For DirectWrite font metrics use the average character width from the TEXTMETRIC structure.

For DirectWrite font metrics use the average character width from the TEXTMETRIC structure.

Currently we fetch the average character width from the GetTextExtentPoint32 API for DirectWrite as skia
does not return this value. The width returned by the GetTextExtentPoint32 API is 1 px larger than the one
reported by the TEXTMETRIC structure causing some dialogs like the bookmark editor to display much larger than that
with GDI.

We now use the average character width from the TEXTMETRIC structure while reporting back the metrics for DirectWrite
fonts. Longer term fix is to find out a way to calculate this from DirectWrite. Does not appear to be straigthforward.
Added a TODO to that effect in the code.

BUG=434588
TEST=Updated the PlatformFontWinTest.Metrics_SkiaVersusGDI test to validate the average character widths.
For manual test, launch Chrome and open the bookmark
editor on Windows 7+. Observe that the window size is larger than the window size when using GDI for fonts.

Committed: https://crrev.com/0852634071b072ab41e1b255075652cf4c29a0df
Cr-Commit-Position: refs/heads/master@{#313801}

[ananta, 2015-01-27 01:51:03]

ananta@chromium.org changed reviewers:
+ asvitkine@chromium.org

[Alexei Svitkine (slow), 2015-01-27 20:10:55]

"Currently we fetch the average character width from the GetTextExtentPoint32
API for DirectWrite as skia
does not return this value. The width returned by the GetTextExtentPoint32 API
is 1 px larger than the one
reported by the TEXTMETRIC structure causing some dialogs like the bookmark
editor to display much larger than that
with GDI."

If this is fixing a user visible issue, can you add that info to the TEST= line?
Also, I'm actually curious why the returned average char width value has such a
big impact on the bookmark editor API. Any ideas?

[ananta, 2015-01-27 20:42:54]

On 2015/01/27 20:10:55, Alexei Svitkine wrote:
> "Currently we fetch the average character width from the GetTextExtentPoint32
> API for DirectWrite as skia
> does not return this value. The width returned by the GetTextExtentPoint32 API
> is 1 px larger than the one
> reported by the TEXTMETRIC structure causing some dialogs like the bookmark
> editor to display much larger than that
> with GDI."
> 
> If this is fixing a user visible issue, can you add that info to the TEST=
line?
> Also, I'm actually curious why the returned average char width value has such
a
> big impact on the bookmark editor API. Any ideas?

The BookmarkEditorView class in its GetPreferredSize function here 
https://code.google.com/p/chromium/codesearch#chromium/src/chrome/browser/ui/...

This function calls the GetLocalizedContentsSize function passing in 70 and 25
for the width and height respectively.
Eventually we end up in PlatformFontWin
https://code.google.com/p/chromium/codesearch#chromium/src/ui/gfx/platform_fo...
The width is used as a multiplier with the average character width to determine
the actual width. In this case using 7 vs 6 causes a difference of 50 pixels in
the actual width.

[Alexei Svitkine (slow), 2015-01-29 20:52:10]

lgtm

https://codereview.chromium.org/875303003/diff/20001/ui/gfx/platform_font_win.cc
File ui/gfx/platform_font_win.cc (right):

https://codereview.chromium.org/875303003/diff/20001/ui/gfx/platform_font_win...
ui/gfx/platform_font_win.cc:513: // TODO(ananta)
Nit:

Usually, there's a : after the TODO(name) and the text follows on the same line.

[ananta, 2015-01-29 21:04:59]

https://codereview.chromium.org/875303003/diff/20001/ui/gfx/platform_font_win.cc
File ui/gfx/platform_font_win.cc (right):

https://codereview.chromium.org/875303003/diff/20001/ui/gfx/platform_font_win...
ui/gfx/platform_font_win.cc:513: // TODO(ananta)
On 2015/01/29 20:52:10, Alexei Svitkine wrote:
> Nit:
> 
> Usually, there's a : after the TODO(name) and the text follows on the same
line.

Done.

[ananta, 2015-01-29 21:12:39]

The CQ bit was checked by ananta@chromium.org

[commit-bot: I haz the power, 2015-01-29 21:13:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/875303003/40001

[commit-bot: I haz the power, 2015-01-29 22:00:50]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-01-29 22:02:26]

Patchset 3 (id:??) landed as
https://crrev.com/0852634071b072ab41e1b255075652cf4c29a0df
Cr-Commit-Position: refs/heads/master@{#313801}