Clean up FreeType library ref counting.

Clean up FreeType library ref counting.

The existing reference counting and code for LCD is difficult to
understand. This cleans it up a bit in preperation for the typefaces
holding the references and faces themselves which will clean this up
quite a bit more.

Committed: https://skia.googlesource.com/skia/+/9dc24686e719566fd9ef054d6d27cff3a12daa71

[bungeman-skia, 2014-12-01 21:28:18]

bungeman@google.com changed reviewers:
+ mtklein@google.com, reed@google.com

[bungeman-skia, 2014-12-01 21:28:19]

This is a bit of a cleanup by itself, but the idea is to eventually push the
FreeType library state owning out to the typefaces since they need the library
anyway for almost everything they do. This will also allow us to simplify the
FT_Face reference counting as well.

[reed1, 2014-12-01 21:38:46]

lgtm

[bungeman-skia, 2014-12-01 21:50:43]

The CQ bit was checked by bungeman@google.com

[commit-bot: I haz the power, 2014-12-01 21:51:40]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/742483004/20001

[commit-bot: I haz the power, 2014-12-01 22:01:37]

Committed patchset #2 (id:20001) as
https://skia.googlesource.com/skia/+/9dc24686e719566fd9ef054d6d27cff3a12daa71

[jcgregorio, 2014-12-02 01:35:45]

On 2014/12/01 22:01:37, I haz the power (commit-bot) wrote:
> Committed patchset #2 (id:20001) as
> https://skia.googlesource.com/skia/+/9dc24686e719566fd9ef054d6d27cff3a12daa71

Was this change expected to have a performance impact?

  https://skiaperf.com/cl/346

For example, one trace from that cluster:

  https://skiaperf.com/#1601