Simplify the NO_SHADOW_OPAQUE_BORDER path with addRoundRect.

Simplify the NO_SHADOW_OPAQUE_BORDER path with addRoundRect.

BUG=231012, 166075
TEST=--enable-new-dialog-style WCMD look the same.
R=wittman@chromium.org
TBR=ben@chromium.org
NOTRY=TRUE

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=198891

[msw, 2013-05-07 20:45:20]

Hey Mike and Ben, please take a look; thanks!

[Mike Wittman, 2013-05-07 20:59:02]

On 2013/05/07 20:45:20, msw wrote:
> Hey Mike and Ben, please take a look; thanks!

Does this produce equivalent results?  When I last looked at the code, the
translation of the round rect to an HRGN on Win32 simply sliced off the corners;
there was no rounding.  It also seemed to produce different results for
different corners when using a uniform radius.

[msw, 2013-05-07 21:19:59]

On 2013/05/07 20:59:02, Mike Wittman wrote:
> On 2013/05/07 20:45:20, msw wrote:
> > Hey Mike and Ben, please take a look; thanks!
> 
> Does this produce equivalent results?  When I last looked at the code, the
> translation of the round rect to an HRGN on Win32 simply sliced off the
corners;
> there was no rounding.  It also seemed to produce different results for
> different corners when using a uniform radius.

It's identical with the 6 pixel corner radius afaict; I've zoomed in a lot on
black and white backgrounds. I also tried a variety of bubble border fill colors
and corner radii to test difference appearances, but the current style seems to
look the best for variable backgrounds. You can see the before/after pics at
http://crbug.com/231012#c11

[Mike Wittman, 2013-05-07 21:35:56]

On 2013/05/07 21:19:59, msw wrote:
> On 2013/05/07 20:59:02, Mike Wittman wrote:
> > On 2013/05/07 20:45:20, msw wrote:
> > > Hey Mike and Ben, please take a look; thanks!
> > 
> > Does this produce equivalent results?  When I last looked at the code, the
> > translation of the round rect to an HRGN on Win32 simply sliced off the
> corners;
> > there was no rounding.  It also seemed to produce different results for
> > different corners when using a uniform radius.
> 
> It's identical with the 6 pixel corner radius afaict; I've zoomed in a lot on
> black and white backgrounds. I also tried a variety of bubble border fill
colors
> and corner radii to test difference appearances, but the current style seems
to
> look the best for variable backgrounds. You can see the before/after pics at
> http://crbug.com/231012#c11

Great, that makes this a lot simpler.  LGTM

[msw, 2013-05-08 00:14:04]

TBR'ing Ben, setting NOTRY=TRUE (Views-related trybots passed, but mac* bots are
behind and linux_rel (GTK) has been borked all day), and landing.

[commit-bot: I haz the power, 2013-05-08 14:15:37]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/msw@chromium.org/14557010/2001

[commit-bot: I haz the power, 2013-05-08 14:26:00]

Change committed as 198891