Clarify kAutohinting is kForceAutohinting.

Clarify kAutohinting is kForceAutohinting.

BUG=skia:2047
R=reed@google.com

Committed: https://code.google.com/p/skia/source/detail?r=13156

[bstell1, 2014-01-15 22:49:19]

Changing the name reflects the change in FreeType. Since 2.4.5 it defaults to
autohinting so we no longer have to force it on.

And forcing it on by default is bad since that prevent native hints from being
used if they are available.

http://sourceforge.net/projects/freetype/files/freetype2/2.4.5/

    CHANGES BETWEEN 2.4.4 and 2.4.5

      ...

      II. IMPORTANT CHANGES

        - If autohinting  is not  explicitly disabled,  FreeType now  uses
          the autohinter if  a TrueType based font doesn't  contain native
          hints.

[bungeman-skia, 2014-01-15 23:17:22]

On 2014/01/15 22:49:19, bstell1 wrote:
> Changing the name reflects the change in FreeType. Since 2.4.5 it defaults to
> autohinting so we no longer have to force it on.
> 
> And forcing it on by default is bad since that prevent native hints from being
> used if they are available.
> 
> http://sourceforge.net/projects/freetype/files/freetype2/2.4.5/
> 
>     CHANGES BETWEEN 2.4.4 and 2.4.5
> 
>       ...
> 
>       II. IMPORTANT CHANGES
> 
>         - If autohinting  is not  explicitly disabled,  FreeType now  uses
>           the autohinter if  a TrueType based font doesn't  contain native
>           hints.

Yes, there is (with this change) no way to state that Skia should use bitcode
hinting but never use autohinting on a modern freetype. The name change here
isn't really because of a change in FreeType, it's simply a clarification of
what the code should have been doing to begin with. All of the flags here in
Skia need to be re-examined to make them more orthogonal (all of them with
regard to hinting and subpixel). As a result, even with this change the code
should still be considered a bit buggy, just differently buggy from before. This
whole file (SkFontHost_FreeType.cpp) is way overdue for updating in general.
There is still some code in here written around the way FreeType 1.x worked.

[bungeman-skia, 2014-01-16 00:05:09]

This change affects more than 5000 blink layout tests so cannot be committed
as-is (in PS3). We will have to consider if other changes should be made first
or if this will simply require a flag. I will investigate more later.

[bungeman-skia, 2014-01-21 19:45:02]

On 2014/01/16 00:05:09, bungeman1 wrote:
> This change affects more than 5000 blink layout tests so cannot be committed
> as-is (in PS3). We will have to consider if other changes should be made first
> or if this will simply require a flag. I will investigate more later.

See the linked bug and https://codereview.chromium.org/138123004/ for a summary
as to why this is causing the Chromium layout tests to fail and a proposed fix.

[bungeman-skia, 2014-01-23 17:14:12]

Since https://codereview.chromium.org/138123004/ has landed, this change can now
land. PTAL.

[reed1, 2014-01-23 18:44:02]

lgtm

[bungeman-chromium, 2014-01-23 19:01:42]

Committed patchset #3 manually as r13156.