Revert 52441 - Revert Win specific elements of 46492 to look for perf impact

There was a regression in the moz page cycler around when
landed. This was ONLY a single core regression, which
suggests it was time spent in the browser becoming 
critical path. This change involved font layout in the
browser, and so it MIGHT be related.

I'll land this, let the per bots start their run, and 
then revert.


TBR=pkasting
Review URL: http://codereview.chromium.org/2895017

TBR=jar@chromium.org
Review URL: http://codereview.chromium.org/3018001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@52456 0039d316-1c4b-4281-b951-d872f2087c98

1 files changed, 3 insertions, 12 deletions

diff --git a/gfx/font_win.cc b/gfx/font_win.cc
index 5660cc7..f9b7243 100644
--- a/gfx/font_win.cc
+++ b/gfx/font_win.cc
@@ -170,18 +170,9 @@ Font Font::DeriveFont(int size_delta, int style) const {
 }
 
 int Font::GetStringWidth(const std::wstring& text) const {
-  int width = 0;
-  HDC dc = GetDC(NULL);
-  HFONT previous_font = static_cast<HFONT>(SelectObject(dc, hfont()));
-  SIZE size;
-  if (GetTextExtentPoint32(dc, text.c_str(), static_cast<int>(text.size()),
-                           &size)) {
-    width = size.cx;
-  } else {
-    width = 0;
-  }
-  SelectObject(dc, previous_font);
-  ReleaseDC(NULL, dc);
+  int width = 0, height = 0;
+  CanvasSkia::SizeStringInt(text, *this, &width, &height,
+                            gfx::Canvas::NO_ELLIPSIS);
   return width;
 }