Move us fully from gfx:: over to skia types for consistency.

Move us fully from gfx:: over to skia types for consistency.


BUG=92085
TEST=BUILD

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=103523

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=103533

[dmac, 2011-09-22 16:36:58]

Don't want to submit this just yet because of the amount of files it touches and
the M15 stability, but wanted to get this reviewed since it is a large change.

[Sergey Ulanov, 2011-09-22 16:50:17]

I suggest that we move all code to gfx types instead. gfx types are simpler and
have fewer dependencies, so it will be easier to port client code to NaCl in the
future. For example SkMath.h has some inline assembler code, which will make it
harder to port it to NaCl.

[dmaclach1, 2011-09-22 16:57:33]

Sergey, we moved to Skia types because we wanted the functionality of
SkRegions for calculating minimum rectangles for updates.

On Thu, Sep 22, 2011 at 09:50,  <sergeyu@chromium.org> wrote:
> I suggest that we move all code to gfx types instead. gfx types are simpler
> and
> have fewer dependencies, so it will be easier to port client code to NaCl in
> the
> future. For example SkMath.h has some inline assembler code, which will make
> it
> harder to port it to NaCl.
>
> http://codereview.chromium.org/7992011/
>

[Sergey Ulanov, 2011-09-22 16:57:59]

On 2011/09/22 16:50:17, sergeyu wrote:
> I suggest that we move all code to gfx types instead. gfx types are simpler
and
> have fewer dependencies, so it will be easier to port client code to NaCl in
the
> future. For example SkMath.h has some inline assembler code, which will make
it
> harder to port it to NaCl.

I take this back. Looked at the gfx code and may be harder to port to NaCl than
Skia, e.g. it depend on some stuff in src/base.

Can you please also update DEPS files to make sure we don't depend on ui/gfx
anymore?

[dmaclach1, 2011-09-22 17:51:35]

> Can you please also update DEPS files to make sure we don't depend on ui/gfx
> anymore?

Which DEPS file are you looking at sergey?

am I going blind? ;-)

[Sergey Ulanov, 2011-09-22 17:59:47]

On Thu, Sep 22, 2011 at 10:51 AM, David Maclachlan <dmaclach@google.com>wrote:

> > Can you please also update DEPS files to make sure we don't depend on
> ui/gfx
> > anymore?
>
> Which DEPS file are you looking at sergey?
>

remoting/base/DEPS and remoting/client/DEPS

Also, remoting/host/DEPS has dependency on src/ui. I'm not sure if we need
it. Probably host code depends on some other code in ui beside ui/gfx.


>
> am I going blind? ;-)
>

[dmac, 2011-09-28 15:58:56]

On 2011/09/22 17:59:47, sergeyu wrote:
> On Thu, Sep 22, 2011 at 10:51 AM, David Maclachlan <dmaclach@google.com>wrote:
> 
> > > Can you please also update DEPS files to make sure we don't depend on
> > ui/gfx
> > > anymore?
> >
> > Which DEPS file are you looking at sergey?
> >
> 
> remoting/base/DEPS and remoting/client/DEPS
> 
> Also, remoting/host/DEPS has dependency on src/ui. I'm not sure if we need
> it. Probably host code depends on some other code in ui beside ui/gfx.
> 
> 
> >
> > am I going blind? ;-)
> >

OK, All fixed up.

[Sergey Ulanov, 2011-09-28 18:09:27]

a few nits. LGTM otherwise.

http://codereview.chromium.org/7992011/diff/10001/remoting/base/decoder_vp8.cc
File remoting/base/decoder_vp8.cc (right):

http://codereview.chromium.org/7992011/diff/10001/remoting/base/decoder_vp8.c...
remoting/base/decoder_vp8.cc:95: ::remoting::Rect remoting_rect =
packet->dirty_rects(i);
Do we need to specify namespace for Rect?

http://codereview.chromium.org/7992011/diff/10001/remoting/base/encoder_vp8.h
File remoting/base/encoder_vp8.h (right):

http://codereview.chromium.org/7992011/diff/10001/remoting/base/encoder_vp8.h...
remoting/base/encoder_vp8.h:19: typedef std::vector<SkIRect> RectVector;
This duplicates typedef in decoder.h. given that it is used only internally in
EncoderVp8 maybe define it as private inside EncoderVp8?

http://codereview.chromium.org/7992011/diff/10001/remoting/host/DEPS
File remoting/host/DEPS (left):

http://codereview.chromium.org/7992011/diff/10001/remoting/host/DEPS#oldcode2
remoting/host/DEPS:2: "+ui",
I think we still need this: remoting/host/event_executor*.cc include
ui/base/keycodes/keyboard_codes.h
Maybe change this to ui/base