gpu_init.cc:131: !command_line.HasSwitch(switches::kHeadless)
&&
> On 2016/12/12 18:21:43, Zhenyao Mo wrote:
> > Can you explain why this is needed?
> 
> In Linux the watchdog thread has a dependency on the X11 display server, which
> won't be there in headless mode. I decided the most straightforward way to
deal
> with that was to turn the watchdog off, since it doesn't really serve any
> purpose in headless mode anyway (we're not currently trying to use any real
GPU
> there).

gpu lgtm

[Sami, 2016-12-13 11:30:02]

The CQ bit was checked by skyostil@chromium.org

[Sami, 2016-12-13 11:30:02]

The patchset sent to the CQ was uploaded after l-g-t-m from
eseckler@chromium.org, sky@chromium.org, jochen@chromium.org, zmo@chromium.org
Link to the patchset: https://codereview.chromium.org/1991953002/#ps380001
(title: "Rebased")

[commit-bot: I haz the power, 2016-12-13 11:30:28]

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, 2016-12-13 13:25:50]

CQ is committing da patch.

Bot data: {"patchset_id": 380001, "attempt_start_ts": 1481628602152550,
"parent_rev": "c44259f35f2e4eeda421beb6620a0f49107145a9", "commit_rev":
"00e6c8fc8172aa8f03380bb315051c24c4d068eb"}

[commit-bot: I haz the power, 2016-12-13 13:26:26]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1991953002
==========

[commit-bot: I haz the power, 2016-12-13 13:26:27]

Committed patchset #20 (id:380001)

[commit-bot: I haz the power, 2016-12-13 13:29:25]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a 
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1991953002
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/eb0f3b0218478f30858dfa2078cb350ea259701f
Cr-Commit-Position: refs/heads/master@{#438152}
==========

[commit-bot: I haz the power, 2016-12-13 13:29:27]

Patchset 20 (id:??) landed as
https://crrev.com/eb0f3b0218478f30858dfa2078cb350ea259701f
Cr-Commit-Position: refs/heads/master@{#438152}

[Sami, 2016-12-13 15:47:51]

A revert of this CL (patchset #20 id:380001) has been created in
https://codereview.chromium.org/2570873002/ by skyostil@chromium.org.

The reason for reverting is: Unexpectedly added a couple of new static
initializers:
https://uberchromegw.corp.google.com/i/chromium/builders/Linux%20x64/builds/2....

[Sami, 2016-12-13 18:26:38]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/eb0f3b0218478f30858dfa2078cb350ea259701f
Cr-Commit-Position: refs/heads/master@{#438152}
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

[Sami, 2016-12-13 18:33:00]

Relanding now that the static initializers have been dealt with
(https://codereview.chromium.org/2577483002/).

[Sami, 2016-12-13 18:33:11]

The CQ bit was checked by skyostil@chromium.org

[commit-bot: I haz the power, 2016-12-13 18:34:14]

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, 2016-12-13 18:42:51]

CQ is committing da patch.

Bot data: {"patchset_id": 380001, "attempt_start_ts": 1481653991969100,
"parent_rev": "9042c078a4ad69191b0bacf253e78f1e101f42c0", "commit_rev":
"c2d8ec1beba3c9c8779f7723a8e80f8aedbbf10c"}

[commit-bot: I haz the power, 2016-12-13 18:43:18]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1991953002
==========

[commit-bot: I haz the power, 2016-12-13 18:43:20]

Committed patchset #20 (id:380001)

[commit-bot: I haz the power, 2016-12-13 18:45:22]

Description was changed from

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/1991953002
==========

to

==========
Implement a runtime headless mode for Linux

This patch implements a runtime headless mode (i.e., a
--headless command line switch) that makes it possible to
use a regular Chrome binary as a headless. When the binary
is launched with this switch, the main entrypoint calls
into the Headless Shell entrypoint, effectively starting
that shell instead of Chrome.

To make this possible we must remove the dependency from
Headless to Ozone, because Ozone is a build-time feature
which is generally not enabled for regular Chrome builds.
In practice this means implementing a new headless-specific
WindowTreeHost and modifying various graphics and input
entrypoints to do something appropriate in headless mode.

Since many of the modifications are in platform-specific
code, this initial patch only adds headless support in Linux.
Other platforms will be added later.

Design doc:
https://docs.google.com/document/d/1aIJUzQr3eougZQp90bp4mqGr5gY6hdUice8UPa-Ys...

BUG=612904
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Committed: https://crrev.com/b354f8865623b8ca8da43f7ef37332bdb586dd82
Cr-Commit-Position: refs/heads/master@{#438239}
==========

[commit-bot: I haz the power, 2016-12-13 18:45:24]

Patchset 20 (id:??) landed as
https://crrev.com/b354f8865623b8ca8da43f7ef37332bdb586dd82
Cr-Commit-Position: refs/heads/master@{#438239}