Reverting, let's see if that makes the bots happier.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/66ba5e7e910565510efe941cee9c1ba3c2db4f0e

commit 66ba5e7e910565510efe941cee9c1ba3c2db4f0e
Author: mkwst <mkwst@chromium.org>
Date: Mon Apr 10 09:26:15 2017

Revert of Roll FreeType to 5a3490e054bda8a318ebde482c7fb30213cab3d9 (patchset #3 id:40001 of https://codereview.chromium.org/2786013002/ )

Reason for revert:
Many CSS layout test failures, starting in https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20Trusty/builds/25380.

I'm guessing that rolling FreeType in https://codereview.chromium.org/2786013002/ might have caused these differences, as it's the only patch in the diff that looks relevant (and because the CQ didn't run layout tests, which is unfortunate: not sure who's responsible for the `analyze` step?).

BUG= 709918 

Original issue's description:
> Roll FreeType to 5a3490e054bda8a318ebde482c7fb30213cab3d9
>
> https://chromium.googlesource.com/chromium/src/third_party/freetype2/+log/66725768cdf758..5a3490e054bda8
>
> Addresses Clang on Windows LLP64 build failure.
>
> BUG= 706725 
>
> Review-Url: https://codereview.chromium.org/2786013002
> Cr-Commit-Position: refs/heads/master@{#463187}
> Committed: https://chromium.googlesource.com/chromium/src/+/395ac4809e4a98b8745753caf2f3f9015b3ca1f1

TBR=bungeman@chromium.org,drott@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 706725 

Review-Url: https://codereview.chromium.org/2810643002
Cr-Commit-Position: refs/heads/master@{#463202}

[modify] https://crrev.com/66ba5e7e910565510efe941cee9c1ba3c2db4f0e/DEPS
[modify] https://crrev.com/66ba5e7e910565510efe941cee9c1ba3c2db4f0e/third_party/freetype/README.chromium
[modify] https://crrev.com/66ba5e7e910565510efe941cee9c1ba3c2db4f0e/third_party/freetype/include/freetype-custom-config/ftoption.h

The described errors haven't been happening on the bot lately so marking as fixed.