Fix font height on Mac with Helvetica Neue at size 16 (ResourceBundle::MediumFont)

Fix font height on Mac with Helvetica Neue at size 16 (ResourceBundle::MediumFont)

Currently, descenders are getting clipped during View layout (see bug).

PlatformFontMac uses -[NSLayoutManager defaultLineHeightForFont:] to
initialize |height_|. However, it has a silly rounding bug. Essentially,
it gives round(ascent) + round(descent). E.g. Helvetica Neue at size 16
gives ascent=15.4634, descent=3.38208 -> 15 + 3 = 18. When the height
should be at least 19.

According to the OpenType specification, these values should simply be
added, so do basically that. But since layout in RenderTextMac is performed
using an integral baseline, use the rounded value of ascent, to ensure all
the descenders fit.

BUG=513570
Committed: https://crrev.com/ef6c19f9665ccfa83c27472d01f543f2bbb20249
Cr-Commit-Position: refs/heads/master@{#340627}

[tapted, 2015-07-24 05:15:26]

Patchset #2 (id:20001) has been deleted

[tapted, 2015-07-24 05:15:41]

Patchset #3 (id:60001) has been deleted

[tapted, 2015-07-24 05:33:59]

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

[tapted, 2015-07-24 06:26:19]

Hi Alexei, please take a look.

There's more tracing on the bug at http://crbug.com/513570

(conveniently, someone already deciphered the NSLayoutManager assembly and put
it on stackoverflow -
http://stackoverflow.com/questions/5511830/how-does-line-spacing-work-in-core...
- and the tracing in patchset 1 is consistent with that).

[Alexei Svitkine (slow), 2015-07-24 15:35:39]

Can you also add a unit test?

[tapted, 2015-07-27 03:58:57]

On 2015/07/24 15:35:39, Alexei Svitkine wrote:
> Can you also add a unit test?

Done.

https://codereview.chromium.org/1252893002/diff/150001/ui/gfx/platform_font_m...
File ui/gfx/platform_font_mac.mm (right):

https://codereview.chromium.org/1252893002/diff/150001/ui/gfx/platform_font_m...
ui/gfx/platform_font_mac.mm:169: ascent_ = ceil([font ascender]);
note I ditched SkScalarCeilToInt here, just because since mac became 64-bit,
CGFloat is now a double. SkScalar remains a float, so I don't think it makes
sense to convert to a float before rounding. (and when SkScalar is a double,
SkScalarCeilToInt is just `(int)ceil(value)`). This wasn't because of any
particular failure, but makes me happier about the float conversions the test is
relying on.

[Alexei Svitkine (slow), 2015-07-27 14:48:04]

lgtm

https://codereview.chromium.org/1252893002/diff/150001/ui/gfx/platform_font_m...
File ui/gfx/platform_font_mac_unittest.mm (right):

https://codereview.chromium.org/1252893002/diff/150001/ui/gfx/platform_font_m...
ui/gfx/platform_font_mac_unittest.mm:64: gfx::Font::FontStyle style[] = {
Nit: styles

[tapted, 2015-07-28 00:38:49]

https://codereview.chromium.org/1252893002/diff/150001/ui/gfx/platform_font_m...
File ui/gfx/platform_font_mac_unittest.mm (right):

https://codereview.chromium.org/1252893002/diff/150001/ui/gfx/platform_font_m...
ui/gfx/platform_font_mac_unittest.mm:64: gfx::Font::FontStyle style[] = {
On 2015/07/27 14:48:04, Alexei Svitkine wrote:
> Nit: styles

Done.

[tapted, 2015-07-28 00:38:58]

The CQ bit was checked by tapted@chromium.org

[tapted, 2015-07-28 00:38:58]

The patchset sent to the CQ was uploaded after l-g-t-m from
asvitkine@chromium.org
Link to the patchset: https://codereview.chromium.org/1252893002/#ps170001
(title: "style -> styles")

[commit-bot: I haz the power, 2015-07-28 00:39:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1252893002/170001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1252893002/170001

[commit-bot: I haz the power, 2015-07-28 01:30:39]

Committed patchset #8 (id:170001)

[commit-bot: I haz the power, 2015-07-28 01:31:27]

Patchset 8 (id:??) landed as
https://crrev.com/ef6c19f9665ccfa83c27472d01f543f2bbb20249
Cr-Commit-Position: refs/heads/master@{#340627}