Make //ui/latency_info a public dep of //ui/events_base

Make //ui/latency_info a public dep of //ui/events_base

This CL also uses export_dependent_settings for the gyp build.

This is a speculative fix for some link issues on Windows builds.

BUG=604034

[Fady Samuel, 2016-04-17 03:25:22]

fsamuel@chromium.org changed reviewers:
+ thakis@chromium.org

[Fady Samuel, 2016-04-17 03:25:58]

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

[commit-bot: I haz the power, 2016-04-17 03:26:09]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1893153002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1893153002/1

[Nico, 2016-04-17 03:31:06]

https://codereview.chromium.org/1893153002/diff/1/ui/events/BUILD.gn
File ui/events/BUILD.gn (right):

https://codereview.chromium.org/1893153002/diff/1/ui/events/BUILD.gn#newcode73
ui/events/BUILD.gn:73: "//ui/latency_info",
event.h (which includes latency_info.h) is in events, not events_base. why isn't
this a public dep of events instead?

Also, maybe Event could have a unique_ptr<LatencyInfo>, then this doesn't have
to be a public dep but could be a private dep maybe?

[Fady Samuel, 2016-04-17 03:38:10]

On 2016/04/17 03:31:06, Nico wrote:
> https://codereview.chromium.org/1893153002/diff/1/ui/events/BUILD.gn
> File ui/events/BUILD.gn (right):
> 
> https://codereview.chromium.org/1893153002/diff/1/ui/events/BUILD.gn#newcode73
> ui/events/BUILD.gn:73: "//ui/latency_info",
> event.h (which includes latency_info.h) is in events, not events_base. why
isn't
> this a public dep of events instead?
> 
> Also, maybe Event could have a unique_ptr<LatencyInfo>, then this doesn't have
> to be a public dep but could be a private dep maybe?

latency_info.cc used to be part of the events.gyp:events_base target.

Maybe this is a bit of a hack but I'm trying to capture the existing semantics
without updating all the targets that depend on events_base. 

If latency_info is a public_dep of events_base then anything that depends on
events_base gets latency_info for "free". I was hoping this would avoid the
whack-a-mole issue.

[Nico, 2016-04-17 03:46:48]

as far as i can tell, event.h is the only header using latencyinfo
(https://code.google.com/p/chromium/codesearch#search/&q=latencyinfo%20file:ui...).
If that stops using it (in the header -- using a unique_ptr), latency_info can
be a private dep – is there a drawback to this? If there is, I think making this
a public_dep of event instead of event_base should work too and is "more
correct".

[Fady Samuel, 2016-04-17 03:54:32]

On 2016/04/17 03:46:48, Nico wrote:
> as far as i can tell, event.h is the only header using latencyinfo
>
(https://code.google.com/p/chromium/codesearch#search/&q=latencyinfo%20file:ui...).
> If that stops using it (in the header -- using a unique_ptr), latency_info can
> be a private dep – is there a drawback to this? If there is, I think
making this
> a public_dep of event instead of event_base should work too and is "more
> correct".

The concern I have is there might be targets that depend on ui/events only to
use latency_info that I missed. By making it a public_dep then that would fix
things. I don't have access to a windows machine at the moment. I'll test on
Monday.

[commit-bot: I haz the power, 2016-04-17 04:36:21]

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

[commit-bot: I haz the power, 2016-04-17 04:36:22]

Dry run: This issue passed the CQ dry run.