headless: Add gn settings for embedded build and headless mode

headless: Add gn setting headless mode

The is_headless setting configures the build in the same
way as the embedded flag in gyp.

Contributions by Alexander Timin <altimin@chromium.org>; and
Alex Clarke <alexclarke@chromium.org>;.

BUG=546953, 318413
Committed: https://crrev.com/a557457caccb3c0f60e5d35ccc8c55af5849d76a
Cr-Commit-Position: refs/heads/master@{#359096}

[Sami, 2015-10-23 11:53:54]

Description was changed from

==========
headless: Add gn settings for embedded and headless modes

BUG=546953
==========

to

==========
headless: Add gn settings for embedded and headless modes

The is_embedded setting configures the build in the same
way as the embedded flag in gyp. The headless setting
currently only checks that the embedded build has been
selected.

BUG=546953,318413
==========

[Sami, 2015-10-23 11:54:21]

Patchset #2 (id:20001) has been deleted

[Sami, 2015-10-23 11:54:28]

Patchset #1 (id:1) has been deleted

[Sami, 2015-10-23 11:55:00]

Description was changed from

==========
headless: Add gn settings for embedded and headless modes

The is_embedded setting configures the build in the same
way as the embedded flag in gyp. The headless setting
currently only checks that the embedded build has been
selected.

BUG=546953,318413
==========

to

==========
headless: Add gn settings for embedded and headless modes

The is_embedded setting configures the build in the same
way as the embedded flag in gyp. The headless setting
currently only checks that the embedded build has been
selected.

Contributions by Alexander Timin <altimin@chromium.org> and
Alex Clarke <alexclarke@chromium.org>.

BUG=546953,318413
==========

[Sami, 2015-10-23 12:51:17]

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

[Sami, 2015-10-23 12:52:37]

Hey John, would you be willing to review this one since you have some context
about the headless project?

[jam, 2015-10-23 16:57:34]

On 2015/10/23 12:52:37, Sami wrote:
> Hey John, would you be willing to review this one since you have some context
> about the headless project?

1) please add more background to the tracking bug
2) can we reduce new terms we introduce into the build? i.e. this is really
about headless, so can we use that in build files instead of embedded? embedded
is a bit vague; and in the project you're working on, it's not really embedded
per se.

[Sami, 2015-10-26 11:56:25]

On 2015/10/23 16:57:34, jam wrote:
> 1) please add more background to the tracking bug

Done. A real design doc is still in the works so I've added a short blurb for
now: https://code.google.com/p/chromium/issues/detail?id=546953

> 2) can we reduce new terms we introduce into the build? i.e. this is really
> about headless, so can we use that in build files instead of embedded?
embedded
> is a bit vague; and in the project you're working on, it's not really embedded
> per se.

The only reason we went with |is_embedded| + |enable_headless| instead of just
|is_headless| is that there's an existing |embedded| setting in gyp-land which
does exactly the same thing we want. I'm not really sure what that mode is used
for[1] and whether we want to support in GN or not. In case we don't, I've now
changed the patch to use |is_headless| throughout.

[1] I only found one FYI bot using it:
http://build.chromium.org/p/chromium.fyi/builders/Ozone%20ECS%20Linux

[Sami, 2015-10-26 11:57:17]

Description was changed from

==========
headless: Add gn settings for embedded and headless modes

The is_embedded setting configures the build in the same
way as the embedded flag in gyp. The headless setting
currently only checks that the embedded build has been
selected.

Contributions by Alexander Timin <altimin@chromium.org> and
Alex Clarke <alexclarke@chromium.org>.

BUG=546953,318413
==========

to

==========
headless: Add gn setting headless mode

The is_headless setting configures the build in the same
way as the embedded flag in gyp.

Contributions by Alexander Timin <altimin@chromium.org> and
Alex Clarke <alexclarke@chromium.org>.

BUG=546953,318413
==========

[jam, 2015-10-26 17:05:32]

https://codereview.chromium.org/1410883007/diff/120001/build/config/features.gni
File build/config/features.gni (right):

https://codereview.chromium.org/1410883007/diff/120001/build/config/features....
build/config/features.gni:29: enable_pdf = !is_android && !is_ios &&
!is_chromecast && !is_headless
it's not clear to me why this should be the cast; i.e. if someone is using
headless for generating previews, maybe they do want to run plugins?

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni
File build/config/ui.gni (right):

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:23: use_ash = (is_win || is_linux) && !is_chromecast &&
!is_headless
curious, why?

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:32: use_aura = is_win || is_linux || is_chromecast ||
is_headless
is this really necessary? i.e. you're targeting linux for now right?

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:36: !is_chromecast && !is_headless
curious why?

https://codereview.chromium.org/1410883007/diff/120001/build/linux/BUILD.gn
File build/linux/BUILD.gn (right):

https://codereview.chromium.org/1410883007/diff/120001/build/linux/BUILD.gn#n...
build/linux/BUILD.gn:14: use_system_fontconfig = !is_chromecast && !is_headless
why?

[Sami, 2015-10-26 18:09:06]

https://codereview.chromium.org/1410883007/diff/120001/build/config/features.gni
File build/config/features.gni (right):

https://codereview.chromium.org/1410883007/diff/120001/build/config/features....
build/config/features.gni:29: enable_pdf = !is_android && !is_ios &&
!is_chromecast && !is_headless
On 2015/10/26 17:05:31, jam wrote:
> it's not clear to me why this should be the cast; i.e. if someone is using
> headless for generating previews, maybe they do want to run plugins?

I suppose you're right. I was thinking NPAPI probably can't run headless but
PPAPI should be fine.

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni
File build/config/ui.gni (right):

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:23: use_ash = (is_win || is_linux) && !is_chromecast &&
!is_headless
On 2015/10/26 17:05:31, jam wrote:
> curious, why?

Is there a reason why we'd want to have a full desktop shell when running
headless? It would seem like window decorations etc. would just get in the way?

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:32: use_aura = is_win || is_linux || is_chromecast ||
is_headless
On 2015/10/26 17:05:31, jam wrote:
> is this really necessary? i.e. you're targeting linux for now right?

Yes, but don't you need to enable Aura to be able to use Ozone in the first
place?

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:36: !is_chromecast && !is_headless
On 2015/10/26 17:05:31, jam wrote:
> curious why?

Similarly to disabling ash, this gets rid of browser chrome which I don't think
we need.

https://codereview.chromium.org/1410883007/diff/120001/build/linux/BUILD.gn
File build/linux/BUILD.gn (right):

https://codereview.chromium.org/1410883007/diff/120001/build/linux/BUILD.gn#n...
build/linux/BUILD.gn:14: use_system_fontconfig = !is_chromecast && !is_headless
On 2015/10/26 17:05:32, jam wrote:
> why?

Right, I suppose that choice is really orthogonal. Removed.

[jam, 2015-10-30 16:15:49]

sorry for the delay, i missed your response.

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni
File build/config/ui.gni (right):

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:23: use_ash = (is_win || is_linux) && !is_chromecast &&
!is_headless
On 2015/10/26 18:09:06, Sami wrote:
> On 2015/10/26 17:05:31, jam wrote:
> > curious, why?
> 
> Is there a reason why we'd want to have a full desktop shell when running
> headless? It would seem like window decorations etc. would just get in the
way?

good point, ok

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:32: use_aura = is_win || is_linux || is_chromecast ||
is_headless
On 2015/10/26 18:09:06, Sami wrote:
> On 2015/10/26 17:05:31, jam wrote:
> > is this really necessary? i.e. you're targeting linux for now right?
> 
> Yes, but don't you need to enable Aura to be able to use Ozone in the first
> place?

if so, then this should key off ozone. but ozone is only on linux, so really
it's by definition that aura is on (i.e. no need for changes to this line)

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:36: !is_chromecast && !is_headless
On 2015/10/26 18:09:06, Sami wrote:
> On 2015/10/26 17:05:31, jam wrote:
> > curious why?
> 
> Similarly to disabling ash, this gets rid of browser chrome which I don't
think
> we need.

if the ui isn't built using views, then which toolkit is used? i think what you
really want is not to build any UI code. but this is changing the toolkit.. i.e.
i'm worried that this introduces another platform to support: linux without
views.

[Sami, 2015-11-05 19:08:17]

Sorry to have left this hanging. PTAL.

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni
File build/config/ui.gni (right):

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:32: use_aura = is_win || is_linux || is_chromecast ||
is_headless
On 2015/10/30 16:15:49, jam wrote:
> On 2015/10/26 18:09:06, Sami wrote:
> > On 2015/10/26 17:05:31, jam wrote:
> > > is this really necessary? i.e. you're targeting linux for now right?
> > 
> > Yes, but don't you need to enable Aura to be able to use Ozone in the first
> > place?
> 
> if so, then this should key off ozone. but ozone is only on linux, so really
> it's by definition that aura is on (i.e. no need for changes to this line)

Ah, got it, is_linux should be enough. Removed.

https://codereview.chromium.org/1410883007/diff/120001/build/config/ui.gni#ne...
build/config/ui.gni:36: !is_chromecast && !is_headless
On 2015/10/30 16:15:49, jam wrote:
> On 2015/10/26 18:09:06, Sami wrote:
> > On 2015/10/26 17:05:31, jam wrote:
> > > curious why?
> > 
> > Similarly to disabling ash, this gets rid of browser chrome which I don't
> think
> > we need.
> 
> if the ui isn't built using views, then which toolkit is used? i think what
you
> really want is not to build any UI code. but this is changing the toolkit..
i.e.
> i'm worried that this introduces another platform to support: linux without
> views.

We've explored this a little more now and I think you're right: the gyp
equivalent this was copied from was there only because they wanted to build a
stripped down version of Content Shell which didn't use Views. Since we're not
going to use Content Shell, there's no need for this setting either. Removed.

[jam, 2015-11-09 16:36:58]

lgtm with one comment

https://codereview.chromium.org/1410883007/diff/170001/chrome/test/BUILD.gn
File chrome/test/BUILD.gn (right):

https://codereview.chromium.org/1410883007/diff/170001/chrome/test/BUILD.gn#n...
chrome/test/BUILD.gn:1113: if (!is_chromeos && use_aura && use_ash) {
you don't need this. use_aura requires use_aura

[altimin, 2015-11-09 17:20:09]

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

[altimin, 2015-11-09 17:20:10]

https://codereview.chromium.org/1410883007/diff/170001/build/config/ui.gni
File build/config/ui.gni (right):

https://codereview.chromium.org/1410883007/diff/170001/build/config/ui.gni#ne...
build/config/ui.gni:36: (is_mac || is_win || is_chromeos || use_aura) &&
!is_chromecast
It doesn't compile in headless mode now without !is_headless here.

[altimin, 2015-11-09 17:20:27]

altimin@chromium.org changed reviewers:
- altimin@chromium.org

[Sami, 2015-11-11 15:47:42]

https://codereview.chromium.org/1410883007/diff/170001/build/config/ui.gni
File build/config/ui.gni (right):

https://codereview.chromium.org/1410883007/diff/170001/build/config/ui.gni#ne...
build/config/ui.gni:36: (is_mac || is_win || is_chromeos || use_aura) &&
!is_chromecast
On 2015/11/09 17:20:10, altimin wrote:
> It doesn't compile in headless mode now without !is_headless here.

I believe this is just an artifact of the way our prototype uses content shell
and should go away once we fix that. If not, let's tweak this later.

https://codereview.chromium.org/1410883007/diff/170001/chrome/test/BUILD.gn
File chrome/test/BUILD.gn (right):

https://codereview.chromium.org/1410883007/diff/170001/chrome/test/BUILD.gn#n...
chrome/test/BUILD.gn:1113: if (!is_chromeos && use_aura && use_ash) {
On 2015/11/09 16:36:58, jam wrote:
> you don't need this. use_aura requires use_aura

Thanks, removed.

[Sami, 2015-11-11 15:47:51]

The CQ bit was checked by skyostil@chromium.org

[Sami, 2015-11-11 15:47:52]

The patchset sent to the CQ was uploaded after l-g-t-m from jam@chromium.org
Link to the patchset: https://codereview.chromium.org/1410883007/#ps210001
(title: "Keep use_cups overridable.")

[commit-bot: I haz the power, 2015-11-11 15:47:59]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1410883007/210001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1410883007/210001

[commit-bot: I haz the power, 2015-11-11 16:24:32]

Committed patchset #10 (id:210001)

[commit-bot: I haz the power, 2015-11-12 19:57:44]

Patchset 10 (id:??) landed as
https://crrev.com/a557457caccb3c0f60e5d35ccc8c55af5849d76a
Cr-Commit-Position: refs/heads/master@{#359096}