Header cleanup in chrome/renderer/.

Header cleanup in chrome/renderer/.

In renderer_webkit_client_impl, several impl classes got moved to the
implementation file.

BUG=none
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=53996

[Elliot Glaysher, 2010-07-27 20:10:59]

Most of this is just moving includes to cc files, but my moving of the impl
classes to the cc file in render_webkitclient_impl.cc also made initialization
lazy, which probably needs real review.

[darin (slow to review), 2010-07-27 20:26:28]

http://codereview.chromium.org/3032027/diff/2001/17
File chrome/renderer/render_process_impl.h (right):

http://codereview.chromium.org/3032027/diff/2001/17#newcode14
chrome/renderer/render_process_impl.h:14: };
nit: no semi-colon on the closing bracket of a namespace

http://codereview.chromium.org/3032027/diff/2001/23
File chrome/renderer/renderer_webkitclient_impl.cc (right):

http://codereview.chromium.org/3032027/diff/2001/23#newcode456
chrome/renderer/renderer_webkitclient_impl.cc:456: if
(!shared_worker_repository_.get())
have you verified that these getters are not accessed from background threads in
webkit?

btw, not very many .cc files include renderer_webkitclient_impl.h.  perhaps
"optimizing" that header is not worth it.

[Elliot Glaysher, 2010-07-27 20:57:00]

http://codereview.chromium.org/3032027/diff/2001/23
File chrome/renderer/renderer_webkitclient_impl.cc (right):

http://codereview.chromium.org/3032027/diff/2001/23#newcode456
chrome/renderer/renderer_webkitclient_impl.cc:456: if
(!shared_worker_repository_.get())
On 2010/07/27 20:26:29, darin wrote:
> have you verified that these getters are not accessed from background threads
in
> webkit?

I wasn't aware there were threading issues. I've changed these so they are built
in the constructor so the semantics should be identical to how construction
worked before.

> btw, not very many .cc files include renderer_webkitclient_impl.h.  perhaps
> "optimizing" that header is not worth it.

My current method of choosing what to optimize is looking at the slowest cc
files in chrome/, and then trying to minimize header
dependencies/un-inline/other stuff.

I didn't notice that this was only referenced from two files. I'd be fine
tossing the changes to this cc/h pair.

[jorlow, 2010-07-28 10:24:18]

LGTM (but probably wait for Darin's LGTM as well)

[darin (slow to review), 2010-07-28 16:55:47]

LGTM