Define cc::RendererCapabilities in its own header, fix up includes

Define cc::RendererCapabilities in its own header, fix up includes

cc::RendererCapabilities was defined in cc/trees/layer_tree_host.h
which had a few subtle implications. First, cc/output/renderer.h needed
to #include layer_tree_host even though cc/output/ logically should not
depend on cc/trees/. Second, since layer_tree_host.h includes pretty
much the entire world, everything that was #including renderer.h was
implicitly pulling in many other headers that it may have used. In the
spirit of IWYU these files should either include or forward declare
types they use.

This puts RendererCapabilities in cc/output/renderer_capabilities and
fixes up includes in many other files.

R=danakj@chromium.org

Committed: https://crrev.com/9b8fda3b7b24371afc2c2ed05d1c6a547cc6a786
Cr-Commit-Position: refs/heads/master@{#320767}

[danakj, 2015-03-13 17:49:26]

LGTM

https://codereview.chromium.org/1007593002/diff/1/cc/BUILD.gn
File cc/BUILD.gn (right):

https://codereview.chromium.org/1007593002/diff/1/cc/BUILD.gn#newcode288
cc/BUILD.gn:288: "output/renderer_capabilities.cc",
Add to the gyp files too please.

[jamesr, 2015-03-16 18:06:03]

The CQ bit was checked by jamesr@chromium.org

[jamesr, 2015-03-16 18:06:03]

The patchset sent to the CQ was uploaded after l-g-t-m from danakj@chromium.org
Link to the patchset: https://codereview.chromium.org/1007593002/#ps20001
(title: "gyp")

[commit-bot: I haz the power, 2015-03-16 18:06:28]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1007593002/20001

[commit-bot: I haz the power, 2015-03-16 19:11:14]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-03-16 19:12:41]

Patchset 2 (id:??) landed as
https://crrev.com/9b8fda3b7b24371afc2c2ed05d1c6a547cc6a786
Cr-Commit-Position: refs/heads/master@{#320767}