Don't repaint tab when TabRenderData hasn't changed.

Don't repaint tab when TabRenderData hasn't changed.
 We are always repainting when new TabData is set even if there is no change.
 This CL reduces tab re-paints by 10~20% on cr48. With 4 tabs, its about 1~1.5seconds reduction.

BUG=none
TEST=This should not change any behavior except for better startup time. I also tested dynamic favicon (http://ajaxify.com/run/favicon).

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=76445

[oshima, 2011-02-26 00:18:29]

http://codereview.chromium.org/6599017/diff/2001/chrome/browser/ui/views/tabs...
File chrome/browser/ui/views/tabs/tab_renderer_data.h (right):

http://codereview.chromium.org/6599017/diff/2001/chrome/browser/ui/views/tabs...
chrome/browser/ui/views/tabs/tab_renderer_data.h:50: favicon.pixelRefOffset() ==
data.favicon.pixelRefOffset() &&
Brett, I'm assuming that chrome does not change
the data in SkPixelRef for favicons. Scott told me that you may know if my
assumption is correct. Let me know if you don't know and I'll find someone else.

[sky, 2011-02-26 01:11:19]

LGTM, but wait for Brett to say if that's the right way to compare SkBitmaps.

  -Scott

[oshima, 2011-02-26 01:15:37]

Yes, i'll wait for his response.

On Fri, Feb 25, 2011 at 5:11 PM, <sky@chromium.org> wrote:

> LGTM, but wait for Brett to say if that's the right way to compare
> SkBitmaps.
>
>  -Scott
>
>
> http://codereview.chromium.org/6599017/
>

[brettw, 2011-02-26 03:26:04]

http://codereview.chromium.org/6599017/diff/2001/chrome/browser/ui/views/tabs...
File chrome/browser/ui/views/tabs/tab_renderer_data.h (right):

http://codereview.chromium.org/6599017/diff/2001/chrome/browser/ui/views/tabs...
chrome/browser/ui/views/tabs/tab_renderer_data.h:50: favicon.pixelRefOffset() ==
data.favicon.pixelRefOffset() &&
Good question. This is *probably* OK but I'm not sure. I'm adding Mike Reed to
the review who will be able to tell us definitively.

[oshima, 2011-03-01 03:32:47]

On 2011/02/26 03:26:04, brettw wrote:
>
http://codereview.chromium.org/6599017/diff/2001/chrome/browser/ui/views/tabs...
> File chrome/browser/ui/views/tabs/tab_renderer_data.h (right):
> 
>
http://codereview.chromium.org/6599017/diff/2001/chrome/browser/ui/views/tabs...
> chrome/browser/ui/views/tabs/tab_renderer_data.h:50: favicon.pixelRefOffset()
==
> data.favicon.pixelRefOffset() &&
> Good question. This is *probably* OK but I'm not sure. I'm adding Mike Reed to
> the review who will be able to tell us definitively.


Mike, could you please take a look?

[reed1, 2011-03-01 22:19:11]

If the two bitmaps have the same pixelRef* and same pixelRefOffset() then I
think you're good. There is also a generationID() on the pixelRef (and therefore
on the bitmap), but that will only tell you when a given pixelref has had its
contents changed.

LGTM

[reed1, 2011-03-01 22:19:45]

BTW - the review was sent to reed@android.com which I no longer use.